Hi Tony, my question is what will be the aperture at. 18mm focal length say for example in a 18-35mm f/1.8 ? i understood that it will be 2.7 at 35mm but what about the aperture value at 18mm. ?
mmj Since the lense can be at f1.8 at any given focal length, this means it will behave like an f2.7 lense, when mounted on a 1.5 crop body, regardless of the focal length
The industry seem to base everything off the established 35mm slr days. So if you take a full frame sensor (35mm) and compare it to a crop sensor then the later is going to be noisier, all other things being equal. What is being demonstrated here is that a crop sensor can deliver an image of equal quality if the crop factor is used correctly with different hardware/settings. Are companies responsible for our ignorance when they sell a lens as 12 - 35mm f/2.8 when its bought for for us with a crop sensor and is therefore equivalent to a 24 - 70mm f/5.6? I think not if the lens is sold as for a full frame body and yes they are being fraudulent if the lens is being sold as for a crop body, totally misleading customers.
Wow Tony! This is my first RUclips comment, and for good reason. I'm absolutely blown away by this video. I just performed a test, as I had to see this for myself. I shot the same scene with two different setups. 1) Canon 6d with a Tamron 24-70 2.8 vc zoomed to 70mm at ISO 4000 @ f4.5 2) Canon sl1 with a Tamron 17-50 2.8 zoomed to about 43mm at ISO 1600 @ f2.8. Both pictures look absolutely indistinguishable from one another. Bokeh and noise are practically identical! I can't believe I never thought of this before. You've changed my life man. Now I better understand the limitations of all of my gear and will govern their use accordingly.
Please correct me when I'm wrong. When it comes to the exposure itself, crop factor does not matter. On every camera the settings of ISO, SS, Aperture will be the same for the same subject under the same light (except when the crop factor crops out area's that are lighter or darker than the average). So a F1.2 lens on a 2x crop is still a fast lens, in the way that it allows you to use a higher shutter speed at the same ISO. However, when it comes to depth of field it will not act the same way. And image quality of course is not the same
Exactly! What Tony talks about applies only to the depth of field. He probably should've mentioned that in this video. It has really created some confusion throughout the industry lol
Hello Tony in the video at the 21:09 mark, you talk about the iris or opening, where do you find this measurement. the example you show nikon 24-70mm 2.8 and the opening of 25mm. I dont know where you get the 25mm from. I looked at dxomark and cant find. Help! I want to be able to make these calculations.
I am a physicist and I've worked on a variety of EM sensors (telescopes and radar) in a career spanning over 20 years. I've only gotten seriously into photography for the past year or so though. Thank you for translating all of the physics in my head into the language of photography. You're absolutely right, camera terminology made a lot of sense when everyone was using film of the same size, but using digital sensors of different sizes makes that terminology confusing. I can see a lot of stuff that I've been doing wrong and can't wait to try my hand at fixing it. Love your work, and I'm definitely subscribing. Cheers!
I wish photographers would stop saying sensor size determines DOF, b/c it doesn't. It might determine your distance to subject, which will in turn determine your DOF, but it doesn't change your DOF. DOF is a result of three variables: Distance to subject, focal length, and aperture. Greg Cazillo has even done a video to prove DOF doesn't change if all three of those variables stay constant, even if you mount a different camera with a smaller sensor.
This is absolutely ground breaking stuff! Tony, I tried applying the crop on my Fuji XT20 and compared the meter readings on my 35mm film AE1. Here are the constants; Iso (1600), shutter speed (1/60th), equivalent focal lengths (28mm). The Fuji's aperture was set to 2.8 and I applied the crop to the AE1 in reverse making it more or less f4. I am getting a results where my meters (both center weighted) are saying that they are getting proper exposure on the respective adjusted f-stops. Am I getting the correct results?
(electrical engineer here) Yes you are doing everything right. However, the issue is, ISO is not a constant. ISO is defined on a sensor-by-sensor basis due to different pixel sizes needing different electrical gains to produce equivalent results. Lets say the brighness of a pixel goes from 1 to 10, if this pixel absorbs 4 photons and the apparent brightness is 8, then the electrical gain would be 2 brighness per 1 photon. If you now take a pixel that is only a 4th of the size, thus only collecting 1 photon, then the gain would be 8 brightness per 1 photon. But still both gains would be called, let's say, ISO 100, since they produce the same result under the same lighting conditions. But what is missed is how the electrical circuitry has to compensate for the smaller pixels.
This guy is by far the most knowledgeable photographer instructor on RUclips. This is the first time when a lot of these photo terms were actually broken down and understood. Bravo subscribed
Firstly P/Sqrt(p) = just the sqrt(p) What happens when photons hit a sensor is that they convert photon energy into photoelectrons by interacting with the silicon layer, but by interacting with the silicon layer they also vary in the timing of when they interact with the silicon layer causing imperfections in the "quality" of photons and therefore the conversion to photonelectrons - this "photon noise" is directly related to the photon energy (light intensity) in an inverse square ratio - i.e. the amount of noise generated by "quality" of photons is the square root of the intensity of the light interacting with the silicon layer on the chip or photon noise (noise) = square root of the photon intensity (signal) But this formula only describes the input part - it doesn't describe the Signal to Noise Ratio, which is the measure of how well the total amount of light "in" (the signal) relates to the total overall noise "out" (of the entire processing system start to finish) that ends up producing the end electronic representation of the photons that hit the sensor and represented by the factors / formula below. 1. The amount of light energy per pixel per second ( like the photon volume or let's call this "P") 2. The efficiency with which the sensor can convert each of the photons into an electrical signal that can be interpreted and read (let's call this "E") - a bad sensor will lose some information which is the same as generating noise 3. The amount time that the sensor takes to accumulate & aggregate the photons into a readable signal (let's call this "T"). 4. The internal current (noise) as a consequence of heat generated from the sensor itself converting photons to electrical signals (call the "H") a cold sensor will have much less introduced false electrons (noise) than a hot sensor 5. The read noise (let's just say the noise generated by the pre-amplifier on the chip for now and call this "R") The formula for SNR is ... P x E x T / (sqrt of (P x E x T + H + R)) 1. Is a natural phenomemon of the input and the same for all cameras 2-5 are consequenses of the manufacturing process of CCD sensors - and the main reason why newer sensors can generate lower noise outputs from inherently noisier photons (low light input).
Well, they only considered the fundamental shot noise (#1), the other factors are almost negligible with modern CMOS sensors. Including all these in a video would make it too complicated for most viewers.
Tony Northrup About that F/2.8 equation that's only half of the truth. "Focal length equivalent" is a short-hand term used to give us uniform information about angle of view of given lens between different formats, but real focal length, real aperture, and what's most important, the brightness of the lens does NOT change. So F/2.8 lens on 35mm camera is as bright as F/2.8 in any other format, smaller or bigger. You can say F/2.8 in MFT is equivalent of F/5.6 in FF only in terms of depth of field, but when you're talking about brightness of the lens, then it's not true, F/2.8 understood as brightness indicator stays F/2.8.
The brightness indicator is calibrated to make the picture appear similarly bright for equal fLen, Aperture, shutter and ISO despite different sensor sizes getting different total light, or alternatively its just indicating light incidence per unit sensor area. Look no further than cell phones to see how light incidence per unit area alone doesn't indicate quality. Taking post crop factor equivalence into ISO, Fstop, and focal length is getting the whole picture, because the goal is finding what it would take to capture an equivalent image.
This is the best crop factor video ever. One more thing, what about the shutter speed? Does sensor size affect shutter speed?. If Fullframe with F/4, ISO 800, and 1 second Shutter speed equal to F/2.8, ISO 350 on APS-C, what shutter speed on APS-C to match with the fullframe?
Hey Tony, AWESOME video. Really great stuff. Question for you: If camera manufacturers are cheating us with their aperture numbers, are medium format lens makers selling themselves short? :)
Hi Tony and Chelsea, I really appreciate your effort in reviewing photographic products in a sensible manner, and Tony, your math rocks , and you know I thought of constantly watching this video time by time to engrave in my mind about the crop factor facts. Otherwise these MFT glass and cameras are so compelling. I was watching these videos since some time now, and I admire the way you critically and scientifically analyze performance. Good job for both of you. I had my first DSLR in 2007 is where the Canon 400D came out to market, bought it with a personal loan! Then I sold it, when I was discouraged by the reach the 90-300mm lens I could afford, and the expensiveness of long lenses. After a while I got this Canon power shot SX-50 claiming to be the world's longest zoom camera for just ~500$. But I realized that this tiny sensor's noise and that it's always starving for light. Then I have ended up in a Canon 700D and I cannot complain about it despite the size and bit high noise. I had almost made up my mind to get a Olympus EM10 which is available body only for 250AUD used, for the compact size of bodies and lenses. But after seeing this video, I thought of stopping a bit and re-looking at getting a APSC ones. So I don't know in APSC mirror-less world is there at least cargo-pant pocketable options with IBIS. I know that A6500 does, but at what cost? So the search continues, and pls give me some tips if you can. Currently I am thinking of these models Olympus OMD EM10 -250AUD used (body only) Any Sony APSC mirror-less Fujifilm X series ILC or fixed lens Canon EOS M6 or M5 etc.
Being an engineering background and that too mechanical, I had numerous doubts and confusion on how all these parameters worked. Sometimes I would think it's a blunder. Watching this video, now I'm aware it's really a marketing blunder.
Tony, cool video man. A real eye opener. I was confused with all the various attributes and numbers i see in my camera and lenses. Now i know what to look for when i need to buy something. You have effectively condensed 100 hours of reading material on the internet to a clean, neat 30 minute video. Thanks again.
I have watched this 3 times now and learn more (or understand more) each time. In my hunt for a 'walk around street lens' I have been looking at the (Pentax) 24mm f2 (~$500) and the 28mm f2.8 (~$190) to be used on the Pentax K-S2. Both lenses have very good ratings. After watching this again I think I'll go with the less expensive lens. The 28 will give me a 42mm f4.2 with a 55 degree angle of view on the APS-C sensor which should do well. I can toss the kit 18-50 f4.5 - 5.6 in my pocket and be covered for most everyday type stuff. Thanks, Tony for helping me better understand the lenses I was considering.
I am doing my post IT career things do learn list. This is the second video I watched from you. The presentation is great and the math is easy. I am just looking for a camera that takes good pictures indoors in a low light school auditorium with kids moving for under 1.2k In any case, this story reminds me of MPG on cars. What the car MFRs never tell you about when they say city/highway MPG are: the average temperature outside, the average number of people (weight) in the car, the average amount of up/down inclines in the road, what tires and inflated PSI was used, and finally, what gas they used. It seems to get the best MPG you must drive with one person at 65 degrees (no heater/AC) on straight flat road to get the best results !!! Thanks again.
Aperture is still the same on any sensor, difference is only in field of view which means you need to recompose image and change distance. And these two steps change resulting depth of field. Not the size of sensor. Lens is still the same, you only take crop out of that image circle. Registration distance is unchanged. ISO in digital era has to take in account the size of each light receptor (pixel), not the size of whole sensor. 16Mpix 1.5x APS-C gathers the SAME amount of light per every pixel as 24Mpix FF sensor and also the noise with same generation of technology is the same. Problem with noise and smaller sensors is, that they reach pretty high resolution, which results in smaller photosites which results in more noise (signal + temperature noise). Cameras like 1Dx and D4 have less noisy images because photosites on their sensors are quite large when compared to APS-C sensor with same resolution. Other issue is with crop-designed lenses, which in most cases have less light transmission (dark corners) than if you use FF lens on crop sensor. If you cut smaller piece of exposed film, it will be still the same film, still the same grain and still the same DOF and bokeh and captured amount of light per square milimeter. The same applies for FF and Crop sensors.
xmeda "Aperture is still the same on any sensor, difference is only in field of view which means you need to recompose image and change distance. And these two steps change resulting depth of field. Not the size of sensor" That is a common myth. First of all you don't need to move closer to recompose you can just zoom in, second if you move closer to recompose you will get a smaller DOF , which will result on a even more blurred background on the FF, which is exactly what we see in the video. The lenses are on the exact same settings but because the sensor size is different the DOF is different. That is the whole point here. No one is saying that the sensor size is changing the state of the lens, at 2.8 the diameter of the "pupil" is still the same, however the angle of view isn't and this affects DOF.
xmeda And why do you zoom young padawan ? Joke aside.I know that, also you are zooming on the FF not the CF. I was just pointing out that "it changes because you move" is a myth. You have to recompose because of the difference in the sensor size and the resulting change in DOF you get by trying to get the same comp is directly proportional to the crop ratio. Don't forget that focal length has the same problem ISO has as Tony mentioned. There projection triangle can be calculated using the focal length, which will be the "height" of the triangle and the base is the sensor size, if we decrease the sensor size we also change the angle of view but keep the height the same. We basically have the same effect if we put a teleconverter except we change the angle of view by changing the height. So in that sense a CF works is the same way as a Teleconverter and what we should be talking about is not the distance where lights converges to a point (focal length) , but about this triangle and you need at least 2 factors out of 3 to do that: angle of view , base (sensor size) or height ( focal lenght).
Rox alcantara welll everything that i said was in agreement with this video. Tony does make a good point in his video. We are using standards that are out-dated and are extremely relative to which sensor we use and aren't really precise (imo enough). The only way to change this is for professionals to be aware of what is going on, most of them aren't. It may seem like one of those conspiracy videos but in this case what he is saying is true. People that buy a 1.4 lens made for a cropped sensor aren't really getting the image quality a 1.4 lens would give to a FF camera, we need more awareness to this.
I fully agree with Tony that deceptive advertising should be pointed out, and the companies that do that should be called on the carpet. Now, when comes to the difference in light gathering ability between smaller sensor and full frame, it's not so easy for the average photographer. Tony pointed out how advantageous larger sensor is improving subject isolation. This can be a big advantage to professional portrait photographers. But does it affect most people? I will posit that for most people it is not *too much* depth of field (DOF) that presents a problem, but *not enough*. Let me give you an example - you have your family at home, they sit around the table, some sit on a couch. You could resit them all in a yearbook style, but you probably want to catch the scene when they are relaxed and behaving naturally. In that case you need *more* DOF. With a full frame, you may need to stop down to f/8 or f/11. That means you will need to slow down your shutter, but that also may present a problem with a large DSLR - mirror slap and shutter slap may force you to use shorter shutter speed. So you end up jacking up your ISO. On the other hand, your Sony RX100 with a CX sensor at f/2 will have a DOF equal (roughly) to f/13 on full frame. You also can handhold the camera more easily at slower shutter speed (no mirror, different shutter mechanism, or electronic shutter). Now you do not have to raise your ISO, and a good part of the full frame advantage has been nullified. This is just one example, but I can come up with many of them, having shot all kinds of sensor (and film) formats for many years. Of course full frame may still produce slightly better quality, but then we get into the convenience or carrying it around instead of having a camera in your pocket. In other words, it's not all black and white, alas. Pick your tools according to your application, and as Tony points out, don't let the vendors sucker you.
+mountainhobo Camera manufacturers always lied about technical specs, also in the film era. Especially DSLR manufacturers like Canon, Nikon, Olympus and so on. For example, they bragged a lot about the shutterspeeds (like 1/4000 which was VERY fast back in the days). When I was studying photography, my teacher brought a shutterspeed tester one day (you removed the film and lens and placed a ultra fast shutterspeed-lightmeter on the place of the film, this way, by pressing the shutter release button, it could accurately measure the true shutterspeed). It all ended up that the fastest shutterspeed was 1/1200 or something and not 1/8000 as the dial of my Nikon F4S showed. It was probably technically not even possible to get a shutter curtain open and closing in 1/8000 of a second even if Nikon stated it could do so. I felt betrayed since that camera used to be the most expensive SLR back in the days (5000 dollars i think back in the days).
+Lutte Kikker That's probably because shutter is not a single piece of metal, but two rolling curtains which follow each other going from top to bottom. Even the highest end cameras take about 1/1000 of a second to actuate, but only expose 1/8th of the at a time in a rolling fashion. Slow mo guys have a recent video about how camera shutters work.
+Arakki I'm afraid you got it the wrong way. If you say the shutter speed is 1/8000, it doesn't mean "let's expose for 1/1000 just a 1/8 of the frame", because this means you expose for 1/1000 sec.
Donato Greco So, what's the real shutterspeed then when all Nikon F4/F4s state 1/8000 of a second? That camera has 2 shutters (2nd curtain). I did shot several sequences (all faster shutterspeeds on high ISO, pretty fast film and broad daylight) and the contact sheets didn't show a real difference in exposure at some point (I think 1/1200 or 1/1500 s, somewhere in that range).
Lutte Kikker You have to think to the shutter curtains at high speed (higher than the flash sync, when the whole sensor is exposed to light) as a horizontal (or vertical) hole in an opaque material. This hole has to grant the exposure set as Shutter Time. If we extremize this concept to a single pixel sensors row exposed one by one for the whole trip of this "hole" along the height (or width) of the sensor, this "hole" will expose every single pixel row for the Shutter Time set in the camera. If this wasn't working this way, you will break the exposure equation (ISO, Aperture, Time).
Fantastic video Tony. No other way to say it but that. You nailed it in the video when you said "there's no such thing as a free lunch". Light is light is light. There is no way that a small sensor with a small lens can inherently gather a "clean" image the same way a large sensor and big lens can. (Although people might argue that the new A7S with its 12mpx and larger photosites that gather more light which makes the ISO "cleaner" at larger numbers, but the math still stands. Well at least I think so...unless I just made a really good point there lol) Anyway, kudos!
Thanks, Dave. A couple of people have brought up the topic of per-pixel noise, which I don't address in this video. It might be worth addressing per-pixel noise in a future video, but we almost always see pictures scaled down to fixed dimensions, rather than looking at individual pixels. Scaling images blends multiple pixels together and makes total light gathered the biggest influence on total noise. The A7S (and the GH4) is designed to do 4k video without blending pixels together, which definitely has its benefits, but isn't really relevant to these concepts.
Yup...while I love my M 4/3 I am often frustrated with it limitations. Max usable ISO and frame rate for example. You just have to know what your getting and make an informed choice. I call the M 4/3 my "Disney" camera since I can walk around with it all day and still get shots I am satisfied with. Better than a point and shoot but less capable than my FF.
TheSunshineSpot Yeah, I love MFT cameras for the same reason. The thing that frustrates me most about the MFT cameras is that the base ISO is 160 (for Panasonic) or 200 (for Olympus). With those small sensors, I need ISO *25* if I'm going to get images as clean as ISO 100 on my full frame cameras! Let me go slower!
Tony Northrup But the sensors have a "native" ISO and going below that will always be a frig, so not always better. For Panasonic (and usually Sony, as Olympus don't do sensors) m43 sensors that's ISO 200 and that's what the pixels will be full up (well, less a margin manufacturers leave for highlight recovery) at that exposure. Unless you can make even deeper pixels (and they are pretty deep already) then you just run out of capacity in the well. (Strictly the Native ISO is when DR is max and noise is min, but it's usually around the full well capacity.) For example an E-M5 pixel can hold 25K electrons (1Dx = 90K, D4 = 118K), the QE is 53% (49%/53%) so basically 50K photons and it's full (but would always leave some margin for highlight recovery, so 40-45K). If an ISO 200 exposure gets you to, say, 30K photons for the brightest areas then that's your lot and a lower ISO isn't possible (as would need too many photons to be stored) except you can throw away some highlight margin and make a somewhat lower ISO. (I believe that the next full ISO under 200 would not have worked and so they go a little less than a stop lower and lose a chunk of the highlight margin.)
Sensor size doesn't have affect for dynamic range or exposure range. It is about different sensor what does the difference for dynamic range, exposure range and ISO. Different sensor == Different result. Same sensor but different size == same image quality with same values but only a different cropping and finally resolution. This can everyone find out by themselves as long they are ready to be open minded and really learn technology. The same thing is exactly with the films. You don't need to adjust shutter speed or aperture when using same film (ISO was tied to film) in different crop factor cameras. Sensor size DOES NOT change depth of field, cause more noise or affect to exposure range. It is difference in other technical features than sensor size. Example: Get a 35mm film camera and load it with 35mm ISO 400 film. Place camera on tripod and leave it there so you don't move it anywhere. Take a photo. Then if you would be a skilled enough (lets say you are), take the non-exposured part of the film and slice a part off so it gives you 1/4 of the 35mm frame size (aka 2x crop factor). Now place the piece of film on the film plate and take a photo with same settings as you did with full frame. Now go to darkroom and develop the film and the piece of the film. Do you claim now that now the piece of film now has grain of ISO 800 or it is somewhat under- or overexposured when compared to full frame? Lets do it with other way on digital manner. Lets walk to manufacturing line up and get a silicone dice what has all the Sony A7r sensors laid down waiting to be cut to components. But lets cut (yes silly, but lets say it would be possible) a single component to 1/4 of its size and with some magical way replace A7r sensor with that our own new custom made 1/4 of the size sensor. So all the other digital parts in our custom Sony A7r is intact, only the sensor size has changed from 36x24mm to 17.3x13mm. Do you claim that now suddenly our custom made Sony A7r has changed so much that ISO noise is worse, you need change shutter speed or aperture to get same exposure or you have 2x depth of field as unmodified stock Sony A7r does when shooting from tripod in same position? Sony, Olympus, Panasonic etc don't lie. They don't cheat you (Well, Canon and Nikon does in many of their cameras).
Tony. Thanks. Are you using the thread size to figure len at wide end. Ie. I'm thinking of getting tamron 70-200 f2.8 has 77mm end. Using d7100 apc body. So it's not really f2.8 when at 200. Because of crop it's 300 divided by 77. Is this right
Very compelling presentation. I was a motion picture lighting/cameraman for 45 years and have the following practical observations: Cine camera lenses use t (transmission) stops, not f-stops to eliminate exposure errors when changing focal length. A lens for 16mm film has to have a much wider full aperture than a 35mm (half-frame) lens to achieve the same shallow dof given the same angle of view and subject distance.
Well, this could be an endless story. Bring medium format and large format into the game and the so called "full frame" cameras (which is a ridiculous name for the 35mm format) and now these are the "cheating" ones as well. This has nothing to do with "cheating" at all. They're just behaving different - each on it's own.
Leif, the cheating part is where manufacturers selectively describe some aspects of their lenses in terms of 35mm, and other aspects in their native format, depending on which makes the lens more appealing. Nobody puts 35mm gear in terms of medium and large format.
Tony Northrup That's a problem the industry created decades ago. When we had medium format and the new "small" 35mm format got more popular they also kept using the focal length as the "measurement" and didn't used the medium format as a reference. Now we see basically the same story again with other smaller sensors. The focal length is the focal length and it has nothing to do with cheating because they're telling us the real focal length of the lens which is true. The only issue is that the focal length is the worst thing to compare lenses because it says nothing about the actual viewing angle we get. But that's a problem the industry created decades ago and photographers got used to it as well. You'll barely find a photo book where they speak about viewing angles instead of focal lengths.
***** Even that wouldn't be untrue because then you could say that Fullframe and Medium Format cameras are cheating because they don't use large format as reference ... and so on. Compare a 50mm Full Frame lens to a 50mm Medium Format lens...the Full Frame lens will behave like a 80mm Medium Format lens. It's just a bad idea to try compare different systems based on those numbers (or at all) in my opinion.
***** The question is what people expect from a 2.8 lens. Are they using the aperture as a reference for the depth of field they get or as a reference for the light sensitivity they can expect on the sensor size they're using? You can turn it in each way..it will always be confusing. A 2.8 lens is a 2.8 lens exposure wise. And then again there would have to be a new reference because full frame can't be the reference either because it's a format that its in the middle and not on the top.
re:Smaller sensor getting less light - you are correct in wording but I believe incorrect in concept / practical application. A Canon crop sensor, for example, will only see the center portion of a 50mm lens compared to a full frame. That means the rest of the light beyond the sensor's borders are discarded (hence the "crop"). Therefore, that light is not being seen, and it could technically be said that it is therefore seeing "less light". However, the light that is hitting the sensor in the middle is of the exact same intensity (ie. same measure of photons) as it would be on a full frame sensor. It absolutely DOES NOT receive less intensity, as that would not make sense on a purely scientific perspective based on the physics of light. It would be simply a full frame image with the edges cut off, at the same intensity for exposure. If you were to open up a 5D and put masking tape over the edges of the sensor to make it only see a 1.6x, the exposure would be the same. It would just be cropped. That's all that happens in a crop camera. So why did your tests reveal more noise at the same ISO? That has much more to do with the size of the photosites on the sensor- that is, the size of each "pixel" being exposed. Remember, a sensor is nothing more than a grid of photosites, each one sending information about the light that it sees to the image processor to turn into an entire picture. The larger the photosite, the more photons it can collect and the more accurate the resulting light sample will be. Smaller photosites that are more densely printed on a sensor will therefore yield more noise than larger ones. You could argue that a larger sensor automatically has larger photosites, but that isn't necessarily true. How large the photosites are is a combination of sensor size to megapixels. The more megapixels a sensor is designed to have, the more photosites need to be crammed onto the surface and therefore the smaller (and less light sensitive) each one will have to be. Take the Sony A7S for example. It is a low light champion because it offers only 12 megapixels compared to the flagship A7R which offers more than 36 megapixels. That means that the A7R has photosites 1/3 the size and therefore will have noisier high ISO. To say it has to do with sensor size alone is terrible misleading. You can have a smaller cropped sensor with lower density photosites (for example, a micro 4/3 with only 12 megapixels) that reasonably outperforms a modern high-density full frame when it comes to noise sampling because each photosite is large enough to gather more light.
So is it not then more about pixel size vs sensor size? As in a 16MP APS-C will have smaller pixels than a 16MP FF. Therefore each pixel on the APS-C will receive fewer photons than the FF?
This video is fantastic, I already knew about crop factor that had to be applied to the aperture as well but somehow it was always considered as a secondary information leading me to forget it when comparing lenses, after this video I'll never convert only the focal lenght without doing the math on the aperture as well, thanks so much Tony!
Can you help me with this. I thought you implied that angle of few and field of view were interchangeable. I thought that angle of few was a function of the lens. Field of view was a function of the sensor. Wide angle lenses will distort faces more than longer focal lengths. The distortion will be present on both cameras but the field of view will show more or less of the scene. One of the reasons I prefer FF over Crop. Better working distance in small rooms w/FF.
Wide angle lenses distort things in the corners of their field of view (= at large angles of view). With small sensors, you can not "see" those corners because they are just outside the sensor. Therefore, you are getting the same distortions at 28mm full frame as at 18mm APS-C. The sensor size really does not matter if you do the right math. The only issue is that lens manufacturers don't make big lenses for small sensors, so we are forced to buy full frame sensors to be able to use big lenses that capture more light and make stronger bokeh
I was in disbelief until doing my own research and running the math. I'm floored. Thank you Tony and friends for putting this together for us. I live in Japan and can only say that my respect for Japanese manufacturers has taken a humungous hit after this. I shoot with Panasonic gear, and used to shoot Canon, but knowing this gives me much more to consider before making any further purchases. Thanks for doing the hard work. I think I'll create a video like this for the Japan market so that folks here can get a better understanding as well. What a damn shame.
Cold hard facts! Love it Tony. I just had a conversation with a client that has worked at a very reputable camera shop for decades and totally shut me down when I talked about crop factor vs aperture. It goes to show that many veterans don't understand the math. Thanks for the great info!
One of the best RUclips photography videos I've seen. Thanks Tony for revealing the truth behind some of these misconceptions in photography and how some manufacturers exploit our ignorance.
Tony Northrup I also recall years ago how some manufacturers were quoting large megapixel counts on their cameras when in fact they were half that amount and then digitally doubled. The same can also be said for the marketing campaigns surrounding ridiculous digital zooms. Unfortunately there will always be a market for people that don't do their research and ans long as there is, then companies will continue to exploit it. Human nature I guess. Thankfully for people like me who love to learn, there is an abundance of research material and people like you that we can learn from. Thanks again Tony. Really enjoy your videos.
This is perhaps the best photography video I've seen on youtube. Wonderfully informative, incredible clear and well documented. This should be required viewing, not only for the novice, but for those of us who started back in the day when film was all there was.
16:15 How did you get the same image size on different crop factors? Zoom adjusted or image crop only? Please explain. Looks like first one was shot with 200mm, second one with 100mm with x2 crop factor. This is wrong in this case and can't be used for the rest of the video here.
Hello Tony, one question, at the end you said that Fuji, Canon and Nikon are not cheating at the moment they specify the aperture in their lenses, do this applies now a day ( I ask because I see this video is from 2014), so for example, the Fuji 16-55 f2.8 is really a 2.8? is that the reason why is bigger and heavier than the new Sony apsc lens 16-55mm 2.8? how do I know they are not cheating? thank you
They are still cheating. Fuji 16-55 f2.8 is really a 2.8, but if you want to compare it with full frame, this lens sadly becomes 24-83 f/4.2. They will just never mention the equivalent aperture.
Nobody is misleading !. A 2.8 lens is 2.8 no matter what size sensor you use it on. The formula for Aperture does not change = Focal / Aperture diameter. When they came up with this for a way to standardize exposure, 35mm film was not even around !
The truth about crop factor(CF) is that it has been applied to all the wrong things to compare crop and full frame(FF) sensors. The only thing that you should apply the CF to is the field of view (FOV). The physics of a lens does not change because it’s on a FF or crop sensor, what changes, and the only thing that changes is the FOV. A crop sensor does not change the focal length or depth of field(DOF) of a lens. Here is where we usually get it wrong, it is when we try to get the FF equivalent on a crop sensor. When we multiply the CF by the focal length we do not in fact change the focal length of the lens but the equivalent FOV. For example when you use the APS-C CF (1.5 or 1.6 for Canon)) to calculate the FF equivalent of 100mm, you get 100mm x CF = (150mm or 160mm). The general assumption that you are getting 150mm or 160mm equivalent focal length is wrong. If you were to compensate for FOV by cropping the FF or using the brenizer method to get the 100mm FOV equivalent on FF and doing a comparison, what you will find is that the DOF is the same, and for any other given focal length it is the same. The marketing on lenses does not make the distinction between focal length FOV equivalent which is why this topic has been so confusing to most of us
I feel like my life is just a big lie right now, and I feel very discouraged as a photographer. However, I am very thankful that you made this video, and opened my eyes on future puchases, so thank you so much, Tony! On an unrelated note, would recommend getting a used 5D Classic?
Thanks! The 5D classic is pretty noisy compared to modern full frame cameras, and it lacks video or live view, but we still have ours and it's a capable camera.
Just a question.. Depth of change on different sensor, even at same equivalent focal lenght, for example you have to double it on mirrorless camera compared to ff. But is it the same for the light on the sensor? 2.8 is 5.6 on 4/3 and so you gain half light with sand shutter and iso settings? Thanks
9:33 That does not make sense. A crop sensor basically crops in on an image that could've been taken by a full frame sensor. Why would an image get darker, just because part of it was cropped out? It doesn't. It stays the same brightness. The amount of light hitting per mm^2 stays the exact same. However, the one thing that probably changed is the size of the photocites. Smaller sensors usually have a higher density of photocites, so individual photocites gather less light, which has to be adjusted for by "secretly turning up the iso".
To see the effect, you need to have the same number of pixels in all sensors, meaning the pixels would have different size indeed. The signal to noise ratio obviously depends on the pixel size. But, I think, including this variable would make the video too complicated and the conclusions would be the same anyways.
right. His Iso theorie is complete nonsense, because he doesnt realy know, how optical systems work. A Camera is not a a bucket in the rain - unless you put it outside without the lens. Or - to explain it correct with the bucket - theorie: if you put 2 buckets in the rain with a lid and the same 2.8 whole in it, you will get the same amount of water in the big and the small bucket.
Tony Northrup I think your example at 31:26 is good... it shows that you are wrong converting aperture. Because, yes, you can shoot sport at 600 mm and f/2.8 with the Lumix and no, it is not equivalent to f/16. You will not need extra high iso but the same iso as the 600mm f/2.8 you would use on a FF. But yes, larger sensor have better image at a given iso so the picture will be better if you go to high iso/high speed. You keep repeating that large sensor receive more light than smaller ones. This is not true if you use the right lense on the right camera. Because the smaller the sensor, the closer it is to the diaphragm. Camera makers are not cheating, f/2.8 is f/2.8. Difference are in DoF and bokeh and I agree they could communicate on that but I think people interested in this are informed.
Tony Northrup The video didn't correct your wrong assumption that more light hits bigger sensor (at a given aperture). You probably believe that rays of lights are parallel to the sensor in a camera? It is evidently not the case because lenses focus light and 'redistribute' it on the sensor (radial beams from the focal point. Smaller sensor are generally closer to the focal point). And that contradicts your theory that smaller sensor would be less luminous and needs conversion to give 'equivalence'. Maybe you should stop 'matching the maths' because I can take the same picture with my RX100III @18,3 mm f/2,8 (50 mm equivalent) at the same speed than with my D300 @50 mm f/2,8... (and you see, I didn't convert aperture) Yes, depth of field/blur will be different. Yes, if I want the same bokeh I have to take the picture at f/7 to have the same result. Yes, it will result in the fact that I will need higher iso to keep the same speed Yes, larger sensor have better high iso performance (for the moment) Yes, larger lenses have better optical performance (and probably for long) But no, f/2,8 is not anything else than f/2,8 on different sensor sizes. ...or maybe manufacturer should also multiply the weight of camera to avoid cheating ? Why they don't? Because weight and sensor size are not mathematically linked... like aperture and sensor size. Aperture (the number written on lenses) is a relative figure. Relative Aperture = Absolute Aperture / Focal length. You evidently never need to convert absolute values. I would say it is 'scalable' (stays the same if geometry is larger). Iso sensibility quality and pixel number are technical parameters, clearly independent from focus and sharpness of a picture and should be treated independently to sensor size and aperture to clarify the discussion. Noise (related to pixel and iso technological treatment - converting light in electrical signals) and blur (optical) are completely independent. I will probably not convince you since you based all your videos on the fact that manufacturer are cheating... but, please, try the Lumix @108mm f/2,8 and your full frame @600mm f/2,8 and tell me if something else than depth of field, noise and optical quality (vignetting, optical aberration) is different. You will see that you will take the picture at the same shutter speed (because the sensor receive the same amount of light).
Anne Hodermi Re-watch the video, you don't really get what he's trying to say coz your statement somehow proves that he was right. You will never get the same result (in term of bokeh and noise) out of 2 lens(for full frame and crop) that has equivalent focal length but with same aperture on different sensor. Of coz you can have same speed, but never same image quality. That's why he said you never want to shoot sports at f16 coz you'll get a noisier image. It not just shutter speed that matters.
mr2014 Bokeh will be different because at the same relative aperture of f/2.8, the real aperture of a Full frame is larger than on smaller sensor cameras (geometrically, beams of lights are distributed differently, better spread with "large holes"; a drawing would be more meaningful) Noise has nothing to do with the "convert the aperture or not" debate. The current best small sensor have less noise than the first full frame sensor. I don't know what is your definition of image quality. Depth of field/bokeh is not the only criteria. If you have a good sensor and big aperture (f/1.8 or 2.8) you will be able to shoot at high enough speed to avoid motion blur. I don't understand why you say I need to shoot at f/16 as the camera can shoot at f/2.8 (and NO, I don't have to convert). The ONLY difference will be: - Bokeh (for geometrical reason) - Distortion (because larger lenses generally have less optical distortion) - Noise (because a (same generation) larger sensors generally have better high iso quality). As the quality of sensor is becoming incredible (iso 3200 with enough light is often acceptable), as optical are sometimes very good and as most of the people don't really care about (subtle difference of) bokeh, I can tell you that the real cheating is to convince people that they need full frame reflex. Let's do a test with two f/2.8 instead of useless theoretical diatribe.
Tony, this was a ridiculously informative video.. a real eye opener. I see all my DX 1.8, 2.8 and similar 4/3 lenses in a whole different way now. And it really explains their performance in comparison to each other. Why are the manufacturers allowed to do this? Thank you for taking the time to explain, and for doing the math. :-) I'm just a hobbyist, but I feel like a much more informed shooter. Thank you. Thomas
@@sakritone Who can asnwer please: in the same condition with FF I take a picture with 50mm at f1.8 at ISO 800 , to have the same exposure on a mft with a 25mm at f1.8 is enough 800 ISO or I have to setup 1600 ISO?
This is one of the Most necessary videos anyone interested in photography should watch..if you can't justify buying FF equipment or dnt know crop factor nature then watch this if you want to make real photos..tony this is one of your Best Videos IMO...Sharing This :)
When I bring up the topic, people look at me like I'm weird or something! :D ... and some even feel insulted by the truth and the facts! They still believe the promoted nonsense rather than facing the math! Then they keep wondering how come their 'super-bright' lenses with their mini-sensor cameras won't work in low-light or fast action! Gee!!! :D
mrgclips Yeah, I experience this, too. Re-teaching a misunderstanding is always difficult, but once people have paid their hard-earned cash based on their own misunderstanding, it becomes emotionally painful for them to admit their mistake... no matter how straightfoward it is to prove.
If you want the same depth of field, you multiply the aperture. But don't multiply aperture to get the same light, because you will just get a brighter image for a cropped sensor
Some of you may find this Wikipedia explanation helpful. Notice that both the focal length and aperture range have been converted, underlining Tony's point. Equivalent aperture range See also: Image sensor format In digital photography, the 35mm-equivalent aperture range is sometimes considered to be more important than the actual f-number. Equivalent aperture is the f-number adjusted to correspond to the f-number of the same size absolute aperture diameter on a lens with a 35mm equivalent focal length. Smaller equivalent f-numbers are expected to lead to higher image quality based on more total light from the subject, as well as lead to reduced depth of field. For example, a Sony Cyber-shot DSC-RX10 uses a 1" sensor, 24-200 mm with maximum aperture constant along the zoom range; f/2.8 has equivalent aperture range f/7.6, which is a lower equivalent f-number than some other f/2.8 cameras with smaller sensors.[20]
I agree with your calculation, although they don't apply to evaluating equivalences. A 50mm full sensor lens is still a 50mm lens lens on a cropped sensor camera, only edges are cut down to make a smaller area picture image covering a smaller, therefor the "equivalent" naming. The lens behaves exactly like like on a full frame camera including f-stops and bokeh of and 50mm lens. The reason is that It does not actually become a 75mm lens, only appears as one.
Agree : exposure doesn't change only depth of field and field of view change
6 лет назад
Laurent Outang Field of view is not gonna change unless you change focusing distance or focal lenght. 50mm has the same depth of field and amount of bokeh on any camera if youre shooting at the same aperture and distance to a subject.
Spot on Tony, the video is great as it enlightens everyone on what the real deal is. Some people will of course argue but I think it's arbitrary. I think one of the reasons marketing people from camera companies would sorta 'trick' consumers on the actual specs is because of the mainstream paradigm on two things 1.) Angle of View (AOV) and 2.) ISO. The standard which you have stressed to have stemmed from an 80yr. old standard that is clearly not applicable in the 'toy' cameras of today (e.g. phone cam to bridge-cam). Unfortunately, the common man has his mind set on these two things only, they couldn't even care less for f-stops. Having had the 135 format as the de-facto standard on photography, it's created a paradigm full of myths and fallacies (too bad for the poor man). So marketing people would rather take a short-cut than coming up with a cheat sheet of the actual specs. After all, it wouldn't make sense for ordinary people to be given a 4.4mm lens camera (1/2" frame ) when the picture looks like a 24mm AOV image (full frame). Ah, they also couldn't care less for frame size and the crop factor that comes with it. Hence these false equivalents are put out as actual specifications. Just my thoughts.
Tony, I'm a little confused by @16:07 on the one hand you show two sensors with the same settings producing an identical image (difference is depth of field). On the other hand @31:20 you talk about f/16, but not for DoF, but rather for light transmission. How can both statements be true? Wouldn't we need to bump up the ISO or increase shutter in the @16:07 example to get the same amount of light?
Never mind, answered my own question watching the Crop Factor 4 video. The ISO setting numbers are the same, however ISO 100 on MFT doesn't gather as much light as ISO 100 on FF. So if you could push a MFT sensor to 1600 ISO and equivalent FF sensor could be pushed to 6400. That explains why my APS-C cameras were really unusable past 3200 ISO, it was the equivalent of 12,800 ISO on FF. Now that I know that and since I'm very often in the 800-1600 ISO range, it seems I should be investing in FF to get rid of noise. THANK YOU TONY!! You'll probably never read this, but this video was a huge help! I was considering MFT, APS-C and FF. Now that I understand the differences, I'll only be going FF.
And one more thing. If the core idea of your video is that all manufacturers lie to us how should they advertize Canon EF lenses for example 70-200mm f2.8. It can be mounted to both EF and EF-S cameras and in your system they should start advertising it as 112-320mm F4.5 lens even though physical quality of lens itself hasn't changed. So even though current system is confusing and marketing deps take advantage of it, it's still easier to figure then the one you came up with.
Well, if we used field of view and total light gathered as metrics, you'd have to apply a crop factor *only* when using lenses designed for larger formats. All native lenses would compare directly without crop factor. Now, anyone who uses anything bigger or smaller than full frame needs to apply the crop factor to understand how lenses compare.
Tony Northrup Interesting thing happens when you are trying to match DoF from MFT to FF. You have that slide in demo but the way you speak of it is that image from FF sensor is the one we wanted to achieve originally but if the situation is opposite and the image from MFT is the target then we end up with a higher F and ISO numbers on FF camera in order to match DoF from smaller sensor. That is actually a common mistake for people who just moved from crop sensors to FF and start shooting wide open just because they did so on coped cameras, everything ends up been out of focus and they don't get why. Their DoF is razor blade thin at F1.4 - 2 on FF, yes they get more light but the situations where you need that shallowness are limited. So basically you are saying that you are sorry for those people shooting with their 12-35mm f2.8 and thinking that they are getting 24-70mm f2.8 but that's not right. It's like you are pointing at their wheelchair which they never knew existed. There's no point in thinking about other camera that you don't shoot with and that EFV line is only for those downscaling the system not for those that think they are missing out something. Marketers take advantage of that in case of those bridge cameras indeed but don't they always do that with other things in our lives? :)
That is total bullshit. A 70-200mm f/2.8 lens does not suddenly drop to f/4.5 while on an EF-S camera. It is still f/2.8. Quit lying to people! I'll prove it to you. Go get a Canon 2x teleconverter and put it on that 70-200. Notice how it still autofocuses, regardless of whether it's on a 5D or 70D? That's because it's an f/5.6 on both cameras now since the Canon 2x teleconverter causes a 2-stop decrease in illuminance. If what you were saying is true, then the 70-200 + 2x converter would not autofocus on a 70D because the F-stop would have dropped from f/4.5 down to f/9.0!! The fact is that a fast f-stop lens still has the same f-stop no matter WHAT camera you put it on. F/2.8 will still project more light per square millimeter on the sensor, no matter what kind of camera it's on. This guy is so full of shit it's not even funny and you people are just LAPPING IT UP.
darkgoob I totally agree with you. Sadly your comment won't be here long... I also showed he was wrong and he blocked my comments. He didn't even had the balls to answer me. My comment was not rude and I asked for anyone to try to prove me wrong in what I showed. Let the truth out!!
Tony Northrup I think I followed most of this video. But I'm just wondering whether you can directly compare the figures of Canon EF lenses to EF-S lenses?? For example, I have an APS-C camera with a EF-S kit lens (18-55 f3.5-5.6) and a EF 50mm f1.4. I understand that the crop factor would apply to both in terms of the focal length, so I have the 35mm equivalents of a 28-88mm lens and an 80mm lens. Does the same apply to the aperture, for depth of field calculation and bokeh?? Do I have the equivalent to a 28-88mm f5.6-8.9 and a 80mm f2.2?? Say if I were to set my EF-S kit lens to effective 80mm and use the lowest f-number, and then match this on my EF lens with the same f-number, should the pictures theoretically be identical for angle of view, exposure, bokeh, etc??
Hi there. I have a question. I just got the D600 with Tamron 24-70 f2.8 and Tokina 11-16 f2.8. Have also the D5200 crop body. My idea is to keep the crop body with a lense with bigger zoom to avoid the lense changing. I know the easiest way but the problem is that I spent almost the whole budget I got buying the full frame body with the both lenses. My question is about the cheapest option which will give me a good F number, because I'm shooting live stage performings which is almost all the time low light scene. I'm sorry for my English and Thank You very much in advance for the help.
Åke Åkesson Right! We totally agree with each other. The problem is that light per mm2 alone doesn't have any impact on your final image quality--it's total light that determines the visible image noise. That's why you use crop factor with the f/stop--to convert light/mm2 to total light and thus predict how your image will appear.
"Total light" is a meaningless and confusing term. Why do I need the same amount of light if my sensor is 4 times smaller ? I don't. The differences you observe are due to the size of the photocell. Think about 16MP on a m43 and 16MP on a full frame. Each photocell in the m43 camera is 4 times smaller than the same photocell on a 35mm camera. The smaller photocell is inherently noisier therefore it has worse performance in low light, and that's why ISO performance degrades much more rapidly with a m43 camera than with a FF one. It's also why m43 camera manufacturers work hard to improve stabilization options. They allow the photographer to expose for more time in low light, at low ISO, in an attempt to compensate for the smaller photocell.
Have you taken a look at the pictures Tony Northrup showed in his video? I think you talk nonesense without any understanding of the topic! ;( What has a good IBIS to do with ISO? Only with lower ISO (25 instead of 200 like most MFT cameras have) you can compensate for the smaller sensor (2 times square)! That's what you call "meaningless" total (amount of) light! :D If you want the same depth of field you'll need f0.6, f0.7 or f0.9 (for the full frame lenses of f1.2, f1.4 or f1.8)! That's what you can take out of Tony's explanations. And what MFT camera manufactures (Olympus and Panasonic) should do! ;) IBIS just helps you to take pictures at a lower shutter speed/aperture without bluring it! IBIS isn't exclusively on MFT cameras (though Olympus at the moment seems to have the best IBIS)! xD
I never said you can get the same depth of field between a m43 and ff camera using the same f-stop, because it can't be done. That's clear and proven. I never argued the results presented are wrong or incorrect. I only argued that the mechanism used for explaining them is not meaningfull. Assume for a second you are attempting to take a handheld picture. Your settings are: f/4, ISO1600, 1/40 exposure time Your camera tells you the picture is 1 step underexposed. Your options are as follows: Increase the aperture or Increase the ISO or Increase the exposure time. On m43, the stabilization (body, lens or both) allows you to increase the exposure time when shooting handheld without getting a smudged picture. There are limits to the technology but it helps. On a ff maybe you would opt to increase the ISO, since the pictures look good at ISO3200 also. On m43 you will want to avoid increasing the ISO since the 4 times smaller photocell in the m43 sensor is much noisier than the photocell in a ff camera. This alone was my point. The observed ISO related results are not due to some "total light" value, it's strictly due to photocell size. The 5DS (50MP) has native ISO limited to 6400 for this reason exactly. The smaller photocell size behaves worse in low light than other ff cameras with smaller number of pixels and therefore larger photocell sizes As mentioned in a dpreview comment: Physics here is very simple. One pixel does not care how much light the whole sensor gets. All it cares about is what it personally gets to deal with its own signal/noise ratio." Cheers.
What happens when i use an adaptor to put a helios 44 on my MFT? It's got more glass so is it actually f2 or f4? And why does the measurements of distance for focal length seem accurate?
Hello Tony. I always enjoy your videos...but also tend to have comments on your methodology as well;) First and foremost, I truly admire the work both you and Chelsea put into your videos. This video is a perfect example and a great video for explaining differences in depth of field between sensor sizes using the same aperture value and the common fallacy in aperture/focal length equivalency between the differing formats and the amount of light gathered by sensors of differing formats. Having said that, there is one major omission which has caused many to doubt your results. I believe it is important to state that your conclusions hold true for sensors of a similar generation with similar pixel size and density. Mathematically you are spot on, however, without this additional information, you end up with paradoxical results such as the e-m1 having similar noise performance to the a6000 and the A7s exceptional high ISO performance vs the A7R. I believe that such explanations would go a long way toward addressing some of the negative comments. in practical terms, contemporary cameras of different sensor sizes should have the same exposure triangle for any given scene, however, they will have different depth of field for the same aperture value and potentially different noise performance depending on pixel size and density. Notwithstanding, this is still consistent with your mathematical explanations of the equivalence of total light based on proportional ISO values for the given aperture and focal length. I agree with you 100%! Great job explaining a very convoluted subject!!!
You're wrong. Light per unit detector (same as unit area) is all that matters (and maybe electronic noise to some extent) if you're interested in the amount of noise. ISO is essentially gain,. It defines sensor performance just as well as film sensitivity. As for DOF and aperture setting i can't quite understand you, so not sure if there's something to contradict. True 2.8 is exactly the same on any true 50 mm lens, Your sensor size has no influence and you can't recalculate it, not the aperture, not the focal length. Same for DOF. The difference arises when you recompose (change distance or focal length) to accomodate the smaller sensor and get the same framing. You're right about not being able to adjust focal length for crop factor, you can't adjust it FOR ANYTHING. It's a physical parameter of an optical system and has nothing to do with the sensor behind it. Really really misleading stuff here, you need to learn some physics yourself. And change your audio settings/mic, it's horrible. I agree with your position though, they're assholes for giving the "equivalent" instead of the real focal length.
I have some questions. If I were shooting a full-frame and an aps-c at ISO100, the total amount of light gathered by the full-frame will be about 2.7x (0.5 stop from f-stop + 1.25 stop from bigger sensor) more than the aps-c's at a set time interval. The 0.5 stop can be eliminated if I were to use a DX lense on an aps-c but the sensor size advantage is still present, which means at a set time full-frame still captures 2.25x more light. Gee, no wonder full frames are so good at low light. But if I were to use a sensor-stabilized apc (like Pentax) with FX prime lenses, the stabilization can help slowing down 3 stops of shutter speed without causing severe motion blur. It partially solves the problem but I still cannot capture freeze motion in low-light with high ISO like a full-frame could. Am I correct?
You are right about ISO, but every generation sensors seem to always be getting better and better. I do not know if there is a limit to how far they can keep pushing sensors. But I would bet a Panasonic Lumix GH4 has better ISO the a Canon 5D Mark 2. So I agree with you about ISO and light but within the same generations of cameras. Also mega pixels play a big part how much light the sensor can get.
I have a 5DmkII and a GH4, the 5DmkII is significantly better at high ISO both for having more Dynamic Range and less Noise. It's a bit over a stop I think, DXO have it closer to a 0.8 stop: www.dxomark.com/Cameras/Compare/Side-by-side/Panasonic-Lumix-DMC-GH4-versus-Panasonic-Lumix-DMC-GH3-versus-Canon-EOS-5D-Mark-II___943_842_483 Go to measurements tab then SNR.
Yeah I guess its going to take GH5 to maybe have better ISO then the Canon 5D Mark 2. When I said Canon 5D Mark 2 I was assuming by now the Panasonic just releasing the GH4 would have better ISO. Guess I was wrong. After seeing how good the Canon 5D Mark 3 is in low light I just guessed the GH4 would be better then the Canon 5D Mark 2.
mp4podcastDOTcom A camera with a smaller sensor will never equally match a camera with a larger sensor in terms of ISO performance. Due to having a smaller surface area a MFT simply cannot absorb the same amount of light as a full frame sensor. On the other side of this, a medium format sensor (larger than full frame) will have better ISO performance than a full frame camera because it will gather more light.
The Canon 6D looks like their best low-light camera (well, plus the 1Dx/1Dc), having a newer generation sensor than the 5DmkIII and is also a fair bit cheaper. I suspect Panasonic will need something in the way of new sensor technology to catch up with the 5DmkII, the 4x size difference is a lot and the 5DmkII sensor has really pretty low read noise for big pixels (3.2 electrons on 64,600 electron capacity pixels).
Zackary Deal You can catch up a fair bit. Can I get a bit mathematical? The 5DmkII converts 33% of the incoming photons to electrons, the pixels can hold 64,600 electrons and have 3.2 electrons read noise. Let's say Panasonic get to 66% efficiency on a 16MP GH5 sensor, which is within the realms as most are over 50% now, plus they can clearly get to lower noise as their GM-1 sensor has 1.7 electron read noise, so not impossible to head towards 5DmkII performance. I think they can get close. However not close to the then-current FF tech of course, just the old stuff, which the OP was talking about.
Hi Tony, after agreeing to most of the points you have made about actual ISO, Aperture & Focal lengths I have a question, what happens to the f-stop number when I use a Nikon 50 mm f/1.8 lens on my Nikon 1 system using the glassless FT1 adaptor, multiply 1.8 by 2.7x, then what about the T-stop of the lens? is that to be also multiplied by the crop factor??? a little help needed !
Tony, does this also apply when putting a full-frame camera into "crop mode?" Does the then cropped image suffer a loss in effective aperture the same way as if it were on a cropped body to begin with?
I can't agree with the idea about total light. Examples in the video makes sense but when you bring sensors with *same area but different resolution*, the story changes. According to the video, because D800 receives the same amount of light as a D4, they should have similar ISO performance. But the D4 is cleaner. This brings us to the point that *per unit area light intensity* is actually what governs noise. D4 has less, but larger pixels so at a given time they can gather more light than each pixel on D800. So they don't have to increase their amplifier gain as much as the D800 to get similar exposure. That results in cleaner images. You could argue that *re-sizing the image* and averaging out the noise will give them similar SNR. And that's true but that defeats the purpose of high resolution and added detail altogether! Anyway, no disrespect intended towards anyone. I like this channel but in this video I had a different opinion so I had to share my views.
ADZCreations Watch the Crop Factor Part 3 video for a discussion on pixel density. There's no overall correlation between noise and pixel density; many cameras have more of one and less of the other.
I am a physicist and I've been telling friends and people arround me exactly the same thing for years. BUT: They often just don't really understand it. They solely look on the label on their camera and believe, what the manufacturer is telling them. Thank you so much for sharing this video. This is just basic education, and if somebody won't believe me in future, I will refer to your video! It's very good!
If you are a physicist, you probably know that the aperture figure written on lenses is a relative quantity (relative aperture = real aperture / focal length). What manufacturers say is correct.
Anne Hodermi The (photographic) aperture of a lens is defined as N=(focal length)/(lens diameter). And I think you really missunderstood the video. Tony is requesting a better and more relevant measurement to compare different camera systems. A 2.8 lens gathers just more light on a full frame sensor, than on a 4/3 sensor, because N=(focal length)/(lens diameter) stays the same. I think the aperture is used for comparison, but it's not meant to compare if you have different sensor sizes. It is true, that the number stays the same, but that is just not relevant for photography. And that is what T.N. says in his video.
pratakk No, a large sensor does not gather more light (per surface) than a smaller one. And that's why the conversion is irrelevant. Normally large sensor (from a same generation) gathers light better (less electronic noise) but not more light. Full frame lenses are generally larger and have also better optical specification (less distortion). These are the only reasons why FF camera give better result than smaller sensor.
Anne Hodermi Think about it again with a very simple approach: a 2.8 lens stays a 2.8 lens. We already discussed that above. (For example 200mm, f2.8 means an aperture opening of 71mm). But what's the difference between the 2.8 lens for full format and micro 4.3? Well, that's very easy as well: - to cover the full frame sensor completely with light, you'll need a big front glass element - to cover a smaller sensor completely with light, you can use a smaller front glass element. That's why a f2.8 lens for micro 4/3 can be smaller. If you compare these two lenses, the only difference is the light collecting ability. The bigger f2.8 full frame lens gathers just more light through the front glass element and the total amount of light gathered is bigger than on the 4/3 sensor. Another example: You can see, what happens if you put a big 2.8 lens for full format on a aps-c sized sensor here: upload.wikimedia.org/wikipedia/commons/f/f5/Full-frame_vs_APS-C.svg What do you think happens to the light, that doesn't hit the smaller sensor - well, it is lost. And the bigger sensor gatheres this light.
pratakk The light that doesn't hit the small sensor is not 'lost' because the small sensor just don't need it (why should the sensor need the light that is outside it). The intensity of light on the small(er) surface is the same as on the larger one (of course, because it is the 'same light' - see your drawing). And it is intensity that matters to have pixel transforming effectively into quality electronic signal. Angle of view are different but amount of light is exactly the same. When you put a lens made for APS-C on Full frame, the same amount (=quantity per surface) and quality of light goes through the lens to hit the center of the sensor as if you put the lens on a APC-S camera (because the optical quality of the lens obviously doesn't change because you change the body). An iphone has a 4,89 x 3,67 mm sensor (factor 7.3). It is a f/2.2. Applying a conversion would lead to what? f/16? If you ever used an iphone, you can see that you don't need to go to very low speed or very high iso to take normal picture. Why? Because the same intensity of light hits the sensor (as for a "FF f/2.2").
One part I don't understand is why you had to change the ISO on the FF to 3200 at @17:35. If the FF equivalent of f/2.8 is f/5.6 on the crop sensor, why don't they give the same luminance at f/2.8 and f/5.6?
Hello Tony, thanks for this clear video. However, what happens when I mount a FF lens on a crop camera, e.g. the Canon "nifty fifty" f/1.8 on a 7D. Focal length will be 80mmFF but how about the DoF and total light gathered? Can you - or anyone else - shed (gathered) light into this? Thanks
oh man i feel so cheated! Wish i saw this a month ago before i ordered my Sony G 18-105 f4, for my A6300.. I knew about the Focal length crop, but the aperture? I'm pretty much a newbie, so i'm getting an F6 instead of an F4!! Damn, Thanks for the heads up..
Cool. Thanks for the info. I'm a bit more researched now so my initial comment was probably not completely accurate. I can also move further away and zoom in to get the same effect as compared to full frame, but i think "compared to full frame" is a mindset i should just eliminate from my thinking.
Yeah you totally should eliminate that thinking. Good photography does not depend on the sensor size in any way. And if you had a fullframe camera, would you then think about how it compares to medium format?
The maximum blur you can get with that is 105 mm / 4 = 26.25 mm i.e about one inch. This is measured on the focal plane. A thumb is one inch wide so if you take a photo of someone with his thumb up and there is light on the background far away the light will show same size as the thumb. Note that the distance does not matter.
I get the impression that there are 3 ranges of zoom lenses for all the manufacturers. Maybe it's a matter of common designs, maybe just marketing convention: * Variable aperture (say 3.5 - 5.6) seem to be the kit/beginner lenses. * Fixed aperture f/4 are enthusiast lenses, more affordable and more compact than the pro lenses * Fixed aperture f/2.8 are the pro lenses. On larger format sensors these tend to be huge. Optical quality is really very good. I use Micro 4/3, which has a "crop factor" of 2. So my 12-40 f/2.8 zoom could be said to be like owning a 24-80 f/5.6 on a full frame system. Sony sell a similar range at f/4 which is a stop faster than my equivalent and a similar size for a similar price. (On a 1.6 crop sensor camera the equivalent to my f/2.8 would be f/3.5) The Olympus f/2.8 is a very nice lens though, and a full frame sensor does not eliminate noise. You can download sample RAW files from the web and explore this. The PRO quality lenses perform very well wide open, certainly sharp enough for my current sensor to resolve. So quite often I can shoot at f/2.8 (or f/1.8 on my primes) where someone else may be stopping down to the equivalent or smaller. The large sensor camera can achieve more, especially when allowed to operate at its base ISO due to sufficient light for the desired shutter speed. (Roll on summer!). The full frame pro lenses are great and there are lovely photos out there taken on a D810 with bright light contrasty skin tones I couldn't hope for. But I don't have £10,000 to spend on a camera and don't want something that won't fit in the rucksack even with the other hiking gear removed.
Tony Northrup Thank you for starting a debate & educating people on issues about manufacturers cheating consumers, on the other side, I think you are stressing too much to prove your point which is half true, the DOF is dependent on the sensor size, so multiplying the f number on a lens with crop factor is correct however, the light intensity projected by a lens is not dependent on the size of the sensor, if you use a FF lens on a FF sensor the whole projection is used & if you use the same lens on 1.5x crop smaller portion of the projected light is used by the sensor but the intensity of light that falls on the crop sensor is still the same as what it would be on a FF sensor's equivalent area ......... intensity of light projected from a f/2.8 is always f/2.8 on any format sensor as long as the lens is used on a sensor that is equal to or smaller than the lens was originally designed for............. about the inconsistency in ISO numbers between different sensor sizes that is an issue about quality & standards, if there is no regulation in the industry then consumers are going to be cheated...... I think you should stress more on pointing out the fact that manufacturers are using different standard ISO on different format/size sensors and cheating consumers! please let me know if I am wrong..............
Purushottam Pokharel yes light intensity for a sq inch is the same across diff sensor sizes (both with lens at same fstop. E.g. f2.8) what Tony is saying is the TOTAL light gathered over whole sensor isn't factored in. This is where DOF and noise levels on smaller sensor cameras is misrepresented. And mis understood.
for a wildlife situation, full frame sensor needs 1.5x more focal length to get the same frame/resolution as a crop sensor would, on the other hand a crop sensor would need an extra stop of aperture opening to get same level of DOF & noise benefits of a FF sensor ??? So eventually one needs to invest more money for focal length if they have FF sensor & in case of not so fast but long lens one needs to invest more money for the FF camera body All of Tony's Effort can't help save money for us :( (bird photographers)
if a narrow FOV is your main concern/priority, then yes, sure, a crop sensor is better (and cheaper)for you. (as long as you don't mind the extra noise of higher ISO using same lens on a crop S body.) but Tony is talking about the same DOF and light gathering ability/noise levels. Those things are very important to a great many photographers. I just finished a video that goes into the science of this subject, along with plenty of illustrations showing exactly what is happening inside the lens... maybe check it out if you are still uncertain about anything. (it explains the same facts, but goes into a bit more detail and demonstrates the principles with illustrations.... also i go into speedsters briefly too)
+Tony Northrup OK I see by working backwards, you mean taking the value of the original lens' highest focal length and the advertised f-stop (aperture) and dividing to get the divisor i.e. Panasonic 12-35mm f/2.8, will be 35/2.8 = 12.5, and this then becomes the divisor used to check any other claim they may be making. Thanks.
Mr. Tony Northrup, i wonder... i have A6000 and sony SEL 50mm f1.8. if someday i buy A7 and put my SEL 50mm f1.8 lens in there and using APS mode(APSC crop) in A7... ist the quality bokeh blur same as if i use fullframe lens 85mm? or not? thanks
All this is true, but you have to decide what is relevant - for YOU. If you're moving from one camera system to another or using 2 systems at once, then all this is essential knowledge and will ensure that you get the best results. However, if you're only shooting with one camera system, you only need to know how your own system perfroms with the ISO and aperture values that system employs - getting to grips with that alone will get you the best results. Stop comparing your one system to another - it's irrelevant.
Baba Wethu If you want to be technical, DoF is a factor of aperture and magnification (the relative size of the subject on the camera's sensor). Focal length and sensor size are factors in magnification, however.
Tony Northrup Wrong. Magnification factor does not depend on sensor size. Depth of field depends on aperture and magnification (magnification depends on focal length and distance between camera and object). Please stop telling nonsense to people.
Caldiar23 Strange that you called him wrong when you're both effectively saying the same thing. The distance between the camera and object is indirectly affected by the sensor size, because if you wish to keep someone's head the same size on an image taken with an APSC camera and a FF camera with the same MP, you would increase the distance between camera and object with APSC.
cheng2006 I'm not saying the same. Magnification is not relative to sensor size. It has absolutely nothing to do with sensor size. AGAIN: Depth of field only depends on aperture and magnification ratio. Tony Northrup How about checking the lens of the Panasonic FZ200 you mention. You will notice the correct numbers are printed on the lens. It's sad so many people believe this nonsense. As I said before, if you don't know how to use focal length then use angle of view. Focal length and numeric aperture can't be converted into anything. AND: You can't use different magnification ratios and same aperture between two shots and then claim the aperture not being correctly stated.
very comprehensive. Thank you Tony, But I have a question, what will happen to the BOKEH if I use an apsc lens on a fullframe body. will the effects be multiplied as well or it will remain shallower because its full frame anyway?
This has to be one of the longest videos that keeps you entertained and informed all the way through. While some comments may bring up technical differences between sensors and megapixel counts, etc. I think the main idea is still totally valid and gives a good idea of what to expect of each sensor type and their respective optics. Thank you!
So, it took us more than 10 years since the appearance of the first APS-C sensor cameras and lenses for someone to come front and explain all the "sex, lies and video tapes". Thank you Tony for making it all clear, lucid and obvious. While we all knew about the field of view change with crop factor sensors from the day one, most of us did not pay attention to the whole equation and all the variables: aperture and ISO. With your help, Tony, it should be understood by anyone interested enough. In order to make things more convenient and for marketing purposes, of course, manufacturers retained the same measurement units when talking about different physical properties. And that caused misinterpretation that lasted until this video was published. Let's be fair, though, and admit that that Panasonic Micro Four Thirds 12-35mm f2.8 lens is indeed f2.8 lens. Just not in Full Frame measurement units, but in Micro Four Thirds measurement units. And that's what makes all the difference. Numerical values are the same but physical implications and light gathering abilities are not. Not without applying the proper math.
@Tony & Chelsea i recently bought a 70-300 4-5.6 full frame lens for my D7200 and am not getting sharp images. i use the option to censor crop an additional 1.3 so that makes my lens a 136.5mm to 585mm f/7.8-10.92. i wanna shoot wild life and am having to shoot at 1/1600sec ISO-1000 F/11 to get a proper exposure should i return the lens and just get the 70-300 nikon dx 4.5-5.6D vr2. or could the softness be the lens or my technique. plz help because this is very frustrating.
Hey Tony, This video is been a good eye opener. No pun intended :) As for camera manufacturers being "evil" and trying to cheat people well they have to sell, don't they? The really good take-away for me at least as a newcomer to DSLR is that if you're stuck with a particular gear setup it sure helps to know its actual range. I personally am a Canon user, there's something about them making their own lenses that I always appreciated and their entry level kits are affordable and easy to use, but from what you explained in this video the same principles, actually no-brainers once you go through the whole explanation, make a whole lot of sense. Even if you got stuck with a pair of lenses and a camera body that's not full-frame with this info you can still adjust it to get the most of it. At least that's my two cents. It will sure help from now to remember that whatever's on the lens (f-stop numbers and focal lengths) to multiply by the unusual 1.6x crop factor that Canon seems to love. I knew ISO is the other piece to play with but now that I know to throw in the crop factor it sure makes sense to get grain-free pics. Also I wanted to say the videos you post with various tutorials are really awesome, keep it up!! ;)
+MuertoInc this is all correct, and 100% consistent with the video. If you think it's inconsistent, you've missed the point. Here's why people need to understand crop factor. 2 lbs of rocks *is* heavier than 2 kg of rocks. 100 mph *is* faster than 100 kph. If you don't use crop factor, you're comparing numbers in different units, and that's misleading.
+Tony Northrup Another way of saying this would be "If you took 2 pictures using a full-frame lens on an APS-C body and a full-frame body, you will not get the same picture if you used the same settings on the two different camera bodies." Correct?
+MuertoInc 2.8 is 2.8 on any camera, but depth of field, the thing that we notice, is different when sensor size is changed. But because depth of field is dependent on aperture, sensor size and focal length, it's hard to think of and use. So it's easier to use 'f/2.8 on a 35mm sensor" as a reference and then we can say that f/2.8 gives the same depth of field on a 35mm sensor as an f/1.9 lens on a 1.5x crop sensor and f/1.4 on a 2x crop sensor.
+MuertoInc the point is, that a small sensor can only make use of a small amount of the light, which is collected by the lens. Because this amount of the light is simply not captured by the sensor - it falls next to the sensor, because it is too small to cover the whole pool of light.
Hi, I love your analysis. I have a problem. I am an eye surgeon and I record surgeries on Sony a6300 at 4k resolution. I need to crop the video by about 70% so that I can show only the central part or just the eye. My question is will a smaller sensor like MFT of Panasonic GH5 (which I intend to buy) instead of APS-C of Sony A6300 at 4 k will have lower signal noise ratio at same ISO considering that MFT sensor image will suffer lower cropping of video instead of the APS-C video.
So if I have a 1.5 crop sensor stated 300mm at aperture 5.6 and iso 1000, would the actual settings be 450mm at 8.4 and iso 1000, meaning I could put a 450 lens on a full frame at aperture 8.4 with ISO 1000 and have the same exposure, this would suggest that for the same focal length, a full frame is no better at gathering light through the lens than a crop sensor, maybe because the same amount of light is exposed to the sensor at a given "actual" focal length in a full frame or crop? If this is true then the only real advantage of the full frame at equivalent zoom, as far as low light is concerned, is its ability to cope with the higher gain of ISO, meaning even though it is the same light received onto the sensor as the crop at a given actual focal length, it is the ability to add and cope with higher ISO that would actually make it better at low light? I am new to photography so I may be getting this all wrong?
Things are not as different now from film as you would have us believe. 35mm film is grainier than 8x10, right? Wrong! Contact print them both and they'll have the same amount of grain. Likewise noise in digital images (all other things being equal). If you "enlarge" images from different sensor sizes the same percent (assuming they have the same image-processing system) they'll have the same amount of noise. Smaller sensors are noisier at the same ISO only because the image is enlarged more. It's not a matter of ISO but of magnification. - Depth of field is a function of four things: subject distance, aperture, focal length, and image magnification. Forget this nonsense about "35mm equivalent". That is meaningless with regard to depth of field, and so is ISO. - I'm not a believer in the idea that we need to shoot film to understand photography, but we do need to understand something about optics, and that is getting lost in the "equivalence" numbers and in ignoring magnification.
The silver halide crystals of a film of 35mm and other of 6x6 of 100ISO have the same size and give the same grain when making contacts; true. The photodiodes (pixels) of a 20Mp FF (24x36) and another 1" sensor do not have the same size (surface). If you compare the image of an FF and a 1" of 20Mp without magnification, the 1" has more noise because the signal /noise ratio of the photodiode is proportional to its surface. The noise is due to the "amplification" of the signal, not mainly the enlargement of the format. What you say would be true if we compare sensors of different formats but proportional resolutions to maintain the same surface of the photodiodes.
Dear Tony, Your conclusions are mathematically correct but the logic behind them is wrong. The ISO standard was devised to provide for equal exposure, not equal picture quality. And yes, 400 ISO film is "noisier" (with larger grains, in fact) than 100 ISO film - in close similarity with the way that digital sensors behave. 8 mm film is much grainier than 36 mm film when viewed on the same screen. All this has also to do with the magnification of the viewed image - a factor that very few tests take into account. Your definition of the term "total gathered light" is useless, no matter how sweet you try to talk us into it. When it comes to noise what really matters is the size of the photo sites (pixels) on the sensor, not the size of the sensor itself. Your "buckets filled by the rain" are relevant to the electronic wells of the pixels, not the entire sensors. It is true that larger sensors tend to have bigger pixels and hence gather more light / produce less noise. It is also true, however, that a 10 MP APS-C sensor will produce less noise than a 36 MP Full Frame sensor (when using the same technology). You are right that equivalent focal lengths have misled quite a few people. The cited apertures, however, express the lenses' ability to transmit light, not the quality of the bokeh or the amount of DoF they produce. As I mentioned above, ISO, shutter speed and aperture values are only about exposure. You can change the DoF by changing the aperture but that's not the only factor that DoF depends on - you'll have to throw in the diameter of the image circle of the lens, the diameter of the circle of confusion (which depends on the size of the pixels on the sensor, not the size of the sensor itself), the magnification of the viewed image, etc. All this are very complex matters that cannot be expressed with a single figure and least of all - with the value of the aperture. It is true that lenses with wider maximum apertures are bigger than those with smaller ones - provided that the lenses produce the same image circle. Things get much more complex when comparing lenses which produce image circles of different diameters - i.e. a 4/3 lens and a FF lens. The size of the lens also depends on its construction (number of optical elements, presence of auto focus (AF), type of AF motor, etc.) and it's not fair to say that a 4/3 F2.8 lens will be the same size as a FF (Full Frame) F5.6 lens. By the way, the ISO conversion tables that you so generously provide are true only if the different sensors (FF, APS-C. 4/3, etc.) have the same pixel density (i.e. they are all 16 MP sensors, for example) and use the same technology (CCD, CMOS, CMOS-BSI, etc.). By the way, again, the Metabones Speed Boosters reduce the image circle of the attached lens and thus decrease the aperture values.Which means that you aperture formula is incomplete. Take care! Chris
Interesting mental exercise, but very little real life value. Fact is, for most people that take pictures with their full frame camera, they don't crank up the ISO to 800 just so it will be equivalent to ISO 100 on their CX camera. On a full frame camera if I have sufficient light to do so, I will take a picture at ISO 100, and adjusting my APS-C camera to ISO 40 to get the equivalent shot is simply not an option. If for instance I have a Nikon D610 (full frame camera), and a D7100 (APS-C Crop Frame camera) they both have a base ISO of 100. The cleanest shots I'll be able to take with either camera is at ISO 100. According to your own calculations, the only way I can get shots that are equivalent is to use ISO 40 on the D7100....which it's not capable of doing, so for all real life intents and purposes the APS-C sensor is indeed noisier. (Not really a myth at all) The other thing is, in terms of the physics of noise, it's actually based on how many photons each individual photocell (pixel) can gather, not just the size of the sensor. That's why you can take two sensors that are exactly the same size (such as a Nikon D3s and a Nikon D3x), both of which are full frame 35mm equivalent size sensors... and the difference in noise between them at the same ISO setting is huge. Because the pixel pitch on the D3x is only 5.9 Microns (barely larger than some APS-C cameras) and the pixel pitch on the D3s is 8.4 Microns. So, while your method of equalizing noise by using the crop factor gives a decent approximation, it doesn't truly describe the actual physics of what's happening nor does it account for the noise differences in cameras with the exact same form factor. Similarly, calculating Depth of Field by multiplying the crop factor once again gives a close enough approximation to work for most people, but also does not accurately account for what's really happening. The calculations for Depth of Field are based solely on 4 different variables: 1.) Focal Length of the lens 2.) Aperture used 3.) Distance between the subject and the focal plane of the camera 4.) The Circle of Confusion for the particular sensor being used. Item #4 is the only variable that is affected by sensor size and crop factor....however, it is the most insignificant part of the equation. Changing the Circle of Confusion numbers from Full Frame to APS-C makes a very very small difference. As a matter of fact, the APS-C actually has a slightly narrower Depth of Field for the exact same Focal Length, Distance and Aperture. For instance using a commonly available Depth of Field Calculator I can punch in the numbers for myself and see the Depth of field at 50mm f/2.8 from 5 feet away on a Nikon D7100 (APS-C crop sensor camera) compared to a D700 (Full Frame 35mm equivalent camera: Same exact settings: 50mm f/2.8 from 5 feet: D7100 = 4.01 inches D700 = 6.02 inches However, because of the crop factor they will not give the equivalent field of view. To get the equivalent field of view I'd have to use a 75mm lens on the D700, so using the same calculator I get: Same field of view settings: D7100 @ 50mm f/2.8 from 5 feet = 4.01 inches D700 @ 75mm f/2.8 from 5 feet = 2.62 inches So, even though the smaller form factor because of it's smaller circle of confusion actually has slightly less Depth of Field.....because the Focal Length is a much bigger factor in the overall equation, to get equivalent fields of view the larger form factor produces the smaller Depth of Field.
Maranatha Photography It's had deep real-world value to us. We use all MFT cameras for our video work. Being familiar with FF results, these simple formulas allowed me to purchase lenses that would give me the noise and depth-of-field that I needed. Some of your concerns, such as pixel density, are addressed here: ruclips.net/video/6Im4W_9blhY/видео.html. I have a new video coming out in the next few days that digs deeper into pixel density, with actual statistics, but my initial findings as discussed in that video were correct. Re: using lower ISOs, that's addressed here to an extent: ruclips.net/video/aJFgDEJTIiI/видео.html Note that "depth of field" has both a technical and real-world meaning. The technical term is what's calculated, but the real-world meaning is more, "how blurry is the background." But to be technical, DoF is determined by only two things: the magnification of the subject and the aperture.
Maranatha Photography Oh, and I'll happily update my videos to correct any mistakes, but I need experimental evidence with actual photographic results proving that various aspects don't matter. You can create a video yourself, or if you don't have the resources, submit a detailed experiment for us to perform at sdp.io/mistake. We'll perform the experiment and make the results public. I've made that offer to thousands of people, and so far, nobody has managed to prove the math wrong.
Tony Northrup First off, thanks for taking the time to respond. Fair enough.......My statement that it has very little real world value is an over-simplification and an over-statement of my case, as you've obviously demonstrated the real world value to you. Also....Just to be clear, I was talking about Pixel Pitch. Pixel Pitch and Pixel Density while closely related and directly correlated are not the same thing. Yes, the number of pixels crammed onto a single sensor does indeed dictate how large each individual pixel can be.....it is indeed the actual size of the individual pixels that determines the full well capacity and exactly how many photons can be gathered (which directly determines the Signal to Noise ratio) The example I already gave is something nobody can/will refute. The D3s and D3x both have the exact same sized sensor but they both have very different noise/ISO characteristics. That alone is all the evidence needed to show that a simple calculation using crop factor can't account for noise characteristics at certain ISOs. I will watch the video you linked to about pixel density before commenting too much deeper about that topic. It's possible we are saying the same thing but using different terminology, so it's not fair to comment until seeing it. As far as the Depth of Field. You are of course right that it boils down to two things, the magnification of the subject and the aperture. However, there are several variables involved in describing the magnification of the subject, including the focal length the size of the circle of confusion, and the distance from the subject. While your method does provide a reasonable approximation for most applications, it is an over-simplification and doesn't really accurately represent the relationship. I will accept your challenge. I don't have the resources to put together quite as nice a video as you do, but I will indeed put together the detailed experiment and the mathematical evidence to prove it. Thanks again for taking your time.
Rich Nicely Also it appears that when I logged in to youtube to respond, that for some reason youtube logged me in with my personal profile "Rich Nicely" rather than my business profile that I used for my original comment "Maranatha Photography".....there isn't any kind of conspiracy to hide my true identity or anything.....just wasn't paying close enough attention to what RUclips account I was logged in as
Tony Northrup Actually, I think I can show this by example without ever having to submit any kind of video or photographic evidence. First off, can we at least agree that Depth of Field calculators that are based on formulas that physicists and optics researchers have agreed on for decades before either of us was ever born, are accurate? If not, and we are dealing with flat earth type of stuff, where I have to prove basic physics, then I'm not sure it's worth the time. If we are at least in agreement at this point..... Here is a calculator that I've seen many prominent photographers use and recommend that's based on those same exact formulas that physicists have agreed on forever: www "." dofmaster "." com/dofjs "." html If you use that calculator it only takes a few examples to show what I'm talking about. Let's say I have a Nikon D700 (Full Frame Camera) and a Panasonic Lumix DMC-GX1 . (MFT....Like you said in the video, the 2X crop factor makes the math easier) If I use my GX1 at 100mm and f/2.8 from 5 feet away then the equivalent field of view on my D700 would be 200mm from 5 feet away. According to the online calculator (and your welcome to use someone else's calculator if you think it's more accurate) The GX1 at 100mm and f/2.8 from 5 feet yields a DoF of .73 inches According to your method, I should get the same DoF with the D700 at 200mm f/5.6 from 5 feet away. Let's see what the calculator says: The D700 at 200mm f/5.6 from 5 feet yields a DoF of .67 inches. OK, you say, that's only a difference of .06 inches. But that's just one example. Let's say you are playing with wider lenses at closer distances The GX1 at 10mm f/2.8 from 10 inches yields a DoF of 2.09 inches The D700 at 20mm f/5.6 from 10 inches yields a DoF of 2.01 inches .08 inches is a little more significant in this range. So, while it's a close enough approximation for most purposes, it's not entirely accurate. Crop factor affects Depth of Field simply because Crop factor forces you to use a shorter focal length on the same camera to get the same Field of View....and focal length directly affects Depth of Field. The fact that multiplying the crop factor by the aperture gets you a very close approximation for purposes of Depth of Field is a happy coincidence.
Hi Tony!! I´m a new follower. Your videos are great! the image, the sound, and the most important, THE CONTENT! I have a question. What about the old film lenses on new APS-C bodies? Is all the same? Example: I have an old Ricoh XR 50mm 1.7 (great lens) on a Pentax K3II (APS-C). I know that the focal length on this body is 75mm but is the F still at 1.7? When you say that the depth of field is larger on APS-C is for APS-C lenses or for all lenses? And the light transmision, is lower on APS-C bodies with film lenses or still the same? Sorry if you talk about it in the video but I dont understand too well english and is not easy to me undertand at all. Thank you!!
Does full trames that have a croo factor option, like nikons dx mode, have the same issue in crop mode? do you still need to apply that factor to those settings?
📚 Buy Our Books on Amazon! 📚
📕Stunning Digital Photography: help.tc/s
📘Lightroom 6 Book: help.tc/l
📙Photoshop Book: help.tc/p
📗Buying Guide: help.tc/b
Your videos are amazingly informative.. on another level
Hi Tony, my question is what will be the aperture at. 18mm focal length say for example in a 18-35mm f/1.8 ? i understood that it will be 2.7 at 35mm but what about the aperture value at 18mm. ?
mmj Since the lense can be at f1.8 at any given focal length, this means it will behave like an f2.7 lense, when mounted on a 1.5 crop body, regardless of the focal length
The industry seem to base everything off the established 35mm slr days. So if you take a full frame sensor (35mm) and compare it to a crop sensor then the later is going to be noisier, all other things being equal. What is being demonstrated here is that a crop sensor can deliver an image of equal quality if the crop factor is used correctly with different hardware/settings.
Are companies responsible for our ignorance when they sell a lens as 12 - 35mm f/2.8 when its bought for for us with a crop sensor and is therefore equivalent to a 24 - 70mm f/5.6? I think not if the lens is sold as for a full frame body and yes they are being fraudulent if the lens is being sold as for a crop body, totally misleading customers.
I bought them
Wow Tony! This is my first RUclips comment, and for good reason. I'm absolutely blown away by this video. I just performed a test, as I had to see this for myself. I shot the same scene with two different setups. 1) Canon 6d with a Tamron 24-70 2.8 vc zoomed to 70mm at ISO 4000 @ f4.5 2) Canon sl1 with a Tamron 17-50 2.8 zoomed to about 43mm at ISO 1600 @ f2.8. Both pictures look absolutely indistinguishable from one another. Bokeh and noise are practically identical! I can't believe I never thought of this before. You've changed my life man. Now I better understand the limitations of all of my gear and will govern their use accordingly.
Thanks, Charles!
Please correct me when I'm wrong. When it comes to the exposure itself, crop factor does not matter. On every camera the settings of ISO, SS, Aperture will be the same for the same subject under the same light (except when the crop factor crops out area's that are lighter or darker than the average).
So a F1.2 lens on a 2x crop is still a fast lens, in the way that it allows you to use a higher shutter speed at the same ISO. However, when it comes to depth of field it will not act the same way. And image quality of course is not the same
You are NOT wrong ! No need for correction.
You are right! And Tony is trying to turn everything upside down.
Exactly! What Tony talks about applies only to the depth of field. He probably should've mentioned that in this video. It has really created some confusion throughout the industry lol
@@fuzzy_pixels wait, crop factor affect aperture, thus it affect exposure too, right?
@@cikenot90 no, your settings don't change, I've tried it
Hello Tony in the video at the 21:09 mark, you talk about the iris or opening, where do you find this measurement. the example you show nikon 24-70mm 2.8 and the opening of 25mm. I dont know where you get the 25mm from. I looked at dxomark and cant find. Help! I want to be able to make these calculations.
I am a physicist and I've worked on a variety of EM sensors (telescopes and radar) in a career spanning over 20 years. I've only gotten seriously into photography for the past year or so though.
Thank you for translating all of the physics in my head into the language of photography. You're absolutely right, camera terminology made a lot of sense when everyone was using film of the same size, but using digital sensors of different sizes makes that terminology confusing. I can see a lot of stuff that I've been doing wrong and can't wait to try my hand at fixing it.
Love your work, and I'm definitely subscribing.
Cheers!
I wish photographers would stop saying sensor size determines DOF, b/c it doesn't. It might determine your distance to subject, which will in turn determine your DOF, but it doesn't change your DOF.
DOF is a result of three variables: Distance to subject, focal length, and aperture. Greg Cazillo has even done a video to prove DOF doesn't change if all three of those variables stay constant, even if you mount a different camera with a smaller sensor.
why are you not the president? you seem like a really nice chap
..seeming to be a really nice chap shouldn't be the criteria for being a president.
It clearly currently is not...
It is
You answered your own question.
Trump already fired him, hahaha!
This is absolutely ground breaking stuff! Tony, I tried applying the crop on my Fuji XT20 and compared the meter readings on my 35mm film AE1. Here are the constants; Iso (1600), shutter speed (1/60th), equivalent focal lengths (28mm). The Fuji's aperture was set to 2.8 and I applied the crop to the AE1 in reverse making it more or less f4. I am getting a results where my meters (both center weighted) are saying that they are getting proper exposure on the respective adjusted f-stops. Am I getting the correct results?
(electrical engineer here) Yes you are doing everything right. However, the issue is, ISO is not a constant. ISO is defined on a sensor-by-sensor basis due to different pixel sizes needing different electrical gains to produce equivalent results. Lets say the brighness of a pixel goes from 1 to 10, if this pixel absorbs 4 photons and the apparent brightness is 8, then the electrical gain would be 2 brighness per 1 photon. If you now take a pixel that is only a 4th of the size, thus only collecting 1 photon, then the gain would be 8 brightness per 1 photon. But still both gains would be called, let's say, ISO 100, since they produce the same result under the same lighting conditions. But what is missed is how the electrical circuitry has to compensate for the smaller pixels.
This guy is by far the most knowledgeable photographer instructor on RUclips. This is the first time when a lot of these photo terms were actually broken down and understood. Bravo subscribed
Firstly P/Sqrt(p) = just the sqrt(p)
What happens when photons hit a sensor is that they convert photon energy into photoelectrons by interacting with the silicon layer, but by interacting with the silicon layer they also vary in the timing of when they interact with the silicon layer causing imperfections in the "quality" of photons and therefore the conversion to photonelectrons - this "photon noise" is directly related to the photon energy (light intensity) in an inverse square ratio - i.e. the amount of noise generated by "quality" of photons is the square root of the intensity of the light interacting with the silicon layer on the chip or photon noise (noise) = square root of the photon intensity (signal)
But this formula only describes the input part - it doesn't describe the Signal to Noise Ratio, which is the measure of how well the total amount of light "in" (the signal) relates to the total overall noise "out" (of the entire processing system start to finish) that ends up producing the end electronic representation of the photons that hit the sensor and represented by the factors / formula below.
1. The amount of light energy per pixel per second ( like the photon volume or let's call this "P")
2. The efficiency with which the sensor can convert each of the photons into an electrical signal that can be interpreted and read (let's call this "E") - a bad sensor will lose some information which is the same as generating noise
3. The amount time that the sensor takes to accumulate & aggregate the photons into a readable signal (let's call this "T").
4. The internal current (noise) as a consequence of heat generated from the sensor itself converting photons to electrical signals (call the "H") a cold sensor will have much less introduced false electrons (noise) than a hot sensor
5. The read noise (let's just say the noise generated by the pre-amplifier on the chip for now and call this "R")
The formula for SNR is ... P x E x T / (sqrt of (P x E x T + H + R))
1. Is a natural phenomemon of the input and the same for all cameras
2-5 are consequenses of the manufacturing process of CCD sensors - and the main reason why newer sensors can generate lower noise outputs from inherently noisier photons (low light input).
Well, they only considered the fundamental shot noise (#1), the other factors are almost negligible with modern CMOS sensors. Including all these in a video would make it too complicated for most viewers.
Tony Northrup About that F/2.8 equation that's only half of the truth. "Focal length equivalent" is a short-hand term used to give us uniform information about angle of view of given lens between different formats, but real focal length, real aperture, and what's most important, the brightness of the lens does NOT change. So F/2.8 lens on 35mm camera is as bright as F/2.8 in any other format, smaller or bigger.
You can say F/2.8 in MFT is equivalent of F/5.6 in FF only in terms of depth of field, but when you're talking about brightness of the lens, then it's not true, F/2.8 understood as brightness indicator stays F/2.8.
The brightness indicator is calibrated to make the picture appear similarly bright for equal fLen, Aperture, shutter and ISO despite different sensor sizes getting different total light, or alternatively its just indicating light incidence per unit sensor area. Look no further than cell phones to see how light incidence per unit area alone doesn't indicate quality. Taking post crop factor equivalence into ISO, Fstop, and focal length is getting the whole picture, because the goal is finding what it would take to capture an equivalent image.
This is the best crop factor video ever.
One more thing, what about the shutter speed? Does sensor size affect shutter speed?.
If Fullframe with F/4, ISO 800, and 1 second Shutter speed equal to F/2.8, ISO 350 on APS-C,
what shutter speed on APS-C to match with the fullframe?
for same exposure and noise you have to use 2 times larger aperture , and 2 times longer shutter and 2 times lower ISO (So F2 400 ISO 2s SS)
Same shutter speed. Either you open the aperture x times or you increase the shutter speed x² times (x = 1.5 or 1.6 for full frame vs APS-C)
Hey Tony, AWESOME video. Really great stuff. Question for you: If camera manufacturers are cheating us with their aperture numbers, are medium format lens makers selling themselves short? :)
That's a good question.
Hi Tony and Chelsea,
I really appreciate your effort in reviewing photographic products in a sensible manner, and Tony, your math rocks , and you know I thought of constantly watching this video time by time to engrave in my mind about the crop factor facts. Otherwise these MFT glass and cameras are so compelling. I was watching these videos since some time now, and I admire the way you critically and scientifically analyze performance. Good job for both of you.
I had my first DSLR in 2007 is where the Canon 400D came out to market, bought it with a personal loan! Then I sold it, when I was discouraged by the reach the 90-300mm lens I could afford, and the expensiveness of long lenses. After a while I got this Canon power shot SX-50 claiming to be the world's longest zoom camera for just ~500$. But I realized that this tiny sensor's noise and that it's always starving for light.
Then I have ended up in a Canon 700D and I cannot complain about it despite the size and bit high noise. I had almost made up my mind to get a Olympus EM10 which is available body only for 250AUD used, for the compact size of bodies and lenses. But after seeing this video, I thought of stopping a bit and re-looking at getting a APSC ones. So I don't know in APSC mirror-less world is there at least cargo-pant pocketable options with IBIS. I know that A6500 does, but at what cost?
So the search continues, and pls give me some tips if you can. Currently I am thinking of these models
Olympus OMD EM10 -250AUD used (body only)
Any Sony APSC mirror-less
Fujifilm X series ILC or fixed lens
Canon EOS M6 or M5 etc.
so my sigma 70 - 200 f2.8 on my D7100 is actually making images like a 70-200 f 4.2 if I have fully understood this?
Thank you for letting me know ;)
It makes images like a 105-300 f/4,2 on a full frame, because of the 1,5 cropfactor.
One of my favorite videos ever...
Being an engineering background and that too mechanical, I had numerous doubts and confusion on how all these parameters worked. Sometimes I would think it's a blunder.
Watching this video, now I'm aware it's really a marketing blunder.
Tony, cool video man. A real eye opener. I was confused with all the various attributes and numbers i see in my camera and lenses. Now i know what to look for when i need to buy something. You have effectively condensed 100 hours of reading material on the internet to a clean, neat 30 minute video. Thanks again.
Vinay Pai Thanks, and you're welcome!
I have watched this 3 times now and learn more (or understand more) each time.
In my hunt for a 'walk around street lens' I have been looking at the (Pentax) 24mm f2 (~$500) and the 28mm f2.8 (~$190) to be used on the Pentax K-S2. Both lenses have very good ratings. After watching this again I think I'll go with the less expensive lens. The 28 will give me a 42mm f4.2 with a 55 degree angle of view on the APS-C sensor which should do well. I can toss the kit 18-50 f4.5 - 5.6 in my pocket and be covered for most everyday type stuff.
Thanks, Tony for helping me better understand the lenses I was considering.
I am doing my post IT career things do learn list. This is the second video I watched from you. The presentation is great and the math is easy. I am just looking for a camera that takes good pictures indoors in a low light school auditorium with kids moving for under 1.2k In any case, this story reminds me of MPG on cars. What the car MFRs never tell you about when they say city/highway MPG are: the average temperature outside, the average number of people (weight) in the car, the average amount of up/down inclines in the road, what tires and inflated PSI was used, and finally, what gas they used. It seems to get the best MPG you must drive with one person at 65 degrees (no heater/AC) on straight flat road to get the best results !!! Thanks again.
Aperture is still the same on any sensor, difference is only in field of view which means you need to recompose image and change distance. And these two steps change resulting depth of field. Not the size of sensor. Lens is still the same, you only take crop out of that image circle. Registration distance is unchanged.
ISO in digital era has to take in account the size of each light receptor (pixel), not the size of whole sensor. 16Mpix 1.5x APS-C gathers the SAME amount of light per every pixel as 24Mpix FF sensor and also the noise with same generation of technology is the same. Problem with noise and smaller sensors is, that they reach pretty high resolution, which results in smaller photosites which results in more noise (signal + temperature noise). Cameras like 1Dx and D4 have less noisy images because photosites on their sensors are quite large when compared to APS-C sensor with same resolution.
Other issue is with crop-designed lenses, which in most cases have less light transmission (dark corners) than if you use FF lens on crop sensor.
If you cut smaller piece of exposed film, it will be still the same film, still the same grain and still the same DOF and bokeh and captured amount of light per square milimeter. The same applies for FF and Crop sensors.
xmeda "Aperture is still the same on any sensor, difference is only in field of view which means you need to recompose image and change distance. And these two steps change resulting depth of field. Not the size of sensor"
That is a common myth. First of all you don't need to move closer to recompose you can just zoom in, second if you move closer to recompose you will get a smaller DOF , which will result on a even more blurred background on the FF, which is exactly what we see in the video. The lenses are on the exact same settings but because the sensor size is different the DOF is different.
That is the whole point here. No one is saying that the sensor size is changing the state of the lens, at 2.8 the diameter of the "pupil" is still the same, however the angle of view isn't and this affects DOF.
tekhiun OMFG... if you "zoom", you change your focal length. Thats why your DOF changes. Not because sensor size. Think first.
xmeda And why do you zoom young padawan ? Joke aside.I know that, also you are zooming on the FF not the CF. I was just pointing out that "it changes because you move" is a myth. You have to recompose because of the difference in the sensor size and the resulting change in DOF you get by trying to get the same comp is directly proportional to the crop ratio.
Don't forget that focal length has the same problem ISO has as Tony mentioned.
There projection triangle can be calculated using the focal length, which will be the "height" of the triangle and the base is the sensor size, if we decrease the sensor size we also change the angle of view but keep the height the same. We basically have the same effect if we put a teleconverter except we change the angle of view by changing the height. So in that sense a CF works is the same way as a Teleconverter and what we should be talking about is not the distance where lights converges to a point (focal length) , but about this triangle and you need at least 2 factors out of 3 to do that: angle of view , base (sensor size) or height ( focal lenght).
I agree with tekhiun . This video is spreading wrong information and taking advantage in attacking Sony Olympus and Panasonic.
Rox alcantara welll everything that i said was in agreement with this video. Tony does make a good point in his video. We are using standards that are out-dated and are extremely relative to which sensor we use and aren't really precise (imo enough). The only way to change this is for professionals to be aware of what is going on, most of them aren't. It may seem like one of those conspiracy videos but in this case what he is saying is true. People that buy a 1.4 lens made for a cropped sensor aren't really getting the image quality a 1.4 lens would give to a FF camera, we need more awareness to this.
I fully agree with Tony that deceptive advertising should be pointed out, and the companies that do that should be called on the carpet. Now, when comes to the difference in light gathering ability between smaller sensor and full frame, it's not so easy for the average photographer. Tony pointed out how advantageous larger sensor is improving subject isolation. This can be a big advantage to professional portrait photographers. But does it affect most people? I will posit that for most people it is not *too much* depth of field (DOF) that presents a problem, but *not enough*.
Let me give you an example - you have your family at home, they sit around the table, some sit on a couch. You could resit them all in a yearbook style, but you probably want to catch the scene when they are relaxed and behaving naturally. In that case you need *more* DOF. With a full frame, you may need to stop down to f/8 or f/11. That means you will need to slow down your shutter, but that also may present a problem with a large DSLR - mirror slap and shutter slap may force you to use shorter shutter speed. So you end up jacking up your ISO.
On the other hand, your Sony RX100 with a CX sensor at f/2 will have a DOF equal (roughly) to f/13 on full frame. You also can handhold the camera more easily at slower shutter speed (no mirror, different shutter mechanism, or electronic shutter). Now you do not have to raise your ISO, and a good part of the full frame advantage has been nullified. This is just one example, but I can come up with many of them, having shot all kinds of sensor (and film) formats for many years.
Of course full frame may still produce slightly better quality, but then we get into the convenience or carrying it around instead of having a camera in your pocket.
In other words, it's not all black and white, alas. Pick your tools according to your application, and as Tony points out, don't let the vendors sucker you.
+mountainhobo Camera manufacturers always lied about technical specs, also in the film era. Especially DSLR manufacturers like Canon, Nikon, Olympus and so on. For example, they bragged a lot about the shutterspeeds (like 1/4000 which was VERY fast back in the days). When I was studying photography, my teacher brought a shutterspeed tester one day (you removed the film and lens and placed a ultra fast shutterspeed-lightmeter on the place of the film, this way, by pressing the shutter release button, it could accurately measure the true shutterspeed). It all ended up that the fastest shutterspeed was 1/1200 or something and not 1/8000 as the dial of my Nikon F4S showed. It was probably technically not even possible to get a shutter curtain open and closing in 1/8000 of a second even if Nikon stated it could do so. I felt betrayed since that camera used to be the most expensive SLR back in the days (5000 dollars i think back in the days).
+Lutte Kikker
That's probably because shutter is not a single piece of metal, but two rolling curtains which follow each other going from top to bottom. Even the highest end cameras take about 1/1000 of a second to actuate, but only expose 1/8th of the at a time in a rolling fashion. Slow mo guys have a recent video about how camera shutters work.
+Arakki
I'm afraid you got it the wrong way.
If you say the shutter speed is 1/8000, it doesn't mean "let's expose for 1/1000 just a 1/8 of the frame", because this means you expose for 1/1000 sec.
Donato Greco So, what's the real shutterspeed then when all Nikon F4/F4s state 1/8000 of a second? That camera has 2 shutters (2nd curtain). I did shot several sequences (all faster shutterspeeds on high ISO, pretty fast film and broad daylight) and the contact sheets didn't show a real difference in exposure at some point (I think 1/1200 or 1/1500 s, somewhere in that range).
Lutte Kikker You have to think to the shutter curtains at high speed (higher than the flash sync, when the whole sensor is exposed to light) as a horizontal (or vertical) hole in an opaque material. This hole has to grant the exposure set as Shutter Time.
If we extremize this concept to a single pixel sensors row exposed one by one for the whole trip of this "hole" along the height (or width) of the sensor, this "hole" will expose every single pixel row for the Shutter Time set in the camera.
If this wasn't working this way, you will break the exposure equation (ISO, Aperture, Time).
Fantastic video Tony. No other way to say it but that. You nailed it in the video when you said "there's no such thing as a free lunch". Light is light is light. There is no way that a small sensor with a small lens can inherently gather a "clean" image the same way a large sensor and big lens can.
(Although people might argue that the new A7S with its 12mpx and larger photosites that gather more light which makes the ISO "cleaner" at larger numbers, but the math still stands. Well at least I think so...unless I just made a really good point there lol) Anyway, kudos!
Thanks, Dave. A couple of people have brought up the topic of per-pixel noise, which I don't address in this video. It might be worth addressing per-pixel noise in a future video, but we almost always see pictures scaled down to fixed dimensions, rather than looking at individual pixels. Scaling images blends multiple pixels together and makes total light gathered the biggest influence on total noise.
The A7S (and the GH4) is designed to do 4k video without blending pixels together, which definitely has its benefits, but isn't really relevant to these concepts.
Yup...while I love my M 4/3 I am often frustrated with it limitations. Max usable ISO and frame rate for example. You just have to know what your getting and make an informed choice. I call the M 4/3 my "Disney" camera since I can walk around with it all day and still get shots I am satisfied with. Better than a point and shoot but less capable than my FF.
TheSunshineSpot Yeah, I love MFT cameras for the same reason. The thing that frustrates me most about the MFT cameras is that the base ISO is 160 (for Panasonic) or 200 (for Olympus). With those small sensors, I need ISO *25* if I'm going to get images as clean as ISO 100 on my full frame cameras! Let me go slower!
Tony Northrup But the sensors have a "native" ISO and going below that will always be a frig, so not always better. For Panasonic (and usually Sony, as Olympus don't do sensors) m43 sensors that's ISO 200 and that's what the pixels will be full up (well, less a margin manufacturers leave for highlight recovery) at that exposure. Unless you can make even deeper pixels (and they are pretty deep already) then you just run out of capacity in the well. (Strictly the Native ISO is when DR is max and noise is min, but it's usually around the full well capacity.)
For example an E-M5 pixel can hold 25K electrons (1Dx = 90K, D4 = 118K), the QE is 53% (49%/53%) so basically 50K photons and it's full (but would always leave some margin for highlight recovery, so 40-45K). If an ISO 200 exposure gets you to, say, 30K photons for the brightest areas then that's your lot and a lower ISO isn't possible (as would need too many photons to be stored) except you can throw away some highlight margin and make a somewhat lower ISO. (I believe that the next full ISO under 200 would not have worked and so they go a little less than a stop lower and lose a chunk of the highlight margin.)
Sensor size doesn't have affect for dynamic range or exposure range. It is about different sensor what does the difference for dynamic range, exposure range and ISO.
Different sensor == Different result. Same sensor but different size == same image quality with same values but only a different cropping and finally resolution.
This can everyone find out by themselves as long they are ready to be open minded and really learn technology.
The same thing is exactly with the films. You don't need to adjust shutter speed or aperture when using same film (ISO was tied to film) in different crop factor cameras.
Sensor size DOES NOT change depth of field, cause more noise or affect to exposure range. It is difference in other technical features than sensor size.
Example: Get a 35mm film camera and load it with 35mm ISO 400 film. Place camera on tripod and leave it there so you don't move it anywhere. Take a photo. Then if you would be a skilled enough (lets say you are), take the non-exposured part of the film and slice a part off so it gives you 1/4 of the 35mm frame size (aka 2x crop factor). Now place the piece of film on the film plate and take a photo with same settings as you did with full frame.
Now go to darkroom and develop the film and the piece of the film.
Do you claim now that now the piece of film now has grain of ISO 800 or it is somewhat under- or overexposured when compared to full frame?
Lets do it with other way on digital manner.
Lets walk to manufacturing line up and get a silicone dice what has all the Sony A7r sensors laid down waiting to be cut to components.
But lets cut (yes silly, but lets say it would be possible) a single component to 1/4 of its size and with some magical way replace A7r sensor with that our own new custom made 1/4 of the size sensor.
So all the other digital parts in our custom Sony A7r is intact, only the sensor size has changed from 36x24mm to 17.3x13mm.
Do you claim that now suddenly our custom made Sony A7r has changed so much that ISO noise is worse, you need change shutter speed or aperture to get same exposure or you have 2x depth of field as unmodified stock Sony A7r does when shooting from tripod in same position?
Sony, Olympus, Panasonic etc don't lie. They don't cheat you (Well, Canon and Nikon does in many of their cameras).
Tony. Thanks. Are you using the thread size to figure len at wide end. Ie. I'm thinking of getting tamron 70-200 f2.8 has 77mm end. Using d7100 apc body. So it's not really f2.8 when at 200. Because of crop it's 300 divided by 77. Is this right
Very compelling presentation. I was a motion picture lighting/cameraman for 45 years and have the following practical observations: Cine camera lenses use t (transmission) stops, not f-stops to eliminate exposure errors when changing focal length. A lens for 16mm film has to have a much wider full aperture than a 35mm (half-frame) lens to achieve the same shallow dof given the same angle of view and subject distance.
Well, this could be an endless story. Bring medium format and large format into the game and the so called "full frame" cameras (which is a ridiculous name for the 35mm format) and now these are the "cheating" ones as well. This has nothing to do with "cheating" at all. They're just behaving different - each on it's own.
Leif, the cheating part is where manufacturers selectively describe some aspects of their lenses in terms of 35mm, and other aspects in their native format, depending on which makes the lens more appealing. Nobody puts 35mm gear in terms of medium and large format.
Tony Northrup That's a problem the industry created decades ago. When we had medium format and the new "small" 35mm format got more popular they also kept using the focal length as the "measurement" and didn't used the medium format as a reference. Now we see basically the same story again with other smaller sensors. The focal length is the focal length and it has nothing to do with cheating because they're telling us the real focal length of the lens which is true. The only issue is that the focal length is the worst thing to compare lenses because it says nothing about the actual viewing angle we get. But that's a problem the industry created decades ago and photographers got used to it as well. You'll barely find a photo book where they speak about viewing angles instead of focal lengths.
***** Even that wouldn't be untrue because then you could say that Fullframe and Medium Format cameras are cheating because they don't use large format as reference ... and so on. Compare a 50mm Full Frame lens to a 50mm Medium Format lens...the Full Frame lens will behave like a 80mm Medium Format lens. It's just a bad idea to try compare different systems based on those numbers (or at all) in my opinion.
***** The question is what people expect from a 2.8 lens. Are they using the aperture as a reference for the depth of field they get or as a reference for the light sensitivity they can expect on the sensor size they're using? You can turn it in each way..it will always be confusing. A 2.8 lens is a 2.8 lens exposure wise. And then again there would have to be a new reference because full frame can't be the reference either because it's a format that its in the middle and not on the top.
***** That's true. What I'm trying to say is that using FF as reference is cheating as well.
re:Smaller sensor getting less light - you are correct in wording but I believe incorrect in concept / practical application. A Canon crop sensor, for example, will only see the center portion of a 50mm lens compared to a full frame. That means the rest of the light beyond the sensor's borders are discarded (hence the "crop"). Therefore, that light is not being seen, and it could technically be said that it is therefore seeing "less light". However, the light that is hitting the sensor in the middle is of the exact same intensity (ie. same measure of photons) as it would be on a full frame sensor. It absolutely DOES NOT receive less intensity, as that would not make sense on a purely scientific perspective based on the physics of light. It would be simply a full frame image with the edges cut off, at the same intensity for exposure. If you were to open up a 5D and put masking tape over the edges of the sensor to make it only see a 1.6x, the exposure would be the same. It would just be cropped. That's all that happens in a crop camera.
So why did your tests reveal more noise at the same ISO? That has much more to do with the size of the photosites on the sensor- that is, the size of each "pixel" being exposed. Remember, a sensor is nothing more than a grid of photosites, each one sending information about the light that it sees to the image processor to turn into an entire picture. The larger the photosite, the more photons it can collect and the more accurate the resulting light sample will be. Smaller photosites that are more densely printed on a sensor will therefore yield more noise than larger ones. You could argue that a larger sensor automatically has larger photosites, but that isn't necessarily true. How large the photosites are is a combination of sensor size to megapixels. The more megapixels a sensor is designed to have, the more photosites need to be crammed onto the surface and therefore the smaller (and less light sensitive) each one will have to be.
Take the Sony A7S for example. It is a low light champion because it offers only 12 megapixels compared to the flagship A7R which offers more than 36 megapixels. That means that the A7R has photosites 1/3 the size and therefore will have noisier high ISO.
To say it has to do with sensor size alone is terrible misleading. You can have a smaller cropped sensor with lower density photosites (for example, a micro 4/3 with only 12 megapixels) that reasonably outperforms a modern high-density full frame when it comes to noise sampling because each photosite is large enough to gather more light.
Mordy Gilden Well said, Sir!
So is it not then more about pixel size vs sensor size? As in a 16MP APS-C will have smaller pixels than a 16MP FF. Therefore each pixel on the APS-C will receive fewer photons than the FF?
This video is fantastic, I already knew about crop factor that had to be applied to the aperture as well but somehow it was always considered as a secondary information leading me to forget it when comparing lenses, after this video I'll never convert only the focal lenght without doing the math on the aperture as well, thanks so much Tony!
Can you help me with this. I thought you implied that angle of few and field of view were interchangeable. I thought that angle of few was a function of the lens. Field of view was a function of the sensor. Wide angle lenses will distort faces more than longer focal lengths. The distortion will be present on both cameras but the field of view will show more or less of the scene. One of the reasons I prefer FF over Crop. Better working distance in small rooms w/FF.
Wide angle lenses distort things in the corners of their field of view (= at large angles of view). With small sensors, you can not "see" those corners because they are just outside the sensor. Therefore, you are getting the same distortions at 28mm full frame as at 18mm APS-C. The sensor size really does not matter if you do the right math. The only issue is that lens manufacturers don't make big lenses for small sensors, so we are forced to buy full frame sensors to be able to use big lenses that capture more light and make stronger bokeh
I was in disbelief until doing my own research and running the math. I'm floored. Thank you Tony and friends for putting this together for us. I live in Japan and can only say that my respect for Japanese manufacturers has taken a humungous hit after this. I shoot with Panasonic gear, and used to shoot Canon, but knowing this gives me much more to consider before making any further purchases. Thanks for doing the hard work. I think I'll create a video like this for the Japan market so that folks here can get a better understanding as well. What a damn shame.
Cold hard facts! Love it Tony. I just had a conversation with a client that has worked at a very reputable camera shop for decades and totally shut me down when I talked about crop factor vs aperture. It goes to show that many veterans don't understand the math. Thanks for the great info!
One of the best RUclips photography videos I've seen. Thanks Tony for revealing the truth behind some of these misconceptions in photography and how some manufacturers exploit our ignorance.
Lee Doyle Thanks, and you're welcome, Lee!
Tony Northrup I also recall years ago how some manufacturers were quoting large megapixel counts on their cameras when in fact they were half that amount and then digitally doubled. The same can also be said for the marketing campaigns surrounding ridiculous digital zooms. Unfortunately there will always be a market for people that don't do their research and ans long as there is, then companies will continue to exploit it. Human nature I guess. Thankfully for people like me who love to learn, there is an abundance of research material and people like you that we can learn from. Thanks again Tony. Really enjoy your videos.
This is perhaps the best photography video I've seen on youtube. Wonderfully informative, incredible clear and well documented. This should be required viewing, not only for the novice, but for those of us who started back in the day when film was all there was.
16:15 How did you get the same image size on different crop factors? Zoom adjusted or image crop only? Please explain. Looks like first one was shot with 200mm, second one with 100mm with x2 crop factor. This is wrong in this case and can't be used for the rest of the video here.
Hello Tony, one question, at the end you said that Fuji, Canon and Nikon are not cheating at the moment they specify the aperture in their lenses, do this applies now a day ( I ask because I see this video is from 2014), so for example, the Fuji 16-55 f2.8 is really a 2.8? is that the reason why is bigger and heavier than the new Sony apsc lens 16-55mm 2.8? how do I know they are not cheating?
thank you
They are still cheating. Fuji 16-55 f2.8 is really a 2.8, but if you want to compare it with full frame, this lens sadly becomes 24-83 f/4.2. They will just never mention the equivalent aperture.
This is a real eye opener and definitely gonna save a lot of money spent on misleading products. Thanks a lot!
Nobody is misleading !. A 2.8 lens is 2.8 no matter what size sensor you use it on. The formula for Aperture does not change = Focal / Aperture diameter. When they came up with this for a way to standardize exposure, 35mm film was not even around !
The truth about crop factor(CF) is that it has been applied to all the wrong things to compare crop and full frame(FF) sensors. The only thing that you should apply the CF to is the field of view (FOV). The physics of a lens does not change because it’s on a FF or crop sensor, what changes, and the only thing that changes is the FOV. A crop sensor does not change the focal length or depth of field(DOF) of a lens.
Here is where we usually get it wrong, it is when we try to get the FF equivalent on a crop sensor. When we multiply the CF by the focal length we do not in fact change the focal length of the lens but the equivalent FOV. For example when you use the APS-C CF (1.5 or 1.6 for Canon)) to calculate the FF equivalent of 100mm, you get 100mm x CF = (150mm or 160mm). The general assumption that you are getting 150mm or 160mm equivalent focal length is wrong. If you were to compensate for FOV by cropping the FF or using the brenizer method to get the 100mm FOV equivalent on FF and doing a comparison, what you will find is that the DOF is the same, and for any other given focal length it is the same.
The marketing on lenses does not make the distinction between focal length FOV equivalent which is why this topic has been so confusing to most of us
I feel like my life is just a big lie right now, and I feel very discouraged as a photographer. However, I am very thankful that you made this video, and opened my eyes on future puchases, so thank you so much, Tony! On an unrelated note, would recommend getting a used 5D Classic?
Thanks! The 5D classic is pretty noisy compared to modern full frame cameras, and it lacks video or live view, but we still have ours and it's a capable camera.
Hey Northrup how many post do you write to yourself?
Now you are being cheated.... Don't fall to that...
Just a question.. Depth of change on different sensor, even at same equivalent focal lenght, for example you have to double it on mirrorless camera compared to ff. But is it the same for the light on the sensor? 2.8 is 5.6 on 4/3 and so you gain half light with sand shutter and iso settings? Thanks
9:33 That does not make sense. A crop sensor basically crops in on an image that could've been taken by a full frame sensor. Why would an image get darker, just because part of it was cropped out? It doesn't. It stays the same brightness. The amount of light hitting per mm^2 stays the exact same. However, the one thing that probably changed is the size of the photocites. Smaller sensors usually have a higher density of photocites, so individual photocites gather less light, which has to be adjusted for by "secretly turning up the iso".
To see the effect, you need to have the same number of pixels in all sensors, meaning the pixels would have different size indeed. The signal to noise ratio obviously depends on the pixel size. But, I think, including this variable would make the video too complicated and the conclusions would be the same anyways.
right. His Iso theorie is complete nonsense, because he doesnt realy know, how optical systems work.
A Camera is not a a bucket in the rain - unless you put it outside without the lens.
Or - to explain it correct with the bucket - theorie: if you put 2 buckets in the rain with a lid and the same 2.8 whole in it, you will get the same amount of water in the big and the small bucket.
"You learn this in second or third grade algebra." You must have gone to a much better school than I did.
Tony Northrup I think your example at 31:26 is good... it shows that you are wrong converting aperture. Because, yes, you can shoot sport at 600 mm and f/2.8 with the Lumix and no, it is not equivalent to f/16. You will not need extra high iso but the same iso as the 600mm f/2.8 you would use on a FF. But yes, larger sensor have better image at a given iso so the picture will be better if you go to high iso/high speed.
You keep repeating that large sensor receive more light than smaller ones. This is not true if you use the right lense on the right camera. Because the smaller the sensor, the closer it is to the diaphragm.
Camera makers are not cheating, f/2.8 is f/2.8. Difference are in DoF and bokeh and I agree they could communicate on that but I think people interested in this are informed.
Anne Hodermi watch the Crop Factor part 3 video; I address all your concerns.
Tony Northrup The video didn't correct your wrong assumption that more light hits bigger sensor (at a given aperture). You probably believe that rays of lights are parallel to the sensor in a camera? It is evidently not the case because lenses focus light and 'redistribute' it on the sensor (radial beams from the focal point. Smaller sensor are generally closer to the focal point).
And that contradicts your theory that smaller sensor would be less luminous and needs conversion to give 'equivalence'.
Maybe you should stop 'matching the maths' because I can take the same picture with my RX100III @18,3 mm f/2,8 (50 mm equivalent) at the same speed than with my D300 @50 mm f/2,8... (and you see, I didn't convert aperture)
Yes, depth of field/blur will be different.
Yes, if I want the same bokeh I have to take the picture at f/7 to have the same result.
Yes, it will result in the fact that I will need higher iso to keep the same speed
Yes, larger sensor have better high iso performance (for the moment)
Yes, larger lenses have better optical performance (and probably for long)
But no, f/2,8 is not anything else than f/2,8 on different sensor sizes.
...or maybe manufacturer should also multiply the weight of camera to avoid cheating ? Why they don't? Because weight and sensor size are not mathematically linked... like aperture and sensor size. Aperture (the number written on lenses) is a relative figure. Relative Aperture = Absolute Aperture / Focal length. You evidently never need to convert absolute values. I would say it is 'scalable' (stays the same if geometry is larger).
Iso sensibility quality and pixel number are technical parameters, clearly independent from focus and sharpness of a picture and should be treated independently to sensor size and aperture to clarify the discussion.
Noise (related to pixel and iso technological treatment - converting light in electrical signals) and blur (optical) are completely independent.
I will probably not convince you since you based all your videos on the fact that manufacturer are cheating... but, please, try the Lumix @108mm f/2,8 and your full frame @600mm f/2,8 and tell me if something else than depth of field, noise and optical quality (vignetting, optical aberration) is different. You will see that you will take the picture at the same shutter speed (because the sensor receive the same amount of light).
Anne Hodermi Re-watch the video, you don't really get what he's trying to say coz your statement somehow proves that he was right. You will never get the same result (in term of bokeh and noise) out of 2 lens(for full frame and crop) that has equivalent focal length but with same aperture on different sensor. Of coz you can have same speed, but never same image quality. That's why he said you never want to shoot sports at f16 coz you'll get a noisier image. It not just shutter speed that matters.
mr2014 Bokeh will be different because at the same relative aperture of f/2.8, the real aperture of a Full frame is larger than on smaller sensor cameras (geometrically, beams of lights are distributed differently, better spread with "large holes"; a drawing would be more meaningful)
Noise has nothing to do with the "convert the aperture or not" debate. The current best small sensor have less noise than the first full frame sensor.
I don't know what is your definition of image quality.
Depth of field/bokeh is not the only criteria.
If you have a good sensor and big aperture (f/1.8 or 2.8) you will be able to shoot at high enough speed to avoid motion blur.
I don't understand why you say I need to shoot at f/16 as the camera can shoot at f/2.8 (and NO, I don't have to convert).
The ONLY difference will be:
- Bokeh (for geometrical reason)
- Distortion (because larger lenses generally have less optical distortion)
- Noise (because a (same generation) larger sensors generally have better high iso quality).
As the quality of sensor is becoming incredible (iso 3200 with enough light is often acceptable), as optical are sometimes very good and as most of the people don't really care about (subtle difference of) bokeh, I can tell you that the real cheating is to convince people that they need full frame reflex.
Let's do a test with two f/2.8 instead of useless theoretical diatribe.
Anne Hodermi Good to see that I'm not the only one who sees the error. An F2.8 lens is an F2.8 lens!
Tony, this was a ridiculously informative video.. a real eye opener. I see all my DX 1.8, 2.8 and similar 4/3 lenses in a whole different way now. And it really explains their performance in comparison to each other. Why are the manufacturers allowed to do this? Thank you for taking the time to explain, and for doing the math. :-) I'm just a hobbyist, but I feel like a much more informed shooter. Thank you. Thomas
Thomas Thanks for the kind words, and glad I could help!
@@sakritone Who can asnwer please: in the same condition with FF I take a picture with 50mm at f1.8 at ISO 800 , to have the same exposure on a mft with a 25mm at f1.8 is enough 800 ISO or I have to setup 1600 ISO?
This is one of the Most necessary videos anyone interested in photography should watch..if you can't justify buying FF equipment or dnt know crop factor nature then watch this if you want to make real photos..tony this is one of your Best Videos IMO...Sharing This :)
When I bring up the topic, people look at me like I'm weird or something! :D ... and some even feel insulted by the truth and the facts! They still believe the promoted nonsense rather than facing the math! Then they keep wondering how come their 'super-bright' lenses with their mini-sensor cameras won't work in low-light or fast action! Gee!!! :D
mrgclips Yeah, I experience this, too. Re-teaching a misunderstanding is always difficult, but once people have paid their hard-earned cash based on their own misunderstanding, it becomes emotionally painful for them to admit their mistake... no matter how straightfoward it is to prove.
You are not alone buddy. LOL!
Best video on RUclips ever!!!
If you want the same depth of field, you multiply the aperture. But don't multiply aperture to get the same light, because you will just get a brighter image for a cropped sensor
Some of you may find this Wikipedia explanation helpful. Notice that both the focal length and aperture range have been converted, underlining Tony's point.
Equivalent aperture range
See also: Image sensor format
In digital photography, the 35mm-equivalent aperture range is sometimes considered to be more important than the actual f-number. Equivalent aperture is the f-number adjusted to correspond to the f-number of the same size absolute aperture diameter on a lens with a 35mm equivalent focal length. Smaller equivalent f-numbers are expected to lead to higher image quality based on more total light from the subject, as well as lead to reduced depth of field. For example, a Sony Cyber-shot DSC-RX10 uses a 1" sensor, 24-200 mm with maximum aperture constant along the zoom range; f/2.8 has equivalent aperture range f/7.6, which is a lower equivalent f-number than some other f/2.8 cameras with smaller sensors.[20]
I agree with your calculation, although they don't apply to evaluating equivalences. A 50mm full sensor lens is still a 50mm lens lens on a cropped sensor camera, only edges are cut down to make a smaller area picture image covering a smaller, therefor the "equivalent" naming. The lens behaves exactly like like on a full frame camera including f-stops and bokeh of and 50mm lens. The reason is that It does not actually become a 75mm lens, only appears as one.
Agree : exposure doesn't change only depth of field and field of view change
Laurent Outang Field of view is not gonna change unless you change focusing distance or focal lenght. 50mm has the same depth of field and amount of bokeh on any camera if youre shooting at the same aperture and distance to a subject.
Spot on Tony, the video is great as it enlightens everyone on what the real deal is. Some people will of course argue but I think it's arbitrary. I think one of the reasons marketing people from camera companies would sorta 'trick' consumers on the actual specs is because of the mainstream paradigm on two things 1.) Angle of View (AOV) and 2.) ISO. The standard which you have stressed to have stemmed from an 80yr. old standard that is clearly not applicable in the 'toy' cameras of today (e.g. phone cam to bridge-cam). Unfortunately, the common man has his mind set on these two things only, they couldn't even care less for f-stops. Having had the 135 format as the de-facto standard on photography, it's created a paradigm full of myths and fallacies (too bad for the poor man). So marketing people would rather take a short-cut than coming up with a cheat sheet of the actual specs. After all, it wouldn't make sense for ordinary people to be given a 4.4mm lens camera (1/2" frame ) when the picture looks like a 24mm AOV image (full frame). Ah, they also couldn't care less for frame size and the crop factor that comes with it. Hence these false equivalents are put out as actual specifications. Just my thoughts.
Tony, I'm a little confused by @16:07 on the one hand you show two sensors with the same settings producing an identical image (difference is depth of field). On the other hand @31:20 you talk about f/16, but not for DoF, but rather for light transmission. How can both statements be true? Wouldn't we need to bump up the ISO or increase shutter in the @16:07 example to get the same amount of light?
Never mind, answered my own question watching the Crop Factor 4 video. The ISO setting numbers are the same, however ISO 100 on MFT doesn't gather as much light as ISO 100 on FF. So if you could push a MFT sensor to 1600 ISO and equivalent FF sensor could be pushed to 6400. That explains why my APS-C cameras were really unusable past 3200 ISO, it was the equivalent of 12,800 ISO on FF. Now that I know that and since I'm very often in the 800-1600 ISO range, it seems I should be investing in FF to get rid of noise.
THANK YOU TONY!! You'll probably never read this, but this video was a huge help! I was considering MFT, APS-C and FF. Now that I understand the differences, I'll only be going FF.
Great info Tony, but I have a question about aperture determine formula. Where can We find iris diameter value?
I want to know too.
+Difference Engine (DE) I believe I have found it. It is written on lense.
+Maigonis Elleris Thanks. :)
Hey I tried to look for it on the lens. Can you be more specific as to where you found it. Thanks!
And one more thing. If the core idea of your video is that all manufacturers lie to us how should they advertize Canon EF lenses for example 70-200mm f2.8. It can be mounted to both EF and EF-S cameras and in your system they should start advertising it as 112-320mm F4.5 lens even though physical quality of lens itself hasn't changed. So even though current system is confusing and marketing deps take advantage of it, it's still easier to figure then the one you came up with.
Well, if we used field of view and total light gathered as metrics, you'd have to apply a crop factor *only* when using lenses designed for larger formats. All native lenses would compare directly without crop factor. Now, anyone who uses anything bigger or smaller than full frame needs to apply the crop factor to understand how lenses compare.
Tony Northrup Interesting thing happens when you are trying to match DoF from MFT to FF. You have that slide in demo but the way you speak of it is that image from FF sensor is the one we wanted to achieve originally but if the situation is opposite and the image from MFT is the target then we end up with a higher F and ISO numbers on FF camera in order to match DoF from smaller sensor.
That is actually a common mistake for people who just moved from crop sensors to FF and start shooting wide open just because they did so on coped cameras, everything ends up been out of focus and they don't get why. Their DoF is razor blade thin at F1.4 - 2 on FF, yes they get more light but the situations where you need that shallowness are limited.
So basically you are saying that you are sorry for those people shooting with their 12-35mm f2.8 and thinking that they are getting 24-70mm f2.8 but that's not right. It's like you are pointing at their wheelchair which they never knew existed. There's no point in thinking about other camera that you don't shoot with and that EFV line is only for those downscaling the system not for those that think they are missing out something.
Marketers take advantage of that in case of those bridge cameras indeed but don't they always do that with other things in our lives? :)
That is total bullshit. A 70-200mm f/2.8 lens does not suddenly drop to f/4.5 while on an EF-S camera. It is still f/2.8. Quit lying to people!
I'll prove it to you. Go get a Canon 2x teleconverter and put it on that 70-200. Notice how it still autofocuses, regardless of whether it's on a 5D or 70D? That's because it's an f/5.6 on both cameras now since the Canon 2x teleconverter causes a 2-stop decrease in illuminance.
If what you were saying is true, then the 70-200 + 2x converter would not autofocus on a 70D because the F-stop would have dropped from f/4.5 down to f/9.0!!
The fact is that a fast f-stop lens still has the same f-stop no matter WHAT camera you put it on. F/2.8 will still project more light per square millimeter on the sensor, no matter what kind of camera it's on.
This guy is so full of shit it's not even funny and you people are just LAPPING IT UP.
If it's an EF(S)... YES! 'S' lens is set up for the APS-C sensors... It doesn't even work correctly on a full frame... - it's cropped -
darkgoob I totally agree with you. Sadly your comment won't be here long... I also showed he was wrong and he blocked my comments. He didn't even had the balls to answer me. My comment was not rude and I asked for anyone to try to prove me wrong in what I showed.
Let the truth out!!
Thank you Tony.
+Fatmir Avdi You're welcome!
Tony Northrup I think I followed most of this video. But I'm just wondering whether you can directly compare the figures of Canon EF lenses to EF-S lenses??
For example, I have an APS-C camera with a EF-S kit lens (18-55 f3.5-5.6) and a EF 50mm f1.4. I understand that the crop factor would apply to both in terms of the focal length, so I have the 35mm equivalents of a 28-88mm lens and an 80mm lens.
Does the same apply to the aperture, for depth of field calculation and bokeh?? Do I have the equivalent to a 28-88mm f5.6-8.9 and a 80mm f2.2?? Say if I were to set my EF-S kit lens to effective 80mm and use the lowest f-number, and then match this on my EF lens with the same f-number, should the pictures theoretically be identical for angle of view, exposure, bokeh, etc??
Hi there. I have a question. I just got the D600 with Tamron 24-70 f2.8 and Tokina 11-16 f2.8. Have also the D5200 crop body. My idea is to keep the crop body with a lense with bigger zoom to avoid the lense changing. I know the easiest way but the problem is that I spent almost the whole budget I got buying the full frame body with the both lenses. My question is about the cheapest option which will give me a good F number, because I'm shooting live stage performings which is almost all the time low light scene. I'm sorry for my English and Thank You very much in advance for the help.
F-number tells you the intensity of the light on each square mm of the sensor - this doesn't change with sensor size
Åke Åkesson Right! We totally agree with each other.
The problem is that light per mm2 alone doesn't have any impact on your final image quality--it's total light that determines the visible image noise. That's why you use crop factor with the f/stop--to convert light/mm2 to total light and thus predict how your image will appear.
Just look at the example pirctures! They proof Tony right! :D
"Total light" is a meaningless and confusing term. Why do I need the same amount of light if my sensor is 4 times smaller ? I don't. The differences you observe are due to the size of the photocell. Think about 16MP on a m43 and 16MP on a full frame. Each photocell in the m43 camera is 4 times smaller than the same photocell on a 35mm camera. The smaller photocell is inherently noisier therefore it has worse performance in low light, and that's why ISO performance degrades much more rapidly with a m43 camera than with a FF one. It's also why m43 camera manufacturers work hard to improve stabilization options. They allow the photographer to expose for more time in low light, at low ISO, in an attempt to compensate for the smaller photocell.
Have you taken a look at the pictures Tony Northrup showed in his video?
I think you talk nonesense without any understanding of the topic! ;(
What has a good IBIS to do with ISO? Only with lower ISO (25 instead of 200 like most MFT cameras have) you can compensate for the smaller sensor (2 times square)! That's what you call "meaningless" total (amount of) light! :D
If you want the same depth of field you'll need f0.6, f0.7 or f0.9 (for the full frame lenses of f1.2, f1.4 or f1.8)! That's what you can take out of Tony's explanations. And what MFT camera manufactures (Olympus and Panasonic) should do! ;)
IBIS just helps you to take pictures at a lower shutter speed/aperture without bluring it! IBIS isn't exclusively on MFT cameras (though Olympus at the moment seems to have the best IBIS)! xD
I never said you can get the same depth of field between a m43 and ff camera using the same f-stop, because it can't be done. That's clear and proven. I never argued the results presented are wrong or incorrect. I only argued that the mechanism used for explaining them is not meaningfull.
Assume for a second you are attempting to take a handheld picture.
Your settings are: f/4, ISO1600, 1/40 exposure time
Your camera tells you the picture is 1 step underexposed.
Your options are as follows: Increase the aperture or Increase the ISO or Increase the exposure time.
On m43, the stabilization (body, lens or both) allows you to increase the exposure time when shooting handheld without getting a smudged picture. There are limits to the technology but it helps.
On a ff maybe you would opt to increase the ISO, since the pictures look good at ISO3200 also.
On m43 you will want to avoid increasing the ISO since the 4 times smaller photocell in the m43 sensor is much noisier than the photocell in a ff camera.
This alone was my point. The observed ISO related results are not due to some "total light" value, it's strictly due to photocell size.
The 5DS (50MP) has native ISO limited to 6400 for this reason exactly. The smaller photocell size behaves worse in low light than other ff cameras with smaller number of pixels and therefore larger photocell sizes
As mentioned in a dpreview comment:
Physics here is very simple. One pixel does not care how much light the whole sensor gets. All it cares about is what it personally gets to deal with its own signal/noise ratio."
Cheers.
Great video, you have opened our eyes. :D
Kevin Seah Thanks!
You are far the best Photografy channel out there! Thanks for your videos and chare so much knowledge!
What happens when i use an adaptor to put a helios 44 on my MFT? It's got more glass so is it actually f2 or f4? And why does the measurements of distance for focal length seem accurate?
akaawol The lens type doesn't affect crop factor, just the snesor size.
So the f2 lens still becomes a f4 even though there is more glass?
Hello Tony. I always enjoy your videos...but also tend to have comments on your methodology as well;) First and foremost, I truly admire the work both you and Chelsea put into your videos. This video is a perfect example and a great video for explaining differences in depth of field between sensor sizes using the same aperture value and the common fallacy in aperture/focal length equivalency between the differing formats and the amount of light gathered by sensors of differing formats.
Having said that, there is one major omission which has caused many to doubt your results. I believe it is important to state that your conclusions hold true for sensors of a similar generation with similar pixel size and density. Mathematically you are spot on, however, without this additional information, you end up with paradoxical results such as the e-m1 having similar noise performance to the a6000 and the A7s exceptional high ISO performance vs the A7R.
I believe that such explanations would go a long way toward addressing some of the negative comments. in practical terms, contemporary cameras of different sensor sizes should have the same exposure triangle for any given scene, however, they will have different depth of field for the same aperture value and potentially different noise performance depending on pixel size and density.
Notwithstanding, this is still consistent with your mathematical explanations of the equivalence of total light based on proportional ISO values for the given aperture and focal length. I agree with you 100%! Great job explaining a very convoluted subject!!!
You're wrong. Light per unit detector (same as unit area) is all that matters (and maybe electronic noise to some extent) if you're interested in the amount of noise. ISO is essentially gain,. It defines sensor performance just as well as film sensitivity.
As for DOF and aperture setting i can't quite understand you, so not sure if there's something to contradict. True 2.8 is exactly the same on any true 50 mm lens, Your sensor size has no influence and you can't recalculate it, not the aperture, not the focal length. Same for DOF. The difference arises when you recompose (change distance or focal length) to accomodate the smaller sensor and get the same framing.
You're right about not being able to adjust focal length for crop factor, you can't adjust it FOR ANYTHING. It's a physical parameter of an optical system and has nothing to do with the sensor behind it.
Really really misleading stuff here, you need to learn some physics yourself. And change your audio settings/mic, it's horrible.
I agree with your position though, they're assholes for giving the "equivalent" instead of the real focal length.
this is very clear.... thanks god, you comment on this :)
No, you're wrong. SNR formula: more total signal = better signal-to-noise ratio.
geezzz the noise gate is aggressive
Bravo to you, Tony, for this excellent informative video! It's like the blinders have been pulled off!
Andrew Sweigart Thanks, Andrew!
I have some questions. If I were shooting a full-frame and an aps-c at ISO100, the total amount of light gathered by the full-frame will be about 2.7x (0.5 stop from f-stop + 1.25 stop from bigger sensor) more than the aps-c's at a set time interval.
The 0.5 stop can be eliminated if I were to use a DX lense on an aps-c but the sensor size advantage is still present, which means at a set time full-frame still captures 2.25x more light.
Gee, no wonder full frames are so good at low light. But if I were to use a sensor-stabilized apc (like Pentax) with FX prime lenses, the stabilization can help slowing down 3 stops of shutter speed without causing severe motion blur. It partially solves the problem but I still cannot capture freeze motion in low-light with high ISO like a full-frame could.
Am I correct?
Thanks for the info.... could i ask one question, my apologies if i missed it. How do i determine what the diameter of the iris is?
Thanks again!!
F-stop settings. The lower the number the wider the iris or Aperture is open.
so if I want a wide angle lens on a ps-c camera i basically have to buy at least a 10mm
You are right about ISO, but every generation sensors seem to always be getting better and better. I do not know if there is a limit to how far they can keep pushing sensors.
But I would bet a Panasonic Lumix GH4 has better ISO the a Canon 5D Mark 2. So I agree with you about ISO and light but within the same generations of cameras.
Also mega pixels play a big part how much light the sensor can get.
I have a 5DmkII and a GH4, the 5DmkII is significantly better at high ISO both for having more Dynamic Range and less Noise. It's a bit over a stop I think, DXO have it closer to a 0.8 stop:
www.dxomark.com/Cameras/Compare/Side-by-side/Panasonic-Lumix-DMC-GH4-versus-Panasonic-Lumix-DMC-GH3-versus-Canon-EOS-5D-Mark-II___943_842_483
Go to measurements tab then SNR.
Yeah I guess its going to take GH5 to maybe have better ISO then the Canon 5D Mark 2. When I said Canon 5D Mark 2 I was assuming by now the Panasonic just releasing the GH4 would have better ISO. Guess I was wrong.
After seeing how good the Canon 5D Mark 3 is in low light I just guessed the GH4 would be better then the Canon 5D Mark 2.
mp4podcastDOTcom A camera with a smaller sensor will never equally match a camera with a larger sensor in terms of ISO performance. Due to having a smaller surface area a MFT simply cannot absorb the same amount of light as a full frame sensor.
On the other side of this, a medium format sensor (larger than full frame) will have better ISO performance than a full frame camera because it will gather more light.
The Canon 6D looks like their best low-light camera (well, plus the 1Dx/1Dc), having a newer generation sensor than the 5DmkIII and is also a fair bit cheaper.
I suspect Panasonic will need something in the way of new sensor technology to catch up with the 5DmkII, the 4x size difference is a lot and the 5DmkII sensor has really pretty low read noise for big pixels (3.2 electrons on 64,600 electron capacity pixels).
Zackary Deal You can catch up a fair bit. Can I get a bit mathematical? The 5DmkII converts 33% of the incoming photons to electrons, the pixels can hold 64,600 electrons and have 3.2 electrons read noise.
Let's say Panasonic get to 66% efficiency on a 16MP GH5 sensor, which is within the realms as most are over 50% now, plus they can clearly get to lower noise as their GM-1 sensor has 1.7 electron read noise, so not impossible to head towards 5DmkII performance. I think they can get close.
However not close to the then-current FF tech of course, just the old stuff, which the OP was talking about.
Went and bought 2 rokinon f1.4 lens for my gh4 to really match the 70-200 f2.8 lens on full frame after this video thanks tony :)
Hi Tony, after agreeing to most of the points you have made about actual ISO, Aperture & Focal lengths I have a question, what happens to the f-stop number when I use a Nikon 50 mm f/1.8 lens on my Nikon 1 system using the glassless FT1 adaptor, multiply 1.8 by 2.7x, then what about the T-stop of the lens? is that to be also multiplied by the crop factor??? a little help needed !
Tony, does this also apply when putting a full-frame camera into "crop mode?" Does the then cropped image suffer a loss in effective aperture the same way as if it were on a cropped body to begin with?
Charles Ingraham Yes, the effect is the same.
I can't agree with the idea about total light. Examples in the video makes sense but when you bring sensors with *same area but different resolution*, the story changes. According to the video, because D800 receives the same amount of light as a D4, they should have similar ISO performance. But the D4 is cleaner. This brings us to the point that *per unit area light intensity* is actually what governs noise. D4 has less, but larger pixels so at a given time they can gather more light than each pixel on D800. So they don't have to increase their amplifier gain as much as the D800 to get similar exposure. That results in cleaner images. You could argue that *re-sizing the image* and averaging out the noise will give them similar SNR. And that's true but that defeats the purpose of high resolution and added detail altogether! Anyway, no disrespect intended towards anyone. I like this channel but in this video I had a different opinion so I had to share my views.
ADZCreations Watch the Crop Factor Part 3 video for a discussion on pixel density. There's no overall correlation between noise and pixel density; many cameras have more of one and less of the other.
This is pretty poisonous for new photographers who don't yet understand the concept that these are simply equivalents and we are not being cheated.
I am a physicist and I've been telling friends and people arround me exactly the same thing for years. BUT: They often just don't really understand it. They solely look on the label on their camera and believe, what the manufacturer is telling them.
Thank you so much for sharing this video. This is just basic education, and if somebody won't believe me in future, I will refer to your video! It's very good!
If you are a physicist, you probably know that the aperture figure written on lenses is a relative quantity (relative aperture = real aperture / focal length).
What manufacturers say is correct.
Anne Hodermi The (photographic) aperture of a lens is defined as N=(focal length)/(lens diameter). And I think you really missunderstood the video. Tony is requesting a better and more relevant measurement to compare different camera systems. A 2.8 lens gathers just more light on a full frame sensor, than on a 4/3 sensor, because N=(focal length)/(lens diameter) stays the same. I think the aperture is used for comparison, but it's not meant to compare if you have different sensor sizes. It is true, that the number stays the same, but that is just not relevant for photography. And that is what T.N. says in his video.
pratakk No, a large sensor does not gather more light (per surface) than a smaller one. And that's why the conversion is irrelevant.
Normally large sensor (from a same generation) gathers light better (less electronic noise) but not more light. Full frame lenses are generally larger and have also better optical specification (less distortion). These are the only reasons why FF camera give better result than smaller sensor.
Anne Hodermi Think about it again with a very simple approach: a 2.8 lens stays a 2.8 lens. We already discussed that above. (For example 200mm, f2.8 means an aperture opening of 71mm).
But what's the difference between the 2.8 lens for full format and micro 4.3? Well, that's very easy as well:
- to cover the full frame sensor completely with light, you'll need a big front glass element
- to cover a smaller sensor completely with light, you can use a smaller front glass element. That's why a f2.8 lens for micro 4/3 can be smaller.
If you compare these two lenses, the only difference is the light collecting ability. The bigger f2.8 full frame lens gathers just more light through the front glass element and the total amount of light gathered is bigger than on the 4/3 sensor.
Another example: You can see, what happens if you put a big 2.8 lens for full format on a aps-c sized sensor here: upload.wikimedia.org/wikipedia/commons/f/f5/Full-frame_vs_APS-C.svg
What do you think happens to the light, that doesn't hit the smaller sensor - well, it is lost. And the bigger sensor gatheres this light.
pratakk The light that doesn't hit the small sensor is not 'lost' because the small sensor just don't need it (why should the sensor need the light that is outside it). The intensity of light on the small(er) surface is the same as on the larger one (of course, because it is the 'same light' - see your drawing). And it is intensity that matters to have pixel transforming effectively into quality electronic signal.
Angle of view are different but amount of light is exactly the same.
When you put a lens made for APS-C on Full frame, the same amount (=quantity per surface) and quality of light goes through the lens to hit the center of the sensor as if you put the lens on a APC-S camera (because the optical quality of the lens obviously doesn't change because you change the body).
An iphone has a 4,89 x 3,67 mm sensor (factor 7.3). It is a f/2.2.
Applying a conversion would lead to what? f/16?
If you ever used an iphone, you can see that you don't need to go to very low speed or very high iso to take normal picture. Why?
Because the same intensity of light hits the sensor (as for a "FF f/2.2").
One part I don't understand is why you had to change the ISO on the FF to 3200 at @17:35. If the FF equivalent of f/2.8 is f/5.6 on the crop sensor, why don't they give the same luminance at f/2.8 and f/5.6?
Hello Tony, thanks for this clear video. However, what happens when I mount a FF lens on a crop camera, e.g. the Canon "nifty fifty" f/1.8 on a 7D. Focal length will be 80mmFF but how about the DoF and total light gathered? Can you - or anyone else - shed (gathered) light into this? Thanks
Found the answer in one of your other videos:
ruclips.net/video/YDbUIfB5YUc/видео.html
oh man i feel so cheated! Wish i saw this a month ago before i ordered my Sony G 18-105 f4, for my A6300.. I knew about the Focal length crop, but the aperture? I'm pretty much a newbie, so i'm getting an F6 instead of an F4!! Damn, Thanks for the heads up..
Cool. Thanks for the info. I'm a bit more researched now so my initial comment was probably not completely accurate. I can also move further away and zoom in to get the same effect as compared to full frame, but i think "compared to full frame" is a mindset i should just eliminate from my thinking.
Yeah you totally should eliminate that thinking. Good photography does not depend on the sensor size in any way. And if you had a fullframe camera, would you then think about how it compares to medium format?
The maximum blur you can get with that is 105 mm / 4 = 26.25 mm i.e about one inch. This is measured on the focal plane. A thumb is one inch wide so if you take a photo of someone with his thumb up and there is light on the background far away the light will show same size as the thumb. Note that the distance does not matter.
I get the impression that there are 3 ranges of zoom lenses for all the manufacturers. Maybe it's a matter of common designs, maybe just marketing convention:
* Variable aperture (say 3.5 - 5.6) seem to be the kit/beginner lenses.
* Fixed aperture f/4 are enthusiast lenses, more affordable and more compact than the pro lenses
* Fixed aperture f/2.8 are the pro lenses. On larger format sensors these tend to be huge. Optical quality is really very good.
I use Micro 4/3, which has a "crop factor" of 2. So my 12-40 f/2.8 zoom could be said to be like owning a 24-80 f/5.6 on a full frame system. Sony sell a similar range at f/4 which is a stop faster than my equivalent and a similar size for a similar price. (On a 1.6 crop sensor camera the equivalent to my f/2.8 would be f/3.5)
The Olympus f/2.8 is a very nice lens though, and a full frame sensor does not eliminate noise. You can download sample RAW files from the web and explore this. The PRO quality lenses perform very well wide open, certainly sharp enough for my current sensor to resolve. So quite often I can shoot at f/2.8 (or f/1.8 on my primes) where someone else may be stopping down to the equivalent or smaller.
The large sensor camera can achieve more, especially when allowed to operate at its base ISO due to sufficient light for the desired shutter speed. (Roll on summer!). The full frame pro lenses are great and there are lovely photos out there taken on a D810 with bright light contrasty skin tones I couldn't hope for. But I don't have £10,000 to spend on a camera and don't want something that won't fit in the rucksack even with the other hiking gear removed.
Tony Northrup Thank you for starting a debate & educating people on issues about manufacturers cheating consumers,
on the other side, I think you are stressing too much to prove your point which is half true, the DOF is dependent on the sensor size, so multiplying the f number on a lens with crop factor is correct
however, the light intensity projected by a lens is not dependent on the size of the sensor, if you use a FF lens on a FF sensor the whole projection is used & if you use the same lens on 1.5x crop smaller portion of the projected light is used by the sensor but the intensity of light that falls on the crop sensor is still the same as what it would be on a FF sensor's equivalent area ......... intensity of light projected from a f/2.8 is always f/2.8 on any format sensor as long as the lens is used on a sensor that is equal to or smaller than the lens was originally designed for.............
about the inconsistency in ISO numbers between different sensor sizes that is an issue about quality & standards, if there is no regulation in the industry then consumers are going to be cheated...... I think you should stress more on pointing out the fact that manufacturers are using different standard ISO on different format/size sensors and cheating consumers!
please let me know if I am wrong..............
Purushottam Pokharel This addresses the light intensity issue: ruclips.net/video/0OtIiwbAZi8/видео.html
Thank you Tony for your response.......
Purushottam Pokharel yes light intensity for a sq inch is the same across diff sensor sizes (both with lens at same fstop. E.g. f2.8) what Tony is saying is the TOTAL light gathered over whole sensor isn't factored in.
This is where DOF and noise levels on smaller sensor cameras is misrepresented. And mis understood.
for a wildlife situation, full frame sensor needs 1.5x more focal length to get the same frame/resolution as a crop sensor would, on the other hand a crop sensor would need an extra stop of aperture opening to get same level of DOF & noise benefits of a FF sensor ???
So eventually one needs to invest more money for focal length if they have FF sensor & in case of not so fast but long lens one needs to invest more money for the FF camera body
All of Tony's Effort can't help save money for us :( (bird photographers)
if a narrow FOV is your main concern/priority, then yes, sure, a crop sensor is better (and cheaper)for you. (as long as you don't mind the extra noise of higher ISO using same lens on a crop S body.) but Tony is talking about the same DOF and light gathering ability/noise levels. Those things are very important to a great many photographers.
I just finished a video that goes into the science of this subject, along with plenty of illustrations showing exactly what is happening inside the lens... maybe check it out if you are still uncertain about anything. (it explains the same facts, but goes into a bit more detail and demonstrates the principles with illustrations.... also i go into speedsters briefly too)
Wow so we're paying $1000 for a pretty slow lens lol shame on you panasonic.
Please where did you you get the 25mm divisor @21:02 from? Excellent presentation.
+Gbedu Malam You basically have to work backwards, because manufacturers don't provide a measurement of the iris.
+Tony Northrup OK I see by working backwards, you mean taking the value of the original lens' highest focal length and the advertised f-stop (aperture) and dividing to get the divisor i.e. Panasonic 12-35mm f/2.8, will be 35/2.8 = 12.5, and this then becomes the divisor used to check any other claim they may be making. Thanks.
Mr. Tony Northrup, i wonder... i have A6000 and sony SEL 50mm f1.8.
if someday i buy A7 and put my SEL 50mm f1.8 lens in there and using APS mode(APSC crop) in A7... ist the quality bokeh blur same as if i use fullframe lens 85mm? or not? thanks
All this is true, but you have to decide what is relevant - for YOU. If you're moving from one camera system to another or using 2 systems at once, then all this is essential knowledge and will ensure that you get the best results. However, if you're only shooting with one camera system, you only need to know how your own system perfroms with the ISO and aperture values that system employs - getting to grips with that alone will get you the best results. Stop comparing your one system to another - it's irrelevant.
You sir, are a great and professional guy. Love your videos. Keep it up
DepthOfMyBrain Thanks!
DoF is actually determined by Aperture and distance to subject.
Baba Wethu If you want to be technical, DoF is a factor of aperture and magnification (the relative size of the subject on the camera's sensor). Focal length and sensor size are factors in magnification, however.
Tony Northrup
Wrong. Magnification factor does not depend on sensor size. Depth of field depends on aperture and magnification (magnification depends on focal length and distance between camera and object). Please stop telling nonsense to people.
Caldiar23 Strange that you called him wrong when you're both effectively saying the same thing. The distance between the camera and object is indirectly affected by the sensor size, because if you wish to keep someone's head the same size on an image taken with an APSC camera and a FF camera with the same MP, you would increase the distance between camera and object with APSC.
cheng2006
I'm not saying the same. Magnification is not relative to sensor size. It has absolutely nothing to do with sensor size. AGAIN: Depth of field only depends on aperture and magnification ratio.
Tony Northrup How about checking the lens of the Panasonic FZ200 you mention. You will notice the correct numbers are printed on the lens. It's sad so many people believe this nonsense. As I said before, if you don't know how to use focal length then use angle of view. Focal length and numeric aperture can't be converted into anything. AND: You can't use different magnification ratios and same aperture between two shots and then claim the aperture not being correctly stated.
Excuse my original comment, I meant to say Aperture, focal length and distance to target.
very comprehensive. Thank you Tony, But I have a question, what will happen to the BOKEH if I use an apsc lens on a fullframe body. will the effects be multiplied as well or it will remain shallower because its full frame anyway?
This has to be one of the longest videos that keeps you entertained and informed all the way through. While some comments may bring up technical differences between sensors and megapixel counts, etc. I think the main idea is still totally valid and gives a good idea of what to expect of each sensor type and their respective optics. Thank you!
So, it took us more than 10 years since the appearance of the first APS-C sensor cameras and lenses for someone to come front and explain all the "sex, lies and video tapes". Thank you Tony for making it all clear, lucid and obvious.
While we all knew about the field of view change with crop factor sensors from the day one, most of us did not pay attention to the whole equation and all the variables: aperture and ISO. With your help, Tony, it should be understood by anyone interested enough.
In order to make things more convenient and for marketing purposes, of course, manufacturers retained the same measurement units when talking about different physical properties. And that caused misinterpretation that lasted until this video was published.
Let's be fair, though, and admit that that Panasonic Micro Four Thirds 12-35mm f2.8 lens is indeed f2.8 lens. Just not in Full Frame measurement units, but in Micro Four Thirds measurement units. And that's what makes all the difference. Numerical values are the same but physical implications and light gathering abilities are not. Not without applying the proper math.
2nd and 3rd grade algebra? i did learn that till 8th lololol
28:35 Oh god Tony this is so funny, I love your nerd ranty videos :)
Christopher O'Grady true
@Tony & Chelsea i recently bought a 70-300 4-5.6 full frame lens for my D7200 and am not getting sharp images. i use the option to censor crop an additional 1.3 so that makes my lens a 136.5mm to 585mm f/7.8-10.92. i wanna shoot wild life and am having to shoot at 1/1600sec ISO-1000 F/11 to get a proper exposure should i return the lens and just get the 70-300 nikon dx 4.5-5.6D vr2. or could the softness be the lens or my technique. plz help because this is very frustrating.
Hey Tony,
This video is been a good eye opener. No pun intended :)
As for camera manufacturers being "evil" and trying to cheat people well they have to sell, don't they?
The really good take-away for me at least as a newcomer to DSLR is that if you're stuck with a particular gear setup it sure helps to know its actual range. I personally am a Canon user, there's something about them making their own lenses that I always appreciated and their entry level kits are affordable and easy to use, but from what you explained in this video the same principles, actually no-brainers once you go through the whole explanation, make a whole lot of sense. Even if you got stuck with a pair of lenses and a camera body that's not full-frame with this info you can still adjust it to get the most of it. At least that's my two cents. It will sure help from now to remember that whatever's on the lens (f-stop numbers and focal lengths) to multiply by the unusual 1.6x crop factor that Canon seems to love. I knew ISO is the other piece to play with but now that I know to throw in the crop factor it sure makes sense to get grain-free pics.
Also I wanted to say the videos you post with various tutorials are really awesome, keep it up!! ;)
Finally after six years I've figured out your secret to success. You're Mr. Rogers for the Millennial generation.
2.8 is 2.8 on any lens. It's the lens NOT the sensor that is "2.8"
Is 2 lbs of feathers heavier than 2 lbs of rocks? Common sense people!
+MuertoInc this is all correct, and 100% consistent with the video. If you think it's inconsistent, you've missed the point.
Here's why people need to understand crop factor. 2 lbs of rocks *is* heavier than 2 kg of rocks. 100 mph *is* faster than 100 kph. If you don't use crop factor, you're comparing numbers in different units, and that's misleading.
+Tony Northrup Another way of saying this would be "If you took 2 pictures using a full-frame lens on an APS-C body and a full-frame body, you will not get the same picture if you used the same settings on the two different camera bodies." Correct?
+MuertoInc 2.8 is 2.8 on any camera, but depth of field, the thing that we notice, is different when sensor size is changed. But because depth of field is dependent on aperture, sensor size and focal length, it's hard to think of and use. So it's easier to use 'f/2.8 on a 35mm sensor" as a reference and then we can say that f/2.8 gives the same depth of field on a 35mm sensor as an f/1.9 lens on a 1.5x crop sensor and f/1.4 on a 2x crop sensor.
+MuertoInc the point is, that a small sensor can only make use of a small amount of the light, which is collected by the lens. Because this amount of the light is simply not captured by the sensor - it falls next to the sensor, because it is too small to cover the whole pool of light.
+Peter Schiffman That is the easy part. The actual point was how one would get a same picture.
P/(P^0.5) = P^0.5 so why not just say SNR = P^0.5?
This formula is BS ...Noise is inherent property of electronics used not square root of LIGHT !
Hi, I love your analysis. I have a problem. I am an eye surgeon and I record surgeries on Sony a6300 at 4k resolution. I need to crop the video by about 70% so that I can show only the central part or just the eye. My question is will a smaller sensor like MFT of Panasonic GH5 (which I intend to buy) instead of APS-C of Sony A6300 at 4 k will have lower signal noise ratio at same ISO considering that MFT sensor image will suffer lower cropping of video instead of the APS-C video.
So if I have a 1.5 crop sensor stated 300mm at aperture 5.6 and iso 1000, would the actual settings be 450mm at 8.4 and iso 1000, meaning I could put a 450 lens on a full frame at aperture 8.4 with ISO 1000 and have the same exposure, this would suggest that for the same focal length, a full frame is no better at gathering light through the lens than a crop sensor, maybe because the same amount of light is exposed to the sensor at a given "actual" focal length in a full frame or crop? If this is true then the only real advantage of the full frame at equivalent zoom, as far as low light is concerned, is its ability to cope with the higher gain of ISO, meaning even though it is the same light received onto the sensor as the crop at a given actual focal length, it is the ability to add and cope with higher ISO that would actually make it better at low light? I am new to photography so I may be getting this all wrong?
Things are not as different now from film as you would have us believe. 35mm film is grainier than 8x10, right? Wrong! Contact print them both and they'll have the same amount of grain. Likewise noise in digital images (all other things being equal). If you "enlarge" images from different sensor sizes the same percent (assuming they have the same image-processing system) they'll have the same amount of noise. Smaller sensors are noisier at the same ISO only because the image is enlarged more. It's not a matter of ISO but of magnification.
-
Depth of field is a function of four things: subject distance, aperture, focal length, and image magnification. Forget this nonsense about "35mm equivalent". That is meaningless with regard to depth of field, and so is ISO.
-
I'm not a believer in the idea that we need to shoot film to understand photography, but we do need to understand something about optics, and that is getting lost in the "equivalence" numbers and in ignoring magnification.
The silver halide crystals of a film of 35mm and other of 6x6 of 100ISO have the same size and give the same grain when making contacts; true.
The photodiodes (pixels) of a 20Mp FF (24x36) and another 1" sensor do not have the same size (surface).
If you compare the image of an FF and a 1" of 20Mp without magnification, the 1" has more noise because the signal /noise ratio of the photodiode is proportional to its surface. The noise is due to the "amplification" of the signal, not mainly the enlargement of the format.
What you say would be true if we compare sensors of different formats but proportional resolutions to maintain the same surface of the photodiodes.
Dear Tony,
Your conclusions are mathematically correct but the logic behind them is wrong.
The ISO standard was devised to provide for equal exposure, not equal picture quality. And yes, 400 ISO film is "noisier" (with larger grains, in fact) than 100 ISO film - in close similarity with the way that digital sensors behave. 8 mm film is much grainier than 36 mm film when viewed on the same screen. All this has also to do with the magnification of the viewed image - a factor that very few tests take into account.
Your definition of the term "total gathered light" is useless, no matter how sweet you try to talk us into it. When it comes to noise what really matters is the size of the photo sites (pixels) on the sensor, not the size of the sensor itself. Your "buckets filled by the rain" are relevant to the electronic wells of the pixels, not the entire sensors. It is true that larger sensors tend to have bigger pixels and hence gather more light / produce less noise. It is also true, however, that a 10 MP APS-C sensor will produce less noise than a 36 MP Full Frame sensor (when using the same technology).
You are right that equivalent focal lengths have misled quite a few people. The cited apertures, however, express the lenses' ability to transmit light, not the quality of the bokeh or the amount of DoF they produce. As I mentioned above, ISO, shutter speed and aperture values are only about exposure. You can change the DoF by changing the aperture but that's not the only factor that DoF depends on - you'll have to throw in the diameter of the image circle of the lens, the diameter of the circle of confusion (which depends on the size of the pixels on the sensor, not the size of the sensor itself), the magnification of the viewed image, etc. All this are very complex matters that cannot be expressed with a single figure and least of all - with the value of the aperture.
It is true that lenses with wider maximum apertures are bigger than those with smaller ones - provided that the lenses produce the same image circle. Things get much more complex when comparing lenses which produce image circles of different diameters - i.e. a 4/3 lens and a FF lens. The size of the lens also depends on its construction (number of optical elements, presence of auto focus (AF), type of AF motor, etc.) and it's not fair to say that a 4/3 F2.8 lens will be the same size as a FF (Full Frame) F5.6 lens.
By the way, the ISO conversion tables that you so generously provide are true only if the different sensors (FF, APS-C. 4/3, etc.) have the same pixel density (i.e. they are all 16 MP sensors, for example) and use the same technology (CCD, CMOS, CMOS-BSI, etc.).
By the way, again, the Metabones Speed Boosters reduce the image circle of the attached lens and thus decrease the aperture values.Which means that you aperture formula is incomplete.
Take care!
Chris
In the above text "same pixel density" should read as "same pixel count". Sorry for the mistake!
Interesting mental exercise, but very little real life value.
Fact is, for most people that take pictures with their full frame camera, they don't crank up the ISO to 800 just so it will be equivalent to ISO 100 on their CX camera.
On a full frame camera if I have sufficient light to do so, I will take a picture at ISO 100, and adjusting my APS-C camera to ISO 40 to get the equivalent shot is simply not an option.
If for instance I have a Nikon D610 (full frame camera), and a D7100 (APS-C Crop Frame camera) they both have a base ISO of 100. The cleanest shots I'll be able to take with either camera is at ISO 100. According to your own calculations, the only way I can get shots that are equivalent is to use ISO 40 on the D7100....which it's not capable of doing, so for all real life intents and purposes the APS-C sensor is indeed noisier. (Not really a myth at all)
The other thing is, in terms of the physics of noise, it's actually based on how many photons each individual photocell (pixel) can gather, not just the size of the sensor.
That's why you can take two sensors that are exactly the same size (such as a Nikon D3s and a Nikon D3x), both of which are full frame 35mm equivalent size sensors... and the difference in noise between them at the same ISO setting is huge.
Because the pixel pitch on the D3x is only 5.9 Microns (barely larger than some APS-C cameras) and the pixel pitch on the D3s is 8.4 Microns.
So, while your method of equalizing noise by using the crop factor gives a decent approximation, it doesn't truly describe the actual physics of what's happening nor does it account for the noise differences in cameras with the exact same form factor.
Similarly, calculating Depth of Field by multiplying the crop factor once again gives a close enough approximation to work for most people, but also does not accurately account for what's really happening.
The calculations for Depth of Field are based solely on 4 different variables:
1.) Focal Length of the lens
2.) Aperture used
3.) Distance between the subject and the focal plane of the camera
4.) The Circle of Confusion for the particular sensor being used.
Item #4 is the only variable that is affected by sensor size and crop factor....however, it is the most insignificant part of the equation. Changing the Circle of Confusion numbers from Full Frame to APS-C makes a very very small difference. As a matter of fact, the APS-C actually has a slightly narrower Depth of Field for the exact same Focal Length, Distance and Aperture.
For instance using a commonly available Depth of Field Calculator I can punch in the numbers for myself and see the Depth of field at 50mm f/2.8 from 5 feet away on a Nikon D7100 (APS-C crop sensor camera) compared to a D700 (Full Frame 35mm equivalent camera:
Same exact settings: 50mm f/2.8 from 5 feet:
D7100 = 4.01 inches
D700 = 6.02 inches
However, because of the crop factor they will not give the equivalent field of view.
To get the equivalent field of view I'd have to use a 75mm lens on the D700, so using the same calculator I get:
Same field of view settings:
D7100 @ 50mm f/2.8 from 5 feet = 4.01 inches
D700 @ 75mm f/2.8 from 5 feet = 2.62 inches
So, even though the smaller form factor because of it's smaller circle of confusion actually has slightly less Depth of Field.....because the Focal Length is a much bigger factor in the overall equation, to get equivalent fields of view the larger form factor produces the smaller Depth of Field.
Maranatha Photography It's had deep real-world value to us. We use all MFT cameras for our video work. Being familiar with FF results, these simple formulas allowed me to purchase lenses that would give me the noise and depth-of-field that I needed.
Some of your concerns, such as pixel density, are addressed here: ruclips.net/video/6Im4W_9blhY/видео.html. I have a new video coming out in the next few days that digs deeper into pixel density, with actual statistics, but my initial findings as discussed in that video were correct.
Re: using lower ISOs, that's addressed here to an extent: ruclips.net/video/aJFgDEJTIiI/видео.html
Note that "depth of field" has both a technical and real-world meaning. The technical term is what's calculated, but the real-world meaning is more, "how blurry is the background."
But to be technical, DoF is determined by only two things: the magnification of the subject and the aperture.
Maranatha Photography Oh, and I'll happily update my videos to correct any mistakes, but I need experimental evidence with actual photographic results proving that various aspects don't matter. You can create a video yourself, or if you don't have the resources, submit a detailed experiment for us to perform at sdp.io/mistake. We'll perform the experiment and make the results public.
I've made that offer to thousands of people, and so far, nobody has managed to prove the math wrong.
Tony Northrup
First off, thanks for taking the time to respond.
Fair enough.......My statement that it has very little real world value is an over-simplification and an over-statement of my case, as you've obviously demonstrated the real world value to you.
Also....Just to be clear, I was talking about Pixel Pitch.
Pixel Pitch and Pixel Density while closely related and directly correlated are not the same thing.
Yes, the number of pixels crammed onto a single sensor does indeed dictate how large each individual pixel can be.....it is indeed the actual size of the individual pixels that determines the full well capacity and exactly how many photons can be gathered (which directly determines the Signal to Noise ratio)
The example I already gave is something nobody can/will refute.
The D3s and D3x both have the exact same sized sensor but they both have very different noise/ISO characteristics.
That alone is all the evidence needed to show that a simple calculation using crop factor can't account for noise characteristics at certain ISOs.
I will watch the video you linked to about pixel density before commenting too much deeper about that topic. It's possible we are saying the same thing but using different terminology, so it's not fair to comment until seeing it.
As far as the Depth of Field. You are of course right that it boils down to two things, the magnification of the subject and the aperture.
However, there are several variables involved in describing the magnification of the subject, including the focal length the size of the circle of confusion, and the distance from the subject.
While your method does provide a reasonable approximation for most applications, it is an over-simplification and doesn't really accurately represent the relationship.
I will accept your challenge. I don't have the resources to put together quite as nice a video as you do, but I will indeed put together the detailed experiment and the mathematical evidence to prove it.
Thanks again for taking your time.
Rich Nicely Also it appears that when I logged in to youtube to respond, that for some reason youtube logged me in with my personal profile "Rich Nicely" rather than my business profile that I used for my original comment "Maranatha Photography".....there isn't any kind of conspiracy to hide my true identity or anything.....just wasn't paying close enough attention to what RUclips account I was logged in as
Tony Northrup Actually, I think I can show this by example without ever having to submit any kind of video or photographic evidence.
First off, can we at least agree that Depth of Field calculators that are based on formulas that physicists and optics researchers have agreed on for decades before either of us was ever born, are accurate?
If not, and we are dealing with flat earth type of stuff, where I have to prove basic physics, then I'm not sure it's worth the time.
If we are at least in agreement at this point.....
Here is a calculator that I've seen many prominent photographers use and recommend that's based on those same exact formulas that physicists have agreed on forever:
www "." dofmaster "." com/dofjs "." html
If you use that calculator it only takes a few examples to show what I'm talking about.
Let's say I have a Nikon D700 (Full Frame Camera) and a Panasonic Lumix DMC-GX1 . (MFT....Like you said in the video, the 2X crop factor makes the math easier)
If I use my GX1 at 100mm and f/2.8 from 5 feet away then the equivalent field of view on my D700 would be 200mm from 5 feet away.
According to the online calculator (and your welcome to use someone else's calculator if you think it's more accurate)
The GX1 at 100mm and f/2.8 from 5 feet yields a DoF of .73 inches
According to your method, I should get the same DoF with the D700 at 200mm f/5.6 from 5 feet away.
Let's see what the calculator says:
The D700 at 200mm f/5.6 from 5 feet yields a DoF of .67 inches.
OK, you say, that's only a difference of .06 inches.
But that's just one example. Let's say you are playing with wider lenses at closer distances
The GX1 at 10mm f/2.8 from 10 inches yields a DoF of 2.09 inches
The D700 at 20mm f/5.6 from 10 inches yields a DoF of 2.01 inches
.08 inches is a little more significant in this range.
So, while it's a close enough approximation for most purposes, it's not entirely accurate.
Crop factor affects Depth of Field simply because Crop factor forces you to use a shorter focal length on the same camera to get the same Field of View....and focal length directly affects Depth of Field.
The fact that multiplying the crop factor by the aperture gets you a very close approximation for purposes of Depth of Field is a happy coincidence.
Hi Tony!! I´m a new follower. Your videos are great! the image, the sound, and the most important, THE CONTENT!
I have a question. What about the old film lenses on new APS-C bodies? Is all the same?
Example: I have an old Ricoh XR 50mm 1.7 (great lens) on a Pentax K3II (APS-C).
I know that the focal length on this body is 75mm but is the F still at 1.7?
When you say that the depth of field is larger on APS-C is for APS-C lenses or for all lenses?
And the light transmision, is lower on APS-C bodies with film lenses or still the same?
Sorry if you talk about it in the video but I dont understand too well english and is not easy to me undertand at all.
Thank you!!
Does full trames that have a croo factor option, like nikons dx mode, have the same issue in crop mode? do you still need to apply that factor to those settings?
P/sqrt(P) = sqrt(P), the P in the denominator is unnecessary