Hi guys! I just uploaded a FOLLOW UP VIDEO, that handles some criticisms about this video. Also flags a couple of small mistakes i made within this vid. Link- ruclips.net/video/fCIWkqcb7FI/видео.html
I'm very glad that I learned something from your video. but pardon me. the DOF diagram illustration is showing in your vedio, from my point of view, after the rays refracted through the lens, the focus circle of further object should be on the Focus Plane's left . Because this point circle gives a narrow angle, thus, the rays though the convex lens can give a easy refraction. Therefore, the focus circle should be on the left of the Focus Plane. Hope I can get your feed back. cheers.
@@yongchaozhao8449 If you mean the magenta dots and lines that appear at about 8:43, you are correct. When the object point moves further from the lens, the image point moves closer to the lens. The video's author does know about the error, and corrected it in his follow-up video.
I have been a photographer for a 5 years now and do a lot of reading and watching on the subject. This is the first time I have had aperture explained to a level where I can understand the mechanics of how this works, not just the effect.
Pete B cool! That was my goal. To explain literally what’s happening in your lens as simply as I could. (But still covering a lot of information... it’s a tricky juggle) Be sure to watch my follow up vid. As I did make a couple of minor mistakes in this vid that I wanted to correct.
I am an engineer so always want to know the specifics of how something works. I didn’t review the FF vs Crop and ISO parts, as I am always fairly confident in that area. Will look out for the update.
Thank you for doing this video. It is hard to find knowledgeable sources for this on the internet. Most people have the “camera for dummies” version of this where it only goes as deep as “iso bigger=brighter” and “aperture smaller number = brighter.” So thank you for explaining the why of it instead of patronizing me with fairly common camera knowledge. 🙏🏻👏👏👏
One important thing I would add to the video though is the fact that when you shoot full frame, you need to get closer to your subject to get the same subject framing, with everything else being the same. So in your image comparison between the FF and the M43, if you moved in closer with the FF to match the framing of the M43, the DOF would be much smaller in the FF than the M43. This is where people realize that when they shoot bigger subjects like people, and they want an even shallower DOF WITH the same framing, a FF sensor suits them better. And yes, you could use a smaller f number on a lens with a cropped sensor to match the FF in this case, but if you are already shooting f1.8 on a cropped sensor, and want an ever shallower DOF, for the same framing, you will be hard pressed to find a lens f 1 or thereabouts. So the key is keeping the FRAMING the same when comparing FF and cropped with everything else being the same. Fantastic video though, I would have loved to see the differences when getting closer with the FF though.
Moving toward or away from the subject completely changes the field of view and introduces background compression or expansion into the equation. It will absolutely change the framing. If you compensate with a longer lens (multiplied by crop factor) on a full frame, you can achieve the actual exact same field of view as a cropped sensor. Crop sensors are ***not*** reaching further. They are just filling the frame with “more subject,” by cropping out the edges of the full frame equivalent. In the photos at 19:35, if you physically moved closer, you would change the relationship of the foreground and background subjects, changing the framing entirely. But what actually happens if you shoot from the same distance with the same lens, is the crop sensor is just a smaller piece of the larger picture. There are cases where either format can be advantageous, crop sensors do have a pixel density advantage which simulates “reach,” but you are correct that full frame does offer more flexibility with shallow depth of field and light intake assuming equivalent lenses.
Definitely! I have been shooting exclusively on the GH4, and have gotten a 'feel' over the years, between my native lenses and speedbooster and Canon glass.... but I've never been able to really grasp it, (or explain it to my colleagues who shoot on the FF cameras like the Ursa Mini) and certainly not to explain it to someone else (which is, in my opinion, the true measure of understanding something). Wonderful video, and a happy new sub from us!
I had to stop this video now. Not to tell you that "I don't understand what you are talking about", but to say that each time I am to some degree confused, you address the subject and get me right back on track. This has seriously! got to be one of the best informative videos I have ever seen. The hat is off to you, sir! Amazing
Excellent. This should (but likely won't) end all the confusion. Math doesn't lie. (And I secretly delight that this vindicates Tony Northrup, who has gotten way too much grief for being right). Thank you for taking the time to put this together!
@@biscuitsalive Not only did I like it, but I learned a lot from it. Been a working pro for over 30 years, but am still constantly learning things, and your video was so precise and clarifying. Your combination of science, math, and real-world examples was extraordinarily helpful. You're helping to clear the fog away!
I know this is a few years after this was first posted and am pleased to see someone get it correct, I have been so frustrated with the amount of posts claiming that sensor size changes depth of field, only subject distance, aperture and focal length change depth of field. I do however feel that there is confusion over the ‘circle of confusion’ and I do not see how it impacts the depth of field with sensor size. As you show in you diagram the point of focus is the size of a pixel and a sharp point, reducing sensor size but retaining the resolution creates the overlap onto other pixels. You state that this impacts depth of field, but I disagree, to me the circle of confusion will have the same effect across the whole sensor regardless of depth of field, it has an impact on the overall sharpness and not depth of field.
I was looking such a good explanation of dof for many years. Exceptional work. Many "photographers" fill up the RUclips servers with stupid videos and all of them say the same obvious things. Aperture , shuter, iso blahblahblah. You took it the core where it needs to be taken. At some point we need to learn why things happen and how in photography. Great work.
I prefer; " more in focus", and less in focus, as you are correct in one sense there is a thin line in "focus" there always could be theoretically a more shape image .... defined or limited by the medium used.. so in reality there is a " field of focus", which is based on an math equation... or ratio .. which after all is where we get f stop and t stop.. so so much nerdiness here LOVE IT
Great stuff, will check the follow up vid, but already a huge thanks for clarifying this complicated subject... been shooting stills for about 25 years (about 10 of those professionally) and only now I have at least a good understanding of the matters (and math) behind all of this! Thank you, Sir!
:D ... look out for a newer more in-depth version of this coming soon... covering ISO equiv and speedboosters etc with more practical tests to back up the maths. :) (I'm half way through it, but its taking a while as i need to keep finding very specific lenses to do it properly)
Very nice explanation. I have sent several M43 users to your video when they try to tell me their Olympus 300mm f/4 is a 600mm f/4. When I tell them it's a 600mm f/8 and they have to boost the heck out of their ISO to match FF...heads explode. The marketing departments of several camera companies feed off the misinformation out there. Small sensors always pay a price.
Yep. They are correct of course when they state their lens/ focal length/Fstops. It’s when they claim their full frame equivalent numbers they often lie on purpose. Some companies even show the full frame equivalence lens next to theirs in the adverts. Not pointing out they the image /depth of field is entirely different.
This is the best explanation I have seen on this topic. Nevertheless, I think it is a mistake to say we make the three adjustments and get equivalence. That's just about as wrong as all the other explanations I see online in one sense. Now, it's true that those three adjustments will make the images very similar. But much of the time it's not even possible to do that. Suppose I use a 85mm f1.2 lens on full frame. Those are common today. Where do you find a f0.6 lens? They don't exist. Let's use an f2.8 100mm lens on full frame. So we need a 50mm lens at f1.4 for mft, which is likely doable. But now, our full frame camera has native ISO of 64. How will we replicate that? And a terrible misunderstanding comes from saying we double the aperture for micro four thirds to get full frame equivalence. That is because aperture is a physical measurement, and using a crop factor multiplication causes another myth because people think that aperture is really halved. But if that were true, we would need to double our exposure. But exposure times are correct. So, while it's true that if we adjust all three of those factors, it's rarely possible to do it in places where it's important. And people often think that statements like "halve the focal length" or "double the exposure" are true stand alone. It is therefore important to explain that these changes are not true when taken one or two at a time and very often they are not achievable. My explanation is that there are ways to make similar images if we intentionally handicap one system or another, but there is not a general way to create equivalence. Sometimes it's the crop sensor with an advantage. Consider a 400mm mft prime lens at f2.8. It has a magnified image so that full frame camera will need a 800mm lens for equivalent field of view. Now it's the full frame lens that can't match because nobody makes an 800mm f4, and if someone did it would be so heavy and expensive that nobody would buy it. True, the bokeh would be different, and there would be more noise. But you could still take pictures in low light because the f2.8 gives real exposure times. Maybe the subject is so complex it can't be made simple, but I have seen a lot of misunderstanding caused by explanations which use all three factors. Don't get me wrong. Your explanation is more clear and correct than any other I have seen. And if we are careful, we can get a correct understanding. But I have seen similar explanations cause as many misunderstandings as the outright myths.
Its a real shame that theres content out there that has a fraction of the effort put into it, and get millions of views where this gets just 89,000. A real shame. Youve explained how the aperture works like none other.
Very nice and easy explanation. The main reason that people say that FF has more bokeh is because the lens for crop sensors are not fast enough to compensate for the crop factor. If they make 25mm f 0.9 with autofocus on MFT it will compete with the standard f 1.8. If they want to compete with the f 1.2 lenses they will have to do some serious engineering to create a lens that has an apperture of f 0.6 . But the speedboosters are a very good solution until they come up with super fast lenses for aps-c and MFT. Let's hope they stop the marketing BS with the crop factors. Best solution I can find to clear the misinformation is to list the focal length as they are but to include the equivalence for the crop factor.
This is the clearest, most concise explanation of this subject I have seen. The key elements are succinctly outlined with easy to understand with illustrated explanations. The pacing is appropriate as well without getting overly bogged down in any of the steps. I've seen many videos and read many articles on this, but this video encapsulates these factors the best. Thank you. If I were to offer any suggestion, then I would recommend: 1) Using a larger colored cursor and moving it much less (the excessive cursor movement is very distracting) 2) For the text heavy slides, I would bring in each text section of the slide separately to help lead the watcher digest the information, instead of overwhelming the watcher all at once. With that said, the video is free and offered as a service, so I can't complain too much. I'm very appreciative for your effort and the time you took to produce it. The video is great educational service to anyone interested in the optics of cameras.
anglewinder thanks! And I agree on your criticisms. I did this video in one take/recording after drawing the illustrations. I should have spent longer on assembly. But I tend not to put lots of work into RUclips when I’m busy with paid work. RUclips is a fun hobby for me. Nothing more. :) But if I have a quiet work period, you will see spikes in quality on my RUclips content, simply as I then have more time to spend.
p.s. if you were getting the rest of the equivalence right, and just not the ISO, then you are still doing better than the majority of people I chat to on the Facebook groups... :) They just do the focal length, get bad results, then say the sensor it crap! (doofasus!)
I think these discussions are much clearer if we focus - no pun intended - on the lens rather than the sensor. The light gathering power of a camera, and its ability to render a shallow depth of field, are determined by the area of the lens’ entrance aperture, and not the sensor size at all. For a fixed depth of field, smaller sensor camera operate at numerically smaller f-stops and iso values - the total amount of light hitting the sensor is independent of the sensor size - and the image noise is typically very similar.
Great, now that I got all that, I will get mad at people that say wrong things about it, even professional photographers on RUclips and such 😑. Your fault. Haha seriously now, thanks a lot for that video!! I actually just came across it, wasn‘t searching for it ... but DOF in relation to crop sensors was definitely something I came a cross with a lot! So happy to understand it now! :)
It gets more obvious, when we consider an even smaller sensor, like the 1/28" sensor of an iPhone 15 pro. The main camera is a 24 mm FF equivalent f1.8 lens, there is the first problem: manufacturers call out the equivalent focal length but not the equivalent aperture. With a crop factor of about 3.5 the iPhone has a 7 mm f1.8 which is equivalent to a 24 mm f6.3. When we set a full frame camera with a 24 mm lens to f6.3 we get the same field of view and the same depth of field as the iPhone. We also have to adjust for exposure: if we have ISO 1600 on the iPhone, we would need ISO 19600 on the FF camera given the same shutter speed. There we see that the best we can do with the iPhone is a shot at f6.3 and ISO 19600 on full frame, but with the FF camera we can open the aperture depending which lens we have to 1.6 for example which would give us 4 stops extra for exposure, we now could either increase the shutter speed by 4 stops, or lower the ISO 4 stops, which will give us ISO 1225, or divide the 4 stops differently. We can't do that with the iPhone, because we would need a 7 mm f0.45 to archive the same with our FF camera at f1.6, f0.5 is the physical limit for any lens, so this lens would be impossible to manufacture, the practical limit is more like f0.7. In conclusion crop sensors appear to have bad light capabilities, but the truth is that the smaller sensor would require lenses that are unobtainable, to get around that the only thing we can do is to increase the focal length, and if we do that we also need to increase the sensor size accordingly. If you want to get the same shot (same depth of field, same field of view) sensor size doesn't matter so much but if you want any shot (aperture wide open, maximum shutter speed, lowest ISO possible) the FF camera will always be better, because you can utilise the higher focal length to achieve a larger aperture and because of that better exposure settings.
I've always liked that analogy about a marching army walking from a hard surface to a soft surface. I think you can take it a step further though. If you imagine each line of people is a different wavelength of light, the higher the wavelength the smaller and faster the gate of those marchers. So all the marchers are marching at the same speed, but on one side you have people taking large slow steps, and on the other side, people taking small fast steps. When they reach the sand, or mud, the people taking big steps get further into the mud before they are slowed down by it, and those taking small steps don't get as far before they are slowed down. That causes the marching lines to diverge from each other. I theory they should begin to converge on the other side of the "lens" but I think that could help explain chromatic aberration.
Your information at 20:35 is spot on! I was JUST about to do a video on this. Everyone explains how to calculate the focal length as far as different sensors go, but they ALWAYS forget one factor. Distance. You don't need to make equivalencies if the phone ISN'T composed the same way :)
Rowan Lacey really appreciate that. With such a contentious and potentially confusing topic, positive comments/feedback really are useful to hear. Then I know I made it all nice and clear. :)
Thank you for creating a wonderfully explained and illustrated video on this topic and doing a Follow-Up video. I'm working on a similar video to create more awareness around what makes the IMAX look and why it's not as unique as it was in its digital infancy.
Excellent video! Great explanation. I found your video looking to more explanations or discussion after the Tony's videos. I simply don't understand why detractors didn't understand or are confused with such basic concepts for a serious photographer. Other interesting aspect is that nobody have provide a single image to show the contrary. Great job.
Claudio Rivera thanks! Yes it's crazy why people still argue this stuff. Especially when it's not difficult to spend a hour or so playing with a couple of cameras and see practical results that Back up the theory fairly easily. As I said in my vid I learnt a fair bit from people like Tony N myself, and I wanted to take it a bit further and look at what's happening inside the lens with my little Illustrations. And it's always good to practically test these things yourself to understand things better.
Very well explained. I knew a lot of the theory, but didn't know why or how. I think the only subject I didn't know about previously was circles of confusion. Your diagrams really helped. I watched a Tony Northrup video on the same subject recently. What he says is basically the same as you, but he doesn't explain why or how. To do this, the diagrams are essential, which is why your video is a better resource. The subject gets very technical and there's a lot to take in. Some people are very keen on the technical stuff purely for the sake of being technical, whereas I just want to have enough knowledge to help me take photos. I now need to watch the follow up video.
Amazing video. Nice to see someone doing more than just scratching the surface of the subject, but at the same time making it easy to understand what he's saying. Subscribing.
At 17:00, the difference between sensor size accounting for the shallower dof on a smaller sensor has nothing to do with pixel size. The difference is because for a relevant comparison (same sized output) the smaller sensor output is magnified more, ergo it demands a smaller coc on the sensor to attain the same output size/coc relationship. Pixel size has nothing to do with any of this, nothing at all. Magnification, and the fact that you need to magnify a smaller sensor output more than a larger sensor has everything to do with it.
Scott Ferris I sort of agree with you. It's not about the pixels really, but I was using a few pixels simply as an example as to why the CoC could be considered OUT of focus on one sensor, and IN focus on another (even with exactly the same lens, same aperture, same distance to subject. ) Let's forget resolution differences between the sensors, and say we are filming 4K video on both sensors. (More like how I would use them in practice anyway) The 4K on the crop sensor obviously has narrower tolerances for what is acceptably in focus ,and what is acceptably out of focus. Due to the CoC. So that's why I used the pixels to explain the difference. (Personally I thought it was a nice clear way to show this. But obviously you disagree.) Anyway I don't think you are arguing with the overall point I was making at all. You just don't like the way I expressed my point.
Of course you agree, my comment is accurate :-) I understand what you were doing, the problem I saw was you never mentioned magnification differences, which is absolutely key to understanding the differences we see, and presented that issue in the terms of pixel size, that will lead to confusion and inaccurate understanding of your otherwise excellent video. The site that explains it all in as much detail as you could ever wish for is here www.josephjamesphotography.com/equivalence/ I am definitely not arguing with your point, just hoping you could add an addendum to point out that critical element of the equation. The magnification 'thing' is so fundamental to photography, and is never talked about, how many people acknowledge or understand that when you crop an image and view at the same size you lose DOF? Once you understand that getting f1.0 dof in your portraits isn't difficult or expensive, just maths :-)
Scott Ferris I think the fact I shrunk the sensor down does clearly show the magnification. (Just sort of inverted) Anyway. Glad you agree in principle. Certainly there are ways I could have described some of this better. I really should have wrote a script. But decided just to reasearch the facts for a few days. Then start recording. :) Clearly not the best way, but it's the quickest. (One take wonder)
Scott Ferris I may make a "proper" version. With tighter structure and fully animated illustrations. (Less power point like :D ) If I have time of course. Probably a weeks work to do a high production value version. So I'm calling this one the alpha version. :)
To be sure, the only reason I made an account and posted for the first time ever is because you specifically ask people to at the end of your video. Not emphasizing the fundamental nature and impact that magnification has in an equivalence explanation is an omission of some note. I applaud your video, and your need for tea at the end :-) but wish, very much, you had included that element.
Very good explanation, thank you for clearing up some misconceptions. The most contentious issue for most will be the noise/iso relationship which will upset a lot of m4/3 and some aps-c owners. You need very much faster lenses to keep the sensor noise under control. As you say, manufacturers are realising that they have to be more honest about lens equivalence not only focal length but also max aperture. Camera and lens manufacturers should make speed boosters that are specific to their larger format lens ranges. I.e. Fujishould make one specifically for mounting full frame legacy Fuji lenses on their aps-c camera bodies.
Howard Maryon-Davis I agree. And the point about the noise vs sensor size has the most variables to factor in. (In regards to each sensor technology/design etc) But I believe it’s safe to say if you took two identical sensors (in terms of the technology, eg. Back lit/cmos/ Same res etc etc) but one is smaller than the other. Then you can use the crop factor squared to equate noise levels. (across the whole image. Bare in mind if res between sensors is different, and you judge image noise at 100% then the formula brakes down as the portion of total Sensor is no longer equal. ) This is why judging 2 x 4k full width sensor images is a good/fair test to reference. As even if one sensor is higher res. The down sampling to 4k helps even things up. (As long as using whole width of each sensor and not cropping like the gh4 does)
Thank you so much man ! You explained it so well accompanied with perfect examples on the go. There’s so much confusion about this stuff but when you think about it from the mathematical point of view, it all starts to make sense. Cheers dude! Keep up the good work.
I am now pointing all my friends that ask me questions about this to your video. This is probably the best lesson on it out of every technical video on RUclips. Thanks so much for taking the time to make it :)
Thank Brian. I may well re do this vid to perfect the message (couple of tiny mistakes) but i think it does a pretty good job of combining everything people need to know.
It does a great job. The depth of field discussion alone makes it more than worthwhile. You may not realize this but in the midst of your physics discussion you showed how depth of field works as a tool for the aesthetics of a given photo. Until aperture is seriously understood like this you can end up fiddling around with 100s of photos before you understand it by accident...LOL. The sooner you realize DoF the quicker you can leverage it in your photography. I was one on the people who spent months fiddling with it without this level of understanding. After much trial and error combined with piecing together bits and pieces from RUclips that is actually self-containted in your video, I finally understood it a few years ago. It's so refreshing to have your video on it that it should be posted everywhere. I'm just glad PetaPixel pointed everyone to it. Bravo sir ;)
High praise indeed. thank you. I make no bones about the fact I learnt much of this from filmmaker IQ and Tony Northrup. But I wanted to try to combine all of those elements, and also help it be more visually literate with the simple illustrations. (For those of us that don't think in math equations) :)
I always think about DOF as subject magnification (distance to subject and focal length) plus aperture. For instance, macro has very low depth of field because subject magnification is higher than if you were say, fitting an entire elephant on the sensor. If you are fitting an entire mountain range on the sensor, subject magnification is very low and depth of field is correspondingly very high, no matter the aperture.
Thanks for taking the time for putting this together, got my first DSLR recently (I’ve only ever used compacts) and really disappointed the sales guy didn’t explain about the cropped sensor vs full frame, and, I can’t understand why Canon won’t label the EFS series of Lenses what they actually are considering they can’t be used on the FF, after watching this video and Tony Northrup I’m now getting better exposure with my pics
rawknroll Paul glad it was useful. Yep a system that shows the FOV angle and the total amount of light let through would be much better than focal length and fstop
50mm (for a full frame lens) is roughly a "sweet spot" in terms of front element size. Assuming you keep the aperture the same, at shorter focal lengths you need a larger front element to accommodate the wider field of view, and at longer focal lengths you need a larger front element to accommodate the larger physical aperture size. Physics is fun! :)
Well, smaller pixels do not give shallower depth of field, because the circle of confusion stays the same. We just use more smaller pixels to show the same size of that circle. I mean if the circle of confusion has diameter ~0.02 mm, it doesn't matter how many pixels we use to show that point (1 pixel with diameter 0.02 mm or 10 pixels with diameter 0.002 mm).
Sameir Ali thank you Sameir. Which mistakes? (Want to add annotations for any to keep the wolves at bay.) I misspoke a couple of tiny times. (Saying complete when I meant compete near end etc.) But I think it's still fairly clear what I meant. But please do point out any specific mistakes so I can amend.
Sameir Ali and I would redo some of the illustrations slightly. (Apple too close to end of lens, and this muddied water for certain points.) Also, the DOF area (yellow dotted line) should be stretching out so more area for "acceptably" in focus behind focus plane then in front. But again that doesn't change the core message... but I should have factored that in.
All the mistakes are acceptable and understandable. By the way, I think you said about square of 200 is 800, that needs to be corrected. What you said about the crop factor was right. So, the ISO is not be be squared. It should be multiplied with 4.
Sameir Ali no what I said was we square the crop factor for the ISO (I will re watch to check I said it correctly later... but certainly that's how I wrote it down for the illustration) So our crop factor for the M43 to FF is X2 So if you get your calculator out and type out, 200 times 2 then press the squared button. This gives us 800 If it was APSC it would be X1.5 (canon is 1.6) So 200 times 1.5 squared = 450 So ISO200 on a apsc cam. And ISO450 on a FF would equal the same total light. And have equivalent noise.
and finally for those new like myself I use the highlight in focus area feature, different names per manufacturer. and then adjust the aperture and watch the field change... Its easier now after a year with this to mentally imagine a wider or narrower dof.... good luck all...
I get it, but I'm trying to find out why my Canon 80D with 18-135 kit lens doesn't have a big depth of field when I have zoomed out. I was suspicious of it not doing this over the last year, so yesterday I did a DOF test on a long row of railings posts, so I compress them by zooming out. I first stood close to them and did around a 60mm focal length, I focussed a third of the distance in to the shot and tried at f8, f11, f16, f20. I know you get refraction with this lens after f11, but it was a DOF test and I still couldn't get everything in focus front to back. It did improve, but still the near and far posts were blurred. I then went around 24ft away from the end post I was stood by and zoomed to around 125 or 130mm and still had the same problem. I thought at least by f11 or f13 everything would be in focus. Is this a fault of this lens, or is this normal, do I have to crop photos every time, or am I doing something wrong. Does this happen with every lens on a Canon crop sensor, because at the moment I can get bigger DOF on my old bridge camera with a tiny sensor, zoomed out at f8. One reason I went to a DSLR was for its much bigger range of aperture values and being able to get everything in focus when I wanted that.
The ISO calculation was really confusing for me in the beginning, however when I thought in stops it all made sense! Since I have to mathematically double the aperture, so from 2 to 4, in stops: 2 stops of light. I now also have to bump up the ISO by 2 stops, which is from ISO 200 to 800. At first I was like "Wait that makes no sense!" but then it all cleared up! Really nice video! Thanks a lot!
Bastian Dittrich thanks dude. Yes everyone thinks differently. But glad you agree, just do it in your head with stops rather than squaring ISO numbers etc. (Hopefully others will spot your comment as it's useful for other camera heads that always think in stops of light.) The main thing to take away is that the ISO number doesn't tell you the total light creating our image. (Total photons hitting whole sensor for shutter duration) So if we can match that total light quantity across diff sized sensors, then the rest of the equation all falls into place perfectly.
It's actually pretty easy to understand once you get the hang of it! The basic principle has to get into your head, once that's done it's pretty easy. Yeah I'm the camera head kinda guy, calculating in stops has become second nature to me...
biscuitsalive Thank you for that! You know what still some people dont believe but at the same time they dont give any evidence just noice. I think you know what kind of childish behavior I mean.
Thanks for the video, very informative. There is only one question that I don’t understand completely. That is how to achieve the same amount of exposure, DOF and NOISE. It is clear why you need to lower the ISO by the square of the crop factor - increasing aperture from 4 to 2 is exactly 2 stops, decreasing ISO from 800 to 200 is also 2 stops (the same would work if one were to decrease the exposure using SS). What I don’t understand is the fact that noise will be the same. If we have 2 sensors with the same amount of pixels, say iso 12800/22750 on M4.3/APS-C should definitely look cleaner than iso 51200 on FF. At least this is what I experienced with my Sony A6500 vs Sony A7IV. Could you please elaborate on this?
Great way of simplifying and therefor clearly explaining the differences and similarities in gear usage and the final images aimed at. Every pro should know. Thanks for sharing your knowledge about the subject!
Great video and excellent explanation. Wish you would have mentioned the lens compression too :). According my understanding it totally depends on the physical distance between the camera and the subject, therefore on an equivalent of 85mm (56mm or so) on crop body the compression would be same as on a real 85mm on full frame body. Please let me know if I am missing something.
Haider Abbas I was just thinking that last night! (That I should have mentioned compression) Yes, although I haven't tested for that personally yet. (Always like to test things physically before accepting anything fully.) I agree that it's just down to distance from subject. Makes sense and it's just the relationship between one object overlapping another from a line of sight. And the FOV of course. As you get further away from the subject. (Say an actor) but the FOV angle shrinks. (To keep framing the same on that actor) So we see less and less of the background scene. And the actor is also actually covering less and less of that background. (Not as important, but worth noting.) This obviously does the same thing to everything that is at difference distances from the camera. (This is why face distorts when using very wides lens, up close. And the same face looks flattened and more orthographic from a great distance with a very long lens. ) Anyway, good point you made. There's a good chance I will revisit this topic again in a future vid, as I have learnt a lot from all the feedback I got from everyone. Both about things I should have said. And ways to make the trickier parts even more clear. Not to mention I would like to add some raw independently gathered data that further proves some of the still disputed statements. (By a small but very vocal minority.) :)
You are the only source on the internet that I've found to explain these things in this amount of detail. Amazing! I was looking for exactly this for weeks!
Great video - but a bit wrong... I have access to a tones of cameras, formats and lenses - and I have tested everything in this video. Greatly used "match" but I have done the tests. A 50mm F1.4 - will get the same amount of light no matter what the sensor size is. The ISO - is so much dependant of sensor technology that I have done the tests only on base ISO. A camera with ISO 800 full frame from 10 years ago compared to a M43 release yesterday will get worse noise level anytime. Basically that bokeh and depth of field is the FEEL - of the image. For example.. if I want to imitate the "feel" of a image shot on a FF 50 1.8 - I need a APSC 35 F1.2 - this will result the same shallow dof and in focus areas - but being a 1.2 will also expose more than the FF 1.8 - on the same settings and base ISO. A 50 1.4 is a 50 1.4 and a 50 1.4 - no matter what sensor is has behind it, the same amount of light will land on the sensor - the image due to the math that can be replicated - that is just a crop story...
I just want to say thanks for the engagement on this video folks! Apart from a very 'shouty' minority on the forums. Most people get it and agree. I think the really clued up viewers already knew this stuff anyway, but I think also a fair chunk of you learned something useful. In which case Im glad the video did its job. And I would like to again remind you to check out Tony Northrups, and FilmmakerIQ channels. As I myself learnt a great deal from them, and hopefully wrapped it all up for you into this one vid for you.
I've seen Tony's take on this and the current video which largely corroborates the same points. It makes sense to me that total light captured is important since we are blowing up pictures to the same display medium (screen or print). It seems like common sense to assume that the light 'wasted' by the cropped sensor not capitalizing on the full image circle projected by the lens would incur a penalty since we are losing "optical resolution". One thing that does intrigue me about the Angry Photographer's take on this is that he insists it is wrong to assume that bigger sensors are like bigger solar panels and consequently no penalty actually exists for cropping (does this mean that the part of the image circle that is captured still yields vastly greater optical resolution than the photosites available to capture it and "wasting" light by cropping stills leaves plentiful optical resolution anyway? - I'm confused...). Angry's point is also that a lens does not care what sensor lies behind it, but I don't think either Tony or biscuits contradict that statement. It would be nice to have them openly debate about this. :-)
Mark Pitsilos it seems to me you get everything you need to know to understand all of the major points in the vid. I'm kind of too tired now tonight to give any any more feedback regarding the some of the minutiae side details. So il leave it there, and pick up any remaining loose threads tomorrow eve..... Night!
just skimming the comments again, regarding that angry guys take on it, I'm confused as hell by his comments, he keeps changing his mind as to exactly what is wrong with the points in my video. And once i realised he was unhinged i have now ceased all communication with him and blocked him on every platform he shouts and spits nonsense in. :) (one of the biggest nutters i have EVER met online. without a close second, he spammed me with so much crap you would not believe it.) but regarding the 'not working like solar panels' argument. This makes no sense to me. The only way his argument would hold any water at all is if, when a manufacturer scales up the sensor size from one type to another, they keep the actual photosites the same size, (now with each 'pixel' having a tiny gap around it). But obviously they don't do that, because they want their sensor to be as efficient as possible. so they cram as much light sensitive surface area into the sensor as possible, wether its big OR small. the signal to noise ratio is what its all about. just like in audio, if you have a weak input and add a ton of gain you get crappy results. (the noise floor is raised etc.) A image sensor simply needs as many photons as possible to give us that strong SNR. and when you make the blooming sensor smaller, it has less photons hitting it.... like, no sh*t sherlock. Even a child could see this as perfectly logical. And considering all of the DXOmark testing data, and all of the practical experience from anyone that uses several sensors sized cameras will tell you. the sensor size has a DIRECT relationship to image noise. As you probably already saw from TonyN vid- 'Crop Factor 4: Debate with a Critic & New Sensor Performance Data' The sensor size has way more sway on image noise than even technology improvements over a few years span. And even when you compare the very worst performing camera to the very best in a particular year. Size still wins that particular arm wrestle time and time again. So why he (angry fat dude camera guy theo or whatever his name is) can't see the wood for the trees despite everyone around him pointing to the mighty oaks all around him and telling him he's in a forest I have no idea. :)
For the consumer, the practical relevance of ISO, and the only one that really matters, is how clean an image is. However every manufacturer develops their own propriety software for processing signal into an image, every manufacturer develops their own sensors with their own methods of generating signal, and every manufacturer chooses their own system for assigning an ISO rating to the results of those systems. These are constantly being evolved. Comparing ISO ratings between different model cameras by the same manufacturer is meaningless, let alone comparing ISO ratings between different manufacturers.
I did catch a few mistakes (you had a few drawings with light bending in a way that would not really happen) but overall this is nicely done. When light passes through a lens element, the lens element does not "know" what is behind it. It doesn't know if it's a big sensor or small sensor. The physics of the lens has the luxury of ignoring the sensor size. Another way to think of it is if you have a slide projector and a projection screen, you can swap a large projection screen for a smaller projection screen. The "projector" does not "know" you did this. But of course the human... DOES know. So the human reacts to the change by adjusting the projector... either changing the physical distance so the image "fits" the screen... or perhaps the projector has a zoom lens and we can adjust the focal length. Basically the short of it is... the actual DoF doesn't _really_ change (strictly in terms of laws of physics)... but the human reaction to the new sensor size (to achieve an equivalent framing) causes the photographer to do things that DO impact the DoF.
hey, thanks, not sure if you watched the follow up vid too, i did flag a couple of silly mistakes in that. (my light ray vectors did go the wrong way at one point) but it doesn't change the overall message, just a silly mistake with my simple illustration. anyway, thanks for input.
thank you for your explanation, it helped a lot and i must say mft sensors are very very good, f. e. em1 II is likly even with a sony a7 iii sensor... but ... FOV + DOF + Aperture is what counts for me and i like the look of FF more. even though i have canon with less accurate sensor.
wow, wonderful! Just a question. The cropped Iso related to the sensor does it also involve the noise level? In other words, in terms of noise level 800FF is equal to the 200 into 4\3?
Great video. Only for ISO is correct but if m43 camera is 10 mpx and FF is 40 mpx that the ISO is the same (pixels per sq. inch). If the resolution is the same, you have to multiply by 4 (2 x 2). For that m43 limit now is 20 mpx (they can not support more for ISO)
The only thing you got wrong is how ISO is calculated. It is light falling on the whole sensor. Not only on a certain area. Otherwise exposure values wouldn't translate between sensor sizes. Not to mention film. Having shot more than a few film formats. 800iso medium format is the same sensitivity of 35mm and smaller films. On film its function of grain size.
You’ve got some of that a bit backwards. The ISO number represents gain added to the signal to boost the signal. (Which give the impression of increasing sensitivity) And the gain is directly proportional to the size of the sensor. Hence if you point camera at a grey wall, and cover part of the sensor. The rest of the sensor doesn’t make the exposed part brighter. Just as the StoR ratio remains equal across any part of the sensor. It’s easy to put this into practice. Just swap a f1.8 full frame lens for a f1.8 apsc lens on a full frame camera. The iso /exposure remains the same. And the image brightness remains the same. But we now have shadow across the outside of the sensor. So obviously there is less TOTAL light hitting than sensor. You see my point? If that sensor was a solar panel it would now be generating less electricity. Even though the light source remained constant. It’s the total amount of photos hitting the sensor is what matters to the image. Not the concentration per square inch.
Think of it this way. Every point in the focusing plane in the real world, is focused down to one point on the sensor. As you watch further that is illustrated
Firstly - thanks for the detailed explanation - 1 point about FF vs. APSC (or MC43), is the recent shift to cropped sensors with high count of MP, i.e., emerging 33MP APSC cameras rumored to come out the next few years - I would like to see this video updated to include the effect of such a shift (more noise - since less light per pixel, but maybe compensated since it makes a difference to the coverage of the CoC ? - And if so,is it pronounced enough to nullify the effect? -Under "low light"/ Normal Light ?
biscuitsalive: Thank you for your comprehensive and useful video. I knew about the DoF and aperture correlation, but with different sensore type it seems to be a complicated concept. So I felt this, but I’ve never put this together. Now i understand better how it works. I would have 2 questions, and i I know this information is in your video, but really just would like to make sure if i get this right. I am asking this from an APS sensor user point of view. For me having a nice background blur on a portrait image is the most important. 1, fuji fx has 1.5 crop factor. If I would like to have the same blur than I have with nikon FF 50mm f1.2, I would need 35mm f0.8. There is no such lens on the market. Right ? Only thing I can do that I use a longer lens and step backward. Am I right ? 2, ISO has nothing to do with other feature of the image except noise level, isn’t it ? I am just asking because I am an APS user now. In order to get similar nose than I got with nikon on 100 ISO, I would need 44 ISO on Fujifilm APS. Again, there is no such sensor. So I guess the main differences show at the “lower end” of the parameter scale. On the other hand there are noise level differences between different type and brand of sensors. So ISO 400 dosn’t result the same level of noise e.g. sony and fuij APS. The noise is not only matter of size. Right?
I think the main reason why these misconceptions exist is because most aps-c cameras dont have lenses that are as fast, relative to their sensor size, compared to lenses designed for full frame cameras. 35mm f0.95 aps-c lens is going to have, roughly the same results as a 50mm f1.4 full frame lens. I wish there were more ultra fast wide lenses for aps-c cameras like a 20mm, or 25mm f0.95!!! That would be AMAZING!!!
Reza Taghipour yep. Your not wrong. All you need is the same size aperture. (Physical size of hole, not f stop number) And things soon start evening off. That means lens have to be about the same size though. (Crop lenses same as FF lense) So, clearly it’s the lenses that are the weak point. And not really the sensor at all.
Theres a youtube explaining refraction, the analogy of soldiers marching from tarmac to sand, its not really whats happening, what is really happening is to do with interference between the light and the fields of the electrons in the glass molecules. Light is an electromagnetic wave, not that it matters and all that, and if I actually understood it, I could explain it, but the light (being part of the electromagnetic spectrum) interferes with the electromagnetic fields of electrons in the glass molecules. Anyway, it doesn't matter, I like your video.
Fantastic video! I enjoyed the follow up as well. I have a quick ? on the circle of confusion and how it effects the other metrics when you take into account when sensor sizes scale, the resolution often changes as well. For instance a GH5 vs an A7RIII.
TheKoop117 yes in this video I was mostly talking about same res at diff sensor sizes. (So 4k video as I mostly use) Higher the pixel density the smaller the circle of confusion becomes. But in terms of end product it still comes down to how it’s viewed. If it’s printed huge and viewed close the allergens depth of field will become narrower. And visa versa. So you do need to take into account the circle of confusion of the sensor/pixel density. AND the final viewing set up. (Tv screen at 3 meters... cinema at 20 meters... billboard at 50 meters etc)
Matthew White ha, well I realise some people just don't care about this technical shit, and just want to see my little Cornwall, kitesurf, mountain bike action vids. :) So get in there first before they hate on it.
![Depth of Field Explained: Ultimate Guide to Camera Focus [Shot List Ep. 4]](/img/1.gif)
Hi guys! I just uploaded a FOLLOW UP VIDEO, that handles some criticisms about this video.
Also flags a couple of small mistakes i made within this vid.
Link-
ruclips.net/video/fCIWkqcb7FI/видео.html
biscuitsalive nice videos, this is how you should learn photography! Please recommend books!
biscuitsalive
I'm very glad that I learned something from your video. but pardon me. the DOF diagram illustration is showing in your vedio, from my point of view, after the rays refracted through the lens, the focus circle of further object should be on the Focus Plane's left . Because this point circle gives a narrow angle, thus, the rays though the convex lens can give a easy refraction. Therefore, the focus circle should be on the left of the Focus Plane. Hope I can get your feed back. cheers.
@@yongchaozhao8449 If you mean the magenta dots and lines that appear at about 8:43, you are correct. When the object point moves further from the lens, the image point moves closer to the lens. The video's author does know about the error, and corrected it in his follow-up video.
holy crap! 4 years of film school, numerous shoots, a ton of bro camera science and only now i finally get it, thanks!
This is the most complete explanation of this yet, been trying to fully understand this for years. Thanks!
Billy nice one Billy! Glad it was useful.
I have been a photographer for a 5 years now and do a lot of reading and watching on the subject. This is the first time I have had aperture explained to a level where I can understand the mechanics of how this works, not just the effect.
Pete B cool! That was my goal. To explain literally what’s happening in your lens as simply as I could. (But still covering a lot of information... it’s a tricky juggle)
Be sure to watch my follow up vid.
As I did make a couple of minor mistakes in this vid that I wanted to correct.
I am an engineer so always want to know the specifics of how something works. I didn’t review the FF vs Crop and ISO parts, as I am always fairly confident in that area. Will look out for the update.
here is the link, its a bit boring, but i like to correct my mistakes if possible - ruclips.net/video/fCIWkqcb7FI/видео.html
Thank you for doing this video. It is hard to find knowledgeable sources for this on the internet. Most people have the “camera for dummies” version of this where it only goes as deep as “iso bigger=brighter” and “aperture smaller number = brighter.” So thank you for explaining the why of it instead of patronizing me with fairly common camera knowledge. 🙏🏻👏👏👏
This has swayed me to keep my m4\3 cameras and lenses. Great explanation!!
Matthew Peer I love both (m43 and FF)
Gh4 prob gets most use.
But a7s comes out at night. :)
Also playing with medium format now too.
Fun to tinker.
One important thing I would add to the video though is the fact that when you shoot full frame, you need to get closer to your subject to get the same subject framing, with everything else being the same. So in your image comparison between the FF and the M43, if you moved in closer with the FF to match the framing of the M43, the DOF would be much smaller in the FF than the M43. This is where people realize that when they shoot bigger subjects like people, and they want an even shallower DOF WITH the same framing, a FF sensor suits them better. And yes, you could use a smaller f number on a lens with a cropped sensor to match the FF in this case, but if you are already shooting f1.8 on a cropped sensor, and want an ever shallower DOF, for the same framing, you will be hard pressed to find a lens f 1 or thereabouts. So the key is keeping the FRAMING the same when comparing FF and cropped with everything else being the same. Fantastic video though, I would have loved to see the differences when getting closer with the FF though.
Moving toward or away from the subject completely changes the field of view and introduces background compression or expansion into the equation. It will absolutely change the framing. If you compensate with a longer lens (multiplied by crop factor) on a full frame, you can achieve the actual exact same field of view as a cropped sensor. Crop sensors are ***not*** reaching further. They are just filling the frame with “more subject,” by cropping out the edges of the full frame equivalent. In the photos at 19:35, if you physically moved closer, you would change the relationship of the foreground and background subjects, changing the framing entirely. But what actually happens if you shoot from the same distance with the same lens, is the crop sensor is just a smaller piece of the larger picture. There are cases where either format can be advantageous, crop sensors do have a pixel density advantage which simulates “reach,” but you are correct that full frame does offer more flexibility with shallow depth of field and light intake assuming equivalent lenses.
Reminds me of the science class on light reflection back in middle school! Thank you so much.
Thank you so much for this, I've been trying to wrap my head around it for years.
Jimber Jam what a perfect comment. Thank you. Let's me know I achieved what I set out to do, and made it nice and clear. :) :) :)
Definitely! I have been shooting exclusively on the GH4, and have gotten a 'feel' over the years, between my native lenses and speedbooster and Canon glass.... but I've never been able to really grasp it, (or explain it to my colleagues who shoot on the FF cameras like the Ursa Mini) and certainly not to explain it to someone else (which is, in my opinion, the true measure of understanding something).
Wonderful video, and a happy new sub from us!
Thanks! I will have a good nose at your uploads later on tonight. (I subbed, so will catch your new ups too)
:| Don't go back too far! Haha!
best explanation and diagrams of all these concepts that I have found on YT so far. thanks big time.
I had to stop this video now. Not to tell you that "I don't understand what you are talking about", but to say that each time I am to some degree confused, you address the subject and get me right back on track. This has seriously! got to be one of the best informative videos I have ever seen. The hat is off to you, sir! Amazing
Excellent. This should (but likely won't) end all the confusion. Math doesn't lie. (And I secretly delight that this vindicates Tony Northrup, who has gotten way too much grief for being right). Thank you for taking the time to put this together!
Peter Lewis glad you liked it. :)
@@biscuitsalive Not only did I like it, but I learned a lot from it. Been a working pro for over 30 years, but am still constantly learning things, and your video was so precise and clarifying. Your combination of science, math, and real-world examples was extraordinarily helpful. You're helping to clear the fog away!
Peter Lewis nice to hear!
This was, put simply, AMAZING! Thanks!!!
Damn, that soldiers analogy... I wonder how physically accurate is it considering the wave form of the light ray. Great video by the way!
My jaw has dropped to the floor. This answers so many questions. I feel like sleeping after having had a long exam.
Well done.
I have a hard time getting my head around something until I understand it intellectually.
Thanks for that.
I know this is a few years after this was first posted and am pleased to see someone get it correct, I have been so frustrated with the amount of posts claiming that sensor size changes depth of field, only subject distance, aperture and focal length change depth of field. I do however feel that there is confusion over the ‘circle of confusion’ and I do not see how it impacts the depth of field with sensor size. As you show in you diagram the point of focus is the size of a pixel and a sharp point, reducing sensor size but retaining the resolution creates the overlap onto other pixels. You state that this impacts depth of field, but I disagree, to me the circle of confusion will have the same effect across the whole sensor regardless of depth of field, it has an impact on the overall sharpness and not depth of field.
Simply brilliant video. No fancy bullshit and down to the point.
Thanks for the feedback, be sure to check the follow up, couple minor mistakes in my illustration which i dress there.
I was looking such a good explanation of dof for many years. Exceptional work. Many "photographers" fill up the RUclips servers with stupid videos and all of them say the same obvious things. Aperture , shuter, iso blahblahblah. You took it the core where it needs to be taken. At some point we need to learn why things happen and how in photography. Great work.
I prefer; " more in focus", and less in focus, as you are correct in one sense there is a thin line in "focus" there always could be theoretically a more shape image .... defined or limited by the medium used.. so in reality there is a " field of focus", which is based on an math equation... or ratio .. which after all is where we get f stop and t stop.. so so much nerdiness here LOVE IT
Thanks i learned a lot from this video! Thank you!
Great stuff, will check the follow up vid, but already a huge thanks for clarifying this complicated subject... been shooting stills for about 25 years (about 10 of those professionally) and only now I have at least a good understanding of the matters (and math) behind all of this! Thank you, Sir!
4 years of doubts were explained today. thanks a lot!!! great work
:D ... look out for a newer more in-depth version of this coming soon... covering ISO equiv and speedboosters etc with more practical tests to back up the maths. :)
(I'm half way through it, but its taking a while as i need to keep finding very specific lenses to do it properly)
Very nice explanation. I have sent several M43 users to your video when they try to tell me their Olympus 300mm f/4 is a 600mm f/4. When I tell them it's a 600mm f/8 and they have to boost the heck out of their ISO to match FF...heads explode. The marketing departments of several camera companies feed off the misinformation out there. Small sensors always pay a price.
Yep. They are correct of course when they state their lens/ focal length/Fstops. It’s when they claim their full frame equivalent numbers they often lie on purpose.
Some companies even show the full frame equivalence lens next to theirs in the adverts.
Not pointing out they the image /depth of field is entirely different.
@@biscuitsalive Why the Engineering depts. always hate the Marketing depts... :P
This is the best explanation I have seen on this topic. Nevertheless, I think it is a mistake to say we make the three adjustments and get equivalence. That's just about as wrong as all the other explanations I see online in one sense. Now, it's true that those three adjustments will make the images very similar. But much of the time it's not even possible to do that. Suppose I use a 85mm f1.2 lens on full frame. Those are common today. Where do you find a f0.6 lens? They don't exist. Let's use an f2.8 100mm lens on full frame. So we need a 50mm lens at f1.4 for mft, which is likely doable. But now, our full frame camera has native ISO of 64. How will we replicate that? And a terrible misunderstanding comes from saying we double the aperture for micro four thirds to get full frame equivalence. That is because aperture is a physical measurement, and using a crop factor multiplication causes another myth because people think that aperture is really halved. But if that were true, we would need to double our exposure. But exposure times are correct. So, while it's true that if we adjust all three of those factors, it's rarely possible to do it in places where it's important. And people often think that statements like "halve the focal length" or "double the exposure" are true stand alone. It is therefore important to explain that these changes are not true when taken one or two at a time and very often they are not achievable. My explanation is that there are ways to make similar images if we intentionally handicap one system or another, but there is not a general way to create equivalence. Sometimes it's the crop sensor with an advantage. Consider a 400mm mft prime lens at f2.8. It has a magnified image so that full frame camera will need a 800mm lens for equivalent field of view. Now it's the full frame lens that can't match because nobody makes an 800mm f4, and if someone did it would be so heavy and expensive that nobody would buy it. True, the bokeh would be different, and there would be more noise. But you could still take pictures in low light because the f2.8 gives real exposure times. Maybe the subject is so complex it can't be made simple, but I have seen a lot of misunderstanding caused by explanations which use all three factors. Don't get me wrong. Your explanation is more clear and correct than any other I have seen. And if we are careful, we can get a correct understanding. But I have seen similar explanations cause as many misunderstandings as the outright myths.
Its a real shame that theres content out there that has a fraction of the effort put into it, and get millions of views where this gets just 89,000. A real shame. Youve explained how the aperture works like none other.
I have searched for this knowledge for a very long time. Thank you very much!
Very nice and easy explanation. The main reason that people say that FF has more bokeh is because the lens for crop sensors are not fast enough to compensate for the crop factor. If they make 25mm f 0.9 with autofocus on MFT it will compete with the standard f 1.8. If they want to compete with the f 1.2 lenses they will have to do some serious engineering to create a lens that has an apperture of f 0.6 .
But the speedboosters are a very good solution until they come up with super fast lenses for aps-c and MFT.
Let's hope they stop the marketing BS with the crop factors. Best solution I can find to clear the misinformation is to list the focal length as they are but to include the equivalence for the crop factor.
This is the clearest, most concise explanation of this subject I have seen. The key elements are succinctly outlined with easy to understand with illustrated explanations. The pacing is appropriate as well without getting overly bogged down in any of the steps. I've seen many videos and read many articles on this, but this video encapsulates these factors the best. Thank you. If I were to offer any suggestion, then I would recommend: 1) Using a larger colored cursor and moving it much less (the excessive cursor movement is very distracting) 2) For the text heavy slides, I would bring in each text section of the slide separately to help lead the watcher digest the information, instead of overwhelming the watcher all at once. With that said, the video is free and offered as a service, so I can't complain too much. I'm very appreciative for your effort and the time you took to produce it. The video is great educational service to anyone interested in the optics of cameras.
anglewinder thanks! And I agree on your criticisms.
I did this video in one take/recording after drawing the illustrations.
I should have spent longer on assembly.
But I tend not to put lots of work into RUclips when I’m busy with paid work.
RUclips is a fun hobby for me. Nothing more. :)
But if I have a quiet work period, you will see spikes in quality on my RUclips content, simply as I then have more time to spend.
Thanks for doing this video! I've done some FF vs M43 DOF tests and didn't know why the ISO was so different for the same exposure, until now.
Nice one Louis, exactly the kind of thing i like to hear. :)
(Lets me know I am being helpful)
p.s. if you were getting the rest of the equivalence right, and just not the ISO, then you are still doing better than the majority of people I chat to on the Facebook groups... :) They just do the focal length, get bad results, then say the sensor it crap! (doofasus!)
Thank you for this video. You nailed it. Thank you for mentioning Filmmaker IQ. It's great. Thank you for the reference.
Thank you very much for this video. You have opened my eyes and brain.
I think these discussions are much clearer if we focus - no pun intended - on the lens rather than the sensor. The light gathering power of a camera, and its ability to render a shallow depth of field, are determined by the area of the lens’ entrance aperture, and not the sensor size at all. For a fixed depth of field, smaller sensor camera operate at numerically smaller f-stops and iso values - the total amount of light hitting the sensor is independent of the sensor size - and the image noise is typically very similar.
Great, now that I got all that, I will get mad at people that say wrong things about it, even professional photographers on RUclips and such 😑. Your fault.
Haha seriously now, thanks a lot for that video!! I actually just came across it, wasn‘t searching for it ... but DOF in relation to crop sensors was definitely something I came a cross with a lot! So happy to understand it now! :)
Fabio Mota glad it was useful. :)
It gets more obvious, when we consider an even smaller sensor, like the 1/28" sensor of an iPhone 15 pro. The main camera is a 24 mm FF equivalent f1.8 lens, there is the first problem: manufacturers call out the equivalent focal length but not the equivalent aperture. With a crop factor of about 3.5 the iPhone has a 7 mm f1.8 which is equivalent to a 24 mm f6.3. When we set a full frame camera with a 24 mm lens to f6.3 we get the same field of view and the same depth of field as the iPhone. We also have to adjust for exposure: if we have ISO 1600 on the iPhone, we would need ISO 19600 on the FF camera given the same shutter speed. There we see that the best we can do with the iPhone is a shot at f6.3 and ISO 19600 on full frame, but with the FF camera we can open the aperture depending which lens we have to 1.6 for example which would give us 4 stops extra for exposure, we now could either increase the shutter speed by 4 stops, or lower the ISO 4 stops, which will give us ISO 1225, or divide the 4 stops differently. We can't do that with the iPhone, because we would need a 7 mm f0.45 to archive the same with our FF camera at f1.6, f0.5 is the physical limit for any lens, so this lens would be impossible to manufacture, the practical limit is more like f0.7.
In conclusion crop sensors appear to have bad light capabilities, but the truth is that the smaller sensor would require lenses that are unobtainable, to get around that the only thing we can do is to increase the focal length, and if we do that we also need to increase the sensor size accordingly. If you want to get the same shot (same depth of field, same field of view) sensor size doesn't matter so much but if you want any shot (aperture wide open, maximum shutter speed, lowest ISO possible) the FF camera will always be better, because you can utilise the higher focal length to achieve a larger aperture and because of that better exposure settings.
I've always liked that analogy about a marching army walking from a hard surface to a soft surface. I think you can take it a step further though. If you imagine each line of people is a different wavelength of light, the higher the wavelength the smaller and faster the gate of those marchers. So all the marchers are marching at the same speed, but on one side you have people taking large slow steps, and on the other side, people taking small fast steps. When they reach the sand, or mud, the people taking big steps get further into the mud before they are slowed down by it, and those taking small steps don't get as far before they are slowed down. That causes the marching lines to diverge from each other. I theory they should begin to converge on the other side of the "lens" but I think that could help explain chromatic aberration.
Grant Pluntze nice one! Yep that works lovely. I could see that working in a animation. :)
Grant Pluntze Dec
Your information at 20:35 is spot on! I was JUST about to do a video on this. Everyone explains how to calculate the focal length as far as different sensors go, but they ALWAYS forget one factor. Distance. You don't need to make equivalencies if the phone ISN'T composed the same way :)
You had me at the opening disclaimer. Haha.
Very useful video, thanks! Learned a bunch of stuff I'd no idea about before :)
Rowan Lacey really appreciate that.
With such a contentious and potentially confusing topic, positive comments/feedback really are useful to hear. Then I know I made it all nice and clear. :)
Thank you for creating a wonderfully explained and illustrated video on this topic and doing a Follow-Up video. I'm working on a similar video to create more awareness around what makes the IMAX look and why it's not as unique as it was in its digital infancy.
Thank you so much for this. It is so accurate.
Excellent video! Great explanation. I found your video looking to more explanations or discussion after the Tony's videos. I simply don't understand why detractors didn't understand or are confused with such basic concepts for a serious photographer. Other interesting aspect is that nobody have provide a single image to show the contrary. Great job.
Claudio Rivera thanks! Yes it's crazy why people still argue this stuff. Especially when it's not difficult to spend a hour or so playing with a couple of cameras and see practical results that
Back up the theory fairly easily.
As I said in my vid I learnt a fair bit from people like Tony N myself, and I wanted to take it a bit further and look at what's happening inside the lens with my little
Illustrations. And it's always good to practically test these things yourself to understand things better.
I watched the mistake film first. A very good real world set of examples.
robbyboyo cheers Robby.
Very well explained. I knew a lot of the theory, but didn't know why or how. I think the only subject I didn't know about previously was circles of confusion. Your diagrams really helped. I watched a Tony Northrup video on the same subject recently. What he says is basically the same as you, but he doesn't explain why or how. To do this, the diagrams are essential, which is why your video is a better resource. The subject gets very technical and there's a lot to take in. Some people are very keen on the technical stuff purely for the sake of being technical, whereas I just want to have enough knowledge to help me take photos. I now need to watch the follow up video.
Amazing video. Nice to see someone doing more than just scratching the surface of the subject, but at the same time making it easy to understand what he's saying. Subscribing.
Nice to hear, thanks.
@@biscuitsalive It's nice to see what's written, or just to read. Hearing that is something else altogether.
At 17:00, the difference between sensor size accounting for the shallower dof on a smaller sensor has nothing to do with pixel size. The difference is because for a relevant comparison (same sized output) the smaller sensor output is magnified more, ergo it demands a smaller coc on the sensor to attain the same output size/coc relationship.
Pixel size has nothing to do with any of this, nothing at all. Magnification, and the fact that you need to magnify a smaller sensor output more than a larger sensor has everything to do with it.
Scott Ferris I sort of agree with you.
It's not about the pixels really, but
I was using a few pixels simply as an example as to why the CoC could be considered OUT of focus on one sensor, and IN focus on another (even with exactly the same lens, same aperture, same distance to subject. )
Let's forget resolution differences between the sensors, and say we are filming 4K video on both sensors. (More like how I would use them in practice anyway)
The 4K on the crop sensor obviously has narrower tolerances for what is acceptably in focus ,and what is acceptably out of focus.
Due to the CoC.
So that's why I used the pixels to explain the difference. (Personally I thought it was a nice clear way to show this. But obviously you disagree.)
Anyway I don't think you are arguing with the overall point I was making at all. You just don't like the way I expressed my point.
Of course you agree, my comment is accurate :-) I understand what you were doing, the problem I saw was you never mentioned magnification differences, which is absolutely key to understanding the differences we see, and presented that issue in the terms of pixel size, that will lead to confusion and inaccurate understanding of your otherwise excellent video.
The site that explains it all in as much detail as you could ever wish for is here www.josephjamesphotography.com/equivalence/
I am definitely not arguing with your point, just hoping you could add an addendum to point out that critical element of the equation. The magnification 'thing' is so fundamental to photography, and is never talked about, how many people acknowledge or understand that when you crop an image and view at the same size you lose DOF? Once you understand that getting f1.0 dof in your portraits isn't difficult or expensive, just maths :-)
Scott Ferris I think the fact I shrunk the sensor down does clearly show the magnification. (Just sort of inverted)
Anyway. Glad you agree in principle.
Certainly there are ways I could have described some of this better.
I really should have wrote a script.
But decided just to reasearch the facts for a few days. Then start recording. :)
Clearly not the best way, but it's the quickest.
(One take wonder)
Scott Ferris I may make a "proper" version. With tighter structure and fully animated illustrations.
(Less power point like :D )
If I have time of course. Probably a weeks work to do a high production value version.
So I'm calling this one the alpha version. :)
To be sure, the only reason I made an account and posted for the first time ever is because you specifically ask people to at the end of your video. Not emphasizing the fundamental nature and impact that magnification has in an equivalence explanation is an omission of some note.
I applaud your video, and your need for tea at the end :-) but wish, very much, you had included that element.
Very good explanation, thank you for clearing up some misconceptions. The most contentious issue for most will be the noise/iso relationship which will upset a lot of m4/3 and some aps-c owners. You need very much faster lenses to keep the sensor noise under control. As you say, manufacturers are realising that they have to be more honest about lens equivalence not only focal length but also max aperture. Camera and lens manufacturers should make speed boosters that are specific to their larger format lens ranges. I.e. Fujishould make one specifically for mounting full frame legacy Fuji lenses on their aps-c camera bodies.
Howard Maryon-Davis I agree.
And the point about the noise vs sensor size has the most variables to factor in. (In regards to each sensor technology/design etc)
But I believe it’s safe to say if you took two identical sensors (in terms of the technology, eg. Back lit/cmos/ Same res etc etc) but one is smaller than the other. Then you can use the crop factor squared to equate noise levels. (across the whole image. Bare in mind if res between sensors is different, and you judge image noise at 100% then the formula brakes down as the portion of total
Sensor is no longer equal. )
This is why judging 2 x 4k full width sensor images is a good/fair test to reference.
As even if one sensor is higher res. The down sampling to 4k helps even things up. (As long as using whole width of each sensor and not cropping like the gh4 does)
Thank you so much man ! You explained it so well accompanied with perfect examples on the go. There’s so much confusion about this stuff but when you think about it from the mathematical point of view, it all starts to make sense. Cheers dude! Keep up the good work.
What matters is entrance pupil of a lens between lens formats to compare equivalences
I am now pointing all my friends that ask me questions about this to your video. This is probably the best lesson on it out of every technical video on RUclips. Thanks so much for taking the time to make it :)
Thank Brian. I may well re do this vid to perfect the message (couple of tiny mistakes) but i think it does a pretty good job of combining everything people need to know.
It does a great job. The depth of field discussion alone makes it more than worthwhile. You may not realize this but in the midst of your physics discussion you showed how depth of field works as a tool for the aesthetics of a given photo. Until aperture is seriously understood like this you can end up fiddling around with 100s of photos before you understand it by accident...LOL. The sooner you realize DoF the quicker you can leverage it in your photography. I was one on the people who spent months fiddling with it without this level of understanding. After much trial and error combined with piecing together bits and pieces from RUclips that is actually self-containted in your video, I finally understood it a few years ago. It's so refreshing to have your video on it that it should be posted everywhere. I'm just glad PetaPixel pointed everyone to it. Bravo sir ;)
High praise indeed. thank you.
I make no bones about the fact I learnt much of this from filmmaker IQ and Tony Northrup. But I wanted to try to combine all of those elements, and also help it be more visually literate with the simple illustrations.
(For those of us that don't think in math equations) :)
I always think about DOF as subject magnification (distance to subject and focal length) plus aperture. For instance, macro has very low depth of field because subject magnification is higher than if you were say, fitting an entire elephant on the sensor. If you are fitting an entire mountain range on the sensor, subject magnification is very low and depth of field is correspondingly very high, no matter the aperture.
Thanks for taking the time for putting this together, got my first DSLR recently (I’ve only ever used compacts) and really disappointed the sales guy didn’t explain about the cropped sensor vs full frame, and, I can’t understand why Canon won’t label the EFS series of Lenses what they actually are considering they can’t be used on the FF, after watching this video and Tony Northrup I’m now getting better exposure with my pics
rawknroll Paul glad it was useful.
Yep a system that shows the FOV angle and the total amount of light let through would be much better than focal length and fstop
50mm (for a full frame lens) is roughly a "sweet spot" in terms of front element size. Assuming you keep the aperture the same, at shorter focal lengths you need a larger front element to accommodate the wider field of view, and at longer focal lengths you need a larger front element to accommodate the larger physical aperture size.
Physics is fun! :)
This is one of the best explanation video I have ever watched. Nice job and thank you !
glad it helped :)
Brilliant video very well explained and easy to understand and grasp
This is amazing! You cleared a bunch of confusions!
Kang Du glad it helped. Lot to digest in there.
Very, very interesting. Thank you.
Well, smaller pixels do not give shallower depth of field, because the circle of confusion stays the same. We just use more smaller pixels to show the same size of that circle. I mean if the circle of confusion has diameter ~0.02 mm, it doesn't matter how many pixels we use to show that point (1 pixel with diameter 0.02 mm or 10 pixels with diameter 0.002 mm).
Thank you so much...
Mistakes are spotted :).
But, thank you so much for this detailed explanation.
Sameir Ali thank you Sameir.
Which mistakes? (Want to add annotations for any to keep the wolves at bay.)
I misspoke a couple of tiny times. (Saying complete when I meant compete near end etc.)
But I think it's still fairly clear what I meant.
But please do point out any specific mistakes so I can amend.
Sameir Ali and I would redo some of the illustrations slightly. (Apple too close to end of lens, and this muddied water for certain points.)
Also, the DOF area (yellow dotted line) should be stretching out so more area for "acceptably" in focus behind focus plane then in front.
But again that doesn't change the core message... but I should have factored that in.
All the mistakes are acceptable and understandable.
By the way, I think you said about square of 200 is 800, that needs to be corrected. What you said about the crop factor was right. So, the ISO is not be be squared. It should be multiplied with 4.
I have subscribed to the channel, and keep looking for more information.
Keep up the good work.
Sameir Ali no what I said was we square the crop factor for the ISO (I will re watch to check I said it correctly later... but certainly that's how I wrote it down for the illustration)
So our crop factor for the M43 to FF is X2
So if you get your calculator out and type out,
200 times 2 then press the squared button.
This gives us 800
If it was APSC it would be X1.5 (canon is 1.6)
So 200 times 1.5 squared = 450
So ISO200 on a apsc cam. And ISO450 on a FF would equal the same total light. And have equivalent noise.
and finally for those new like myself I use the highlight in focus area feature, different names per manufacturer. and then adjust the aperture and watch the field change...
Its easier now after a year with this to mentally imagine a wider or narrower dof.... good luck all...
I get it, but I'm trying to find out why my Canon 80D with 18-135 kit lens doesn't have a big depth of field when I have zoomed out. I was suspicious of it not doing this over the last year, so yesterday I did a DOF test on a long row of railings posts, so I compress them by zooming out. I first stood close to them and did around a 60mm focal length, I focussed a third of the distance in to the shot and tried at f8, f11, f16, f20. I know you get refraction with this lens after f11, but it was a DOF test and I still couldn't get everything in focus front to back. It did improve, but still the near and far posts were blurred. I then went around 24ft away from the end post I was stood by and zoomed to around 125 or 130mm and still had the same problem. I thought at least by f11 or f13 everything would be in focus. Is this a fault of this lens, or is this normal, do I have to crop photos every time, or am I doing something wrong. Does this happen with every lens on a Canon crop sensor, because at the moment I can get bigger DOF on my old bridge camera with a tiny sensor, zoomed out at f8. One reason I went to a DSLR was for its much bigger range of aperture values and being able to get everything in focus when I wanted that.
Incredible thank you.
Very clear, worked for me!
Franco Salomón (thumbs up icon)
Hats off bro! Nice and concise. Keep it up. Need more info like this.
Finally, a diagram at 13:38 that explains what's going on!!
The ISO calculation was really confusing for me in the beginning, however when I thought in stops it all made sense!
Since I have to mathematically double the aperture, so from 2 to 4, in stops: 2 stops of light. I now also have to bump up the ISO by 2 stops, which is from ISO 200 to 800.
At first I was like "Wait that makes no sense!" but then it all cleared up!
Really nice video! Thanks a lot!
Bastian Dittrich thanks dude.
Yes everyone thinks differently. But glad you agree, just do it in your head with stops rather than squaring ISO numbers etc. (Hopefully others will spot your comment as it's useful for other camera heads that always think in stops of light.)
The main thing to take away is that the ISO number doesn't tell you the total light creating our image. (Total photons hitting whole sensor for shutter duration)
So if we can match that total light quantity across diff sized sensors, then the rest of the equation all falls into place perfectly.
It's actually pretty easy to understand once you get the hang of it! The basic principle has to get into your head, once that's done it's pretty easy.
Yeah I'm the camera head kinda guy, calculating in stops has become second nature to me...
Very nice! A lot of work you have done respect!
Some Thing thanks! I got fed up with arguing with people in camera groups about this stuff so thought I would try to clearly explain everything here.
biscuitsalive Thank you for that! You know what still some people dont believe but at the same time they dont give any evidence just noice. I think you know what kind of childish behavior I mean.
best vid i've seen on this so far. great job!
Thanks William, glad it ticked a few boxes for you :)
Thanks for the video, very informative. There is only one question that I don’t understand completely. That is how to achieve the same amount of exposure, DOF and NOISE. It is clear why you need to lower the ISO by the square of the crop factor - increasing aperture from 4 to 2 is exactly 2 stops, decreasing ISO from 800 to 200 is also 2 stops (the same would work if one were to decrease the exposure using SS). What I don’t understand is the fact that noise will be the same. If we have 2 sensors with the same amount of pixels, say iso 12800/22750 on M4.3/APS-C should definitely look cleaner than iso 51200 on FF. At least this is what I experienced with my Sony A6500 vs Sony A7IV. Could you please elaborate on this?
Great way of simplifying and therefor clearly explaining the differences and similarities in gear usage and the final images aimed at. Every pro should know. Thanks for sharing your knowledge about the subject!
Beeldproducent - film-video-foto I tip my hat to you sir!
So good
Great video and excellent explanation. Wish you would have mentioned the lens compression too :). According my understanding it totally depends on the physical distance between the camera and the subject, therefore on an equivalent of 85mm (56mm or so) on crop body the compression would be same as on a real 85mm on full frame body. Please let me know if I am missing something.
Haider Abbas I was just thinking that last night!
(That I should have mentioned compression)
Yes, although I haven't tested for that personally yet. (Always like to test things physically before accepting anything fully.) I agree that it's just down to distance from subject.
Makes sense and it's just the relationship between one object overlapping another from a line of sight. And the FOV of course.
As you get further away from the subject.
(Say an actor) but the FOV angle shrinks. (To keep framing the same on that actor)
So we see less and less of the background scene. And the actor is also actually covering less and less of that background. (Not as important, but worth noting.)
This obviously does the same thing to everything that is at difference distances from the camera.
(This is why face distorts when using very wides lens, up close. And the same face looks flattened and more orthographic from a great distance with a very long lens. )
Anyway, good point you made.
There's a good chance I will revisit this topic again in a future vid, as I have learnt a lot from all the feedback I got from everyone.
Both about things I should have said.
And ways to make the trickier parts even more clear.
Not to mention I would like to add some raw independently gathered data that further proves some of the still disputed statements. (By a small but very vocal minority.)
:)
Great video. Very clear and teaching. Thanks.
Glad it was helpful :)
You are the only source on the internet that I've found to explain these things in this amount of detail. Amazing! I was looking for exactly this for weeks!
Great video - but a bit wrong...
I have access to a tones of cameras, formats and lenses - and I have tested everything in this video. Greatly used "match" but I have done the tests. A 50mm F1.4 - will get the same amount of light no matter what the sensor size is. The ISO - is so much dependant of sensor technology that I have done the tests only on base ISO. A camera with ISO 800 full frame from 10 years ago compared to a M43 release yesterday will get worse noise level anytime. Basically that bokeh and depth of field is the FEEL - of the image. For example.. if I want to imitate the "feel" of a image shot on a FF 50 1.8 - I need a APSC 35 F1.2 - this will result the same shallow dof and in focus areas - but being a 1.2 will also expose more than the FF 1.8 - on the same settings and base ISO. A 50 1.4 is a 50 1.4 and a 50 1.4 - no matter what sensor is has behind it, the same amount of light will land on the sensor - the image due to the math that can be replicated - that is just a crop story...
I just want to say thanks for the engagement on this video folks! Apart from a very 'shouty' minority on the forums. Most people get it and agree.
I think the really clued up viewers already knew this stuff anyway, but I think also a fair chunk of you learned something useful. In which case Im glad the video did its job.
And I would like to again remind you to check out Tony Northrups, and FilmmakerIQ channels.
As I myself learnt a great deal from them, and hopefully wrapped it all up for you into this one vid for you.
I've seen Tony's take on this and the current video which largely corroborates the same points.
It makes sense to me that total light captured is important since we are blowing up pictures to the same display medium (screen or print).
It seems like common sense to assume that the light 'wasted' by the cropped sensor not capitalizing on the full image circle projected by the lens would incur a penalty since we are losing "optical resolution".
One thing that does intrigue me about the Angry Photographer's take on this is that he insists it is wrong to assume that bigger sensors are like bigger solar panels and consequently no penalty actually exists for cropping (does this mean that the part of the image circle that is captured still yields vastly greater optical resolution than the photosites available to capture it and "wasting" light by cropping stills leaves plentiful optical resolution anyway? - I'm confused...).
Angry's point is also that a lens does not care what sensor lies behind it, but I don't think either Tony or biscuits contradict that statement.
It would be nice to have them openly debate about this. :-)
Mark Pitsilos it seems to me you get everything you need to know to understand all of the major points in the vid. I'm kind of too tired now tonight to give any any more feedback regarding the some of the minutiae side details.
So il leave it there, and pick up any remaining loose threads tomorrow eve..... Night!
just skimming the comments again, regarding that angry guys take on it, I'm confused as hell by his comments, he keeps changing his mind as to exactly what is wrong with the points in my video.
And once i realised he was unhinged i have now ceased all communication with him and blocked him on every platform he shouts and spits nonsense in. :)
(one of the biggest nutters i have EVER met online. without a close second, he spammed me with so much crap you would not believe it.)
but regarding the 'not working like solar panels' argument.
This makes no sense to me.
The only way his argument would hold any water at all is if, when a manufacturer scales up the sensor size from one type to another, they keep the actual photosites the same size, (now with each 'pixel' having a tiny gap around it). But obviously they don't do that, because they want their sensor to be as efficient as possible. so they cram as much light sensitive surface area into the sensor as possible, wether its big OR small.
the signal to noise ratio is what its all about.
just like in audio, if you have a weak input and add a ton of gain you get crappy results. (the noise floor is raised etc.)
A image sensor simply needs as many photons as possible to give us that strong SNR.
and when you make the blooming sensor smaller, it has less photons hitting it.... like, no sh*t sherlock.
Even a child could see this as perfectly logical.
And considering all of the DXOmark testing data, and all of the practical experience from anyone that uses several sensors sized cameras will tell you. the sensor size has a DIRECT relationship to image noise.
As you probably already saw from TonyN vid-
'Crop Factor 4: Debate with a Critic & New Sensor Performance Data'
The sensor size has way more sway on image noise than even technology improvements over a few years span.
And even when you compare the very worst performing camera to the very best in a particular year.
Size still wins that particular arm wrestle time and time again.
So why he (angry fat dude camera guy theo or whatever his name is) can't see the wood for the trees despite everyone around him pointing to the mighty oaks all around him and telling him he's in a forest I have no idea. :)
Awesome info you shared here. Thank you.....
glad you liked it. :)
Great video. Thanks and respect!
For the consumer, the practical relevance of ISO, and the only one that really matters, is how clean an image is. However every manufacturer develops their own propriety software for processing signal into an image, every manufacturer develops their own sensors with their own methods of generating signal, and every manufacturer chooses their own system for assigning an ISO rating to the results of those systems. These are constantly being evolved. Comparing ISO ratings between different model cameras by the same manufacturer is meaningless, let alone comparing ISO ratings between different manufacturers.
Saving this to link to anyone who keeps correcting me and can't take envelope drawings and math.
Axel Alexson :)
I did catch a few mistakes (you had a few drawings with light bending in a way that would not really happen) but overall this is nicely done.
When light passes through a lens element, the lens element does not "know" what is behind it. It doesn't know if it's a big sensor or small sensor. The physics of the lens has the luxury of ignoring the sensor size. Another way to think of it is if you have a slide projector and a projection screen, you can swap a large projection screen for a smaller projection screen. The "projector" does not "know" you did this. But of course the human... DOES know. So the human reacts to the change by adjusting the projector... either changing the physical distance so the image "fits" the screen... or perhaps the projector has a zoom lens and we can adjust the focal length.
Basically the short of it is... the actual DoF doesn't _really_ change (strictly in terms of laws of physics)... but the human reaction to the new sensor size (to achieve an equivalent framing) causes the photographer to do things that DO impact the DoF.
hey, thanks, not sure if you watched the follow up vid too, i did flag a couple of silly mistakes in that. (my light ray vectors did go the wrong way at one point) but it doesn't change the overall message, just a silly mistake with my simple illustration. anyway, thanks for input.
thank you for your explanation, it helped a lot and i must say mft sensors are very very good, f. e. em1 II is likly even with a sony a7 iii sensor...
but ...
FOV + DOF + Aperture is what counts for me and i like the look of FF more.
even though i have canon with less accurate sensor.
You explained wonderfully..
Someone liked first, anyway im second to like. This is real nerdy stuff but still willing to learn and thanks for sharing!
Thanks for that :)
Why would the apple at 13:47 only use the middle of the lens?
Well Done! Thanks a bunch!
Wes Baker glad it was useful.
wow, wonderful! Just a question. The cropped Iso related to the sensor does it also involve the noise level? In other words, in terms of noise level 800FF is equal to the 200 into 4\3?
Great video. Only for ISO is correct but if m43 camera is 10 mpx and FF is 40 mpx that the ISO is the same (pixels per sq. inch). If the resolution is the same, you have to multiply by 4 (2 x 2). For that m43 limit now is 20 mpx (they can not support more for ISO)
Regarding MP.
I work in video. So most of my calculations are presuming we are looking at 2 cameras at same resolution.
(Both shooting 4K )
The only thing you got wrong is how ISO is calculated. It is light falling on the whole sensor. Not only on a certain area. Otherwise exposure values wouldn't translate between sensor sizes. Not to mention film. Having shot more than a few film formats. 800iso medium format is the same sensitivity of 35mm and smaller films. On film its function of grain size.
You’ve got some of that a bit backwards.
The ISO number represents gain added to the signal to boost the signal. (Which give the impression of increasing sensitivity)
And the gain is directly proportional to the size of the sensor.
Hence if you point camera at a grey wall, and cover part of the sensor. The rest of the sensor doesn’t make the exposed part brighter.
Just as the StoR ratio remains equal across any part of the sensor.
It’s easy to put this into practice. Just swap a f1.8 full frame lens for a f1.8 apsc lens on a full frame camera.
The iso /exposure remains the same. And the image brightness remains the same.
But we now have shadow across the outside of the sensor.
So obviously there is less TOTAL light hitting than sensor.
You see my point?
If that sensor was a solar panel it would now be generating less electricity. Even though the light source remained constant.
It’s the total amount of photos hitting the sensor is what matters to the image.
Not the concentration per square inch.
2:14 why would the image be focused to singular point rather than whole sensor ? I feel this question is stupid be I would like to know
Think of it this way.
Every point in the focusing plane in the real world, is focused down to one point on the sensor.
As you watch further that is illustrated
Firstly - thanks for the detailed explanation - 1 point about FF vs. APSC (or MC43), is the recent shift to cropped sensors with high count of MP, i.e., emerging 33MP APSC cameras rumored to come out the next few years - I would like to see this video updated to include the effect of such a shift (more noise - since less light per pixel, but maybe compensated since it makes a difference to the coverage of the CoC ? - And if so,is it pronounced enough to nullify the effect? -Under "low light"/ Normal Light ?
biscuitsalive: Thank you for your comprehensive and useful video. I knew about the DoF and aperture correlation, but with different sensore type it seems to be a complicated concept. So I felt this, but I’ve never put this together. Now i understand better how it works. I would have 2 questions, and i I know this information is in your video, but really just would like to make sure if i get this right. I am asking this from an APS sensor user point of view. For me having a nice background blur on a portrait image is the most important. 1, fuji fx has 1.5 crop factor. If I would like to have the same blur than I have with nikon FF 50mm f1.2, I would need 35mm f0.8. There is no such lens on the market. Right ? Only thing I can do that I use a longer lens and step backward. Am I right ? 2, ISO has nothing to do with other feature of the image except noise level, isn’t it ? I am just asking because I am an APS user now. In order to get similar nose than I got with nikon on 100 ISO, I would need 44 ISO on Fujifilm APS. Again, there is no such sensor. So I guess the main differences show at the “lower end” of the parameter scale. On the other hand there are noise level differences between different type and brand of sensors. So ISO 400 dosn’t result the same level of noise e.g. sony and fuij APS. The noise is not only matter of size. Right?
So the only problem is that 24mm f1.4 ff lens exists, but 12mm f0.7 mft lens still does not
I think the main reason why these misconceptions exist is because most aps-c cameras dont have lenses that are as fast, relative to their sensor size, compared to lenses designed for full frame cameras. 35mm f0.95 aps-c lens is going to have, roughly the same results as a 50mm f1.4 full frame lens. I wish there were more ultra fast wide lenses for aps-c cameras like a 20mm, or 25mm f0.95!!! That would be AMAZING!!!
I just found a website called dof simulator and it is a great way to see how all this works out!!!
Reza Taghipour yep. Your not wrong.
All you need is the same size aperture. (Physical size of hole, not f stop number)
And things soon start evening off.
That means lens have to be about the same size though. (Crop lenses same as FF lense)
So, clearly it’s the lenses that are the weak point. And not really the sensor at all.
Theres a youtube explaining refraction, the analogy of soldiers marching from tarmac to sand, its not really whats happening, what is really happening is to do with interference between the light and the fields of the electrons in the glass molecules. Light is an electromagnetic wave, not that it matters and all that, and if I actually understood it, I could explain it, but the light (being part of the electromagnetic spectrum) interferes with the electromagnetic fields of electrons in the glass molecules. Anyway, it doesn't matter, I like your video.
You’re a legend!
perspective is controlled by distance so you cannot get the same picture
Great and clear explanation. Tnx
timmerdude thanks. glad to hear that. :)
Fantastic video! I enjoyed the follow up as well. I have a quick ? on the circle of confusion and how it effects the other metrics when you take into account when sensor sizes scale, the resolution often changes as well. For instance a GH5 vs an A7RIII.
TheKoop117 yes in this video I was mostly talking about same res at diff sensor sizes.
(So 4k video as I mostly use)
Higher the pixel density the smaller the circle of confusion becomes.
But in terms of end product it still comes down to how it’s viewed.
If it’s printed huge and viewed close the allergens depth of field will become narrower. And visa versa.
So you do need to take into account the circle of confusion of the sensor/pixel density. AND the final viewing set up.
(Tv screen at 3 meters... cinema at 20 meters... billboard at 50 meters etc)
get your inner nerd on, I prefer geek! love the humor in the warning !
Matthew White ha, well I realise some people just don't care about this technical shit, and just want to see my little Cornwall, kitesurf, mountain bike action vids. :)
So get in there first before they hate on it.
I hear you
and obey! haha