They Taught You WRONG: Falloff, Roll-off, Rendering, and Micro-Contrast
HTML-код
- Опубликовано: 20 май 2024
- There are some terms that have cropped up a LOT recently in marketing langauge, and they're often used in misleading ways to manipulate an uninformed public. Let's talk about what these terms are, what they really mean, and how they're sometimes used deceptively. If you want to hear me on FULL rant mode, this video is for you. As always, I could be massively wrong, so feel free to incinerate me in the comments.
Yep, I really hate it when someone has a lot of respect from all the or the majority of people on a social platform and uses these buzz words to put something down and try to sound like they are saying something objective.
really love this video
just as you were explaining rolloff a light went off in my head and i thought wait a second
something digital is either clipped or it isn't
You got it!! A+ work :-)
Bingooo
Hahaha, I love this video, you forgot to address 3D pop! 😂
Omg exactly right.
I think what people perceive as micro-contrast is similar to the effect of clarity or mid-detail of image processing. Essentially a spacial contrast effect.
As for roll-off, some sensors, especially CCD sensors exhibit smearing and blooming, which of course can alter the rendering (there, I said it 😄) of highlights.
In general I agree with your observations, especially when used as marketing BS.
Rolloff is just a term that people started using during the film era. When we say rolloff on digital cameras we are describing the emulation of gradation of light
love your sense of accuracy! very well presented! really enjoyed it and learned a lot
Thank you Lars!
The only time I really heard the term rendering used when in VFX work, not in lens selection.
i think a lot of this just comes from professionals or people wanting to seem more professional being afraid of just saying “I like how this thing looks” that they feel they have to pull out all the buzzwords to justify their belief, despite having no understanding of what any of it actually means.
I find it really quite frustrating to hear hugely successful DPs and directors repeat this kind of thing because all of the information is only a google away, and in all likely hood their ACs or technicians probably know lots more than them about all of this
Just a quick note about the use of micro-contrast, in my experience a lot of people basically just use this when they mean sharpness. So i guess you could have a lens with overall low contrast but then high “micro-contrast” through being sharp, though again in my experience the onset LUT would influence this more
Ding ding ding, we have a winner! People in EVERY field are reluctant to say those three scary words: "I don't know." So they just repeat fancy words or phrases hoping they sound smart. And even though I'm being probably too snarky by half in every video, I try really hard to make clear that I could be wrong, that I don't know everything, that I'm still learning.
This is so informative. Thanks for making this video!
Great video.
To me mico-contrast means that a lens has contrast between different values not just bright and dark things. I guess that comes under contrast.
So for me it is when I have a subject in focus and the subject is not very different from the background (colour and brightness wise) but it is quite easy to spot the edges of things. As people like to say, it makes things pop off the screen.
But again it can be explained with contrast, so fair point.
Maybe it's useful in differentiating between lenses that have severe contrast issues like vintage lenses. So my pentacon has bad contrast (blacks don't exist under f2.8, it just grays) but my 7artisans has ok contrast (blacks are black always) but very bad microcontrast (it difficult to differentiate between different objects)
Idk, I'm just sharing my opinion here. And I'm quite aware that I'm wrong lol. I'll look up on it a bit more when I'm free
Fall-Off: I have not done deep thinking on this one. However, Peter Karbe, a Leica optical designer, talks about fall off. I feel every time I add a link to a comment RUclips deletes my comment. You can search on RUclips for an interview with him where he talks about fall off. He's the head of Leica's optical design. I don't think he would talk about it if it wasn't really a thing. Also, it has been claimed that the fall off between a spherical and aspherical lens design behaves differently. Let me know if there is a way to send you links without having my comment deleted.
camera conspiracies = Micro Contrast - Love the content btw
Micro contrast is the darling of the Leica crowd
@@evocati6523 I don`t care how they call it, its great.
I’ll bite on focus fall off. We use this a bunch in terms of qualities we like or don’t like. But mainly as it relates to sharp modern lenses. I think very sharp lenses wide open have a very abrasive “fall off” in focus due to focus plane in relation to wide open sharpness. It’s one of the main things we start with first to try to make that transition more subtle by softening lenses wide open and correcting for that. That does affect more of the image than the focus transition qualities, but especially in sharp sensors like raptor/venice, you really notice how distracting it can be. I don’t know of a simpler way to state that quality when explaining a look to someone, but focus falls off has worked well in quickly identifying that quality. On board with the rest though.
Also one word I despise is “character” when describing a lens with say more optical flaws. A lens being sharp and neutral is also character. Just different character.
So far, all I hear you saying is that you think some lenses are too sharp. I don’t hear you defining any new quality unless I’m missing something.
@@nick_salazar its not that its a new quality, but there is a relationship between sharpness and t stop and specific properties at those t stops that affect that one quality. when all of those properties come together, they create a unique appearance of that transition. There are some sharp lenses like the xenons that do this really nice and subtle and some like the CNE, where as you pull focus you can almost see a plane of focus shifting dramatically.
@@NorthwestCameraCo maybe I'm just too thick to understand what you're talking about, but I still don't hear you describing a difference between one lens and another. Isn't the focal plane supposed to move as we rack focus? That's ... what pulling focus is, right? Maybe one lens is just sharper. If you can find me two lenses with roughly the same sharpness, and roughly equal bokeh qualities, AND demonstrate some difference in how their out of focus areas behave, then we've got something to talk about. But when you say "specific properties" without being able to show me what those properties ARE, then I have a hard time believing that we're talking about something other than the qualities we DO have names for (sharpness, bokeh, abberations, etc). This is my whole beef with "focus falloff" - if we have to say "it affects specific properties" but we can't demonstrate them, then is it really a thing? Also, please don't take me to be too aggressive. I respect your expertise and I'm happy to be wrong! Maybe I just need to have it explained to me like I'm a five-year-old.
@@nick_salazar Its very easy to demonstrate if you were say trying to show them off. and yes focus pulling moves the plane, but on say a CNE wide open, its so sharp wide open that you almost can see it front to back moving since the plane is very shallow vs other lenses say a fd 55 at 1.4 that focus plane appearance is much less shallow due to it being less soft, more bloomy wide open. to me, this one trait is almost my entire basis for judging lenses anymore as a lens that exhibits less abrasive focus transition to me is much more attractive as it feels a little less digital since film had more halation and softness by nature. To me if i want something to feel or appear less digital im looking for lenses that Wide open have some amount of softness, aberation, bloom to make that transition more subtle. also they make pulling focus a little more forgiving as well. That being said, it's not hard to put into words, but it's hard to put into FEW words and I think coining a term like this keeps it easy enough to digest without having to deep dive. That being said, Sharpness, Bokeh, T-stop all play into that specific quality of how that transition behaves. Find two similar lenses and that might be the same. But also 2 35mm lenses that have same sharpness and same t stop may behave totally different based on the lenses construction in terms of elements, glass quality, etc.
Like say you made 3 cookie recipes. Same amount of butter, same amount of flour, same amount of choc chips but all the ingredients were made by different manufacturers. they would all be the same, but would most likely taste different or be slightly different in appearance.
Great video, thanks for the info.
Ooh I inspired a YT video! Appreciate the time dude 🙏
Thank you, especially for micro-contrast. I have lost respect for reviewers who suddenly started using this term with no real explanation or definition. It makes me feel as those they are just making terms up to sound intelligent or something.
Unfortunately, there is a bit of a sycophantic relationship with reviewers who rely on incoming product for their continued existence. They will often just repeat the marketing language of those sending them the product, because that’s what keep them publishing and popular.
I'm certainly guilty of some of these but for the most part I try not to use vocab I don't fully understand for fear of ending up in a conversation with people like you. And I don't mean that as a negative, but as an acknowledgment that my path has been different from yours. However, people misuse words all the time and the only problem with being right is that it's only relevent as long as there's someone around that knows better... The classic idiocracy problem. Education helps, but you're talking about people that are self taught, or picked up things around a set, even a proper education isn't a guarantee you'll get these kinds of details right, hell I've modified my understanding of things several times over the years because the crew around me held a different understanding of the meaning of something than I did. If your goal is to communicate a peice of info to someone else, either observational or functional, than the most successful path towards meeting that goal will be the one with the fewest blocks ;)
Oh boy am I incredibly glad to see someone who finally knows what they are talking about!!
I came to these exact same conclusions after using magic lantern for a while now. The sensor I use is over a decade old, possibly made in 2009 and it's no different than modern sensors except in having off-die Analog to Digital Converters causing noise that lowers the dynamic range by 1.5 stops of it's potential.
I've learned the only improvements other than I mentioned already are noise performance (BSI) leading to more DR, higher readout speeds and much more powerful processors. My camera can achieve 90% of the things that raptor can (14bit linear RAW, not log) only issues being icy one reads out at t's 4MP max while the raptor is over 30MP, higher DR with the Raptor which can be solved by shaping light and not shooting high contrast scenes, the Global Shutter and most annoyingly the fact that the HDMI is a measly 1080i which makes me pull out my hair sometimes. Try pulling focus on that.
Other than that my camera does everything well.
Where they do differ however is that camera is a workhorse meant to take a beating and speed up your production and not slow it down, mine is an utter annoyance. Most problem could be solved by rehousing it but nothing will fix the fact that it's a decade old budget camera and was seriously cheap when it came out.
What seems to now make the most difference for me is the lenses, the location, the lighting, the set design, the actors, sound, editing and so on. The 4MP images it produces can be upscaled to 8MP (4k) without most people ever noticing it's any less than that and is fully HDR compatible which the HDR image blows peoples minds when I show them.
Stop being influenced by marketers lying to you. My next upgrade will be a SIGMA FP since it's the closest to what I want from a camera and my productions are no where near requiring a RED Komodo yet. I can't justify buying a camera like that unless the productions are over 10 or 20k per project.
Well said. Question, I have an one image using a Canon or Sony mirrorless camera, and I have another image using the RED Raptor. Both are using same lens, shooting the same subject. the RED raptor image looks more "filmic" and prefessional than the image in the mirrorless. This may be subjective, but I can totally see the difference. The mirrorless cameras look to "digital" or "camcorder" like. What do you attribute that to? is it the dynamic range, is it the highlight roll-off? What makes one image look more professional than the other? Thank!
Well, sometimes people mean different things when they say that they like one image versus another. In my experience, the biggest difference image wise is that the V-raptor has a very strong OLPF, which is a piece of glass in front of the sensor that reduces high frequency light, and results in a slightly softer but smoother image you might say. In my mind that is the largest difference you’ll see between those two cameras. Yes, the dynamic range is more on one than the other, but if you shoot properly on each camera and color match the images, the biggest difference apart from the OLPF effects is that there should be more noise in the deeper shadows on the mirrorless. Each camera’s “color science” can make a difference, but if you properly color match then those differences will be minimal.
@@nick_salazar Thanks for the reply bud 👍
You’re welcome! Take with a grain of salt though - I am still learning like everyone else :-) don’t be afraid to tell me I’m wrong.
On set, the "Rendering" always irked me.. C'mon man, I'm a 3D artist. Don't use this term for lens in front of me 😐.
Don't get me started about rolloff, given similar freedom of stops (dynmaic range) you can match look of any camera to any other camera. Sensor doesn't give a shit about your highlights and shadows it just tries to capture the maximum possible it can.
Your job as a colorist/compositor/editor is to "set" that rolloff to the art director'e liking.
Very interesting, I've been trying to disentangle some of these pseudo-concepts in my head lately and my simple conclusion was always, "it's more complex than (insert favourite truism)..." because all the image pipeline elements, both hardware and software, interact. What I think you are implying, is that often people are pointing to the affect an image has on them (emotionally, evocatively, associatively) and then attributing a simple (simplistic?) explanation. Moreover, taste gets confused with technique.
On micro-contrast, I saw recently a video on photoshop tools that was the simplest explanation of differences in types of contrast - and what I took away was that it's worth testing tools to see what they do. There is no doubt that different tools are doing different work, and that the effects are discernible. I liked the explanation you gave of focus planes and the effect of curvature of optics, and that is a good example of where misunderstandings creep in - I recently read some write ups on a forum of an old Pentax 67 lens and found some people describing "soft edges" and others talking about the curvature and how they had to work within the lens' design limits/features to find an optimal focus for a shot...
So what I'm basically saying is that let's speak in plan language. If we don't understand a concept, let's just say so, instead of trying to hide behind a fancy word. Regarding contrast, let's be careful to distinguish lens contrast from the contrast tools in a software like photoshop. Yes, they may be organized around the same concept, but they might not work quite the same. In either case, I like your idea of "testing tools to see what they do." Simple curiosity is a great way to make art.
@@nick_salazar I'm on the same page, but of course on RUclips a lot of people feel the need to project expertise even when they're still learning... and 'plain language' is more often spoken by the true experts who are confident in what they're saying. Yes of course there is contrast from a lens that is different from the software settings, my point was that it is easier to isolate the conceptual aspects of contrast when broken down in terms of software tools and their effects, than it is to understand these things from a talking head video. Seeing is believing as they say, and lens designs and coatings etc are less visible to the average user - it takes a more trained eye to understand what they are seeing, and pin point where a look is coming from: the lens, the sensor, the processing, etc...
Anyway, nice channel and look forward to more of your thoughts... something I just realised, provoked by your video, is that there are so many videos about "achieving 'that' look"... for example, getting the medium format look on a full frame camera... there's something there that needs unpacking, don't you think?
Great video Nick, just wanted to ask a question about roll-off. I was under the impression that it was not quite the same as DR since something like film shot at the same exposure as a digital camera will have much more roll-off, and while you can match them in post, the way that it's distributed as a converted rec709 file will look way different! Different log curves and different conversions will give different looks, and that's what people refer to - despite being able to match them in post. The same way different lenses and colour sciences will give differnt looks that you can also match in post. Have I been completely wrong about this the whole time? 🤯 please let me know if you have an answer, thanks!
Dynamic range is the size of your bucket. How much light it can take before it spills over (clipping). Roll-off is what you do with the water - how you map values from your sensor onto a finished image.
Those buzzwords don't annoy me honestly. I get that they are misused and it would be better to be more accurate but not a huge deal. The buzzwords that I actually get annoyed with are the marketing ones like "game changer" and "next level".
Game changer is the worst! It's most commonly used as a clickbait so I skip any video using it.
Micro-contrast I’ve heard used interchangeably with edge contrast. And I mostly hear it when people refer to the sharpness of a lens. For lower price ranges, lens manufacturers will oversharpen the lens to increase the edge contrast as a way to make a cheap lens appear sharp.
I’ve observed this in some lenses, like the arri signature primes vs a sigma art lenses. I don’t know if the terms are correct, but it’s something that can be observed for sure.
Thanks for the education!
I'm not sure I'd agree with the popular concept that "budget lenses are artificially oversharpened." Rather, all lens design is compromise. Sharpness, aperture, transmission, abberation correction, vignetting, size/weight, cost, etc, these are all constraints that cannot all be simultaneously optimized. It has become easier and less expensive to mass-produce sharp lenses, so if sharpness and cost are your two priorities, you can make such a lens. If other factors are your priority, maybe you let go of some sharpness along the way. There's nothing wrong with a lens whose cost is low or is mass produced, if it achieves the characteristics you're looking for. There are a ZILLION copies of the Helios 58mm and Jupiter-9 85mm lenses out there, and those were just used to make Dune 2!!!
Micro-contrast: There is some legit confusion on what this term means, but it is easiest to see when you compare Zeiss and Leica. Both Zeiss and Leica lenses tend to be high global contrast lenses. It will vary by lens, but Leica can have a slight advantage for global contrast and color saturation. However, the shadow areas do not completely behave the same. Generally, Zeiss will display darker, blacker shadows and Leica the shadows will be lighter and with more detail. When it comes to the contrast in smaller details Zeiss will have the slight advantage here. You can see this in leaves on a tree or buildings with texture, or even on things like pores on skin.
Do you know of any clear examples which would demonstrate this? I’ve heard these kinds of explanations, but they always seem to be a bit hand wavy. No one has ever shown me an “A vs B” demonstration which shows a quality called microcontrast.
@@nick_salazar Yes. Like I said in my other comment I feel any link I add RUclips deletes my comment. Do you know how I can send you this information?
@@nick_salazar Try your luck with finding this. There is a blog called the other side of bokeh. It's a wordpress site. There is a comparison Leica vs Zeiss Episode #2. Look at the bricks and shadows in the close up crop of the subway. This is not the best example as the light is flat and dull and low contrast. Light also determines the contrast of the scene. The lens varies in how much it can transfer, which is always less than the scene. Maybe instead search for an article called Micro Contrast and the Zeiss Pop by Lloyd Chambers. He has comparison photos.
I know a great lens when i see it, no marketing bullshit has ever convinced me otherwise.
They can call it whatever they like, i don`t care.
You might aswell advise to buy the lens that`s most visually pleasing.
That`s why i like Leica, Zeiss and Voightlander.
Many of these terms are used to convince absolute beginners to buy cheap shit with marketing terms and misleading info.
Like your cheap china or sigma or kitlens suddenly can compete with great glass.
I think this video might have been living in your head for too long! This happens to me too, becomes a bit of a semantics wormhole. Side note: GREAT Tittle, right up my alley for YT strategy.
Thanks Mark! Hope to see your footage on the Arles primes soon.
Use of these words is not the most offensive thing a person can do, particularly when trying to lock in the deal memo. Outward facing communications with the money bags sometimes requires a finessed communication style akin to a 6th grade educational level. And sometimes, having an amenable posture can encourage new relationship work. In this new digital age, your definitions are spot. Technical terms and definitions are really productive in communicating with members of the camera production crew, colorists, and VFX venders.
I'd encourage you to just write and speak plainly.
What's the relationship between contrast and sharpness of a lens? I always thought that what people meant by 'micro-contrast' is both very high ccontrast and sharpness.
Perhaps that is what they mean. But you can have sharp and low contrast lenses, or soft and high contrast lenses. They are somewhat independent.
To be clear, I'm not an optical designer, and I imagine they might say there is a relationship between the two in terms of actually making a lens. I'm just observing that there are lenses that have high/low contrast, and high/low sharpness, and that one does not always go with the other.
I like to think some lenses have certain magical qualities but I don't really know what I'm talking about half the time tbh. But nobody knows, we're all talking the same nonsense, no one has called me out yet. Still, I'm wondering how would I explain why pick a Cooke vs Zeiss using technically accurate terms.
I empathize with this question of yours. On the one hand, lenses are a collection of technical elements to produce an image. On the other hand, different lenses make us feel a certain way and we do not always understand what the technical reasons are that they did. The best we can do is try to understand the technical side, and admit defeat when we’ve reached the end of our understanding. Better to just say, “I like how it looks, but I don’t know why“ than to wrongly attribute it to some magical property like microfunk or jumble-rings.
You are right, but there is one little thing about roll off that seem sort of valid and thats LOG. All cameras have different log profiles. If you see the s curve of the dynamic range of a camera, will be different to another, so there will be some difference in highlight roll off. Maybe marginal but still, there is such thing, dont you think?
LOG is simply an intermediate mapping between raw sensor data and our final image. Just another mapping. So yes you can find a different log curve with each brand best suited to their sensor. It still does not change what the sensor collected.
YOU SIR! Know what you're talking about
ha! I love this so much bullshit hype buzzword out there, right up there with "I like to rig my gear out" 🤢 I am your subscriber No: 450 :)
Thanks Cinenimus! Now let's rig out the comments section!
Keep hearing leica and micro contrast heh
I can't stand your sound good talkers that say shit just to sound good and smart when in reality like you said... don't exist and other stuff that makes no sense lol. Don't get me started... "cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic cinematic " ... So freakin annoying! Thank you for speaking up on this! About time someone else brings this up, when I do, people have a problem like bro.. hush with the terminologies that serves no purpose. Just say the damn video is decent lol
This reminds me of when people talk about a lenses “resolution” like some lenses are higher resolution than others. This doesn’t make any sense to me and falls into the made up category of lens “Rendering” to me. It’s my understanding that lenses are either sharp or soft or somewhere in between. Maybe they have better contrast and less flaring, hazing and defraction and what not than other lenses but I’m not sure how this could be considered resolution. Is this something real? Am I just misinformed?
Hey Galen - that's definitely a fair question. "Resolution" is how many line pairs you are able to see through a lens. They make charts for this, with thin black lines spaced closer and closer together at different parts of the chart. A lens with higher resolution can produce an image where you can see finer and finer spacings of those lines. Eventually either the lens reaches its limit, or otherwise it outresolves the sensor, and the resulting image just shows a grey blur where there should be very fine sets of black lines. Does that make sense?
@@nick_salazar Sure, that makes sense. However, how does is this different from just sharpness? When I watch lens review they often use the focus chart to test it and it often has the line test on it. Is it because the line tests are done by projecting light through the lens onto a surface?
We are getting beyond my depth here, so some of this is my speculation and should be taken as such. But I believe that photographing a focus chart vs using a projector to test sharpness, are just two different ways of doing the same thing. But the projector route has certain advantages, like being able to adjust back focus without using shims, so that you can determine exactly how many shims you need on your particular set up. Now, as to resolution versus sharpness, I think sharpness is going to be how the image behaves at the edge of two line pairs, the values that you will see between them or across the border so to speak. this may be what some people mean when they use that term micro contrast. Others have suggested this in the comments.
Resolution should be considered one component of 'sharpness' just as contrast or flare control are components of sharpness. Except that resolution is maybe 80% of sharpness vs the other factors.
@@Catonfire88 I don't think that's correct. Resolution is a specific measurement of detail. Flare and contrast are different phenomena and not related. You can have very high-resolution lenses with or without significant flare, with or without high contrast. The attributes are not necessarily co-dependent.
Since you seem to understand stuff like this, I have a question!
If you switch lenses on a smaller sensor camera to match the framing and DOF on a 35mm sensor (without moving the subject or camera), does it actually match or is there a difference in the distortion of the subject?
I’ll give an example, but I might get some of the math wrong, so please allow for mistakes:
If in the exact same spot I have a full frame sensor with a 56mm lens at 2.8, and I match it with an APS-C sensor with a 35mm lens at 1.8 (or whatever will display the same apparent DOF); will the subject look identical or will their features be slightly more elongated in the 35mm lens because it’s a 35mm?
I have never heard someone address this, just FOV and DOF. Thank you!
Great question. If the field of view is the same, and you go one stop faster on the smaller sensor, they will be functionally identical. The only differences left would be any individual characteristics different to each lens (distortion, vignetting, bokeh characteristics, flare, etc). But yes the image, depth of field, EVERYTHING will be identical. This concept is called "Equivalency of Lenses" and it can take a while to get your head around it.
I'm actually planning a video about this, because it's still so widely misunderstood. There is no "compression" or "wide angle look" that is inherent to a given focal length. It is 100% about your field of view, your distance to subject/background, and your imaging circle/sensor size. This gets VERY misunderstood when people talk about anamorphics, and say things like "field of view of ABC with the compression of XYZ" which is just wrong.
I’m here for that video!
You should try the Canon RF 85 f1.2 DS for the lovely bokeh
Oh yes, I owned that lens and LOVED it. After a year or so I swapped it for the non-DS version. Still not sure that was the right move. That lens is extraordinary.
Well Roll-off is difficult on a linear raw signal that is a non issue but when LOG and manufacturer Luts mess everything up i see where that comes from. But it is fixable
Not really. We have perfect mathematical transforms to get back from log … known as inverse log! Also log is just meant to save space, as the upper stops have more data than needed. Compressing to log gives them equal data on each level. It doesn’t matter how you record the data, roll off is how you present it in final form.
@@nick_salazar That is not what i meant. I see it on the user side and partly on the manufacturer side, why people judge roll off differently. For Example i discussed often on set with directors or clients that what they see is either a a monitor lut or a unfitting lut etc. Red for example offered different tonemappings while other manufacturers offer lut that to not fit the camera at all. I had so many sony cameras in hands that all recorded slog 3 but what came out of the camera was totally different and even on cameras with the same sensor since i had the fx3 and fx6.
how woud you describe the famous "zeiss 3d pop"?
Hype. They’re lovely lenses, but in my opinion they don’t have any special separation or “pop” than other comparable ones do.
@@nick_salazar haha so all Zeiss lenspire articles about 3d pop and micro contrast (which from my understanding they describe as basically edge sharpness) is all a fraude
@@ShortStoryInspiration Zeiss is a fantastic company with legendary lenses and a rich history in the world of cinema. While I have not read all of their articles, I would imagine that they contain a lot of great information, as well as a little bit of overblown marketing. It's important not to think poorly of a company just because of one or two misplaced marketing terms, but rather to evaluate the products on their own merits. In the case of Zeiss, the products are absolutely the gold standard of excellence. It's ALSO possible that I'm dead wrong, and "microcontrast" is a real thing. I've just never seen it adequately explained or demonstrated, beyond merely contrast and sharpness.
Check out MTF charts and maybe the ZEISS Pdf about them all they show is Sharpness and Microcontrast i sent you the link to the paper via email. (i can not post links here)
That Dp on the interview I THINK , was trying to simplify the terms it for the interviewer . At least i hope so
They were specifically talking about how "the sensor responds to contrast in the shadows" which is quite meaningless. In addition, the particular camera they were talking about does NOT outperform the other major in DR or shadow noise, if that's what they meant. I think it was just a good artist who is a bad technician. That can happen. MANY artists will spout false facts about their tools. Ultimately it's about creating a feeling, and if they think their tool creates it, that's a powerful enough placebo that it can translate to the audience just as if it were true.
@@nick_salazar that is very true, most of the stuff makes sense what you said in the video , essentially these are (or most ) just fancy or plain wrong terms . I dont know enough to call all of them “fake” but im sure about the rendering and rolloff and micro contrast 😂😭 , they are bullshit .
You could interpret that just as them referring to the shadow detail… cause I feel like saying something has contrast in a luminance range sort of just means there’s more data? I agree of the micro contrast phrase though..
Great video. But there is something I don't know if I agree about "roll-off" (and I see a lot of people are reluctant to accept what you say on this). Let me put it like this: In the same way that there are different levels of noise in the shadows, in the different under-exposed regions in different sensors, wouldn't it be reasonable to assume that even if you transpose the way the dynamic range is used in a slightly "worse" sensor to match that of an Alexa sensor (exposing for the highlights), the quality of the upper zones of latitude may not be the same? In a sense, a good comparison to this would be audio bit-depth, or even color bit-depth, in the way it can produce more gradations, through different regions of a digital signal. It's an approximate comparison, or even a metaphor here, but I am basically inquiring about the fact that not all dynamic range is created equal - something we are much more accustomed to think about, and recognize, in terms of the different kinds and intesities of noise in the shadows, than in the different kinds of "noise" in the nighlights, so to speak. Sorry if my langugage is not precise, but I hope I was able to convey my argument, and really hope you can address this point.
So when you say "the way it can produce more gradations" - that is precisely the measurement that we call dynamic range. A stop is a stop is a stop. It doesn't change across the range, it's a ratio of brightness. One stop means twice the brightness. The sensor captures all that information, and then it gets processed (either internally in the camera, or in your computer if you recorded RAW). How you process that information afterward is called color grading, or color science, etc. The sensor itself has no idea how to "roll off" the highlights. It's literally like a bucket, just reporting how full it is. Each pixel says something like "my red is 80% full, my green is 82% full, and my blue is 53" full." And then it's up to your color grading pipeline to decide how each raw pixel gets mapped into a final image.
@@nick_salazar My friend, I know perfectly well what dyamic range and color grading is. I don't know what in my question has given you the impression that you need to explain the basics. That is not what I am writing you about (I have been in the business for over 20 years, have shot on negative film etc). I am under the impression you are repeating your conclusions and your metaphors over and over without really paying attention to what people are responding to you, and asking. Please, re-read my question if you are interested in digging a bit deeper into conversation with strangers - which is what you seem to be suggesting you want to do, in your video. I will reiterate my main point: I believe you may be glossing over the fact that the zones, or the areas of dynamic range, at both ends of the spectrum, MAY NOT BE EQUALLY RICH in terms of the information they are able to gather, in different sensors. For example: look at how in the darker areas, at the low end of dynamic range, in cheaper cameras, there is much more grain than in the similar regions of higher end cameras. Now think about the highlights. Wouldn't it be reasonable to think this has implications in the way the gradations (yes, standard as they may be in terms on brightness values) feel in the highlights? This is what I am throwing back to you. Would be lovely to see you actually address this.
@@ernatcho6409 so that grain in the shadows -- that's noise. That's precisely the enemy of dynamic range. You can graph it and see it in a chart. In the lowest usable stops, it is about 1/2 the magnitude of the stop itself. At the upper stops, it's far less (on the order of 1/100th). You can think of that noise as an overlay on top of the recorded information. But at the brighter stops, it's so small as to have virtually no effect. Also, you keep implying that different stops have different "range" which is why I repeated the definition of what a stop is. You seem to misunderstand what it is. I'm not trying to talk down to you, I'm trying to explain something that you don't seem to grab yet. Think of it this way - a ruler gives you the ability to measure distance, right? And with lager gaps, it's easier for you to be more precise (I'm certain this is about 3cm ... definitely not 4cm, definitely not 1cm). At smaller measurements, your "noise" comes in (not sure if this is 0.5mm or 0.7mm). That's exactly how the noise in a sensor works. The signal (which is a measurement) is less accurate down by the noise floor, much more accurate with some headroom above it.
When you talk about "gradations" perhaps you are thinking of how fine of a difference a sensor may record between two similar colors. That's something we call "color bit depth" and most sensors today measure in 16-bit depth, which is usually compressed down to 12, 13, or 14 bits using a log curve that doesn't add any noise (exploiting that difference in noise as described above). When we color grade, we eventually reduce it down to 8-10 bits for delivery, which is enough to describe more colors than we can see with our eyes anyway. In any event, NONE of this really affects roll-off. Just noise, clipping, or potentially artifacts like color banding (in the case of very low-quality sensors like a GoPro or lower-quality recording formats like 8-bit h264).
@@nick_salazar Dude, you are the one who is misunderstanding what I am saying and not being able to grasp something. I don't want to think you are being disingenuous, but it's borderline. It's crazy how you keep running away from what I am presenting to you and constantly going back to implying there is something I don’t know. Why would you do that? Why not accept you are talking to someone who knows perfectly well what a stop is… I never implied that different stops have different “ranges”! That would be stupid! Why would you say that, and where in my response is that? And why would say “…so that grain in the shadows - that's noise" ? I obviously know what noise is! And I used grain momentarily to refer to it, as many of us sometimes do... It’s almost like you are using this weird patronizing tone because you are lazily trying to avoid the main issue, which you have finally ever so slightly rubbed against. I will quote you: "At the upper stops, it's far less (on the order of 1/100th). You can think of that noise as an overlay on top of the recorded information. But at the brighter stops, it's so small as to have virtually no effect". Besides your clearly lacking understanding of the nature of noise when you talk about an overlay (there, take some of your own poison), this thing is exactly my point: maybe it’s not “so small as to have virtually no effect”. To be more specific: maybe the kind of degradation equivalent to lowlight noise, but that happens in the highlights, which we have less vocabulary to talk about, is part of a real issue concerning how different sensors manifest "noise"/deterioration at the extreme stops of their dynamic ranges! And I am suggesting this could result in different ways that a sensor will influence in the production of an image in areas in which highlights become clipped ie. “roll-off”. Listen: I have no dog in this fight, I normally don’t even use the word roll-off either. But I am just rationally thinking this argument through after your video (or at least trying... as you are clearly really hard to reason with), because I am also honestly interested in more rigorous description of the tools we work with, and I hope you actually latch on to what I am saying instead of insisting there is something I am not understanding here. Which there’s not.
Hi Nick. I like you have a "fetish" for word accuracy, mostly because all things are given by their distinctions in language and human beings have the capacity to CREATE from nothing, innovate and deepen our perceptions all due to the gift of our ability to create distinctions in language. Said more formally...
"Observation is a phenomenon of distinction in the domain of language..." -Ludwig Wittgenstein-
Without which, there would be NOTHING. No love, no hate, no good, no bad, no context and no content, no personal history, no strongly felt opinions - nothing... Because it all "lives" in the domain of language. Language affords everything.
What we need (as a global human race), is to develop a faculty for the tolerance of detail. But I'm afraid not everyone shares that interest and love...
Thank you! Yes, precision of language is important to me. But we should also remember that the fundamental quality of language is that … it changes over time! So we should be careful not to be too stingy as we cling to our ideas of right and wrong. Dangling prepositions, split infinitives, and the like, may offend traditional sensibilities but work fine in modern vernacular. Or as Jefferson said, “In matters of principle, stand like a rock; in matters of taste, swim with the current.“
@@nick_salazar I love that quote! Yep, I'm not trying to be a snob or "better than"; as a matter of fact I love that I can benefit from having an abundance of mentors :-)
Another reason why I love the way your channel is evolving. Keep taking me to school Nick! LOL!
I agree with most of this, but in some areas I felt as if you were perhaps misunderstanding or ignoring the colloquial meaning of these words, at least how I understand them. For example, roll off is a term adopted as a shorthand to quickly explain the camera or lens’ standard transition between color and light values, not just how it handles the brightest and darkest areas. I would argue that this is probably more effected by bit-rate or photosites than DR. Lenses also have different reactions to light levels that can create a different out-of-the-box, visible “roll off” that can give something like a person’s cheek a more “three dimensional” look because you are seeing a more subtle difference in the various color and light values. Can all of these things be corrected in post with appropriate skill and planning? Sure. Look at The Creator, it looks great. But on set, in the monitor, in front of investors and clients, when critical creative choices are being made… you kind of are reliant on what the camera natively captures. Also, DR might be 13 stops in two different cameras, but one camera might have far more info recoverable in the highlights and the other better in the shadows, so it’s not really the same for every situation. I own both a RED Gemini and FX6, and while they both can be made to look basically the same, it’s not like all that it takes is putting them both through the same CST
And for “rendering” I don’t know what to tell you there, other than I have done tests between different lenses of objectively similar specs, and been left with entirely different feeling images. And I’m not talking pretentious “oh I see this one is a hair sharper if I squint” type stuff. As a short hand, saying that this lens, again with nothing done to it, gives a more pleasing look than this one because of who knows how many minuscule details, shortening all that to saying “I prefer the way this lens renders skin tones” is easier than trying to figure out what exactly is happening within the lens, or saying that in 100% of lighting situations I can trust that my colorist/VFX will give me the same feeling this other lens gives me out of the box
I’m not sure if you missed some of the video, or if I did a bad job explaining. But what you’ve written in here virtually all corroborates what I’m stating in the video above.
Sounds like you just say “rendering” when you don’t know what the differences are. That’s what I'm arguing against here. My suggestion is that if you don’t know why you like the image, just say “I don’t know” instead of using a word designed to pretend that you do know.
Really appreciate the replies, Nick. I went back and watched those two sections again. You are right, we are saying the same thing about “rendering” in that it has become a colloquial term for “looks better to me”. Apologies for not hearing that the first time. I guess that I would argue the casual use of “rendering” has much more utility for persuasion than “I don’t know” so it does actually serve a purpose beyond just sounding erudite. Totally agree that “rendering” is not a technical term.
But if the truth is that you don’t know, then you shouldn’t be searching for a word with better “utility for persuasion” as you say. You should just … admit that you don’t know. This is my entire complaint about the word.
11:40 no i do not think so check out the following lenses when making portraits or at best with a LensAlign MkII Focus Calibration System and you will see it: "Bad" or strange Falloff has for example the MR.Ding 50mm 1.1 OR the Canon 50mm L 1.2 and good falloff has the Sony GM 50 1.2 both at 1.2 or 1.1. I can see that even while focussing the mr Ding :D.
Are you able to describe what this property is, or is it just one of those “you’ll know it when you see it“ things which is my whole problem here. No one has ever described in simple terms what that is supposed to mean.
@@nick_salazar i sent you picture in the email :) it is really hard to explain. I sent you pictures of a angled focus chart. With one number in focus the other lines and numbers will show you what i mean. On one lens the other numbers lokks super soft almost like gaussian blur on the other lens it is sharper, hazier, it looks mistier and not as dens. that effect is not only on focus charts but on fine details when shooting portraits wide open or on nature shot and makros
I’m sorry but I’m still not seeing a new quality separate from sharpness and contrast. Perhaps you should consider that if a certain quality is just too hard to describe or demonstrate, maybe it doesn’t actually exist.
@@nick_salazar No i do not think "not existing" is a realistic option since you can clearly see it, i think just the definition is not clear enough so you see the difference. Sharpness is described as detectable lines and higher frequency than contrast. lines can be visible with low contrast without affecting the over capability of a lens to show bigger contrast. If you have calibrated lenses before or used focus charts you will see why. Because the Resolution section of these charts have such fine lines that it look grey on the first glance. If the lens can detect lines to a certain point this indicates sharpness. If these lines are hazy blurred and are zoomed in not black anymore you lost micro contrast. If the bigger charts have nice blacks you are in macro contrast territory.I can send you more examples but at this point i am not sure if that makes sense
bruh...lol
You gotta be kidding.
This seems more like a rant than trying to disprove any of these buzzwords. Even RED has „highlight rolloff“ options to choose from in their menu settings. It’s not like RED decides to just make up some settings in order to appear more interesting. I would have appreciated if you would have at least tried to make some comparison or be a bit more scientific about it.
RED’s “highlight rolloff” settings are just different LUTs, how the sensor data is mapped onto a final image. Exactly as described in the video above :-)
I’m just trying to help people use the words correctly. SENSORS don’t have roll off. LUTs or mappings do.
@@nick_salazar The RED's highlight rolloff setting is not a LUT. It controls how the camera transitions from properly exposed areas to overexposed areas, affecting highlight smoothness and detail retention. This setting is part of the camera's image processing pipeline, whereas LUTs are used in post-production for color grading and transforming color values.
Yes it’s a LUT. It’s applied as a preview for monitoring, but when you get back to grade the raw footage, you can pick which one you use. There are twelve of them, corresponding to the three contrast and four highlight rolloff settings you described above. However you want to describe it, it’s a mapping from raw values the sensor actually records into a finished image.