Maximize Your Camera's Dynamic Range: Secrets Revealed!
Вставка
- Опубліковано 10 вер 2024
- Have you ever wondered how dynamic range impacts the quality of your images? In this video, I'll explain dynamic range and how it affects your camera's sensor performance. Discover practical tips on how to harness the full potential of your camera's dynamic range to enhance image quality iusing negative ISO.
Dynamic range refers to the range of brightness your camera can capture-from the darkest shadows to the brightest highlights. Understanding this can significantly improve how you handle exposure, particularly in high contrast scenes. We'll explore how cameras typically capture less dynamic range than the human eye and why maximizing this range is crucial for professional-quality results.
Learn about the exposure techniques that utilize the dynamic range to its fullest, like setting the correct exposure for middle gray and adjusting ISO to balance noise and detail in your images. We'll also discuss how different ISO settings can shift your camera’s dynamic range and the impact of shooting in native versus non-native ISO settings.
Whether you're shooting in low light or trying to manage high contrast scenes, understanding dynamic range is key to avoiding clipped highlights and crushed blacks. This video is packed with insights and tips, backed by real test footage to show the practical application of these techniques.
Join us as we break down complex concepts into actionable advice that can transform your filming technique and help you capture stunning visuals that stand out.
✅ Don't forget to download your free camera settings cheat sheet: www.cnomadic.c... ✅
It’s nothing new that shooting with a low ISO will produce the best picture quality with low noise in the shadows. Let me take you back a few years to film, we really only had choice from around 50 ASA (ISO)) to 400 ASA (ISO) and if you wanted to increase the ISO you had to push process the film in development. This resulted in a grainy (noise) negative nothing could be done about this and we just accepted it. The picture quality from even entry level cameras is remarkable. Too many “Photographers” get way too obsessed with a little bit of noise in a photo. Enjoy being creative and go out and produce some great photos, noises or not.
ISO is definitively not 'gain', that's according to the ISO standard. It simply sets the relationship between exposure (as in amount of light at the film/sensor) and lightness (how light or dark the final image is). In analog video, with no digital computation, this was called 'gain' because it actually was the voltage gain of an amplification stage. For people with a background in video it seems natural to conflate the two, and the manufacturers have adopted this because they think it makes the transition easier. However, thinking of it as gain brings out a number of misconceptions. I think this video is caught up in some of these. One misconception is that so called 'ISO noise' comes from 'amplification'. It doesn't - it's photon shot noise caused by low exposure. So, setting what you call 'negative ISO' (a technical impossibility) only creates lower noise by increasing the exposure. The general rule for minimising noise is to maximise exposure, which means a low 'EI' (a tautology, since ISO is an EI by definition). That's all this video is about. Use a big exposure, which is OK if your conditions and requirements allow it. The fact that this simple message gets obscured in a load of often incorrect technicalities is down to the confused mess that the camera manufacturers have made of setting exposure on digital cameras. And you should make it clear that you are talking about video - stills photographers are going to get hopelessly confused trying to follow this.
Even worse, by saying negative ISO this could be seen as suggesting setting the camera to extended ISO values lower than the base ISO settings
For example, the alpha cameras that allow you to extend the ISO range down to say ISO 50
These are digitally pulled ISO values that have reduced dynamic range compared to the default values
@@Supercon57 Agreed, though in general the below base settings don't have reduced DR - they have the same. The exceptions (mostly Sony)are where the design engineers have made some very odd gain decisions and actually increase gain for those settings. In usual practice those settings will give a higher DR in the shot because the set the metered exposure bigger, so the ratio of maximum actual signal (exposure) to noise floor is bigger than had you not used the lo ISO setting. The risk you're taking is blowing the highlights.
Apparently some cameras such as the Hasselblad H3D's native ISO is 50 and gain is applied by internal processing and post processing in Phocus. This was reported by independent testing. Perhaps this is a behavior which is unique to CCD sensors, I do not know.
@@crawford323 You're missing the argument. The point is that ISO is not 'gain'. Saying that it is, but gain might be applied in non gainy ways doesn't make much sense.
thanks for reading from the Manual. good job.
Same applies the other way - when shooting clouds - raising the ISO will give more detail in the highlights. It seems like it's the wrong thing to do because it's usually really bright in the sky - this is where proper ND filters come into play.
Yes that is accurate. The same shift happens going above native.
You should begin by saying you are talking about video. This discussion is just confusing for still cameras. The second you refer to IRE level, you have left still photography behind, as still photo levels are not measured on an IRE scale. Also, while you properly talk about the upper end of dynamic range as clipping level, you fail to mention that the lower end of dynamic range is taken at a certain signal to noise ratio, which is an arbitrary (or testing method standard) level, not an actual clipping, because digital sensors are linear down to zero light. In film, the underexposure latitude was determined by an industry standard contrast gradient in the toe of the characteristic curve. In analog video cameras, it was determined by the camera gamma correction curve, which was linear below a specified level, and therefore determined the level below which the CRT gamma would produce a toe.
I think you may be misunderstanding the details
By lowering the ISO you are having to brighten the shutter and/or aperture to compensate for the darker ISO which results in more light hitting the sensor giving a cleaner image
I find when working with cinema cameras is better to think in ISO and not gain
My first introduction into this whole middle grey changing based on ISO was when canon released the C100 with CLog
Having a higher ISO in bright situations can result in better detail retention in bright scenes and a lower ISO can have better shadow retention im dark scenes
Its not about changing ISO for exposure its more about shifting middle grey for changing the dynamic range of the shadows vs the highlights
In general, people worry too much about raising the ISO. I shot in dark conditions on an old 16mp m43 camera, and by "over" exposing, results were perfectly acceptable at 1600 ISO and 3200 with tweaks in post for 18" prints. If subjects are evenly illuminated it's easier than large areas of blown highlights, like skies.
Thanks this explains why i have so limited headroom in portrait photography when i expose faces at 18% grey using ISO 100 as i was thought in college. The answer is underexposing my subject at ISO 100 or moving up in ISO. I can't find the same type of graph which shows the dynamic range distribution above and under 18% grey vs ISO for my Canon EOS RP camera, but it looks like this graph is more or less generic for all sensors giving an even distribution at ISO 3200. A graph like this should actually be included in the camera manual.
It would be helpful to add Video in your title, since us photo enthusiasts were all thinking Negative iso, what?
recently did some similar tests with FX3 and FX30 and I'll be likely shooting 1 stop overexposed (1 stop negative ISO in your terms) more often. Glad you're putting this info out there. These cameras have gotten good enough that we often don't need absolute maximum dynamic range and gaining some cleaner, more defined shadow range can be well worth the loss of a stop or even more at the top end
IE doesn’t change iso. It just change brightness of picture on your screen. Due to this you add light or open aperture for correct exposure . That’s why you got more clean picture.
With my Sony taking still images in raw, the best dynamic range is definitely at base iso 100. I pay no attention to mid grey and always expose to just retain highlights. There is a little headroom when the histogram just clips, so the highlights are safe as can be seen in processing. Unfortunately no camera so far meters from raw, so we don’t know just how much margin there is for each exposure. My understanding of dropping the iso below native is that you actually just over-expose and hope that the highlights don’t clip! Each stop below base iso actually loses a stop of dynamic range if your highlights are just bordering on clipping. If you actually have a couple of stops of headroom because your highlights are well below clipping, then dropping the iso brings them into play and you are actually exposing to the right, as we should in digital raw. This gives the lowest noise in the shadows. Dropping two stops in flat light may work, but in a high contrast situation there is likely to be clipping. Dropped iso can be useful to get a longer exposure, eg to smooth out water movement. An ND filter is better.
For JPEGs the image needs to look right in the viewfinder, but for raw it may look washed out or dark depending upon the scene. All is revealed in processing and in base iso and with highlights just below clipping there is the maximum image quality to work with.
I'm not sure if I'm mistaken, but I really don't mind whether middle gray is at 41-IRE or 32-IRE. To me, middle gray is just that-middle gray. My light meter and camera, along with the lenses, should match in tone based on the settings I use. Of course, we always need to keep dynamic range in mind. In very bright situations, instead of lowering the ISO, we actually need to raise it.
I agree with you. For my exposing for my subject is more important. I think the point is that different formats have different requirements of exposure and as cinematographers or photographers we need to understand what they are in order for us to get the most performance out of our equipment and format we select to work with.
landscape: longer shutter, portrait, sport,wildlife, event: wider aperture, bigger sensor
This is why I never became a pro photographer.... Whoosh.. so far over my head I never even felt it go past... seriously didn't understand a word of this .. lol ... back to being a dummy.
How do you determine the native ISO of your camera? I don't see anything on Canon's site or the manual. Only the ISO expressed as a range
Native ISO as used here is just the manufacturer's choice of a default value. Still cameras have in the past often defined "native ISO" as the lowest available setting. Even then, it's the manufacturer's decision as to how much highlight latitude is available before clipping, what value in the raw output corresponds to mid gray, whether mid gray corresponds 13% or 18% diffuse reflectance in the scene, whether that will be scaled to 127 or something else in an 8-bit jpg file, etc,. etc.
Fuji X-T100 with the XC 15-45 mm gives me pictures with clipped highlights easily. But when I put on the 50-230 mm not so much. Why? Because the 50-230 lets in less light? When I use the Nikon D5000 the highlights dont blow out. Why? I dont think it has to do with DR. There is something else going on
The only explanation I can come up with is your aperture setting. Aperture will let more or less light in.
@@cnomadic No. Its the sensor settings in D5000 done by Nikon that is causing the highlights not to blow out. Some say its how the camera is set to read a grey card, neutral grey. But this is something I dont understand or know. About the X-T100 it seams the 50-230 works as an ND filter in some strange way, but I dont understand that either. Most of the time I neither under or overexpose with aperture. At P mode or auto it stays at zero. D200 underexposes to preserve highlights. I dont think D5000 does.
Very informative, thank you 👍
Thank you. Very useful insight.
Hi, thank you for that informative video, can you tell me where can I find those diagrams for a7iv ?
Unfortunately these diagrams come from the Sony Cine website but that only covers their Cine line Venice, fx9, fx6…. The A7 IV is outside of that line of product so I’m not sure Sony produced similar charts for that camera. Let me know if you find something similar I’d love to see it
This video makes no sense for a photographer.
definitely sent me into introspection ....
Interesting vid. Thx.
I tend to simplify things by treating the native ISO as "box speed" on old film emulsions. Because in RAW. You are essentially only working with that native ISO. All the other settings are pushing and pulling that native ISO, to use the old vernacular.
So. In shots that live in the shadows. I can pull the native ISO. For my BMPCC6K. that can be the higher native ISO setting of 3200 being used in the camera. But I say to it to treat it as ISO1250 when showing it to me. That way, the middle grey is a stop and a bit higher up on the file. Which means I am technically overexposing it by that much. So I get shadow detail that's so much further from the noise floor and clip as much from the highlight details. When viewed with the compensating setting, I get clean shadows at the expense of highlights
Conversely if my scene has little shadow detail worth saving and most is in the highlights. I can do the opposite. By pushing the film/file. I tell the system to consider middle grey to be lower on the recorded image. Resulting in a file or negative that's underexposed. This means, when viewing is compensated. I get worse shadow detail. Because the viewer had to draw from details closer to the noise floor. But highlight details are preserved as those are just as much further from clipping. With my cameras dual native ISO. I can exploit that quirk by using ISO1000. This records the native ISO400 image underexposed by a stop and a bit. Saving a lot of highlight details while sacrificing as much shadow details to the gods of photography.
On a traditional single native ISO800 camera for example, you can do the same. Set it to something like 1600 or 3200 to underexpose the file, saving highlights. Or set it to 400 or 200 or thereabouts to overexpose the file to save shadow detail from the noise floor, sacrificing highlight details that'll be more clipped.
Yeah what you are describing is basically how the EI (exposure index) works. The Black Magic Cinema dynamic range chart is very different from the other manufacturers, with middle gray moving up or down based on the ISO setting. I found it confusing since all the other manufacturers clearly show middle gray staying at the same exposure level. Great post!
Impressive stuff! Thanks for sharing.
Interesting. Thanks. Reminds me of Anselm Adam's Zone System.
Yes! I use the EL Zone system and it would be a great tool to measure the added stops when using EI
Hold on. You introduce the term IRE without defining what it is. Not good.
You can learn more about IRE in this past video: ua-cam.com/video/VkZdV3arsG8/v-deo.htmlsi=v7ZOn2_3vU5EeFcm
So where can you get these charts from? e.g. Nikon D500
I found those charts doing a simple search for Sony FX6 Dynamic range test chart. Same with canon , Arri and Red. I did not find a Nikon one when prepping for this video.
strange thing, i was taught that doubling/halving the signal is equal to 3dBi inrease/decrease of signal.
weird, I just double check to make sure, but everything I found is double = 6db...
That's true... for a the gain of a linear signal like audio. Volume or amplification of light is +/- 6db for 1 stop.
A digital sensor is a photon counter. A stop increase is doubling the number of photons, therefore the electronic voltage doubles.
Let me know if i'm wrong, but a lot of cameras have native ISO's that are around 100 or 200. Most won't let you go below 100... so, how do you go to negative ISO?
Yeah I believe you are correct. The only way I know of is by shooting in log profile which often uses a higher native ISO
Is this about video or photography? UA-camrs should be making this clear when they do these videos.
Everything he said is equally applicable to both video and still photography.
@@Secession1900 This became untrue the moment he referred to IRE level.
3 minutes in and this is where I disagree. There is no such thing as negative ISO except in dual gain output cameras. Slog3 at ISO 800 is infact just ISO 100. To map out the entire dynamic range properly, slog3 exposes the image 3 stops brighter. But the sensor is infact still at ISO 100 or 0db gain.
Even the dual native iso of 12800 in the fx6 is just iso 1600. Don't believe me? Take raw photos at iso 100 and slog3 clips at iso 800. Now grade both to match each other without changing the exposure except the shadows. Tadaa! They match perfectly.
All EI does, is simply let you digitally in the preview expose the image up or down by several stops. You cannot drive the sensor below 0db gain. All you're doing at ISOs below 100 or beloe 800 in Slog3 is pull the entire image down below the 0 IRE. It's like taking the entire waveform and pulling it down equally.
I dont fault you though, this stuff took me a long while to understand
Well I agree with just about everything you said. Funny thing is different manufacturers show a -6db gain on their dynamic range chart when going 1 stop below native. Also if you keep the lens cap on and record some frames, and look at the file on a waveform monitor you can actually see the noise floor of the signal. That same signal at 2 stop below native will show significantly less noise floor than at native ISO. To me this tells me something does happen to the signal not just because of exposure. Every camera applies a gain and a noise reduction to the signal and I think the waveform is telling me that bellow native the signal is treated differently.
@@cnomadicWell -6db in the case of slog3 is just ISO 400, not EI400. You'll notice the clipping point moving down from it's usual point in the wave from it you try to clip the image. So you're reducing the overall dynamic range by about a stop. You'll get one stop worth of improved noise performance but you'll lose it in the highlights. CINEEI can retain the entire dynamic range without sacrificing dynamic range.
Well again I agree except the Sony chart shows specifically EI values with the dynamic range shifting down.
@@cnomadic it doesn't really shift down. The dynamic range always stays the same in CINE EI. I'm not very good at using words to explain complex ideas but I really recommend watching *Max Sautner* video on CINE EI. He has by far the best explanation of how it works on youtube. The reason why it might "shift down" has more to do with your exposure and the exposure compensation that's happening with the monitor.
Like I said, highly recommend watching that video
I will definitely watch it. And I really appreciate this conversation. That’s why I do those videos to learn and discuss complex concepts. So thank you for the information and your willingness to share what you have learned. I hope we get to discuss more stuff in future videos
So essentially you're saying if you keep ISO lower, you will have less grain/moe details in the shadows. Is this new news?
I think this is going to require an increase in light. But, at a lower ISO, those areas exposed in lower zones will have better detail. I don't believe a lower ISO will compensate for inadequate light. Low light equals noise.
I understand your point here. I think what is different is the introduction of native ISO and the fact that cameras like the Sony FX6 actually doesn't give you an easy option to change your ISO away from native. Using Sony's CineEI, I can only shoot in low base 800 or high base 12800. EI is not actually changing the ISO, it is recorded as a metadata. Only since the last update of DaVinci Resolve and Adobe Premiere are the software now able to apply the Metadata to affect the image.
@@Effin_the_Chat That is correct. You still need to expose middle gray to the same value. So yes in my example I had to compensate using my aperture to make sure everything was exposed correctly. This technique becomes valuable when shooting a scene that has a majority of shadows, but one where you still control light level.
He said in the video that even if you lower the iso (in order to get more detail/less noise in the shadows), you still have to expose correctly. That would mean to perhaps add light. I guess his point is that you don’t actually have to shoot at base iso all the time. Sometimes you actually get a better image when not doing so.
@@DrWasim That is correct, you would use this technique if your scene is mostly in shadows, but you control light level. Middle gray should still be exposed correctly.
very interesting, i'd be interested to know what @simon_dentremont thinks of this! i'm not knowledgeable enough in either the science or the technology to compare and contrast the advice you both give on the use of ISO settings in digital cameras in low light settings. it might even be an interesting 'in conversation' type video
What you are essentially saying and should have been explained is that if you raise your iso, your shutter speed goes up, and you are, in result, underexposing and thus prevent highlights from clipping.
You mean shutter speed goes up?
@@MusaonYT Thanks for catching the error, I corrected it.
Sorry I don't follow. My shutter speed is independent of my ISO setting. You should always exposed for middle gray. So during my little test, when I lowered the EI (ISO) I opened my aperture to bring in more light to make sure middle gray was still exposed at 41 IRE.
@cnomadic To simplify the explanation: Imagine you have ISO 100, a set aperture, and you choose the shutter speed for 41 IRE. Now you raise your ISO to 200. This means the camera will pretend that the sensor captures twice the amount of light. In the preview screen, the image will be accordingly amplified to appear twice as bright as the original data suggests. The data stored in the raw on the other hand will not change, just the information added that it should be interpreted as twice as bright. The preview, however, is used for calculating the IRE value in the camera. Thus, the brighter preview will lead to a higher IRE value displayed. When you now adjust the exposure time to get to the IRE value of 41, you have to half it. By halving it, you underexpose the darker parts, but you provide more dynamic range for the highlights (in other words, it prevents clipping and moves the brightness levels of the higlights into a range that the sensor can capture with more accuracy). That's exactly what the graph is showing.
@@maggnet4829 yes that is correct, with the exception of gain. When doubling ISO you apply 6db of gain to the signal, generating noise. But you are correct by raising ISO you shift the dynamic range into the highlights giving you more stops to protect highlights. And I have seen people shoot that way for very bright or contrast scene. You just need to be ok with the added noise (which should be minimal at 6db)
Pourquoi aller dans les iso négatif alors que la plage du capteur est beaucoup plus performante dans les hautes lumières ?
Pour la photo il vaut mieux optimiser son exposition à +1 IL ou +1.33 IL pour avoir une meilleur gamme dynamique.
Alors pourquoi aller dans les iso négatif en sachant ça ?
Je suis Francais mais je n'ai jamais discupte photographie en francais, alors j'essais mais pardonez moi si je fais des erreurs. L'idée ici c'est d'utiliser le maximum de la gamme dynamique. Même dans les hautes lumières, la gamme dynamique reste a peut pres pareile avec 6 arrêts dans les reflets et 9 arrêts dans les ombres. En utilisant les ISO négatifs on peut avoir plus d'arrêts dans les ombres. Mais évidemment il faut toujours exposer pour Middle Grey donc soit rajouter de la lumière ou comme dans mon cas changer l'ouverture pour exposer correctement. Cette technique n'est pas pour filmer une scène de nuit quand il n'y a pas assez de lumière, dans ce cas la passer sur la base haute lumière sera mieux. Je crois que le capteur a des performances bien différentes entre le mode photographie et cinéma. Je ne suis pas photographe, mais après plusieurs conversations avec des photographes, il est évident que le capteur utilise une procédure différente pour traiter le signal. En tout cas merci de votre réflexion intéressante.
@@cnomadic félicitation votre français est bien plus français que certaines personnes 😂.
Il est dommage que ce ne soit pas préciser que la gamme dynamique soit différente entre le mode photo et le mode film.
Du coup si j’ai bien compris en mode film en basse lumière il vaut mieux sous exposer pour avoir une meilleur plage dynamique et ainsi capter plus de détails dans les ombres et aussi avoir moins de grains ?
Je connais beaucoup de photographe qui applique cette méthode en photographie et ils disent qu’il y a plus détails mais je ne suis pas sûr de leurs réflexions.
l’ETTR (Expose to the Right) permet généralement de capturer un plus grand nombre de couleurs utiles, notamment dans les zones sombres de l’image, en réduisant le bruit et en maximisant la qualité des données capturées.
Dites moi ce que vous en pensez ?
Merci 🙏
Whatever the hell "middle grey" even is, objectively. 0:57
You know there are different ‘skin tones’ right?
I am sure he was talking about average Caucasian skin tone, for which this is true. Obviously different skin tones produce a different result.
I am always open to constructive criticism and suggestions for improvement, but I do not appreciate your comment. I try to bring the most accurate information possible to help all of us improve the quality of our work. I did not create those references and the charts presented, those are industry-standard tools aimed at providing a general framework for cinematographers and photographers. While these references might not cover every nuance of skin tone, they serve as a starting point for professional calibration.Sad to see this conversation happening here, I hope you continue to enjoy my free content without feeling the need to voice unnecessary criticisms that doesn't move the conversation forward.
Word salad.
I laughed so hard at this comment! Thank you!!! You are totally right, if only I knew how to explain this using real words…. Thanks again 👍
@cnomadic type it out. Read it at a rubber duck as you fake record. This will allow you to put the words right, and work on flow. Changing one word can force a rewrite of several paragraphs. Make sure what you write is how you will spaleak it. So only a few cues will make the words flow.
Make some kinds of teleprompter. (UA-cam it, several ways of doing it.)
Why don’t you, as a pro photographer, have a printed gray-scale chart. It’s a basic tool lol!