Canon shooter here: we can go below native ISO, it shows a little “L” next to the ISO. I’ve experienced the same thing, I’ll lose stops in the highlights. But, if my scene doesn’t have any bright highlights, or is naturally a darker scene (night shot, moodier, or whatever) then purposely going below native definitely gives cleaner images on the R5. The C70 doesn’t make as a big of a difference because the DGO is pretty good at keeping clean shadows anyhow. This was all just by eye though, so I’m thankful for your testing with some real equipment and science!
You can do it two ways.. lowering iso bellow native, lets say to 400 = you lose 1 highlight stop but gain 1 shadow stop thx to lower noise OR just simply keep 800 iso and expose 1 stop beyond, you lose that 1 stop again, but gain more light = less noise = 1 shadow stop more... same result. You will always have that lets say 12 stops no matter what ISO you go.. But one more thing I found last week and its pretty crazy.. I shot one scene with 800 iso and then the same scene with 10 000 ISO, when I brought down the 10 000 ISO footage to 800 ISO exposure level.. guess what.. the 10K one was MUCH cleaner!!! WHAT THE HELL??? So I can even use ISO for ETTR?!? Can you explain @Gerald Undone ? Much love 〽
“I brought down the 10,000 ISO footage to 800 ISO exposure level” Can you say that in a different way? I don’t understand what you mean. Are you saying when you adjusted exposure in your editing program? How do you know you brought it down to 800 ISO? Did you lower exposer by 3.67?
This is the type of content I live for on this channel: the most random, in depth thing that no one else will ever probably look into but yet super helpful/important to know & honest cereal reviews are great too :D
My understanding from the photography standpoint is this: The Native ISO (100 for photo) gives you regular Highlights and Shadows. If you go below native in the extended Range you clip your Highlights MUCH quicker but you gain clean shadows. Basically "shifting" your dynamic range to allow heavier shadow recovery if needed with the compromise of clipping highlights way sooner.
Yes, this is generally the case. If your scene doesn’t have bright highlights you want to preserve lowering ISO is a perfectly fine method to achieve desired exposure if you want to do it in camera. For creative control overexposing the scene and then correcting in post will always produce cleaner results in the shadows. But honestly, most pro level cameras delivery plenty of shadow details these days.
Spot on - I always remember: the native and above require you to give the camera less light, preserving the highlights; and the below-natives make you give the camera more light, cleaning up the shadows. Both useful at the right time--thanks for the tests!
Gerald I have learned so much from your videos. But sometimes im confused as to what I should pay attention to and what I shouldnt as a novice. You should make a videography 101 course for sony cameras
You’ve been a big part of my camera journey and thank you! I find myself back to your channel when caught with technical questions in the middle of the night.
It's basically EI, using lower ISO means lower EI baked into the image. The analog gain remains same at native when shooting at lower ISO. You compensate for lower ISO's darker image by opening iris and lowering shutter, resulting an overexposed image but already corrected that overexposure for you when recorded, so it may have better shadow information (lower noise) but less highlight information (clipping faster).
@@geraldundoneyou’re baking the mluts exposure In but not the lut itself. Which in a way is sortof more convenient for post. Because if you’re shooting at EI 200 while at the base 640 or 800 depending which camera you’re using, the benefit to using the iso way is that when you get to post you don’t have to manually bring down the exposure. Whereas with the ei version, your shot will look two stops too bright, which you just manually bring down in post. I don’t really see the benefit there as you’re getting the same benefit with both and the same trade off in the highlights using either method. Seems like ei might just be giving you more work in post for the same effect on your image. Is there any real benefit to using ei then?
@jonathanmasters2178 - That's a good point and an interesting idea. I believe the point and benefit of EI in those circumstances is consistency. Increasing/decreasing gain can impact saturation, white balance, luminance of practicals, etc. So when you hand footage off to the colorist, the results are more consistent with a constant EI and that skilled professional can adjust exposure as needed without worrying about those other changes I mentioned.
@@jonathanmasters2178 exactly what I do.. I rather overexpose and lose stops rather than lowering clipping point with lower ISO.. so I dont have to bring clipping point back to 100 IRE in post.. right?
This is great. I feel like when I am in my shop filming I have started going below native for my last couple of video and felt like I was giving something up, but now I know in the nice controlled environment it is totally worth it and actually simplifies my filming a lot.
This answers so many questions for me. I would just drop below native whenever it was too bright and it was driving me crazy trying to figure out why some shots had a lower white point. I have always appreciated your videos, but this one got me to subscribe to your channel.
Yeah, I would say the basic rule is signal to noise ratio. The more light you can expose the sensor to, the lower the noise floor will be. The higher the ISO, the more gain the sensor has, including gain to the noise floor, and the reason lower ISOs result in cleaner images is not because of the ISO, but because of the increased light hitting the sensor.
well, I am an old film and paper guy, so when I come across this, it surprises me that anyone has broken it down to this point. I remember Ansel Adams and his Zone System analysis, and pushing and pulling film to gain control of contrast and detail in the shadows and highlights.Digital has far more latitude, but there are some similarities. Ansel Adams had hoped to live long enough to see the things that you are talking about, but he only saw the beginnings of the digital revolution. But anyway, this is really good work!
Thanks! This was very helpful. Most people just like to repeat what they hear, but it's really nice when someone actually takes the time to study and test what happens, thanks for sharing!
Craziest thing - At 10:34am I was doing some rather unscientific tests with shooting below the Native ISO of 800 on my Canon R7 with Clog3 and Cinema Gamut... Just happy to know that I'm not the only person who lie's awake at night pondering these important questions.. 😂 LOVE your channel, brother.. Cheers from Milton, Ontario..
I also found on my R6 that if you push exposure to the right with ISO and lowering exposure in post, you get much cleaner video. What I am saying is that I shot same scene with 800 ISO and 10 000 ISO, I brought back down the 10 000 ISO one in post and it was much more or more like noise free video.. I really dont know what is going on. But try for yourself.
Thank you, I've always wondered about this! (Btw, I would KILL for a video like this explaining superwhites and what's the point of having them and how are you supposed to use them, and also if the 16-255 RGB range that some cameras have has anything to do with that)
Super-whites and sub-blacks (and all of their other names) can be great at times. I've once had a very underexposed video but was able to save it because there was that little bit of extra information in the 0-63 range that showed as black but actually had recoverable detail. That was on my Fuji X-T3 and X-T4 in FLog. Generally, I just keep the files in full data-level(0-1023) instead of video data-level (64-959) when editing or re-encoding camera source files. For delivery I chose video data-level as that's the norm.
@@DesertCookie Do you talk about RAW or 10bit or ALL-Intra or what those numbers (0-1023 and 64-959) means, and where to set it in camera? like Shooting into RAW (into atomos/odyssey recorder with sony cam) because i can restore some stops in edit later, if i miss exposed shot and its overexposed or underexposed.
Another example of why you are the best at what you do! I am a run and gun event shooter and have always defaulted to just drop ISO below when I had too...and never really noticed a difference...now I know why. Thank you!
As a general rule, when I'm filming a bright scene I shoot at native or a little over to help with highlight roll off, when shooting a dark scene, I usually go a little below native to retain as much shadow detail. Counter intuitive, but it's all about what you need to retain most... and requires lights
I’ve been waiting for you to do this video for years. I’ve always noticed (anecdotally) that I was losing dynamic range, but I never knew how much I was losing. So I often stop down my aperture and raise my iso. Obviously there are certain situations where I want shallow depth of field but for the most part I can get away with stopping down the aperture.
One thing not discussed is if shooting log and using a LUT or colour managed workflow the LUT or other colourspace transform is going to expect certain things to be a certain levels. For example as the nominal "correct" exposure for S-Log3 puts white at 61% and you will clip at 94% most LUT's will have a highlight rolloff that starts at 61% input and map 94% input to 100% output. But if you record at a lower than base ISO the clip point becomes lower, so the first issue is the LUT's output won't ever reach 100%. Assuming you expose so the mid range is similar to "normal" then the rolloff range from the LUT will also be much reduced (white gets closer to clip, so rolloff range is reduced). Whether you grade before or after the LUT to get back to a 100% output you will need to stretch out the highlights which is not ideal. For the best results you will want LUT's that are designed for the altered and reduced input range.
As an a7siii shooter I really appreciate you taking the time to figure this out for us! I always wondered and heard it was BADDDD hahah but I never tested it. Thanks for doing all the hard work! You’re the best!
I absolutely love the technical breakdown in great detail. The most difficult part for me is... I shoot run and gun most of the time, it would be impossible to maintain the proper adjus😊tments while filming live action.
(I think?) Over exposing Log and then correcting it in post relies on increased light through scene, shutter speed or aperture rather than raising the ISO. Thank you for the run through, very insightful and straight to the point as usual :)
Thanks so much for your in-depth tests! I was just wondering this very topic this week! Now i know below native i don't recover highlights! Less dynamic range in the highlights but less noise in the shadows.
I was lost for a lot of this video, but got it at the end... thank you for this deep dive and explanation, it was really helpful. My take-away is: 1. I don't absolutely need to use ND's while shooting outdoors in SLOG 2. I'm gonna start using my zebra's to aim for nailing exposure for skin in SLOG, so in theory, even when going below the native ISO, I should still get good exposures as long as I have a good range on the lower zebra limit.
I needed this video! Thank you... I shoot a lot outdoors. Sometimes, I do not have appropriate ND filters and need to drop down ISO to expose properly. Now I will not be to bothered about ruining the footage at lower ISO.
Great video, unique topic. The reason why shooting with Sony is so frustrating is due to the presence of noise in their base ISO. The base ISO set by the manufacturer should be determined considering the signal-to-noise ratio and maximum dynamic range, not solely based on dynamic range. This is why pure digital cinema cameras are excellent and unlikely to be replaced in the near future. With Alexa, you can expose 2 stops over and under and still recover everything, whereas with Sony, it needs to be done perfectly.
Nicely done! It’s great having the information. I use the a1 and a7C and haven’t had any issues with losing highlights or shadow details. Setting the zebras and understanding what exposure settings to use and when is the first step. The amount of dynamic range on these sensors is really impressive. Unless someone is really lost on proper exposure techniques, I don’t understand how they would have issues with clipping highlights. I guess it’s more about the possibilities than a normal issue. Thankfully the base ISO on the a1 is 100. I remember the days of bracketing to make sure you got the image. There’s really no reason for that anymore except maybe a rare situation. These cameras are absolutely amazing.
Super appreciate the time and effort you put into these tests and illustrating them to those who care. This channel really deserves more subs than some others. Keep it up Gerald!
This was really educational, and It was great you gave a real scenario example of when you might use it right after all the chart and tech info! Thanks for this video! 👏🏼👏🏼
Damn, you are extremely thorough! What a great video. I was recently in Ghana and was afraid of my footage because I had to shoot almost everything at way lower ISO. Footage was fine, and this video confirmed it. Thanks Gerald!
In my experience with Canon & RED cameras I find I only need to increase the iso for the stops of light I can't control in a scene. Most if not all mirrorless cameras currently have sensor's designed for still applications and produce the best results at iso 100. Cameras like the R5 are 14bit stills images, but Canon has to compress (remove 2 bits of data) for a 12bit video file and cram it into a log curve. No one has really talked about this yet, would love to see you take it on
I clicked on a video in hopes that I would learn something about Native SIO. Everything flew over my head. So I really I have learned that there is much I don't know.
It’s simple but counterintuitive. When shooting a scene that has most of its tone below a neutral grey, reduce your iso and get down to LIGHTING like the DP you fantasize that you are. When shooting scenes that have bright light, up the ISO and break out your ND filters and bounce to start sculpting. Dynamic range is a sliding scale when you have ISO control. Even if you’re shooting raw, this method ensures that you’re lighting the scene correctly for the camera to have what it needs for the cleanest image.
literally yesterday I dived deep into this topic as it was at the back of my head for super long.. And today, boom, your video with perfect explanation. Thanks a lot! Great work as always 🙌
if you raise ISO from 800 to 1600 instead of lowering it to 400 you should gain a stop that is not clipped instead of losing one. Raising ISO protects highlights and lowering ISO protects shadows. (building picture).
The “bigger picture” is one of scene contrast (measured in stops) vs. the practical range (measured in f/stops) of the sensor. My test for practical range is to photograph a test target consisting of white and black terry wash rags draped over a gray card and color chart. Photographing the target in flat, cross and back lighting in different lighting conditions (clear, cloudy, overcast) with exposure keyed to keeping the non specular highlights in the white towel below clipping will provide a practical understanding of how a camera sensor is able to cope with the contrast of those different lighting conditions. Regardless of what ISO setting is used when the contrast of the scene exceeds the optimal (native ISO) range of the sensor scene detail will need to be sacrificed in either the highlights or shadows, or supplemental lighting will be needed via reflectors or artificial sources. My background is still portrait photography and the ideal lighting scenario outdoors for portraits is to put the sun behind the subject as rim light, setting exposure to keep the highlights it creates below clipping. Then in the front of the subject use fill and key lights as in a studio: Start with fill and raise it until detail is seen in the front of the black towel target, then add the key over it at the same downward / sideways angle of the skylight until the highlights on the “shaded” front of the white towel are lifted to just below those of the sun-lit highlights. But before setting the lights in front look critically at how the skylight is modeling the faces. If the subject is looking at the horizon their brows will always shade the eye sockets. No matter how much artificial fill and key lighting is added the eyes will always look dull compared to nose/cheeks/chin where the skylight reaches. So the first step is to get the subjects looking up into the sky at about a 30° angle which will require raising the POV of the camera - as done with gantries in the big budget Hollywood movies. Raising the POV of the camera to keep plane of sensor parallel with plane of face eliminates perceptive on the faces and helps eliminate distracting backgrounds. Don’t have a gantry? Mount the camera head to a board and clamp it to the top of a step ladder. When the subject’s faces are lifted to get the ambient light in the eyes and turned to get the desired lighting pattern the key light needs to be added from the same downward angle and direction to match or the lighting pattern on the face will become a crossed-shadow muddle. That will require taller stand than an indoor studio setting. With the sunlight used as rim light, with key and fill sources in front to match the angle of the skylight modeling and fill it is possible to match the range of the scene to that of the sensor and record the maximum dynamic range the camera is capable of. But most of the time that is not practical.
Fantastic explanation and demonstration. Thank you Gerald! Intuitively, I figured you’d lose highlight stops. But I was completely unaware that she shadows actually clean up. So I’ll definitely keep that in mind. 😊
Exposing for middle grey and lowering ISO implicitly means that you increase the actual exposure (ie you move from f/4 to f/2.8 or from 1/60s to 1/30s etc) So those dark regions actually emit more photons that the sensor picks up to work with. On the other side a greater area of the scene emits more photons than the sensor can measure, thus it is clipped.
Close..it's not that more photons are emitted (a constant light source is constant) its that the wider aperture gathers more photons [f4 to f2.8] or the exposure duration is longer (1/60 to 1/30) doubling the number photons acquired. Consequently more "signal" is collected.
That is a very interesting result. I just tried to avoid it so far since I was not quite sure what exactly I was trading or giving up. Basically this means that in a controlled home scene you might as well just shoot with the low iso value anyways since you have sufficient controlled light, don't need the extra noisy DR stops anyways and can lower the background noise even further.
That is useful information. I was always scared, because even if I didn't see much difference in my testing, I figured as soon as I tried to use it while shooting for a client it would wreck me somehow. So this is great knowing exactly how it works. Thank you for putting the time in!
(Long ass reply here).. I guarantee you, your client cares zero about your ISO, unless they themselves are photographers/judges. The client, unless they have the technical know-how, just want to know that whatever they asked you to photograph, looks good. That's it. And it needs to look good at first glance. If you can pass the first glance test, that's it, everything else (technically) is just improving your product (but has nothing that the client will care for). In fact, when your client is trying to make a decision between you and someone else, they won't be thinking, "this image has more/less white dots". They don't even know to look for that. They are going to first think about the purpose of the photo/product and ask themselves, can this photographer convey this purpose or give me what I want (by looking at your portfolio)? I find all this technical detail, and care, is much ado about nothing. Hardly any of this makes or breaks a photo, unless at drastic degrees (eg: very high ISO). The important things that make a photo great is not in the scientific details (I say this as someone with a major in General Physics and in Computer Science...science is my business, but my thoughts come from the artistic side of me as someone born to parents who are painters, sculptor and designers). Again, clients want to know, "can this photographer give me what I want?". 10/10 times what they want is a specific look and feel, not technical greatness. I've seen some shitting photos, with shitty retouching (I hate retouching at least the majority I've seen outside the fashion industry), and the client is just goo-goo gaa-gaa over it. Because they aren't seeing it from a photographers eyes, and that's what we need to understand. Improve your craft (artistic creativity) but don't get bogged down in the technicalities. Use the technicalities to exploit towards your creativity.
Leveraging one's technical knowledge to benefit what a client wants and expects is important, i.e., knowing about the relevant techniques, gear selection and usage and artistry are a synergistic package. I do agree that most clients are simply interested in results that support their needs, thus usually best to avoid offering up unasked for techie talk with them unless they ask a question about such, and then carefully and briefly titrate one's reply.
Fuji has their lowest ISO at 80, which is one below the native of 160. It is marked with an L on the dial (on those cameras that still have a dial). In video you'll find other "native" ISOs in which the noise cleans right up (800 and 200, for FLog on the X-Trans 3 sensor of the X-T3, X-T4, ...) that don't really affect dynamic range though (as far as I've noticed).
@@neramp559 The X-T3 and X-T4 (and possibly other models with the same sensor and image processor) have multiple pre-amp stages. They behave a lot like dual native ISOs but technically aren't. For the end user the result is the same: Lower image noise at certain ISO settings. I shot my first project at ISO 640 as that was the lowest FLog allowed; a month after shooting I found out that there is another pre-amp-stage at ISO 800 that would've resulted in significantly less noise. I now only shoot FLog on my X-T3 at ISO 800 and 2000 or switch to HLG as that uses the pre-amps for Rec.709 which means at ISO 3200 it is about as clean as FLOG at ISO 2000; I trade some dynamic range capabilities for a cleaner image. I've done a lot of tests myself and threw a few comparisons together here: nc.rhprivat.de/s/W8NG3JpPZ2MzCKD
My brain hurts. I've watched this twice. I'll have to watch it again to really understand it and take it in. But, it's great information and is challenging my knowledge gaps. Thank you for putting it out there.
I'm a newbie to photography (1 year now I own the Sony A7IV) - Even if I'm not understanding the subject in its full extension: In any case, it's still interesting to watch your "lessons" - watching you share your experiences for us to refer to!
Same is applied in all cameras. I've learned it from my BMPCC6K. It's like counterintuitive for beginners, because when You lower the ISO You bring the exposure down, but the rule is "lower ISO - more information in shadows, higher - more in highlights". So we have to choose ISO if we want range in the shadows or highlights and then use aperture, filters and speed/angle.
I’ve always shot at base or higher, so it’s good to know that in a pinch I could go down to the extended ISO range won’t be as terrible as I had been told.
This was an amazing video, very in-depth and I am a sucker for going as low as possible but not realizing im looing in DR. Thank you and it makes me reconsider how I shoot.
This is a good illustration of how lower ISO values should be used for shooting in low light and higher ISO values should be used in bright settings. At higher ISO values, your clipping point is as close as possible to 100 IRE. And as you push the ISO higher, it shifts middle grey up towards the highlights, creating more artificial dynamic range. And although that would technically add noise, it doesn't matter, since it's all hot exposure values. Just as long as you don't have any larges areas of shadow in your shot too. And for the reverse, you can push your ISO down below native when shooting in dark settings. Then just add as much light as possible so that your shadows will be extremely clean and, just as before, you'll have artificial shadow dynamic range, since middle grey is pushed down toward black. And you won't miss the loss of dynamic range in the highlights unless you actually have a bright point of light in your shot. So if you're shooting at night, all you have to worry about and streetlights, practicals, etc. Rather than worrying about the sun and the bright midday sky.
You actually taught me some new stuff. Signal-to-noise ratio seems to be the most important point of selecting the ISO. How much can be recovered with Topaz or some other noise reducing software would be an interesting question. Some people have extreme opinions about the subject with no data to present.
I find that this affects and applies to live video productions as well. I regularly use full-frame mirrorless cameras for live broadcast situations with very dynamic lighting. Aperture and overall look is way more of a priority than minute differences in dynamic range or noise in the shadows. We often have to dip into the extended ISO to match exposure camera to camera on the fly. Experientially, I've found that doing this doesn't have a noticeable affect on the overall image quality and it's reassuring to have this deep dive on the topic that proves out what I've felt was true: IE using extended ISO correctly doesn't have a significant affect on image quality.
Mr. Gerald , Since you have hundreds of thousands of viewers the odds of you seeing this...but I'm optimistic. Your info is amazing and , when I watch it a few times , always on point. Just wondering if you have seen the 7 artisan cine lenses. Some of the images Roman Hense gets with them is amazing. He's also a great colorist, so there's that. I gather they are made to mimic eis lens? Thanks for the time and effort you spend on these videos. The whole community really appreciates you
Great video. It seems counterintuitively that, in the end, you should use lower ISO for darker indoor scenes with controlled lighting, and native ISO for brighter outdoor scenes with larger dynamic range.
The lower you go with the iso the more detail the camera captures in the shadows and the higher you go with the iso (practically speaking) the more detail it captures in the highlights. That's why the iso 200 is able to beat the iso 4000, the best result would've been lighting like youre shooting for 100 and shoot at 200 to get the best result in shadow noise, loosing a little bit of the highlights.
Thanks Gerald for the insightful video! It seems like shooting below native ISO is a convenient substitute for ETTR and grading back down to reduce noise.
Canon shooter here: we can go below native ISO, it shows a little “L” next to the ISO. I’ve experienced the same thing, I’ll lose stops in the highlights. But, if my scene doesn’t have any bright highlights, or is naturally a darker scene (night shot, moodier, or whatever) then purposely going below native definitely gives cleaner images on the R5. The C70 doesn’t make as a big of a difference because the DGO is pretty good at keeping clean shadows anyhow. This was all just by eye though, so I’m thankful for your testing with some real equipment and science!
Thanks for chiming in. Great comment!
You can do it two ways.. lowering iso bellow native, lets say to 400 = you lose 1 highlight stop but gain 1 shadow stop thx to lower noise OR just simply keep 800 iso and expose 1 stop beyond, you lose that 1 stop again, but gain more light = less noise = 1 shadow stop more... same result. You will always have that lets say 12 stops no matter what ISO you go..
But one more thing I found last week and its pretty crazy..
I shot one scene with 800 iso and then the same scene with 10 000 ISO, when I brought down the 10 000 ISO footage to 800 ISO exposure level.. guess what.. the 10K one was MUCH cleaner!!! WHAT THE HELL??? So I can even use ISO for ETTR?!? Can you explain @Gerald Undone ?
Much love 〽
@@getmarked really? on what camera did you do it, any references?
@@hannestell canon r6. I would love to post it somewhere. Yeah. Its crazy
“I brought down the 10,000 ISO footage to 800 ISO exposure level”
Can you say that in a different way? I don’t understand what you mean. Are you saying when you adjusted exposure in your editing program? How do you know you brought it down to 800 ISO? Did you lower exposer by 3.67?
This is the type of content I live for on this channel:
the most random, in depth thing that no one else will ever probably look into but yet super helpful/important to know
& honest cereal reviews are great too :D
My understanding from the photography standpoint is this:
The Native ISO (100 for photo) gives you regular Highlights and Shadows.
If you go below native in the extended Range you clip your Highlights MUCH quicker but you gain clean shadows.
Basically "shifting" your dynamic range to allow heavier shadow recovery if needed with the compromise of clipping highlights way sooner.
Yes, this is generally the case. If your scene doesn’t have bright highlights you want to preserve lowering ISO is a perfectly fine method to achieve desired exposure if you want to do it in camera. For creative control overexposing the scene and then correcting in post will always produce cleaner results in the shadows. But honestly, most pro level cameras delivery plenty of shadow details these days.
Solid explanation dude! That’s super interesting about the cleaner shadows at the “extended”settings
Spot on - I always remember: the native and above require you to give the camera less light, preserving the highlights; and the below-natives make you give the camera more light, cleaning up the shadows. Both useful at the right time--thanks for the tests!
great tip, thanks !
You called?
Haha. Oh that's right, I teased you for shooting below native during your bag tour, didn't I? 🤓💜
@@geraldundone I still remember it vividly🥲
I cant really find info in the onlije manual of a7iv. Sony doesnt talk about it but many people on youtube are talking about dual iso and stuff
Gerald I have learned so much from your videos. But sometimes im confused as to what I should pay attention to and what I shouldnt as a novice. You should make a videography 101 course for sony cameras
@@rumorscamerasSony’s standards are too high to call them dual native iso
You’ve been a big part of my camera journey and thank you! I find myself back to your channel when caught with technical questions in the middle of the night.
It's basically EI, using lower ISO means lower EI baked into the image. The analog gain remains same at native when shooting at lower ISO. You compensate for lower ISO's darker image by opening iris and lowering shutter, resulting an overexposed image but already corrected that overexposure for you when recorded, so it may have better shadow information (lower noise) but less highlight information (clipping faster).
Yeah, I had a similar thought too, but with the MLUT baked, yeah.
@@geraldundoneyou’re baking the mluts exposure In but not the lut itself. Which in a way is sortof more convenient for post. Because if you’re shooting at EI 200 while at the base 640 or 800 depending which camera you’re using, the benefit to using the iso way is that when you get to post you don’t have to manually bring down the exposure. Whereas with the ei version, your shot will look two stops too bright, which you just manually bring down in post. I don’t really see the benefit there as you’re getting the same benefit with both and the same trade off in the highlights using either method. Seems like ei might just be giving you more work in post for the same effect on your image. Is there any real benefit to using ei then?
Why I didnt know this earlier! 😭😭
@jonathanmasters2178 - That's a good point and an interesting idea. I believe the point and benefit of EI in those circumstances is consistency. Increasing/decreasing gain can impact saturation, white balance, luminance of practicals, etc. So when you hand footage off to the colorist, the results are more consistent with a constant EI and that skilled professional can adjust exposure as needed without worrying about those other changes I mentioned.
@@jonathanmasters2178 exactly what I do.. I rather overexpose and lose stops rather than lowering clipping point with lower ISO.. so I dont have to bring clipping point back to 100 IRE in post.. right?
This is great. I feel like when I am in my shop filming I have started going below native for my last couple of video and felt like I was giving something up, but now I know in the nice controlled environment it is totally worth it and actually simplifies my filming a lot.
"If you turn down the ISO below a native ISO, you don't recover highlights." Learned something new today - thank you!
This is a very educational video. At 63, I still learned something, Well Done, Gerald!
36, saluting this attitude. 😄❤️
The interesting thing with the creative world is that there's endless knowledge. If you even thing you've learned it all - you 're out.
This answers so many questions for me. I would just drop below native whenever it was too bright and it was driving me crazy trying to figure out why some shots had a lower white point. I have always appreciated your videos, but this one got me to subscribe to your channel.
Yeah, I would say the basic rule is signal to noise ratio. The more light you can expose the sensor to, the lower the noise floor will be. The higher the ISO, the more gain the sensor has, including gain to the noise floor, and the reason lower ISOs result in cleaner images is not because of the ISO, but because of the increased light hitting the sensor.
well, I am an old film and paper guy, so when I come across this, it surprises me that anyone has broken it down to this point. I remember Ansel Adams and his Zone System analysis, and pushing and pulling film to gain control of contrast and detail in the shadows and highlights.Digital has far more latitude, but there are some similarities. Ansel Adams had hoped to live long enough to see the things that you are talking about, but he only saw the beginnings of the digital revolution. But anyway, this is really good work!
Thanks! This was very helpful. Most people just like to repeat what they hear, but it's really nice when someone actually takes the time to study and test what happens, thanks for sharing!
Craziest thing - At 10:34am I was doing some rather unscientific tests with shooting below the Native ISO of 800 on my Canon R7 with Clog3 and Cinema Gamut... Just happy to know that I'm not the only person who lie's awake at night pondering these important questions.. 😂
LOVE your channel, brother.. Cheers from Milton, Ontario..
I also found on my R6 that if you push exposure to the right with ISO and lowering exposure in post, you get much cleaner video. What I am saying is that I shot same scene with 800 ISO and 10 000 ISO, I brought back down the 10 000 ISO one in post and it was much more or more like noise free video.. I really dont know what is going on. But try for yourself.
Thank you, I've always wondered about this! (Btw, I would KILL for a video like this explaining superwhites and what's the point of having them and how are you supposed to use them, and also if the 16-255 RGB range that some cameras have has anything to do with that)
Is for HDR
Super-whites and sub-blacks (and all of their other names) can be great at times. I've once had a very underexposed video but was able to save it because there was that little bit of extra information in the 0-63 range that showed as black but actually had recoverable detail. That was on my Fuji X-T3 and X-T4 in FLog.
Generally, I just keep the files in full data-level(0-1023) instead of video data-level (64-959) when editing or re-encoding camera source files. For delivery I chose video data-level as that's the norm.
@@DesertCookie Do you talk about RAW or 10bit or ALL-Intra or what those numbers (0-1023 and 64-959) means, and where to set it in camera?
like Shooting into RAW (into atomos/odyssey recorder with sony cam) because i can restore some stops in edit later, if i miss exposed shot and its overexposed or underexposed.
Another example of why you are the best at what you do! I am a run and gun event shooter and have always defaulted to just drop ISO below when I had too...and never really noticed a difference...now I know why. Thank you!
Gerald - you are so technically knowledgable - it is a pleasure to listen to your explanations. Thanks for all the hard work :)
As a general rule, when I'm filming a bright scene I shoot at native or a little over to help with highlight roll off, when shooting a dark scene, I usually go a little below native to retain as much shadow detail. Counter intuitive, but it's all about what you need to retain most... and requires lights
I’ve been waiting for you to do this video for years. I’ve always noticed (anecdotally) that I was losing dynamic range, but I never knew how much I was losing. So I often stop down my aperture and raise my iso. Obviously there are certain situations where I want shallow depth of field but for the most part I can get away with stopping down the aperture.
One thing not discussed is if shooting log and using a LUT or colour managed workflow the LUT or other colourspace transform is going to expect certain things to be a certain levels. For example as the nominal "correct" exposure for S-Log3 puts white at 61% and you will clip at 94% most LUT's will have a highlight rolloff that starts at 61% input and map 94% input to 100% output. But if you record at a lower than base ISO the clip point becomes lower, so the first issue is the LUT's output won't ever reach 100%. Assuming you expose so the mid range is similar to "normal" then the rolloff range from the LUT will also be much reduced (white gets closer to clip, so rolloff range is reduced). Whether you grade before or after the LUT to get back to a 100% output you will need to stretch out the highlights which is not ideal. For the best results you will want LUT's that are designed for the altered and reduced input range.
As an a7siii shooter I really appreciate you taking the time to figure this out for us!
I always wondered and heard it was BADDDD hahah but I never tested it.
Thanks for doing all the hard work!
You’re the best!
I absolutely love the technical breakdown in great detail. The most difficult part for me is... I shoot run and gun most of the time, it would be impossible to maintain the proper adjus😊tments while filming live action.
(I think?) Over exposing Log and then correcting it in post relies on increased light through scene, shutter speed or aperture rather than raising the ISO. Thank you for the run through, very insightful and straight to the point as usual :)
Thanks so much for your in-depth tests! I was just wondering this very topic this week!
Now i know below native i don't recover highlights! Less dynamic range in the highlights but less noise in the shadows.
I was lost for a lot of this video, but got it at the end... thank you for this deep dive and explanation, it was really helpful. My take-away is:
1. I don't absolutely need to use ND's while shooting outdoors in SLOG
2. I'm gonna start using my zebra's to aim for nailing exposure for skin in SLOG, so in theory, even when going below the native ISO, I should still get good exposures as long as I have a good range on the lower zebra limit.
I've always wondered about this. Thanks for covering this Gerald!
Same. My OCD thanks you, Gerald.
I have a quick video for you today and it's 15 minutes.
I love you Gerald
perfect way to keep your shutter to 180 degree if your ND is not strong enough on bright lenses. Thanks for sharing Gerald!
I struggled for a long time to actually understand how ISO really works, this video really helped thank you :)
I’m a self thought videographer and I use A7S 3… just wanted to say you’re a wizard and I love your videos. This was very helpful!
Dude, you are a freaking genius. I have no idea how you get this nerdy on this stuff and make it clear to us. Thanks Gerald!!
I needed this video! Thank you... I shoot a lot outdoors. Sometimes, I do not have appropriate ND filters and need to drop down ISO to expose properly. Now I will not be to bothered about ruining the footage at lower ISO.
Considering the stops lost in highlights when dropping down ISO, I might still need ND filters, especially outdoors requiring more highlight DR.
Great video, unique topic. The reason why shooting with Sony is so frustrating is due to the presence of noise in their base ISO. The base ISO set by the manufacturer should be determined considering the signal-to-noise ratio and maximum dynamic range, not solely based on dynamic range. This is why pure digital cinema cameras are excellent and unlikely to be replaced in the near future. With Alexa, you can expose 2 stops over and under and still recover everything, whereas with Sony, it needs to be done perfectly.
Nicely done! It’s great having the information. I use the a1 and a7C and haven’t had any issues with losing highlights or shadow details. Setting the zebras and understanding what exposure settings to use and when is the first step. The amount of dynamic range on these sensors is really impressive. Unless someone is really lost on proper exposure techniques, I don’t understand how they would have issues with clipping highlights. I guess it’s more about the possibilities than a normal issue. Thankfully the base ISO on the a1 is 100.
I remember the days of bracketing to make sure you got the image. There’s really no reason for that anymore except maybe a rare situation. These cameras are absolutely amazing.
The most useful video on ISO on the entire internet.
Super appreciate the time and effort you put into these tests and illustrating them to those who care. This channel really deserves more subs than some others. Keep it up Gerald!
I love how you speak to your audience :) Thank you for figuring this out/sharing for us! :D
Thank you! ❤
This was really educational, and It was great you gave a real scenario example of when you might use it right after all the chart and tech info!
Thanks for this video! 👏🏼👏🏼
Damn, you are extremely thorough! What a great video. I was recently in Ghana and was afraid of my footage because I had to shoot almost everything at way lower ISO. Footage was fine, and this video confirmed it. Thanks Gerald!
Great video, I was always curious about this. You answer all my questions mate, amazing
THIS is the specific type of content I am looking for... ❤
In my experience with Canon & RED cameras I find I only need to increase the iso for the stops of light I can't control in a scene. Most if not all mirrorless cameras currently have sensor's designed for still applications and produce the best results at iso 100. Cameras like the R5 are 14bit stills images, but Canon has to compress (remove 2 bits of data) for a 12bit video file and cram it into a log curve. No one has really talked about this yet, would love to see you take it on
We Need Gerald Always!!!! It was Awesome Meeting you in New York Sony Creative Space!!!!!!
I clicked on a video in hopes that I would learn something about Native SIO. Everything flew over my head. So I really I have learned that there is much I don't know.
It’s simple but counterintuitive. When shooting a scene that has most of its tone below a neutral grey, reduce your iso and get down to LIGHTING like the DP you fantasize that you are. When shooting scenes that have bright light, up the ISO and break out your ND filters and bounce to start sculpting. Dynamic range is a sliding scale when you have ISO control. Even if you’re shooting raw, this method ensures that you’re lighting the scene correctly for the camera to have what it needs for the cleanest image.
Yep, I use lower ISO specifically for less noise in darker areas of scenes where high dynamic range is less relevant to my needs
Thanks for doing this test for us, another useful bit of knowledge to have in the toolkit while shooting!
literally yesterday I dived deep into this topic as it was at the back of my head for super long.. And today, boom, your video with perfect explanation. Thanks a lot! Great work as always 🙌
if you raise ISO from 800 to 1600 instead of lowering it to 400 you should gain a stop that is not clipped instead of losing one. Raising ISO protects highlights and lowering ISO protects shadows. (building picture).
The “bigger picture” is one of scene contrast (measured in stops) vs. the practical range (measured in f/stops) of the sensor. My test for practical range is to photograph a test target consisting of white and black terry wash rags draped over a gray card and color chart. Photographing the target in flat, cross and back lighting in different lighting conditions (clear, cloudy, overcast) with exposure keyed to keeping the non specular highlights in the white towel below clipping will provide a practical understanding of how a camera sensor is able to cope with the contrast of those different lighting conditions.
Regardless of what ISO setting is used when the contrast of the scene exceeds the optimal (native ISO) range of the sensor scene detail will need to be sacrificed in either the highlights or shadows, or supplemental lighting will be needed via reflectors or artificial sources.
My background is still portrait photography and the ideal lighting scenario outdoors for portraits is to put the sun behind the subject as rim light, setting exposure to keep the highlights it creates below clipping. Then in the front of the subject use fill and key lights as in a studio:
Start with fill and raise it until detail is seen in the front of the black towel target, then add the key over it at the same downward / sideways angle of the skylight until the highlights on the “shaded” front of the white towel are lifted to just below those of the sun-lit highlights.
But before setting the lights in front look critically at how the skylight is modeling the faces. If the subject is looking at the horizon their brows will always shade the eye sockets. No matter how much artificial fill and key lighting is added the eyes will always look dull compared to nose/cheeks/chin where the skylight reaches. So the first step is to get the subjects looking up into the sky at about a 30° angle which will require raising the POV of the camera - as done with gantries in the big budget Hollywood movies. Raising the POV of the camera to keep plane of sensor parallel with plane of face eliminates perceptive on the faces and helps eliminate distracting backgrounds. Don’t have a gantry? Mount the camera head to a board and clamp it to the top of a step ladder.
When the subject’s faces are lifted to get the ambient light in the eyes and turned to get the desired lighting pattern the key light needs to be added from the same downward angle and direction to match or the lighting pattern on the face will become a crossed-shadow muddle. That will require taller stand than an indoor studio setting.
With the sunlight used as rim light, with key and fill sources in front to match the angle of the skylight modeling and fill it is possible to match the range of the scene to that of the sensor and record the maximum dynamic range the camera is capable of. But most of the time that is not practical.
honestly such a fantastic video, so clearly explained
Fantastic explanation and demonstration. Thank you Gerald! Intuitively, I figured you’d lose highlight stops. But I was completely unaware that she shadows actually clean up. So I’ll definitely keep that in mind. 😊
Exposing for middle grey and lowering ISO implicitly means that you increase the actual exposure (ie you move from f/4 to f/2.8 or from 1/60s to 1/30s etc) So those dark regions actually emit more photons that the sensor picks up to work with. On the other side a greater area of the scene emits more photons than the sensor can measure, thus it is clipped.
Close..it's not that more photons are emitted (a constant light source is constant) its that the wider aperture gathers more photons [f4 to f2.8] or the exposure duration is longer (1/60 to 1/30) doubling the number photons acquired. Consequently more "signal" is collected.
That is a very interesting result. I just tried to avoid it so far since I was not quite sure what exactly I was trading or giving up. Basically this means that in a controlled home scene you might as well just shoot with the low iso value anyways since you have sufficient controlled light, don't need the extra noisy DR stops anyways and can lower the background noise even further.
That is useful information. I was always scared, because even if I didn't see much difference in my testing, I figured as soon as I tried to use it while shooting for a client it would wreck me somehow. So this is great knowing exactly how it works. Thank you for putting the time in!
(Long ass reply here)..
I guarantee you, your client cares zero about your ISO, unless they themselves are photographers/judges. The client, unless they have the technical know-how, just want to know that whatever they asked you to photograph, looks good. That's it. And it needs to look good at first glance. If you can pass the first glance test, that's it, everything else (technically) is just improving your product (but has nothing that the client will care for).
In fact, when your client is trying to make a decision between you and someone else, they won't be thinking, "this image has more/less white dots". They don't even know to look for that. They are going to first think about the purpose of the photo/product and ask themselves, can this photographer convey this purpose or give me what I want (by looking at your portfolio)?
I find all this technical detail, and care, is much ado about nothing. Hardly any of this makes or breaks a photo, unless at drastic degrees (eg: very high ISO). The important things that make a photo great is not in the scientific details (I say this as someone with a major in General Physics and in Computer Science...science is my business, but my thoughts come from the artistic side of me as someone born to parents who are painters, sculptor and designers).
Again, clients want to know, "can this photographer give me what I want?". 10/10 times what they want is a specific look and feel, not technical greatness. I've seen some shitting photos, with shitty retouching (I hate retouching at least the majority I've seen outside the fashion industry), and the client is just goo-goo gaa-gaa over it. Because they aren't seeing it from a photographers eyes, and that's what we need to understand.
Improve your craft (artistic creativity) but don't get bogged down in the technicalities. Use the technicalities to exploit towards your creativity.
@@thecsciworker291 yup
Leveraging one's technical knowledge to benefit what a client wants and expects is important, i.e., knowing about the relevant techniques, gear selection and usage and artistry are a synergistic package. I do agree that most clients are simply interested in results that support their needs, thus usually best to avoid offering up unasked for techie talk with them unless they ask a question about such, and then carefully and briefly titrate one's reply.
Fuji has their lowest ISO at 80, which is one below the native of 160. It is marked with an L on the dial (on those cameras that still have a dial). In video you'll find other "native" ISOs in which the noise cleans right up (800 and 200, for FLog on the X-Trans 3 sensor of the X-T3, X-T4, ...) that don't really affect dynamic range though (as far as I've noticed).
dual gain iso xt3?!
@@neramp559 The X-T3 and X-T4 (and possibly other models with the same sensor and image processor) have multiple pre-amp stages. They behave a lot like dual native ISOs but technically aren't. For the end user the result is the same: Lower image noise at certain ISO settings. I shot my first project at ISO 640 as that was the lowest FLog allowed; a month after shooting I found out that there is another pre-amp-stage at ISO 800 that would've resulted in significantly less noise. I now only shoot FLog on my X-T3 at ISO 800 and 2000 or switch to HLG as that uses the pre-amps for Rec.709 which means at ISO 3200 it is about as clean as FLOG at ISO 2000; I trade some dynamic range capabilities for a cleaner image.
I've done a lot of tests myself and threw a few comparisons together here: nc.rhprivat.de/s/W8NG3JpPZ2MzCKD
My brain hurts. I've watched this twice. I'll have to watch it again to really understand it and take it in. But, it's great information and is challenging my knowledge gaps. Thank you for putting it out there.
And as always, AMAZING! Thanks for the thorough explanation!
Always fascinating. Thanks for sharing Gerald!
I've been wondering this question for so long! Thanks for the video GU !
Thank you for covering this! Especially in your testing first way.
I'm a newbie to photography (1 year now I own the Sony A7IV) - Even if I'm not understanding the subject in its full extension: In any case, it's still interesting to watch your "lessons" - watching you share your experiences for us to refer to!
Great video. This is exactly how the BMPCC cameras work when you shoot in BRAW. Great info.
This is awesome. Ive been needing this video. I think about it all the time.
Excellent evaluation. Exactly what I was looking for. Thanks for doing all the heavy lifting for Us!
This is insightful, thanks Gerald. I like this video format also, casual and experimental.
I've always wondered about this! Thank you Gerald!
Same is applied in all cameras. I've learned it from my BMPCC6K. It's like counterintuitive for beginners, because when You lower the ISO You bring the exposure down, but the rule is "lower ISO - more information in shadows, higher - more in highlights". So we have to choose ISO if we want range in the shadows or highlights and then use aperture, filters and speed/angle.
ISO doesn't actually change the exposure. The ISO is after the sensor capture. It only changes the gain.
@@JimRobinson-colors You're right, but You know what I mean...
Just in time to start learning how to use my new camera. Amazing timing Gerald!!
Thank you for showing the differences in detail.
I’ve always shot at base or higher, so it’s good to know that in a pinch I could go down to the extended ISO range won’t be as terrible as I had been told.
Loved this! You are an amazing source of knowledge Gerald. 😊
This was an amazing video, very in-depth and I am a sucker for going as low as possible but not realizing im looing in DR. Thank you and it makes me reconsider how I shoot.
Thank you! Love this kind of videos!
This is a good illustration of how lower ISO values should be used for shooting in low light and higher ISO values should be used in bright settings. At higher ISO values, your clipping point is as close as possible to 100 IRE. And as you push the ISO higher, it shifts middle grey up towards the highlights, creating more artificial dynamic range. And although that would technically add noise, it doesn't matter, since it's all hot exposure values. Just as long as you don't have any larges areas of shadow in your shot too.
And for the reverse, you can push your ISO down below native when shooting in dark settings. Then just add as much light as possible so that your shadows will be extremely clean and, just as before, you'll have artificial shadow dynamic range, since middle grey is pushed down toward black. And you won't miss the loss of dynamic range in the highlights unless you actually have a bright point of light in your shot. So if you're shooting at night, all you have to worry about and streetlights, practicals, etc. Rather than worrying about the sun and the bright midday sky.
CZcams has cursed Gerald’s sub count. Just let this man grow🥹.
You actually taught me some new stuff. Signal-to-noise ratio seems to be the most important point of selecting the ISO.
How much can be recovered with Topaz or some other noise reducing software would be an interesting question.
Some people have extreme opinions about the subject with no data to present.
Thanks for explaining all this as well as you did and then wrapping up the end. It useful. Thanks!
I find that this affects and applies to live video productions as well. I regularly use full-frame mirrorless cameras for live broadcast situations with very dynamic lighting. Aperture and overall look is way more of a priority than minute differences in dynamic range or noise in the shadows. We often have to dip into the extended ISO to match exposure camera to camera on the fly. Experientially, I've found that doing this doesn't have a noticeable affect on the overall image quality and it's reassuring to have this deep dive on the topic that proves out what I've felt was true: IE using extended ISO correctly doesn't have a significant affect on image quality.
I’ve always wondered this. Thanks for covering it. Very interesting and useful to know.
Mr. Gerald , Since you have hundreds of thousands of viewers the odds of you seeing this...but I'm optimistic. Your info is amazing and , when I watch it a few times , always on point. Just wondering if you have seen the 7 artisan cine lenses. Some of the images Roman Hense gets with them is amazing. He's also a great colorist, so there's that. I gather they are made to mimic eis lens? Thanks for the time and effort you spend on these videos. The whole community really appreciates you
Great video. It seems counterintuitively that, in the end, you should use lower ISO for darker indoor scenes with controlled lighting, and native ISO for brighter outdoor scenes with larger dynamic range.
Fascinating and educational. Thanks much for posting !
The lower you go with the iso the more detail the camera captures in the shadows and the higher you go with the iso (practically speaking) the more detail it captures in the highlights. That's why the iso 200 is able to beat the iso 4000, the best result would've been lighting like youre shooting for 100 and shoot at 200 to get the best result in shadow noise, loosing a little bit of the highlights.
Thanks Gerald for this! I was wondering that, but of course never tested it, so this helps a lot 💪
Always wondered how it looks. Thanks for the video!
Thank you!!
I’ve been curious about this for years. Been keeping it at base/native to play it safe 😅
Thanks, great to clear this up - '"Aaaright, I'm Done" never gets old :)
Gerald don’t miss.
This was a very interesting breakdown. Nice job as always!
THANK YOU
Thanks Gerald for the insightful video! It seems like shooting below native ISO is a convenient substitute for ETTR and grading back down to reduce noise.
yes
Thanks for testing this Gerald! This is super helpful!
Our favorite nerd is back! Always fun to watch ;)
Very helpful, and very interesting. Thanks, Gerald
Such a difficult concept to explain 😂
As always you handle it with meticulous ease! Thanks
I'm Gerald Undone and it's Morbin' Time