Why ETTR Is Wrong - Do This Instead

[Abinoone]

Another great video, Steve! I learned things from this video that I didn't know - now I'll be the life of family parties as I explain how ISO doesn't really affect noise. In any event, thanks for taking the time and effort to put such an informative video together, and keep up the great work!

 

[Cameron T]

Another great video, Steve! I learned things from this video that I didn't know - now I'll be the life of family parties as I explain how ISO doesn't really affect noise. In any event, thanks for taking the time and effort to put such an informative video together, and keep up the great work!

Iso is like artifical amplification on a mic too. I've used that analogy plenty of times.

 

[Marcelo_R]

Iso is like artifical amplification on a mic too. I've used that analogy plenty of times.

LOL! I use the old TV antenna and the "snow" on the screen, add an amp to increase the weak reception and yes, you get to barely see the image... and a lot more snow too!

 

[Cameron T]

I run into more audio guys locally at the photo store asking.

Also a younger crowd, from after that time (and the time of the stone tablets )

 

[ajrmd]

Love the disruption (or irruption?) over ETTR. With ISO invariant sensors, Steve's observations regarding noise largely hold true. The only reason to strive to ETTR today is to assist AF (particularly with Nikon).

 

[Brian L]

Great video Steve, destroying the ETTR fallacy, at least when using ISO. Since I usually shoot in Manual with Auto-ISO, using exposure comp to push the histogram right is useless; or left for that matter. Translating it for my simple mind, SS and aperture are “optical” adjustments; ISO is a digital - or sensor - adjustment. Therefore, it’s useless to overexpose using this digital variable.

A couple of years ago, I had a wonderful Teton wildlife workshop with another photographer (the other Steve). The second day he suggested I try shooting aperture priority with manual ISO rather than full manual with auto-ISO. At the time I didn’t understand why it differed from my technique - thinking it was just two different journeys to the same exposure.

After watching your video, I now see why his technique is perhaps better at managing noise. By adjusting ISO manually in aperture priority, the moving variable is shutter speed. This means I’m paying closer attention to the prospect of noise - when I shoot manual + auto-ISO I’m less attentive to the noise indicator as it floats as I change shutter speed. For example, if I ETTR in aperture priority my exposure variable is a slower SS, thereby reducing optical noise.

Of course, there’s always a trade - how low goes the shutter before I blur focus.

in any case, as always, I really appreciate your thoughts, Steve. Thanks.

 

[Box Brownie]

Hi Steve

An interesting video....................I will have to watch it a again to fully get to grips with it.

I have been an Olympus now OM user for a few years and as you likely know the mFT sensor does exhibit some marked the higher the ISO you go.....though the DxO PL is a tamer of the noise present.

Now, I went to the Photos to Photons and picked my two bodies the OM1 and the OM1 mk2 and see this

You can only see EXIF info for this image if you are logged in.

There is a very slight kink at at ISO 1006 but to my eye for all its worth the line is linear.

I am unsure what to make of this in regard to your commentary re: dual gain or not, and would appreciate your insight?

 

[Steve]

Hi Steve

An interesting video....................I will have to watch it a again to fully get to grips with it.

I have been an Olympus now OM user for a few years and as you likely know the mFT sensor does exhibit some marked the higher the ISO you go.....though the DxO PL is a tamer of the noise present.

Now, I went to the Photos to Photons and picked my two bodies the OM1 and the OM1 mk2 and see this

View attachment 106274There is a very slight kink at at ISO 1006 but to my eye for all its worth the line is linear.

I am unsure what to make of this in regard to your commentary re: dual gain or not, and would appreciate your insight?

I don't think the OM 1 series uses a dual gain sensor. It doesn't look like it on the chart, either.

 

[Box Brownie]

I don't think the OM 1 series uses a dual gain sensor. It doesn't look like it on the chart, either.

Ah! so in regard to your video and where you mentioned both the Z8 & 9 and the Sony A1(?) having such a dual gain sensor and picking the ISO appropriately.....what would be the your advice re: OM users?

TIA

 

[ajrmd]

You can only see EXIF info for this image if you are logged in.

Hi Steve

An interesting video....................I will have to watch it a again to fully get to grips with it.

I have been an Olympus now OM user for a few years and as you likely know the mFT sensor does exhibit some marked the higher the ISO you go.....though the DxO PL is a tamer of the noise present.

Now, I went to the Photos to Photons and picked my two bodies the OM1 and the OM1 mk2 and see this

View attachment 106274There is a very slight kink at at ISO 1006 but to my eye for all its worth the line is linear.

I am unsure what to make of this in regard to your commentary re: dual gain or not, and would appreciate your insight?

It’s difficult to say whether the OM1 is a dual gain system or not. If one looks at the shadow noise versus ISO which can be used to gauge whether a sensor is ISO invariant, it appears that at lower ISO’s the sensor is not whereas at some point the sensor becomes relatively ISO invariant. By comparison, the Z8 plot demonstrates nearly true ISO invariance with a dual gain sensor. So, ant low ISO’s shadow noise is a function of ISO, whereas it becomes less important when the main gain amplifier is applied which is around ISO 1000.. Unfortunately, my knowledge of the OM system is limited, though perhaps someone can add additional information.

 

[Steve]

Ah! so in regard to your video and where you mentioned both the Z8 & 9 and the Sony A1(?) having such a dual gain sensor and picking the ISO appropriately.....what would be the your advice re: OM users?

TIA

Looking it up, it seems that it's not ISO invariant, so you don't consider dual gain, just base ISO if you want to do any kind of ETTR. That said, if you're fighting backlight, just use the lowest ISO you can (like, keep it under 400~500, ideally more like 200 or less) and get the sky close to or just starting to clip - that'll get the job done

 

[Box Brownie]

It’s difficult to say whether the OM1 is a dual gain system or not. If one looks at the shadow noise versus ISO which can be used to gauge whether a sensor is ISO invariant, it appears that at lower ISO’s the sensor is not whereas at some point the sensor becomes relatively ISO invariant. By comparison, the Z8 plot demonstrates nearly true ISO invariance with a dual gain sensor. So, ant low ISO’s shadow noise is a function of ISO, whereas it becomes less important when the main gain amplifier is applied which is around ISO 1000.. Unfortunately, my knowledge of the OM system is limited, though perhaps someone can add additional information.

I am not sure which chart you are showing as the the title is the same as the one I used and Steve indicated to use in his video

Edit ~ ooops I see the the right chart....the second one down the list but your scaling is different as the graph is way flatter in my 'view' i.e. without the uptick that you have shown?

Shadow Improvement of Photographic Dynamic Range versus ISO Setting

 

[Neil H.]

When I first learned about ETTR I understood that this provided you with the largest (best?) raw file. So I placed my camera on a tripod and started with a dark exposure with no darks clipped. (This is best done with a complex image with lots of tones and colours). I then moved the histogram to the right taking images as I went. The last ETTR image was 6MB larger than the first dark image. It was explained to me that the last (right) histogram box contains vastly more data than the first (left) box of the histogram. So if you want to capture the best white feather detail you should ensure that you have information in that right hand box AT CAPTURE. If you do not capture this detail I was told you can not get it by brightening a dark image. So the question of noise was secondary with ETTR, the main benefit is capturing the most highlight detail and a significantly larger file. You often find that bird images will have nothing in that right hand box which is the box that captures the most information compared to all the other boxes. I certainly know that the file size is significantly bigger - 6MB more in the highlights is not to be sniffed at. So was that advice to get a bigger (better?) file (by using ETTR) with more highlight detail nonsense?

 

[Box Brownie]

Looking it up, it seems that it's not ISO invariant, so you don't consider dual gain, just base ISO if you want to do any kind of ETTR. That said, if you're fighting backlight, just use the lowest ISO you can (like, keep it under 400~500, ideally more like 200 or less) and get the sky close to or just starting to clip - that'll get the job done

Lower ISO on the mFT sensor is always as clean as whistle for any given shutter speed & aperture.

I need to, as mentioned, re-watch your video and consider my configurations
NB I usually shoot Tv with shutter speed (up to 1/3200 or tad higher) & wide open for fast moving birds in flight or at take off) to suit the subject and situation and use Auto ISO though with quite a high upper limit because as I said above I have to date tamed the noise in post.

 

[Ed Erkes]

I may be confused, but to me the concept of ETTR has changed with modern cameras and sensors that are largely ISO invariant. ETTR, to me, means capturing the maximum amount of data (signal) through adjustments to shutter speed and aperture. I don't really consider ISO part of the actual exposure process any more. So brightening the image with ISO is NOT ETTR, since you are NOT adjusting exposure. You are moving the histogram to the right, but not through an exposure adjustment. I admit I did not watch the entire video, because once he started changing the ISO on his examples, I thought that is not ETTR, since he is not changing the exposure. Am I way off base in my thinking?

 

[DRwyoming]

Am I way off base in my thinking?

Watch the rest of the video, that's exactly the point Steve makes and if you watch more you'll see he shows the limited value of ETTR even when light gathering via aperture or shutter speed is reasonably adjusted.

IOW, sure gather as much light as the situation and needs including DoF and stopping action require but then ISO can just be dropped for a normal exposure with little need to push the histogram to the right only to correct it in post. Yup, getting more light to the sensor is still great but then just drop ISO for a nominal exposure with potentially small exceptions when working near dual gain sensor switching points.

But basically you're not really missing anything but do watch all of the video as that's where it goes.

 

[JAJohnson]

So the question of noise was secondary with ETTR, the main benefit is capturing the most highlight detail and a significantly larger file. You often find that bird images will have nothing in that right hand box which is the box that captures the most information compared to all the other boxes. I certainly know that the file size is significantly bigger - 6MB more in the highlights is not to be sniffed at. So was that advice to get a bigger (better?) file (by using ETTR) with more highlight detail nonsense?

I once had the problem of clipping white and light colored feather detail on certain birds because of my tendency to be too aggressive with ETTR. I was advised to use of exposure comp to lower exposure a bit to capture the detail I was clipping. I’m much happier with my bird images (for great egrets, snowy egrets, spoonbills, gulls, etc) since then.

I think the advice you mentioned is correct but be careful not to overdo it.

 

[arbitrage]

ETTR using SS and/or aperture can be of some benefit to increase the signal to the shadows. ETTR using ISO is pointless.
I've come used to using the dual-gain ISO as my base ISO but that is mainly because I'm always chasing faster SS for BIF. So I will bring ISO down to 500 and then any extra light I have I will up my SS to get better BIF results.
Because I shoot Sony (most of the time) I have the benefit of using the Zebras in stills and have it custom set to +107 or +109. Therefore I really don't have to worry about accidentally clipping things. I also shoot full M all the time so I don't have to worry about the camera blowing my exposure during a BIF sequence when the background brightness and/or size of the bird vs the background changes.

 

[Krish]

Brilliant Explanation - and timely. So many of us have not adjusted fully to what modern tech needs us to adapt to.

 

[Activert]

Thank you for another great video. I don’t try to expose to the right, I only check the histogram to see if my exposition is correct. Most of the time I shoot wide open and I adjust my shutter speed according to the situation and then I adjust my iso.. I try to keep my ISO low when it is possible.

 

[DavidT]

Interesting video. I have never used a histogram. I pay attention to the zebras in Sony cameras, expose to what appears to be proper for the image I’m trying to take. Typically I’m only adjusting less than a 1/3-1/2 stop in post.
It’s what the mirrorless cameras have given us that’s worth switching to them. Being able to see live exposure has been a game changer. I’m old school learning on film. I always try and nail exposure in camera.

 

[RichF]

Another great video that explains complicated concepts, in very simple terms. Bottom line seems to be expose the same way we did with film.

 

[Hooligans Imagery]

Thanks for this great learning.
So proper exposure.
So if the SS is your primary variable in wildlife depending on the situation, in flight, perched, running, sleeping etc., and you are looking at your histogram, you want the largest aperture without clipping correct, factoring in DOF decisions? Am I stating this correctly? Nikon doesn't have the zebras for still images so we need the histogram (forgetting that picture control made for this purpose. )

I realize there are situations where the aperture is the primary and that would suggest the opposite.

 

[DRwyoming]

So if the SS is your primary variable in wildlife depending on the situation, in flight, perched, running, sleeping etc., and you are looking at your histogram, you want the largest aperture without clipping correct, factoring in DOF decisions? Am I stating this correctly?

If the image has a lot of bright highlights that demand pushing up against clipping then sure but that's not really the best general guideline. A lot of photos shot in softer more diffuse light won't push you up against clipping so I wouldn't use that as your proper exposure criteria unless all of your shooting demands it.

Several takeaways from Steve's video but one big one is to not artificially push yourself up against clipping via ISO on hopes of lowering noise. Expose properly as the scene and lighting demand but if you're worried about noise the key is to open up aperture and lower shutter speed as much as you think you can get away with but at that point there's no need to set ISO high or leave ISO set relatively high to keep the histogram shifted up to the right if that isn't the best final exposure. In that situation where you've gathered as much light as you think you can get away with via shutter speed and aperture, just drop or set ISO for the final desired exposure.

But sure if you shoot a lot in high contrast light your subjects and scenes have a lot of tonal range like bright skies and dark subjects then yeah proper exposure will likely drive you right up just shy of clipping the brightest highlights. But shoot more in softer diffused light under high thin clouds or near the edges of day and proper exposure may very well not drive you right up to clipping so in those situations the key is best final exposure and not automatically driving every image as bright as you can get it.

Discussions like these get tangled fast as we talk about working in different automated exposure modes like Aperture or Shutter Priority. I find it a lot easier to discuss these things when shooting in full manual with manual ISO or from there manual with Auto ISO. Basically:

- Pick a shutter speed that stops or blurs action as desired
- Pick an aperture that gives the necessary DoF which in most wildlife action cases means wide open or stopped down no more than a stop or so especially when isolating subjects with long lenses.

- Then pick an ISO for proper final exposure or when shooting in Auto ISO let the camera set ISO taking into account any exposure compensation

Once you've done that you've gathered as much light as you think you can and if you want more you have to risk a slower shutter speed and possible subject blur, open up aperture for less DoF or some combo of both but even once you've done those things just pick an ISO for proper final exposure and there's little to nothing to be gained by leaving ISO higher than necessary and driving the histogram to the right unless that's what proper exposure means for that particular subject, scene and lighting.

 

[Steve]

Interesting video. I have never used a histogram. I pay attention to the zebras in Sony cameras, expose to what appears to be proper for the image I’m trying to take. Typically I’m only adjusting less than a 1/3-1/2 stop in post.
It’s what the mirrorless cameras have given us that’s worth switching to them. Being able to see live exposure has been a game changer. I’m old school learning on film. I always try and nail exposure in camera.

That's how I do it as well with Sony. I watch my brighter areas and keep them just under zebra clipping. If there's no white area or highlight area, I just judge the brightness in the viewfinder and fix it later - I'm usually pretty close. (Psst - don't tell anyone I don't rely that much on the histogram )