Is ISO 160 the "Native" ISO on the 6D and 5D3?

I've heard some people mention before that ISO 160 is the real native ISO of these cameras instead of ISO 100. Is that true? I really don't notice much of a difference between the two personally, but most of my shots are at ISO 100

I Was told by Canon CPS it was on the 7D mk I and the 5D3 - no idea on the 6D .

No but for those shooting in camera JPEG's or video then it is the ISO value that may offer the highest DR, although at the cost of a stop of highlight headroom. This is due to the fact that ISO 160 is created digitally by exposing at ISO 200 (loosing the stop of highlight headroom) and pulling the exposure by 1/3 stop (improving the SNR in the dark tones). If you were to expose ISO 100 as ISO 80 and perform the same exposure pull in post it would produce a result even better than ISO160. This is an example of the advantages of exposing to the right, you keep the ISO 100 headroom and improve the dark SNR with the pull. The thing is that you can only do this if you shoot RAW. The thing is if you choose to use the ISO 100 full stop multiples, and expose to the right to maximise the recorded signal, and thus SNR, you then find that you actually have far more control over the image processing, than simply pulling the exposure in the way that the camera does. Well it does if you use a RAW developer other than DPP (at least up to 3.x not too sure about 4.x not having a supported camera). The other thing to remember is that the digital pull/push the 1/3rd stop ISO's perform is done to the data before it is written to the RAW file, so is cooked in there.

For the most headroom with the most DR with Canon DSLR sensors the best strategy is to always shoot to the right, so that needed highlight detail is recorded at the right hand edge of the histogram. It is also best practice that if you can no longer extend the shutter speed or open the aperture due to other image constraints to increase the ISO in full stops until you can place the wanted highlight detail where you want it to be. You can then normalise the image in post to accurately reflect the actual brightness of the scene. Generally your RAW conversion program will have far more control of this, as it will have tools to allow you to adjust the White Point/Highlights/mid tones/shadows/black point at least semi independently. When the camera does this as part of the digital ISO control, it simply pulls all parts of the exposure by the same amount.

Alan

BigAl007 wrote in post #17481368
No but for those shooting in camera JPEG's or video then it is the ISO value that may offer the highest DR, although at the cost of a stop of highlight headroom. This is due to the fact that ISO 160 is created digitally by exposing at ISO 200 (loosing the stop of highlight headroom) and pulling the exposure by 1/3 stop (improving the SNR in the dark tones). If you were to expose ISO 100 as ISO 80 and perform the same exposure pull in post it would produce a result even better than ISO160. This is an example of the advantages of exposing to the right, you keep the ISO 100 headroom and improve the dark SNR with the pull. The thing is that you can only do this if you shoot RAW. The thing is if you choose to use the ISO 100 full stop multiples, and expose to the right to maximise the recorded signal, and thus SNR, you then find that you actually have far more control over the image processing, than simply pulling the exposure in the way that the camera does. Well it does if you use a RAW developer other than DPP (at least up to 3.x not too sure about 4.x not having a supported camera). The other thing to remember is that the digital pull/push the 1/3rd stop ISO's perform is done to the data before it is written to the RAW file, so is cooked in there.

For the most headroom with the most DR with Canon DSLR sensors the best strategy is to always shoot to the right, so that needed highlight detail is recorded at the right hand edge of the histogram. It is also best practice that if you can no longer extend the shutter speed or open the aperture due to other image constraints to increase the ISO in full stops until you can place the wanted highlight detail where you want it to be. You can then normalise the image in post to accurately reflect the actual brightness of the scene. Generally your RAW conversion program will have far more control of this, as it will have tools to allow you to adjust the White Point/Highlights/mid tones/shadows/black point at least semi independently. When the camera does this as part of the digital ISO control, it simply pulls all parts of the exposure by the same amount.

Alan

So let's say I ettr 2/3rd of a stop. When I put this file in Lightroom, what are the basic steps you would take to develop the photo.

Does that mean I can't happily shoot my 6D at 25,600 anymore? Jeez, in good light I can shoot 12,800 JPG. I hardly think 100/125/160/200 makes much difference.

I always shoot RAW and usually use ISO 100 for all of my landscape work. I guess I'll try using multiples of 100 and see if it makes any difference. I commonly expose to the right and sometimes the highlights are recovered just fine from an image shot at +2 EV with my 6D.

The whole changing the ISO thing when shooting ETTR or one of it's derivatives, is that if you can increase the exposure using the aperture or shutter speed to move the wanted highlight detail to the right hand edge of the histogram while using base ISO you do so. If though you run out of available aperture or shutter speed adjustment, because the camera/lens doesn't offer the values you need, or other artistic needs get in the way, such as DoF, subject motion blur or even camera shake in hand held shots then you should use the ISO options to move the histogram. What we are doing here is using the optimum value of ISO and exposure for any particular shot.

Also with shooting ETTR it is not simply a case of throwing on say 2/3 stop or whatever EC, what you really should be doing is checking the histogram and highlight blinkies by chimping. This is really the only way to ensure that you are maximising the possible exposure. Although this is effectively what the camera does with the ISO 160 series of ISO values. It is simply exposing 1/3 stop to the right of the normal exposure for ISO 200 and pulling all values back linearly afterwards. This was the approach taken with ETTR originally, and is the "By the numbers" answer.

Personally when I am developing an image I will be doing it with the intention of achieving the artistic results I am looking for. So for example I might pull back the highlight detail by quite a bit, to get the maximum detail into the highlights. This may of course actually end up compressing the highlight details. On the other hand I might not pull back some of the shadow areas at all. This is an effective way of boosting the shadows without actually having to digitally touch the image data, which of course would be adding additional noise to the image. By using the full dynamic range of the sensor I can then effectively tone map the image as I see fit, with the minimum of degradation of the image in the process. I usually use Adobe process version 2012 for this. I like the way it works, even though it is doing some hidden tricks of it's own in the process. If one really needs to do things by the numbers though it would be better to use a different RAW converter. PV2010 is still good for this, or there are other very good RAW converters available.

Doing ETTR properly really requires that you pay a lot of attention to the actual exposure, so that you can get it fully optimised. Any level of underexposure from the optimum loses you dynamic range, while you should always be right at the limit for overexposure in the wanted highlights. Because you are always looking for the optimum exposure you will always be having to optimise the post processing applied to the image. For those that want somewhere to start when working in LR and PV2012 I would usually start by bringing the highlights down, then the midtones,which use the exposure slider, then the shadows. As the sliders are interdependent, and overlap quite a bit, you may actually find that to get back the shadow detail that you want you have to apply a positive boost to the shadows slider, remember that often you are only bringing the brightness back to the level that the image was exposed at, so there is no net boost to the wanted image data or the noise.

Alan

BigAl007 wrote in post #17481814
The whole changing the ISO thing when shooting ETTR or one of it's derivatives, is that if you can increase the exposure using the aperture or shutter speed to move the wanted highlight detail to the right hand edge of the histogram while using base ISO you do so. If though you run out of available aperture or shutter speed adjustment, because the camera/lens doesn't offer the values you need, or other artistic needs get in the way, such as DoF, subject motion blur or even camera shake in hand held shots then you should use the ISO options to move the histogram. What we are doing here is using the optimum value of ISO and exposure for any particular shot.

Also with shooting ETTR it is not simply a case of throwing on say 2/3 stop or whatever EC, what you really should be doing is checking the histogram and highlight blinkies by chimping. This is really the only way to ensure that you are maximising the possible exposure. Although this is effectively what the camera does with the ISO 160 series of ISO values. It is simply exposing 1/3 stop to the right of the normal exposure for ISO 200 and pulling all values back linearly afterwards. This was the approach taken with ETTR originally, and is the "By the numbers" answer.

Personally when I am developing an image I will be doing it with the intention of achieving the artistic results I am looking for. So for example I might pull back the highlight detail by quite a bit, to get the maximum detail into the highlights. This may of course actually end up compressing the highlight details. On the other hand I might not pull back some of the shadow areas at all. This is an effective way of boosting the shadows without actually having to digitally touch the image data, which of course would be adding additional noise to the image. By using the full dynamic range of the sensor I can then effectively tone map the image as I see fit, with the minimum of degradation of the image in the process. I usually use Adobe process version 2012 for this. I like the way it works, even though it is doing some hidden tricks of it's own in the process. If one really needs to do things by the numbers though it would be better to use a different RAW converter. PV2010 is still good for this, or there are other very good RAW converters available.

Doing ETTR properly really requires that you pay a lot of attention to the actual exposure, so that you can get it fully optimised. Any level of underexposure from the optimum loses you dynamic range, while you should always be right at the limit for overexposure in the wanted highlights. Because you are always looking for the optimum exposure you will always be having to optimise the post processing applied to the image. For those that want somewhere to start when working in LR and PV2012 I would usually start by bringing the highlights down, then the midtones,which use the exposure slider, then the shadows. As the sliders are interdependent, and overlap quite a bit, you may actually find that to get back the shadow detail that you want you have to apply a positive boost to the shadows slider, remember that often you are only bringing the brightness back to the level that the image was exposed at, so there is no net boost to the wanted image data or the noise.

Alan

Thanks a lo for your reply. Your posts are very informative.