DR100% vs 200/400%
RoamingScott
Offline
Activity
[X]
Details
p.2 #1DR100% vs 200/400%
This thread is a fantastic microcosm of how hard it is to have an intelligent conversation on this forum.
Jan 23, 2026 at 08:52 PM
p.2 #2DR100% vs 200/400%
Measurebating. It's a word that actually made the wiki dictionary. This thread is a perfect example of it.
To the OP: You and I had an interchange about DR 200-400 in another thread. I told you how several of us have parsed out how it worked, but you disregarded it claiming it was the end all solution to gaining DR. It is as both Dan and Scott have alluded above: nothing more that an automated under-exposure to preserve highlights, then a bump to normalize gamma for a properly "exposed" image. An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image. The actual recipe of how Fuji actually accomplishes it is not documented anywhere that I could find, ditto Nikon's version -- but a gamma curve adjustment post capture will not show any change in exposure OR ISO... Regardless, if you underexpose then bump exposure in a post routine via any methodology, you will either lose total DR or gain noise, or both -- there is no free lunch in physics. The only "magic" here is that very dark region noise gains are far less visible issues in images than clipped highlights -- protect the highlights, then gain a little noise in the darkest shadows areas you have to lift and nobody notices anyway, voilà, a "better looking" image...
Also, I agree ISO 500 is an odd choice on the GFX 100. It and 320/400 generate more noise than 800. In fact, ISO 500 is about equivalent noise to 1200 on the GFX 100 sensor. Granted, DR is a bit better at the lower ISO's, but with the noise offset, it turns out in real use, 320 through 1200 are all but equal in performance. So a not horrible Auto ISO practice would be ISO 100 for static, well lit and/or tripod images; Auto 100-1600 for most general "walk-around" imaging; and 100-6400 for broad range imaging -- and here go ahead and take it up to 12,800 at night if you don't mind adding post NR tools for the images that get taken there. This system works like a champ.
And so, IMHO: Auto ISO IS actually a Godsend to modern imaging, even on the GFX and more so on the fatter-pixel bodies... When used properly, it allows using a single MANUAL exposure and relying on ISO across a range of lighting situations that we would have needed multiple films to cover even half as well back in the day (let alone far more flexibility combined with A S P modes). Of course you need to implement it properly and not go overboard in your top ISO choice. But there even 3200 remains very usable on the GFX 100... Okay, /rant, measurebate away...
To the OP: You and I had an interchange about DR 200-400 in another thread. I told you how several of us have parsed out how it worked, but you disregarded it claiming it was the end all solution to gaining DR. It is as both Dan and Scott have alluded above: nothing more that an automated under-exposure to preserve highlights, then a bump to normalize gamma for a properly "exposed" image. An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image. The actual recipe of how Fuji actually accomplishes it is not documented anywhere that I could find, ditto Nikon's version -- but a gamma curve adjustment post capture will not show any change in exposure OR ISO... Regardless, if you underexpose then bump exposure in a post routine via any methodology, you will either lose total DR or gain noise, or both -- there is no free lunch in physics. The only "magic" here is that very dark region noise gains are far less visible issues in images than clipped highlights -- protect the highlights, then gain a little noise in the darkest shadows areas you have to lift and nobody notices anyway, voilà, a "better looking" image...
Also, I agree ISO 500 is an odd choice on the GFX 100. It and 320/400 generate more noise than 800. In fact, ISO 500 is about equivalent noise to 1200 on the GFX 100 sensor. Granted, DR is a bit better at the lower ISO's, but with the noise offset, it turns out in real use, 320 through 1200 are all but equal in performance. So a not horrible Auto ISO practice would be ISO 100 for static, well lit and/or tripod images; Auto 100-1600 for most general "walk-around" imaging; and 100-6400 for broad range imaging -- and here go ahead and take it up to 12,800 at night if you don't mind adding post NR tools for the images that get taken there. This system works like a champ.
And so, IMHO: Auto ISO IS actually a Godsend to modern imaging, even on the GFX and more so on the fatter-pixel bodies... When used properly, it allows using a single MANUAL exposure and relying on ISO across a range of lighting situations that we would have needed multiple films to cover even half as well back in the day (let alone far more flexibility combined with A S P modes). Of course you need to implement it properly and not go overboard in your top ISO choice. But there even 3200 remains very usable on the GFX 100... Okay, /rant, measurebate away...
Jan 23, 2026 at 11:17 PM
p.2 #3DR100% vs 200/400%
gdanmitchell wrote:
An extreme example, but, it turns out that ISO is not exactly irrelevant. Who knew?!
https://gallery.gdanmitchell.com/gallery/var/albums/HumanWorld/Unclassified/Random-Images/Screenshot%202026-01-23%20at%201_59_43%E2%80%AFPM.jpg
I never said it's irrelevant, just has less effect on the image than SS and A.
Use the EC dial and compromise on your ideal SS and A to get ISO where you want it, bearing in mind dual gain sensor characteristics as well as noise etc.
Gerry
Jan 24, 2026 at 02:01 AM
p.2 #4DR100% vs 200/400%
chez wrote:
Except in the most benign and perfectly lit situations either we or the camera still need to come up with the situationally-best combination of ISO, shutter speed, and aperture and we consider subject motion, DOF, noise, and other things to make the choice.With today’s noise reduction…I don’t worry much about iSO. If your shutter is too slow and you get a blurry image…you are toast. If you dof is too narrow and you get a fuzzy background…you are toast. Those are the two things I worry about and just let the iso figure itself out.
I worry a whole lot less about it, especially now with the Adobe AI Denoise function in ACR. But I still prefer to make optimal choices for all settings that affect the image.
As I wrote more than once, I can see situations in which the automated Dynamic Range settings could make sense (and also in which the AUTO ISO option could be useful), but I feel that in much of my photography I’d rather make intentional decisions about ISO and how to handle the potential for blown highlights and blocked shadows.
YMMV.
- - -
RoamingScott wrote:
This thread is a fantastic microcosm of how hard it is to have an intelligent conversation on this forum.
As always, thank you for your helpful, informative, and “intelligent” contributions on the subject of the thread.
- - -
gyoung143 wrote:
https://gallery.gdanmitchell.com/gallery/var/albums/HumanWorld/Unclassified/Random-Images/Screenshot%202026-01-23%20at%201_59_43%E2%80%AFPM.jpgI never said it's irrelevant, just has less effect on the image than SS and A.
Use the EC dial and compromise on your ideal SS and A to get ISO where you want it, bearing in mind dual gain sensor characteristics as well as noise etc.
Gerry
No, you did not say that. :-)
gdanmitchell wrote
which, in a roundabout way, confirms my point about it not being irrelevant. ;-)Except in the most benign and perfectly lit situations either we or the camera still need to come up with the situationally-best combination of ISO, shutter speed, and aperture and we consider subject motion, DOF, noise, and other things to make the choice.
I worry a whole lot less about it, especially now with the Adobe AI Denoise function in ACR. But I still prefer to make optimal choices for all settings that affect the image.
As I wrote more than once, I can see situations in which the automated Dynamic Range settings could make sense (and also in which the AUTO ISO option could be useful), but I feel that in much of my photography I’d rather make intentional decisions about ISO and how to handle the potential for blown highlights and blocked shadows.
YMMV.
- - -
RoamingScott wrote:
This thread is a fantastic microcosm of how hard it is to have an intelligent conversation on this forum.
As always, thank you for your helpful, informative, and “intelligent” contributions on the subject of the thread.
- - -
gyoung143 wrote:
gdanmitchell wrote
An extreme example, but, it turns out that ISO is not exactly irrelevant. Who knew?!https://gallery.gdanmitchell.com/gallery/var/albums/HumanWorld/Unclassified/Random-Images/Screenshot%202026-01-23%20at%201_59_43%E2%80%AFPM.jpg
Use the EC dial and compromise on your ideal SS and A to get ISO where you want it, bearing in mind dual gain sensor characteristics as well as noise etc.
Gerry
No, you did not say that. :-)
Jan 24, 2026 at 09:59 AM
p.2 #5DR100% vs 200/400%
Jack, some of your claims disagree with my experience with the GFX100S II.
I shall first quote the claims, then respond.
"DR. It is ... nothing more that an automated under-exposure to preserve highlights, then a bump to normalize gamma for a properly "exposed" image."
"Shoot a raw + jpeg with DR400 on, then shoot the same shot with it off. Compare the raws and the dr400 will be 2 stops under-exposed. ...So all it does is under expose the raw, but internally processes the jpeg for expanded dr, predominantly seen in the jpeg highlights. It’s a JPEG process tool, and does zero to benefit a raw."
What you claim is correct at the base ISO with DR100%.
When, for whatever reason, ISO is at or above 320 with DR100%, then switching to Dr200% or DR400% has no effect on the exposure. The shutter speed and the aperture remain unaffected, so is the exposure.
Another claim, quote: "ISO 500 is an odd choice on the GFX 100. It and 320/400 generate more noise than 800. In fact, ISO 500 is about equivalent noise to 1200 on the GFX 100 sensor. Granted, DR is a bit better at the lower ISO's, but with the noise offset, it turns out in real use, 320 through 1200 are all but equal in performance."
The PDR for my camera, GFX100S II isn't available from photonstophotos. When looking at the closely related GFX100 II, the PDR values are as follows:
12.55 (ISO80) - 11.4 (ISO200) - 10.79 (ISO320) - 10.49 (ISO400) - 10.91 (ISO500) -10.62 (ISO640) - 10.33 (ISO800).
These suggest that the dual-gain kicks in at ISO500. One may want to skip ISO320 and 400.
To confirm whether my GFX100S II behaves like the bigger "brother", I collected dark frames at different ISO values, then processed from raw by overexposing by 4 stops. The following is a compilation of the central crops at 400% magnification.
There is a visible change in the amount of noise with more gain. The changes are also reflected in the disk sizes of the crops. The larger is the size, the more noise there is in the frame.
These dark frames confirm that the dual gain on the GFX100S II is implemented at the same ISO500 as in GFX100 II. ISO500, 640, and 800 appear to have less noise than at ISO320 and 400.
I mentioned this before: on my camera, the custom set 1 is limited to ISO80 - 200, then the custom set 2 is from ISO500 to 12800. I switch from 1 to 2 when the shutter speed becomes too long for a subject.
I shall first quote the claims, then respond.
"DR. It is ... nothing more that an automated under-exposure to preserve highlights, then a bump to normalize gamma for a properly "exposed" image."
"Shoot a raw + jpeg with DR400 on, then shoot the same shot with it off. Compare the raws and the dr400 will be 2 stops under-exposed. ...So all it does is under expose the raw, but internally processes the jpeg for expanded dr, predominantly seen in the jpeg highlights. It’s a JPEG process tool, and does zero to benefit a raw."
What you claim is correct at the base ISO with DR100%.
When, for whatever reason, ISO is at or above 320 with DR100%, then switching to Dr200% or DR400% has no effect on the exposure. The shutter speed and the aperture remain unaffected, so is the exposure.
Another claim, quote: "ISO 500 is an odd choice on the GFX 100. It and 320/400 generate more noise than 800. In fact, ISO 500 is about equivalent noise to 1200 on the GFX 100 sensor. Granted, DR is a bit better at the lower ISO's, but with the noise offset, it turns out in real use, 320 through 1200 are all but equal in performance."
The PDR for my camera, GFX100S II isn't available from photonstophotos. When looking at the closely related GFX100 II, the PDR values are as follows:
12.55 (ISO80) - 11.4 (ISO200) - 10.79 (ISO320) - 10.49 (ISO400) - 10.91 (ISO500) -10.62 (ISO640) - 10.33 (ISO800).
These suggest that the dual-gain kicks in at ISO500. One may want to skip ISO320 and 400.
To confirm whether my GFX100S II behaves like the bigger "brother", I collected dark frames at different ISO values, then processed from raw by overexposing by 4 stops. The following is a compilation of the central crops at 400% magnification.
There is a visible change in the amount of noise with more gain. The changes are also reflected in the disk sizes of the crops. The larger is the size, the more noise there is in the frame.
These dark frames confirm that the dual gain on the GFX100S II is implemented at the same ISO500 as in GFX100 II. ISO500, 640, and 800 appear to have less noise than at ISO320 and 400.
I mentioned this before: on my camera, the custom set 1 is limited to ISO80 - 200, then the custom set 2 is from ISO500 to 12800. I switch from 1 to 2 when the shutter speed becomes too long for a subject.
Jan 24, 2026 at 11:56 AM
p.2 #6DR100% vs 200/400%
I purchased RawDigger and looked at the raw histograms of the three files, recorded with DR100%, DR200%, and DR400% (see the opening post for the jpeg of the entire photographed scene). The exposure settings - shutter speed and the aperture - were the same between the three raw files, as well as was the ISO (at 500).
Uploaded below is a compilation of the raw histograms, together with the percentages of underexposed pixels measured by RawDigger. The DR100% raw file had overexposed pixels in the green channels. There are no overexposed pixels in the DR200% and DR400% raw files, according to RawDigger.
What these histograms clearly show is 1 stop and 2 stop shift to the left with DR200% and DR400%, respectively.
To the best of my understanding, I guess that the analogue gain was reduced proportionally with these settings. I don't think that the camera manipulated and actually reduced the exposure (by "secretly" increasing the shutter speed or by closing the aperture). There seems to be no evidence in the histograms to suggest that GFX100S II employs some form of dual gain output (that is, a low gain toward the strong signal and a normal gain toward the weak signal).
What I also think (and would welcome a correction, if I am wrong) is that the reduced analogue gain equally affected the signal, the shot and the read noises. Thus, I expect that the S/N ratio (that is the dynamic range) didn't change as a result.
Uploaded below is a compilation of the raw histograms, together with the percentages of underexposed pixels measured by RawDigger. The DR100% raw file had overexposed pixels in the green channels. There are no overexposed pixels in the DR200% and DR400% raw files, according to RawDigger.
What these histograms clearly show is 1 stop and 2 stop shift to the left with DR200% and DR400%, respectively.
To the best of my understanding, I guess that the analogue gain was reduced proportionally with these settings. I don't think that the camera manipulated and actually reduced the exposure (by "secretly" increasing the shutter speed or by closing the aperture). There seems to be no evidence in the histograms to suggest that GFX100S II employs some form of dual gain output (that is, a low gain toward the strong signal and a normal gain toward the weak signal).
What I also think (and would welcome a correction, if I am wrong) is that the reduced analogue gain equally affected the signal, the shot and the read noises. Thus, I expect that the S/N ratio (that is the dynamic range) didn't change as a result.
Jan 26, 2026 at 07:46 PM
p.2 #7DR100% vs 200/400%
One useful way to see what is happening is to shoot a gray scale chart, perhaps one with something like 8 or 16 patches. It will show narrow peaks for each patch, and it is easy to see how (if) the exposure is being manipulated.
Edited on Jan 27, 2026 at 10:19 AM · View previous versions
Edited on Jan 27, 2026 at 10:19 AM · View previous versions
Jan 26, 2026 at 11:10 PM
p.2 #8DR100% vs 200/400%
Jack Flesher wrote:
An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image.
Except you expose exactly the same (shutter speed and aperture) for the same highlight retention. The only difference is that the camera raises shadows instead of you doing it in PP, and many times the former looks much more natural than the latter.
You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.
An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image.
Except you expose exactly the same (shutter speed and aperture) for the same highlight retention. The only difference is that the camera raises shadows instead of you doing it in PP, and many times the former looks much more natural than the latter.
You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.
Jan 27, 2026 at 12:16 AM
p.2 #9DR100% vs 200/400%
Makten wrote:
Jack Flesher wrote
An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image.You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.
This is possibly confusing. Naturally, gain (ISO) cannot change exposure, simply because it is applied toward an already captured signal, after the sensor had been exposed to light. However, in aperture or shutter priority, exposure compensation or ISO are implemented by the camera manufacturers to be linked to either the shutter speed or the aperture. E.g., in the aperture priority case, increase ISO gain by a stop, and the shutter speed doubles. Thus, the exposure is reduced by a stop.
My interest in how DR200% and DR400% work is all to get an understanding of whether these settings affect exposure or not, when used in the light when auto-ISO is above 320 with DR100%.
If the answer is "no", as it seems to me, then this is better than using exposure compensation in the A or S modes.
Theoretically, in full manual and with ISO decoupled from the shutter speed and aperture, one can play with the gain manually, like the camera seems to be doing with DR200% and DR400%. Note that when ISO is two stops above the base (320 vs 80 for GFX100S II), the sensor is not at FWC: there is a roughly 2-stop headroom for playing with the gain. In this regime, when highlights clipping happens, this is caused exclusively by too much analogue gain (prior to ADC).
It seems that DR200 and DR400 are methods that use 1 and 2 stop less gain than the normal auto-ISO (when the ISO is 1 and 2 stops above the base ISO, respectively). This make more sense than exposure compensation in the A or S modes.
Jan 27, 2026 at 07:46 AM
p.2 #10DR100% vs 200/400%
Makten wrote:
Except you expose exactly the same (shutter speed and aperture) for the same highlight retention. The only difference is that the camera raises shadows instead of you doing it in PP, and many times the former looks much more natural than the latter.
You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.
As I indicated in the previous thread on this, it works great for in-cam JPEG shooters. However it is absolutely detrimental to the RAW. While it does seem to work behind the scenes and produce natural-looking in-camera JPEG’s, it absolutely under-exposes the RAW file by 1 (DR200) or 2 (DR400) stops, and in that compresses shadows (significantly) beyond what a normal, proper exposure does. With the added bit-depth from RAW, I rarely run into a situation where the normal matrix metered exposure will blow highlights in the RAW capture — yes it can happen, but an experienced shooter will understand immediately when it is happening — and just like for the JPEG shooter’s VF/LCD display showing the better shadows with DR200/400 on, with DR200/400 OFF it will similarly show blown highlights at the same time for the RAW shooter, allowing the knowledgable photographer to compensate accordingly.
Working with raw, I have far more latitude to manage global exposure, HDR sliders and gamma in post for a far superior output compared to the JPEG. And I suspect the camera is doing this via a preset algorithm since it’s using raw sensor data to generate the jpeg.
In conclusion, if you shoot only JPEG, I think it’s an effective tool, however if you shoot RAW or RAW+JPEG, it is generally detrimental to the RAW.
Jack Flesher wrote
An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image.You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.
As I indicated in the previous thread on this, it works great for in-cam JPEG shooters. However it is absolutely detrimental to the RAW. While it does seem to work behind the scenes and produce natural-looking in-camera JPEG’s, it absolutely under-exposes the RAW file by 1 (DR200) or 2 (DR400) stops, and in that compresses shadows (significantly) beyond what a normal, proper exposure does. With the added bit-depth from RAW, I rarely run into a situation where the normal matrix metered exposure will blow highlights in the RAW capture — yes it can happen, but an experienced shooter will understand immediately when it is happening — and just like for the JPEG shooter’s VF/LCD display showing the better shadows with DR200/400 on, with DR200/400 OFF it will similarly show blown highlights at the same time for the RAW shooter, allowing the knowledgable photographer to compensate accordingly.
Working with raw, I have far more latitude to manage global exposure, HDR sliders and gamma in post for a far superior output compared to the JPEG. And I suspect the camera is doing this via a preset algorithm since it’s using raw sensor data to generate the jpeg.
In conclusion, if you shoot only JPEG, I think it’s an effective tool, however if you shoot RAW or RAW+JPEG, it is generally detrimental to the RAW.
Jan 27, 2026 at 08:39 AM
|
p.2 #11DR100% vs 200/400%
Jack Flesher wrote:
As I indicated in the previous thread on this, it works great for in-cam JPEG shooters. However it is absolutely detrimental to the RAW. While it does seem to work behind the scenes and produce natural-looking in-camera JPEG’s, it absolutely under-exposes the RAW file by 1 (DR200) or 2 (DR400) stops, and in that compresses shadows (significantly) beyond what a normal, proper exposure does. With the added bit-depth from RAW, I rarely run into a situation where the normal matrix metered exposure will blow highlights in the RAW capture — yes it can happen, but an experienced shooter will...Show more →
This might just be a semantic issue, but I don't believe that the shadows are "compressed" (i.e., an actual loss of dynamic range in the shadows).
While I didn't look at the charts posted in the thread, I don't think underexposing is necessarily detrimental to a RAW file. Even experienced shooters usually rely on the camera to do the metering, and if it messes up, which is sometimes does, the DR feature can be a life saver. And sometimes even experienced shooters mess up the exposure on our own. I know it's been helpful to me personally, even when processing RAW files. But I agree, it's intended to be used for JPGs.
As you can see online, there are many opinions about whether it's better to slightly underexpose a raw file or slightly overexpose it (assuming you cannot nail the exposure). I'm in the "rather underexpose than overexpose" camp myself.
As I indicated in the previous thread on this, it works great for in-cam JPEG shooters. However it is absolutely detrimental to the RAW. While it does seem to work behind the scenes and produce natural-looking in-camera JPEG’s, it absolutely under-exposes the RAW file by 1 (DR200) or 2 (DR400) stops, and in that compresses shadows (significantly) beyond what a normal, proper exposure does. With the added bit-depth from RAW, I rarely run into a situation where the normal matrix metered exposure will blow highlights in the RAW capture — yes it can happen, but an experienced shooter will...Show more →
This might just be a semantic issue, but I don't believe that the shadows are "compressed" (i.e., an actual loss of dynamic range in the shadows).
While I didn't look at the charts posted in the thread, I don't think underexposing is necessarily detrimental to a RAW file. Even experienced shooters usually rely on the camera to do the metering, and if it messes up, which is sometimes does, the DR feature can be a life saver. And sometimes even experienced shooters mess up the exposure on our own. I know it's been helpful to me personally, even when processing RAW files. But I agree, it's intended to be used for JPGs.
As you can see online, there are many opinions about whether it's better to slightly underexpose a raw file or slightly overexpose it (assuming you cannot nail the exposure). I'm in the "rather underexpose than overexpose" camp myself.
Jan 27, 2026 at 09:10 AM
p.2 #12DR100% vs 200/400%
This whole discussion has been very helpful. Thanks Ruthenium for all your testing and thank you to all the other contributors.
Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.
Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.
Jan 27, 2026 at 09:41 AM
p.2 #13DR100% vs 200/400%
InFocus2014 wrote:
This whole discussion has been very helpful. Thanks Ruthenium for all your testing and thank you to all the other contributors.
Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.
I think the dual gain sensor switches from low gain to high gain at ISO800. But you can use DR400 as low as ISO640.
This whole discussion has been very helpful. Thanks Ruthenium for all your testing and thank you to all the other contributors.
Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.
I think the dual gain sensor switches from low gain to high gain at ISO800. But you can use DR400 as low as ISO640.
Jan 27, 2026 at 10:05 AM
p.2 #14DR100% vs 200/400%
Jack Flesher wrote:
Except you expose exactly the same (shutter speed and aperture) for the same highlight retention. The only difference is that the camera raises shadows instead of you doing it in PP, and many times the former looks much more natural than the latter.
You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.As I indicated in the previous thread on this, it works great for in-cam JPEG shooters. However it is absolutely detrimental to the RAW. While it does seem to work behind the scenes and produce natural-looking in-camera JPEG’s, it absolutely under-exposes the RAW file by 1 (DR200) or 2 (DR400) stops, and in that compresses shadows (significantly) beyond what a normal, proper exposure does. With the added bit-depth from RAW, I rarely run into a situation where the normal matrix metered exposure will blow highlights in the RAW capture — yes it can happen, but an experienced shooter will...Show more →
You keep insisting on that DR200 or DR400 settings "under-expose(!) the RAW file by 1 (DR200) or 2 (DR400) stops"
Can you confirm that you have evidence for this at ISO above 320?
Can you confirm that it is the EXPOSURE that is reduced with DR200 and DR400, and not the analogue GAIN (when ISO is above 320)?
There is a big difference between the two scenarios.
I agree with you that the UNDEREXPOSURE is "detrimental to the RAW" (in the shadows).
This is different with the GAIN.
With a practically ISOless sensor, there is no increased noise penalty in post, if the analogue gain was lower than the optimal. For sensors that are not ISOless, the SNR is slightly improved by using a higher gain (up to the point of highlights clipping), however, I expect that the effect of using a lower than optimal GAIN is less detrimental to the RAW than the UNDEREXPOSURE.
Would you agree?
Makten wrote
Jack Flesher wrote
An internal processing gimmick for those that don't know how to properly expose and process out a normal-looking image.You will of course get the exact same amount of noise in the (lifted) shadows in both cases, because you let the exact same amount of light hit the sensor (no, ISO doesn't change "real" exposure).
Edit: Not to mention that the shadows won't be all black in the EVF when composing the image. Added bonus.
You keep insisting on that DR200 or DR400 settings "under-expose(!) the RAW file by 1 (DR200) or 2 (DR400) stops"
Can you confirm that you have evidence for this at ISO above 320?
Can you confirm that it is the EXPOSURE that is reduced with DR200 and DR400, and not the analogue GAIN (when ISO is above 320)?
There is a big difference between the two scenarios.
I agree with you that the UNDEREXPOSURE is "detrimental to the RAW" (in the shadows).
This is different with the GAIN.
With a practically ISOless sensor, there is no increased noise penalty in post, if the analogue gain was lower than the optimal. For sensors that are not ISOless, the SNR is slightly improved by using a higher gain (up to the point of highlights clipping), however, I expect that the effect of using a lower than optimal GAIN is less detrimental to the RAW than the UNDEREXPOSURE.
Would you agree?
Jan 27, 2026 at 10:09 AM
p.2 #15DR100% vs 200/400%
According to pal2tech the DR200/400 settings do something before the data goes to the analog to digital converter:
I'm not sure where he gets this information, but he does show that there is a difference in the raw files.
I'm not sure where he gets this information, but he does show that there is a difference in the raw files.
Jan 27, 2026 at 10:18 AM
p.2 #16DR100% vs 200/400%
mdude85 wrote:
Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.I think the dual gain sensor switches from low gain to high gain at ISO800. But you can use DR400 as low as ISO640.
Dual gain is implemented at different ISO in different cameras; thus, this is camera-specific.
For example, GFX100, GFX100S, and GFX100 II all employ the dual gain at ISO 500 (according to photonstophotos)
Another example: GFX50S II doesn't use the dual gain at all, according to the same source.
InFocus2014 wrote
This whole discussion has been very helpful. Thanks Ruthenium for all your testing and thank you to all the other contributors.Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.
Dual gain is implemented at different ISO in different cameras; thus, this is camera-specific.
For example, GFX100, GFX100S, and GFX100 II all employ the dual gain at ISO 500 (according to photonstophotos)
Another example: GFX50S II doesn't use the dual gain at all, according to the same source.
Jan 27, 2026 at 10:18 AM
p.2 #17DR100% vs 200/400%
Makten wrote:
…the camera raises shadows instead of you doing it in PP, and many times the former looks much more natural than the latter.
I see it pretty much the opposite way. The automatic in-camera shadow raising may well be good enough in some cases and doesn’t require any input from the photographer — so it can be easier.
However, you can certainly duplicate that canned effect in post, but you are not limited to that.
For example, it isn’t unusual to want to bring up shadows in one particularly dark area of the frame while keeping that adjustment away from other parts of the image. That’s what masks are for. (In LR or ACR/PS you can, for example, select an area in many ways, including luminosity levels, etc. and then adjust only that area.)
The in-camera method (like shooting jpg) applies the camera’s adjustments uniformly to the entire image. (As I and others have pointed out, this dynamic range option is likely most useful to jpg shooters who aren’t interested in more sophisticated post-processing or doing things like the digital equivalent of zone system style exposure.)
ruthenium wrote:
You keep insisting on that DR200 or DR400 settings "under-expose(!) the RAW file by 1 (DR200) or 2 (DR400) stops"
Can you confirm that you have evidence for this at ISO above 320?
Can you confirm that it is the EXPOSURE that is reduced with DR200 and DR400, and not the analogue GAIN (when ISO is above 320)?
I don’t see any other way that this could work. To my understanding, what is going on is that the file is being “under-exposed” by one or two “stops” equivalent in order to retain highlight detail, and then a curve is applied that raises the levels at the dark end of the luminosity range.
You can’t actually apply a curve to the sensor input or its registration. Each photosite still receives and records whatever luminosity level strikes it, no matter what the setting. The sensor records what the sensor records. But you can manipulate the data from the sensor after the fact. And, while working with the raw image data (either in-camera before the camera produces a jpg, or in post) the shadows can be lifted while retaining good image quality.
The concept is roughly analogous to what film photographers did with the zone system back in the day. If you had a subject with a very wide dynamic range (determined by spot-metering the bright and dark areas) you would expose for the best compromise that wold give you sufficient detail in the highlights (densest part of the negative) and shadows (most transparent part) to work with in post. This meant that your exposure could create a negative that made a pretty awful straight (not dodged or burned) contact print, but which contained sufficient image data to allow you to extract the best detail in post. (Here “post” refers to film development variations and to darkroom techniques like dodging and burning.)
With digital, unless we want to let them blow out (sometimes OK) the most critical thing is protecting the highlights. When they blow out there is no way to restore them in post — all highlight areas simply get the same maximum luminosity value, e.g. pure white. As a result, in wide dynamic range scenes, the shadows can end up looking too dark. However, with wide dynamic range cameras today, there is sufficient good quality detail in those shadows so that we can bring back shadow detail quite effectively. Motto: Protect the highlights and restore the shadows in post.
Edited on Jan 28, 2026 at 09:29 AM · View previous versions
…the camera raises shadows instead of you doing it in PP, and many times the former looks much more natural than the latter.
I see it pretty much the opposite way. The automatic in-camera shadow raising may well be good enough in some cases and doesn’t require any input from the photographer — so it can be easier.
However, you can certainly duplicate that canned effect in post, but you are not limited to that.
For example, it isn’t unusual to want to bring up shadows in one particularly dark area of the frame while keeping that adjustment away from other parts of the image. That’s what masks are for. (In LR or ACR/PS you can, for example, select an area in many ways, including luminosity levels, etc. and then adjust only that area.)
The in-camera method (like shooting jpg) applies the camera’s adjustments uniformly to the entire image. (As I and others have pointed out, this dynamic range option is likely most useful to jpg shooters who aren’t interested in more sophisticated post-processing or doing things like the digital equivalent of zone system style exposure.)
ruthenium wrote:
You keep insisting on that DR200 or DR400 settings "under-expose(!) the RAW file by 1 (DR200) or 2 (DR400) stops"
Can you confirm that you have evidence for this at ISO above 320?
Can you confirm that it is the EXPOSURE that is reduced with DR200 and DR400, and not the analogue GAIN (when ISO is above 320)?
I don’t see any other way that this could work. To my understanding, what is going on is that the file is being “under-exposed” by one or two “stops” equivalent in order to retain highlight detail, and then a curve is applied that raises the levels at the dark end of the luminosity range.
You can’t actually apply a curve to the sensor input or its registration. Each photosite still receives and records whatever luminosity level strikes it, no matter what the setting. The sensor records what the sensor records. But you can manipulate the data from the sensor after the fact. And, while working with the raw image data (either in-camera before the camera produces a jpg, or in post) the shadows can be lifted while retaining good image quality.
The concept is roughly analogous to what film photographers did with the zone system back in the day. If you had a subject with a very wide dynamic range (determined by spot-metering the bright and dark areas) you would expose for the best compromise that wold give you sufficient detail in the highlights (densest part of the negative) and shadows (most transparent part) to work with in post. This meant that your exposure could create a negative that made a pretty awful straight (not dodged or burned) contact print, but which contained sufficient image data to allow you to extract the best detail in post. (Here “post” refers to film development variations and to darkroom techniques like dodging and burning.)
With digital, unless we want to let them blow out (sometimes OK) the most critical thing is protecting the highlights. When they blow out there is no way to restore them in post — all highlight areas simply get the same maximum luminosity value, e.g. pure white. As a result, in wide dynamic range scenes, the shadows can end up looking too dark. However, with wide dynamic range cameras today, there is sufficient good quality detail in those shadows so that we can bring back shadow detail quite effectively. Motto: Protect the highlights and restore the shadows in post.
Edited on Jan 28, 2026 at 09:29 AM · View previous versions
Jan 27, 2026 at 10:25 AM
p.2 #18DR100% vs 200/400%
ruthenium wrote:
Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.I think the dual gain sensor switches from low gain to high gain at ISO800. But you can use DR400 as low as ISO640.Dual gain is implemented at different ISO in different cameras; thus, this is camera-specific.
For example, GFX100, GFX100S, and GFX100 II all employ the dual gain at ISO 500 (according to photonstophotos)
Another example: GFX50S II doesn't use the dual gain at all, according to the same source.
Correct, I was multitasking and I should have noted that! On earlier Fuji cameras I think it's 800. I guess you will have to experiment with different ISOs to see which ones give the best results for you at various ISOs and DR settings.
mdude85 wrote
InFocus2014 wrote
This whole discussion has been very helpful. Thanks Ruthenium for all your testing and thank you to all the other contributors.Based on other posts on this subject, I have, in the past, dismissed this feature. One thing I didn't take into account was the characteristics of the dual gain sensor. I think I will now try it out and compare to my traditional procedure of watching for highlight clipping and then raising the shadows in post.
For example, GFX100, GFX100S, and GFX100 II all employ the dual gain at ISO 500 (according to photonstophotos)
Another example: GFX50S II doesn't use the dual gain at all, according to the same source.
Correct, I was multitasking and I should have noted that! On earlier Fuji cameras I think it's 800. I guess you will have to experiment with different ISOs to see which ones give the best results for you at various ISOs and DR settings.
Jan 27, 2026 at 10:32 AM
p.2 #19DR100% vs 200/400%
mivadep wrote:
According to pal2tech the DR200/400 settings do something before the data goes to the analog to digital converter:
I'm not sure where he gets this information, but he does show that there is a difference in the raw files.
This is not that straightforward.
The key question is where the highlights are clipped - on the sensor or in the ADC?
There are two distinct scenarios:
1) The light is abundant, the ISO is at the base value, e.g. ISO80 for my camera, the GFX100S II (the base ISO can be different for different cameras). In this scenario, the sensor is near its full-well-capacity (FWC), and there is a risk of the highlights being clipped because of overexposure.
A reasonable strategy to deal with this is to add exposure compensation (-1/3, -2/3, -1, as much as needed according to the histogram). This shall make the camera increase the shutter speed in the A mode, or close the aperture in the S mode.
DR200 and DR400 settings will do exactly the same, but in a more "crude" way: by reducing exposure by -1 and -2 stops, respectively.
2) In the distinct second scenario, the light is less abundant and the auto-ISO is at or above the value of 2 x base ISO (e.g., at or above ISO320 on my camera GFX100S II). This is when the sensor is not at FWC and isn't overexposed, except in the special case of dealing with moderately lit scenes where there are points of bright light. A trivial example of the latter is a moderately lit Christmas tree with bright lights on it. But this is a special case. Outside of the special cases, when the highlights are clipped, this happens in the analogue-to-digital converter (not on the sensor) if the pre-amplifier happened to apply a stronger analogue gain toward the signal from the sensor.
It is in this distinct scenario, when the problem can be helped by reducing the gain, without changing the exposure.
Cameras like GFX100S II (and the related GFX100 bodies) have the base ISO at 80 and the dual gain at ISO 500. This suggests a practical strategy of using two custom sets: one with the ISO from 80 to 200 for best exposure, and the other from 500 to 12800 (or a lower upper ISO limit if so is desired) for any practical reason when the best exposure cannot be used.
Then, the photographer would switch from custom set 1 to 2 as needed.
It makes sense to have DR at 100% in customer set 1.
I think that custom set 2 may have DR set to 200%. Another option is AUTO, however, I have no idea how smart is the camera, and how it decides when DR should be switched from 100% to 200%. This needs to be tested.
Understandably, my strategy with the two custom sets is for the GFX100S II, and is decided by the knowledge of the base ISO (80) and the ISO when the dual gain is employed by the camera (500). For other cameras, my strategy may not be applicable if there is no dual gain, or when the dual gain is implemented at a higher ISO than 500.
Added: I don't mean to contradict "he does show that there is a difference in the raw files" - this is correct. Indeed, the raw files are affected by the DR settings. This is important for those who shoot raw.
According to pal2tech the DR200/400 settings do something before the data goes to the analog to digital converter:
I'm not sure where he gets this information, but he does show that there is a difference in the raw files.
This is not that straightforward.
The key question is where the highlights are clipped - on the sensor or in the ADC?
There are two distinct scenarios:
1) The light is abundant, the ISO is at the base value, e.g. ISO80 for my camera, the GFX100S II (the base ISO can be different for different cameras). In this scenario, the sensor is near its full-well-capacity (FWC), and there is a risk of the highlights being clipped because of overexposure.
A reasonable strategy to deal with this is to add exposure compensation (-1/3, -2/3, -1, as much as needed according to the histogram). This shall make the camera increase the shutter speed in the A mode, or close the aperture in the S mode.
DR200 and DR400 settings will do exactly the same, but in a more "crude" way: by reducing exposure by -1 and -2 stops, respectively.
2) In the distinct second scenario, the light is less abundant and the auto-ISO is at or above the value of 2 x base ISO (e.g., at or above ISO320 on my camera GFX100S II). This is when the sensor is not at FWC and isn't overexposed, except in the special case of dealing with moderately lit scenes where there are points of bright light. A trivial example of the latter is a moderately lit Christmas tree with bright lights on it. But this is a special case. Outside of the special cases, when the highlights are clipped, this happens in the analogue-to-digital converter (not on the sensor) if the pre-amplifier happened to apply a stronger analogue gain toward the signal from the sensor.
It is in this distinct scenario, when the problem can be helped by reducing the gain, without changing the exposure.
Cameras like GFX100S II (and the related GFX100 bodies) have the base ISO at 80 and the dual gain at ISO 500. This suggests a practical strategy of using two custom sets: one with the ISO from 80 to 200 for best exposure, and the other from 500 to 12800 (or a lower upper ISO limit if so is desired) for any practical reason when the best exposure cannot be used.
Then, the photographer would switch from custom set 1 to 2 as needed.
It makes sense to have DR at 100% in customer set 1.
I think that custom set 2 may have DR set to 200%. Another option is AUTO, however, I have no idea how smart is the camera, and how it decides when DR should be switched from 100% to 200%. This needs to be tested.
Understandably, my strategy with the two custom sets is for the GFX100S II, and is decided by the knowledge of the base ISO (80) and the ISO when the dual gain is employed by the camera (500). For other cameras, my strategy may not be applicable if there is no dual gain, or when the dual gain is implemented at a higher ISO than 500.
Added: I don't mean to contradict "he does show that there is a difference in the raw files" - this is correct. Indeed, the raw files are affected by the DR settings. This is important for those who shoot raw.
Jan 27, 2026 at 11:24 AM
p.2 #20DR100% vs 200/400%
mdude85 wrote:
Working with raw, I have far more latitude to manage global exposure, HDR sliders and gamma in post for a far superior output compared to the JPEG. And I suspect the camera is doing this via a preset algorithm since it’s using raw sensor data to generate the jpeg.
In conclusion, if you shoot only JPEG, I think it’s an effective tool, however if you shoot RAW or RAW+JPEG, it is generally detrimental to the RAW.This might just be a semantic issue, but I don't believe that the shadows are "compressed" (i.e., an actual loss of dynamic range in the shadows).
While I didn't look at the charts posted in the thread, I don't think underexposing is necessarily detrimental to a RAW file. Even experienced shooters usually rely on the camera to do the metering, and if it messes up, which is sometimes does, the DR feature can be a life saver. And sometimes even experienced shooters mess up the exposure on our own. I know it's been helpful to me personally, even when processing...Show more →
The natural brightness range evident in a 'scene' is more than can be accommodated on any media we might want to 'finish' our pictures, paper, or screen, espevially in bright sunshine, even if we could record it via sensor to a file, so compromise is surely necessary even with digital. On b&w negative it could be compensated for, by exposure, development and choice of paper grade (applied sensutometry) and was I'm sure used by many even before Ansel Adams put a fancy label on it as 'zone system'. In some ways he was a precursor the U tube warriors, oversimplified known physics when a good understanding of the process is all that is required.
On colour neg, and transparency there wasn't much to be done about it, at least until Fuji came up with three 'grades' of slide film, I still have some Astia and Provia in the freezer.
With an EVF we have a good tool to get it 'right', and with the knowledge that 'under exposed shadows' has far less impact on finished image IQ than burnt out and lost forever highlights, we can make our own decisions at the point of taking the photos.
The DR 200 and 400 settings seem to me to be another Gimmick for those wanting to avoid post processing, modifying the jpegs to give similar effect. It seemed to me that increasing the minimum ISO to just below the dual gain change is the worst possible thing to do if noise is the 'enemy', so I dismissed it and went back to making my decision at the point 9f setting the exposure, how much to 'inderexpose' to keep the highlights that I want and what ISO to use to keep noise as low as possible.
Gerry
Jack Flesher wrote
As I indicated in the previous thread on this, it works great for in-cam JPEG shooters. However it is absolutely detrimental to the RAW. While it does seem to work behind the scenes and produce natural-looking in-camera JPEG’s, it absolutely under-exposes the RAW file by 1 (DR200) or 2 (DR400) stops, and in that compresses shadows (significantly) beyond what a normal, proper exposure does. With the added bit-depth from RAW, I rarely run into a situation where the normal matrix metered exposure will blow highlights in the RAW capture — yes it can happen, but an experienced shooter will understand immediately when it is happening — and just like for the JPEG shooter’s VF/LCD display showing the better shadows with DR200/400 on, with DR200/400 OFF it will similarly show blown highlights at the same time for the RAW shooter, allowing the knowledgable photographer to compensate accordingly.Working with raw, I have far more latitude to manage global exposure, HDR sliders and gamma in post for a far superior output compared to the JPEG. And I suspect the camera is doing this via a preset algorithm since it’s using raw sensor data to generate the jpeg.
In conclusion, if you shoot only JPEG, I think it’s an effective tool, however if you shoot RAW or RAW+JPEG, it is generally detrimental to the RAW.
While I didn't look at the charts posted in the thread, I don't think underexposing is necessarily detrimental to a RAW file. Even experienced shooters usually rely on the camera to do the metering, and if it messes up, which is sometimes does, the DR feature can be a life saver. And sometimes even experienced shooters mess up the exposure on our own. I know it's been helpful to me personally, even when processing...Show more →
The natural brightness range evident in a 'scene' is more than can be accommodated on any media we might want to 'finish' our pictures, paper, or screen, espevially in bright sunshine, even if we could record it via sensor to a file, so compromise is surely necessary even with digital. On b&w negative it could be compensated for, by exposure, development and choice of paper grade (applied sensutometry) and was I'm sure used by many even before Ansel Adams put a fancy label on it as 'zone system'. In some ways he was a precursor the U tube warriors, oversimplified known physics when a good understanding of the process is all that is required.
On colour neg, and transparency there wasn't much to be done about it, at least until Fuji came up with three 'grades' of slide film, I still have some Astia and Provia in the freezer.
With an EVF we have a good tool to get it 'right', and with the knowledge that 'under exposed shadows' has far less impact on finished image IQ than burnt out and lost forever highlights, we can make our own decisions at the point of taking the photos.
The DR 200 and 400 settings seem to me to be another Gimmick for those wanting to avoid post processing, modifying the jpegs to give similar effect. It seemed to me that increasing the minimum ISO to just below the dual gain change is the worst possible thing to do if noise is the 'enemy', so I dismissed it and went back to making my decision at the point 9f setting the exposure, how much to 'inderexpose' to keep the highlights that I want and what ISO to use to keep noise as low as possible.
Gerry
Jan 28, 2026 at 04:32 AM