gyoung143 wrote:
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
I just think people who shot in the film era have different rules for their tools, so to speak.
These days, the noise between ISO 100 and ISO 500 is virtually indistinguishable, even if the image is "technically cleaner" at ISO 500 (when dual gain switches the base ISO) than the ISO just below it. One really doesn't need to keep noise as low as possible -- the images are already super clean even at very high ISOs, and post-processing is able to eliminate pretty much any noise if you want it to.
I don't agree that D-Range is a gimmick. It helps you achieve an image with apparent higher dynamic range without much post-processing. That is a valuable tool for a lot of people.
As usual, this is a overly-wrought mental gymnastics exercise with 0 real world examples provided. All hang wringing and no pudding.
Go out in the world and shoot a challenge scene at DR100, 200, and 400 and come back here and post your findings, as you so often ask others to do, Dmitri.
mdude85 wrote:
I just think people who shot in the film era have different rules for their tools, so to speak.
These days, the noise between ISO 100 and ISO 500 is virtually indistinguishable, even if the image is "technically cleaner" at ISO 500 (when dual gain switches the base ISO) than the ISO just below it. One really doesn't need to keep noise as low as possible -- the images are already super clean even at very high ISOs, and post-processing is able to eliminate pretty much any noise if you want it to.
I don't agree that D-Range is a gimmick. It helps you achieve an image with apparent higher dynamic range without much post-processing. That is a valuable tool for a lot of people. ...Show more →
I don't disagree with any of that, even before dual gain sensors the noise performance at up to 500 was far, far better than, f9r instance, Provia 400, and it takes a fair bit of looking to see any noise at thatlevel now. And I do in practice let ISO wander up there if the compromises on aperture and SS would be too great. It's a matter of judgement..
I usually shoot to raw only, the jpeg is a 'backup and stays in the camera normally. I did try DR200 & 400 when I got my first Fuji, an Xpro2, and I couldn't see any advantage over my 'normal' method of exposing for highlights and raising shadows in post processing, as well as the limitations on ISO settings available it didn't seem to work on the RAF file.
Might indeed be useful if/when shooting for jpegs out of camera, which I have done occasionally when pictures needed to a tight deadline, its limitations quite small to add to those already accepted. Fortunately I don't need to do it very often anymore!
RoamingScott wrote:
As usual, this is a overly-wrought mental gymnastics exercise with 0 real world examples provided. All hang wringing and no pudding.
Go out in the world and shoot a challenge scene at DR100, 200, and 400 and come back here and post your findings, as you so often ask others to do, Dmitri.
Oh and provide the RAW files while you're at it.
Everyone here deeply appreciates the extensive effort you put into crafting your careful, thoughtful, and informative replies — and the helpful real-world examples you share from your photography to illustrate your points. We cannot overlook your long history of sharing your raw files here to back up your claims on FM. We particularly appreciate that in this post asking others to share proof in the form of raw files that you shared your own raw files demonstrating your claims about DR100, DR200, and DR400 files.
It is, indeed, difficult to argue with any of the evidence you shared.
mdude85 wrote:
D-Range… helps you achieve an image with apparent higher dynamic range without much post-processing. That is a valuable tool for a lot of people.
I think that’s a fair comment and it is precisely in line with what I’ve been saying about the feature.
gyoung143 wrote:
I did try DR200 & 400 when I got my first Fuji, an Xpro2, and I couldn't see any advantage over my 'normal' method of exposing for highlights and raising shadows in post processing, as well as the limitations on ISO settings available it didn't seem to work on the RAF file.
Might indeed be useful if/when shooting for jpegs out of camera...
Gerry
Yes, the advantage is specifically when shooting for JPGs out of camera (or JPGs requiring minimal post-processing) so that you don't have to spend time in post-processing pulling up the shadows.
At the same time, I've had images that have been saved by DR even when shooting RAW.
mdude85 wrote:
Yes, the advantage is specifically when shooting for JPGs out of camera (or JPGs requiring minimal post-processing) so that you don't have to spend time in post-processing pulling up the shadows.
At the same time, I've had images that have been saved by DR even when shooting RAW.
Are you saying that it should be effective with the RAF file as well? I couldn't see that when I tried it . If si I will have a go with the Xt5 and see what happens!
RoamingScott wrote:
As usual, this is a overly-wrought mental gymnastics exercise with 0 real world examples provided. All hang wringing and no pudding.
Go out in the world and shoot a challenge scene at DR100, 200, and 400 and come back here and post your findings, as you so often ask others to do, Dmitri.
gdanmitchell wrote:
I think that’s a fair comment and it is precisely in line with what I’ve been saying about the feature.
My camera has two settings:
DYNAMIC RANGE
D RANGE PRIORITY
To avoid a misunderstanding, my posts in this thread are concerned exclusively with the first one and the options therein (DR100, DR200, and DR400). This should not be confused with the D RANGE PRIORITY.
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.
Nope, not at all.
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....Show more →
You are not understanding when and how to use it. If you use the same shutter speed and aperture to get the same highlight coverage, there is absolutely no loss using the DR modes with raw files. The sensor captures the exact same amount of photons. The only thing that happens is that the camera lifts the shadows for you, and in my case it looks way better than if I do it in PP.
Edit: As stated in the post above, "D range priority" is something else and _that_ I'd say is only meaningful when shooting JPG (even if I don't at all like the results), because all that happens to the raw file is that it's darker overall. Exactly like if you would ETTR for a bright sky or whatever.
In my testing using my preferred raw processor, all DR400 does to the raw is trim a stop or so off the highlights; or more precisely takes 255 highlights down to around 225. It does it in a non-linear way so that is at least a plus, but it sometimes leaves a weird highlight shadow around the specular which is unnatural, a lot like tone mapping halos which I despise. And then it’s difficult to make that highlight look natural, hence my dislike of it. If it only affected the jpeg, I might use it, but for me it messes up the raw highlights to an unnatural appearance and why I don’t like it. YMMV. It then also appears to lift the shadow a bit on the jpeg, probably another stop, but not on the raw, likely due to the added bit depth.
FWIW my high contrast test scene was an indoor shot with widow view in full daylight and so cast specular highlights on inside objects.
My conclusion is simple. Great it works for you guys and you love it, but it doesn’t work for me.
FWIW, I just did a little experiment with XT5 raw files using the 100%, 200%, and 400% settings. I photographed an eight-patch gray scale chart at ISO 800 and f/5.6, letting the camera select the shutter speed.
As expected, the histograms for the test images show eight sharp peaks separated by deep valleys, with each peak corresponding to one of the eight gray values from the chart.
If the dynamic range of the raw file were expanded by choosing 200% or 400% we would expect the peaks to move closer together as the dynamic range expanded. E.g. those eight patches would cover a smaller percentage of an expanded 200% or 400% dynamic range.
That's not what happened. In order to compare, I copied the histograms from the three files and superimposed and aligned them as layers in Photoshop, where I could adjust their transparency in order to see two at once. The raw file curves looked virtually identical with their peaks exactly the same distance apart on the histogram. (In my somewhat informal test, my 400% image was darker with the whole intact histogram shifted very slightly toward darker values, but that could be a result of my handholding the camera, etc.)
I expect that in jpg files the 200% and 400% apply a post-capture curve that raises the brightness at the low luminosity end of the scale — which would move the peaks closer together and to the right. (I took Fujifilm's claim for granted and did not test the jpg files.)
But it appears to me that the raw file is unaffected by these settings. Whatever is happening, I don't see any indication that the raw file DR is somehow expanded by 200% (one stop) or 400% (two stops).
"But it appears to me that the raw file is unaffected by these settings. Whatever is happening, I don't see any indication that the raw file DR is somehow expanded by 200% (one stop) or 400% (two stops).
gdanmitchell wrote:
But it appears to me that the raw file is unaffected by these settings. Whatever is happening, I don't see any indication that the raw file DR is somehow expanded by 200% (one stop) or 400% (two stops).
Compare specular reflection highlights with it on vs off. I can see visually and dropper the difference in the raw. And in the raw it’s difficult to “undo.”
Jack Flesher wrote:
Compare specular reflection highlights with it on vs off, even in the raw.
I did not do that. My intent was just to look for the effect of any curve being applied to the raw file, which would have shown up as displacement (or changes to the relatives positions) of the peaks for each of the patches on the grayscale chart. There wern’t any.
What I did see, though my test could not control for this, was the overall the 400% DR raw file was a little darker. There was no change in the spacing of the peaks. That could mean only of several things:
1. My exposure changed for reasons related to the casual nature of my test. (I might have positioned the center area a little different in each test.)
2. Fujifilm is lowering the overall exposure a bit at the 400% setting even with the raw file.
If the latter, it would obviously affect specular highlights.
I’ll try to retest under more controlled circumstances: camera on the tripod, entirely manual exposure with exactly the same settings for each frame, and so forth. (Those things would not change my observations about the historgram consistency.)
Dan, whether your "exposure changed for reasons related to the casual nature" of the test should be indicated by the shutter speed. If the shutter speed didn't changed (you haven't mentioned whether it did or didn't), then the exposure couldn't change (unless the camera has changed that secretly, while pretending it hadn't).
I don't know why you thought that the camera dynamic range could expand with DR200 or DR400. Considering that the dynamic range is defined as a ratio of the strongest signal to the noise (or to an arbitrarily set level corresponding to the weakest detectable signal, as in the photographic dynamic range from Bill Claff), then the only way to increase the dynamic range is by increasing the signal by increasing the exposure.
The exposure can be increased only by reducing the shutter speed and/or by opening the aperture. This is why the dynamic range inversely depends on the ISO, because the ISO changes with exposure like temperature changes with the amount of heat (only, ISO scales as an inverse of exposure, while temperature increases with heat).
ruthenium wrote:
Dan, whether your "exposure changed for reasons related to the casual nature" of the test should be indicated by the shutter speed. If the shutter speed didn't changed (you haven't mentioned whether it did or didn't), then the exposure couldn't change (unless the camera has changed that secretly, while pretending it hadn't).
I did not check because it was a quick, informal test… and because it would not affect what I was looking for, namely the pattern of peaks aligning with grayscale patches on my target.
But when I get a chance, as I mentioned earlier, I’ll test again with all setting determined manually and kept the same from shot to shot.
I don't know why you thought that the camera dynamic range could expand with DR200 or DR400
Maybe because of that naming that suggests that you should get “200%DR” or “400% DR?” ;-)
quote] Considering that the dynamic range is defined as a ratio of the strongest signal to the noise (or to an arbitrarily set level corresponding to the weakest detectable signal, as in the photographic dynamic range from Bill Claff), then the only way to increase the dynamic range is by increasing the signal by increasing the exposure. [/quote
That is not making sense to me. Increasing exposure does not increase the potential recorded dynamic range. It merely shifts the luminosity range of the image to a brighter or less bright level. The sensor itself can record whatever dynamic range it can record.
The exposure can be increased only by reducing the shutter speed and/or by opening the aperture. This is why the dynamic range inversely depends on the ISO, because the ISO changes with exposure like temperature changes with the amount of heat (only, ISO scales as an inverse of exposure, while temperature increases with heat).
Sorry, you are not making sense. The object of these settings is to capture a wider dynamic range without losing shadows or blowing our highlights while keeping the exposure the same. That’s literally what the 200%DR and 400%DR settings claim to do.
I believe that they do not affect the actual raw signal — other than possibly suppressing the exposure a bit at the 400% setting to protect highlights as Jack thinks. And I’m not certain about that.
Looking at the histogram, the graphical representation of the intensity at each step of the gray scale image, the distance between peaks and more or less the shapes of those peaks does not really change. If the sensor were somehow recording a greater dynamic range either the dark tones would have to be lightened (by far the most likely on digital cameras) or somehow the highest peak luminosities would ahve to be reduced. Either of those would necessarly compress the range of the grayscale chart image so that it takes up a smaller percentage of the now-expanded raw file’s dynamic range.
That did not happen.
On the other hand, I do believe (though I did not test it, instead taking Fujifilm at their word) that when using the 400% and 200% settings the camera’s software is applying some kind of curve to the in-camera output of the raw signal from the sensor and (most likely, by far) bumping up the luminosity at the dark end of the scale by one or two stops.
As has been mentioned several times in the thread, that could be useful to jpg-only shooters who are not all that interested in post-processing and who want to take greater advantage of the cameras ability to deal with larger dynamic ranges. On the other hand, those with more sophisticated needs for image development and processing are probably going to shoot raw and not see much if any advantage from using these settings instead of simply exposing to protect highlights and recovering shadows in post as needed.
Dan, "exposure" and "dynamic range" are well-defined parameters. I am using these as defined.
Regarding "Increasing exposure does not increase the potential recorded dynamic range", this is incorrect. In fact, there is a simple relationship between the two. An added stop of light is expected to increase the dynamic range by close to a stop. For example, the measured photographic dynamic range of the GFX100 II changes from 11.4 at ISO200 to 12.3 at ISO100. For another pair of values: PDR is 10.0 and 9.0 at ISO 1000 and 2000, respectively.
DR200 and DR400 don't imply an extended dynamic range. These settings are intended to prevent clipping of the highlights. This can never increase or expand the dynamic range.
ruthenium wrote:
Dan, "exposure" and "dynamic range" are well-defined parameters. I am using these as defined.
Regarding "Increasing exposure does not increase the potential recorded dynamic range", this is incorrect. In fact, there is a simple relationship between the two. An added stop of light is expected to increase the dynamic range approximately by close to a stop. For example, the measured photographic dynamic range of the GFX100S II changes from 11.4 at ISO200 to 12.3 at ISO100. This is an oversimplified interpretation, of course, because the shot noise also changes with increased exposure as a square root of the signal.
DR200 and DR400 don't imply an extended dynamic range. These settings are intended to prevent clipping of the highlights. This can never increase or expand the dynamic range. ...Show more →
If you increase exposure to the point highlights are burnt out any further exposure will reduce tge dynamic range recorded as shadows increase closer to the maximum that can be recorded.
gyoung143 wrote:
If you increase exposure to the point highlights are burnt out any further exposure will reduce tge dynamic range recorded as shadows increase closer to the maximum that can be recorded.
Gerry
Yes, but if you increase exposure to the point where highlights are just not clipping, you have gained DR compared to exposing with the highlights a bit lower (which the camera metering in many cases will do unless you force it not to).
For this very reason, lifting an image shot at ISO 100 +1 stop in PP, is effectively the same as having shot it at ISO 200. A lot of people don't know this, unfortunately.
No matter the settings, you want to expose the sensor to as much light as possible (without clipping) if you want to maximize DR.
