p.7 #1 · Utility of dynamic range beyond a certain point
snapsy wrote:
All of which are processing options available to the photographer. For example, here's the window brighter and room darker:
With the brighter window… perhaps you would also want to maintain or even increase the brightness of the light coming through that window that strikes surfaces in the room.
Your brighter window casts less light into the room and your darker window seems to cast more! ;-)
(Tricky stuff, these post-processing changes, no?)
p.7 #2 · Utility of dynamic range beyond a certain point
gdanmitchell wrote:
With the brighter window… perhaps you would also want to maintain or even increase the brightness of the light coming through that window that strikes surfaces in the room.
Your brighter window casts less light into the room and your darker window seems to cast more! ;-)
(Tricky stuff, these post-processing changes, no?)
Hehe, @chiron wanted a brighter window and a darker room and that's what he got!
p.7 #3 · Utility of dynamic range beyond a certain point
gdanmitchell wrote:
With the brighter window… perhaps you would also want to maintain or even increase the brightness of the light coming through that window that strikes surfaces in the room.
Your brighter window casts less light into the room and your darker window seems to cast more! ;-)
(Tricky stuff, these post-processing changes, no?)
---------------------------------------------
snapsy wrote:
Hehe, @chiron@ wanted a brighter window and a darker room and that's what he got!
The room in the brighter window/darker room version might benefit from more contrast, which might address Dan's issue, which I don't disagree with.
p.7 #4 · Utility of dynamic range beyond a certain point
gdanmitchell wrote:
Mine is “a file with the maximum usable image data for expected use in post.” (Ansel didn’t say it that way, but that was his approach.)
To generalize, with digital that usually means protecting the highlights and potentially recovering shadows, at least with scenes that have large dynamic ranges.
I’ve shared a few examples of what you can do at my website. Here’s one done with a camera that is over a decade old. (The article is over 10 years old, and I would process this photograph a bit differently today.)
In other news, I was at the National Gallery in London today, and I thought back to an earlier discussion here about dynamic range in paintings. I photographed a few examples that I’ll share later — after I get home and have access to a real computer to prepare the files for sharing. The short story is that they nicely illustrate how landscape (and seascape) painters compress a very wide dynamic range in the original scene into a much narrower range in the painting....Show more →
Dan, that’s a great tutorial, I enjoyed seeing your process.
p.7 #5 · Utility of dynamic range beyond a certain point
Ross Martin wrote:
Dan, that’s a great tutorial, I enjoyed seeing your process.
Thanks, Ross.
As you perhaps noticed, that tutorial was written over a decade ago… and some of the adjustments I used are no longer the best options… or even available in current versions of the software! And as I reread that article of mine and looked at the photograph I realized that I would interpret it a bit differently today.
One of the biggest changes is that now I can do even more in ACR before bringing the smart object into a layer in PS. In fact, sometimes I barely do any editing in Photoshop, especially with the recent addition and improvement of the powerful masking features in ACR.
At some point I probably need to either rewrite that article or else write a new one with a more recent photograph example.
Of course, the main underlying point remains: I don’t generally expose for the best SOOC image but for the best/most image data that I can bring innto post to get the final result I’m after.
p.7 #6 · Utility of dynamic range beyond a certain point
Daran wrote:
I don't think that is true. You generally need to be careful when comparing DR numbers, as there are multiple incompatible definitions. E.g. photographic DR is relative to a small image sample area spot, which has no universally accepted size. Or the digital engineering DR, which has a universal definition but tends to be much better for low resolution sensors and isn't applicable to film at all.
The Hurter-Driffield curve (film) and SNR (digital) methodologies are not identical, so yes I agree that the numerical difference between 15 stop film and the 12.55 stop A7RVI should not be interpreted as exactly 2.45 stops.
However, both metrics attempt to characterize usable exposure range. A film stock measured at roughly 15 stops and a sensor measured at 12.55 stops provide evidence that the film likely has greater exposure latitude. The methodological differences justify caution about the magnitude of the advantage, not dismissal of the comparison altogether.
When making prints in the darkroom that you can really see the film's exposure range by looking at the projected image. The projected image looks far better than any print you can make because the dynamic range of the paper is limited.
Until recently we have been limited to seeing 10 stops of dynamic range on our monitors, because an IPS display with 1000:1 contrast can only display 10 stops.
Now with OLED monitors with 1,500,000:1 contrast ratio, the screens can now display 20 stops of information.
Therefore instead of editing our images for "zone 3 to zone X", we can in theory do "zone 3 to zone 20" if the OLED monitor is the display medium.
However at 12.55 stops, even the A7RVI is yet not up to the task of matching what the displays can do.
p.7 #7 · Utility of dynamic range beyond a certain point
Jack Flesher wrote:
I’d appreciate seeing a real world example of where properly-exposed images from a camera with 13 stops of DR makes a notably better image than one with only 12. Seriously, when cameras passed 10 stops of DR was when I started editing black back into most of images. Today for me, that’s a de-rigueur adjustment.
Are you excluding landscape shots into the sun?
For me, I expect to take advantage in things like night cityscapes (after blue hour), moon shots with landscape with timing maybe 15 min after sunset or later, abandoned buildings (or I guess the more common task for photographers would be real estate shots), cave shots, etc. These latter use cases are very niche, so maybe only apply to me.
p.7 #8 · Utility of dynamic range beyond a certain point
Not shure if this was mentioned yet, but if you really need to over/underexpose a pic and pull back all the DR possible - shoot un compressed! Many modern cameras (I know Sony) will allow for +5 stops of post processing with minimal noise (relatively)
p.7 #9 · Utility of dynamic range beyond a certain point
aCuria wrote:
The Hurter-Driffield curve (film) and SNR (digital) methodologies are not identical, so yes I agree that the numerical difference between 15 stop film and the 12.55 stop A7RVI should not be interpreted as exactly 2.45 stops.
I stand corrected. I knew BW film was better than color film, but for some reason I thought it was similar to what BW sensors have over color sensors.
What a rabbit hole trying to research this. Loads of tradeoffs and few hard measurements to discuss. It would seem that the better BW sensors out there can roughly match analog BW, though I could not find anything definitive on this. There is a lot of discussion as to how film gives you a pleasing non-linear response, but then you can always use transfer curves in digital to get the same and digital certainly opens more ways to deal with the result in post.
However, both metrics attempt to characterize usable exposure range. A film stock measured at roughly 15 stops and a sensor measured at 12.55 stops provide evidence that the film likely has greater exposure latitude. The methodological differences justify caution about the magnitude of the advantage, not dismissal of the comparison altogether.
Uh, that isn't even an argument? Unless you have a way to quantify an upper limit on the difference, it could be nearly anything.
Until recently we have been limited to seeing 10 stops of dynamic range on our monitors, because an IPS display with 1000:1 contrast can only display 10 stops.
You can't see DR beyond what the medium allows, but you can map things non-linearly. Most want DR in digital in order to lift shadows in post while maintaining highlights with details. That works well with less DR in the resulting medium.
p.7 #10 · Utility of dynamic range beyond a certain point
aCuria wrote:
The Hurter-Driffield curve (film) and SNR (digital) methodologies are not identical, so yes I agree that the numerical difference between 15 stop film and the 12.55 stop A7RVI should not be interpreted as exactly 2.45 stops.
However, both metrics attempt to characterize usable exposure range. A film stock measured at roughly 15 stops and a sensor measured at 12.55 stops provide evidence that the film likely has greater exposure latitude. The methodological differences justify caution about the magnitude of the advantage, not dismissal of the comparison altogether.
When making prints in the darkroom that you can really see the film's exposure range by looking at the projected image. The projected image looks far better than any print you can make because the dynamic range of the paper is limited.
Until recently we have been limited to seeing 10 stops of dynamic range on our monitors, because an IPS display with 1000:1 contrast can only display 10 stops.
Now with OLED monitors with 1,500,000:1 contrast ratio, the screens can now display 20 stops of information.
Therefore instead of editing our images for "zone 3 to zone X", we can in theory do "zone 3 to zone 20" if the OLED monitor is the display medium.
However at 12.55 stops, even the A7RVI is yet not up to the task of matching what the displays can do....Show more →
This is why AI will never suceed. people have no imagination to ask the correct questions. If you placed a grey letter 5 mm in the center of a white monitor with 20 stops of light difference your eyes would not be able to see what the letter is due to glare from your eyes. a camera could also not take an image of the letter due to veiling glare and pixel bleeding from the sensor.
p.7 #11 · Utility of dynamic range beyond a certain point
aCuria wrote:
The Hurter-Driffield curve (film) and SNR (digital) methodologies are not identical, so yes I agree that the numerical difference between 15 stop film and the 12.55 stop A7RVI should not be interpreted as exactly 2.45 stops.
However, both metrics attempt to characterize usable exposure range. A film stock measured at roughly 15 stops and a sensor measured at 12.55 stops provide evidence that the film likely has greater exposure latitude. The methodological differences justify caution about the magnitude of the advantage, not dismissal of the comparison altogether.
When making prints in the darkroom that you can really see the film's exposure range by looking at the projected image. The projected image looks far better than any print you can make because the dynamic range of the paper is limited.
Until recently we have been limited to seeing 10 stops of dynamic range on our monitors, because an IPS display with 1000:1 contrast can only display 10 stops.
Now with OLED monitors with 1,500,000:1 contrast ratio, the screens can now display 20 stops of information.
Therefore instead of editing our images for "zone 3 to zone X", we can in theory do "zone 3 to zone 20" if the OLED monitor is the display medium.
However at 12.55 stops, even the A7RVI is yet not up to the task of matching what the displays can do....Show more →
I think that it is safe to say that some form of monitor is usually the final display of a majority of images most of us capture. As OLED and Micro-RGB monitor technology moves into the mainstream, why are we continuing to shun HDR image workflows?
p.7 #12 · Utility of dynamic range beyond a certain point
sbay wrote:
Are you excluding landscape shots into the sun?
For me, I expect to take advantage in things like night cityscapes (after blue hour), moon shots with landscape with timing maybe 15 min after sunset or later, abandoned buildings (or I guess the more common task for photographers would be real estate shots), cave shots, etc. These latter use cases are very niche, so maybe only apply to me.
Yes. While I have occasionally done "Sun in my landscape, I let the Sun blow and go for the Sunstar because that's the only way it looks right to me, YMMV.
I've shot a lot of dimmer indoors with bright outdoor windows. Bodie interiors are a good example for me if you've ever shot there. Here I had no problems with Nikon Z7ii, Z9, Leica SL3 or Fuji XH2 raw files keeping enough detail in a single raw to be able to tune the highlight slider down to recover outside details in the windows, and raise the shadow slider enough to maintain great shadow detail in the interiors without generating excessive noise. This was using C1 as my raw converter and I've since switched to Sony and abandoned C1, so I am only assuming ACR will allow me to do similar. If it doesn't, I will be back to C1 in a nano...
p.7 #13 · Utility of dynamic range beyond a certain point
Jack Flesher wrote:
Yes. While I have occasionally done "Sun in my landscape, I let the Sun blow and go for the Sunstar because that's the only way it looks right to me, YMMV.
I've shot a lot of dimmer indoors with bright outdoor windows. Bodie interiors are a good example for me if you've ever shot there. Here I had no problems with Nikon Z7ii, Z9, Leica SL3 or Fuji XH2 raw files keeping enough detail in a single raw to be able to tune the highlight slider down to recover outside details in the windows, and raise the shadow slider enough to maintain great shadow detail in the interiors without generating excessive noise. This was using C1 as my raw converter and I've since switched to Sony and abandoned C1, so I am only assuming ACR will allow me to do similar. If it doesn't, I will be back to C1 in a nano...
I actually looked at some of those shots and it seems the highlights are quite often just a mass of pure white…no details. I will take any extra dynamic range I can get just for those special cases. YMMV.
p.7 #14 · Utility of dynamic range beyond a certain point
chez wrote:
I actually looked at some of those shots and it seems the highlights are quite often just a mass of pure white…no details. I will take any extra dynamic range I can get just for those special cases. YMMV.
Hmmm. Perhaps you could link to them so we can discuss directly what you see as blown?
p.7 #15 · Utility of dynamic range beyond a certain point
A74me wrote:
This is why AI will never suceed. people have no imagination to ask the correct questions. If you placed a grey letter 5 mm in the center of a white monitor with 20 stops of light difference your eyes would not be able to see what the letter is due to glare from your eyes. a camera could also not take an image of the letter due to veiling glare and pixel bleeding from the sensor.
I don’t understand what you are trying to get at. Did you see my example images in my previous post? I’m not talking about a grey letter in a field of white. I can’t imagine such a subject being interesting photographically.
Image B: Single image "highlight recovery" test. Foreground exposure was correct in camera, and Highlights set to -100 in lightroom (max slider) to "recover highlights" as much as possible. Highlights have been clipped and cannot be fully recovered. Foreground looks correct.
Image C: Single image "shadow recovery" test. Scene exposed to as to preserve highlights. Overall exposure raised in POST so the sky looks correct. Highlights set to -100 to undo raised exposure in sun area. Shadows +100 to brighten foreground.
Differentiation in light / dark areas of foreground cannot be recovered from this raw file. They have been crushed to a similar brightness.
The difference in exposure between images B and C is 2 2/3 stops.
Single image "highlight recovery" test. Foreground exposure was correct in camera, and Highlights set to -100 in lightroom (max slider) to "recover highlights" as much as possible. Highlights have been clipped and cannot be fully recovered. Foreground looks correct.
Single image "shadow recovery" test. Scene exposed to as to preserve highlights. Overall exposure raised in POST so the sky looks correct. Highlights set to -100 to undo raised exposure in sun area. Shadows +100 to brighten foreground.
Differentiation in light / dark areas of foreground cannot be recovered from this raw file. They have been crushed to a similar brightness.
...Show more →
That's not a problem of the shadow areas, but only your error in editing. Lowering higlights and raising shadows for the whole image is NOT the correct method. The flatness of the rocks is only because of this flawed method.
Correct method: Have masks for sky, rocks, water; edit exposure, white point, black point, contrast etc. for each mask individually.
p.7 #18 · Utility of dynamic range beyond a certain point
hasenbein wrote:
That's not a problem of the shadow areas, but only your error in editing. Lowering higlights and raising shadows for the whole image is NOT the correct method. The flatness of the rocks is only because of this flawed method.
Correct method: Have masks for sky, rocks, water; edit exposure, white point, black point, contrast etc. for each mask individually.
Thats what I tried at first, but the water does not look right like this. The exposure of the water must be changed as a gradient for it to look right, and the rocks still look kinda flat. Effort level to edit with the masks is easily 100x that of doing the HDR image stack, which took like 10 seconds of messing with the sliders.
p.7 #19 · Utility of dynamic range beyond a certain point
aCuria wrote:
I don’t understand what you are trying to get at. Did you see my example images in my previous post? I’m not talking about a grey letter in a field of white. I can’t imagine such a subject being interesting photographically.
the point is we can see a scene in real life because our eyes move around the scene and adjusts to see the larger DR. but a monitor is small compared to the actual scene and our eyes dont have the view seperation to actually see 20 stops of DR all at once. so i dont see any point in moniotors with an extreme DR.
p.7 #20 · Utility of dynamic range beyond a certain point
A74me wrote:
the point is we can see a scene in real life because our eyes move around the scene and adjusts to see the larger DR. but a monitor is small compared to the actual scene and our eyes dont have the view seperation to actually see 20 stops of DR all at once. so i dont see any point in moniotors with an extreme DR.
You do have a point, but do note that a standard monitor has less DR than what the eye can see all at once.
I think its still possible to appreciate a 20 stop HDR image, even if some parts are "clipped" by the eye...
For example, I do appreciate a 42mm bin even as the image is brighter than my 21mm optic, even though my pupil may already be the limiting factor.
