chiron
Offline
Upload & Sell: On
|
 Re: Utility of dynamic range beyond a certain point | |
snapsy wrote:
chiron wrote:
snapsy wrote:
Jack Flesher wrote:
snapsy wrote:
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.
What's your definition of "properly-exposed"?
Fair question. Non-specular highlights not blown. And I consider the Sun or any other direct light source in the frame essentially a specular highlight.
Here's an image from my S1R. The highlights in the window are metered just at raw clipping (ISO 100 f/8 1/20). I have a second exposure that's 1EV brighter (ISO 100 f/8 1/10). These are without noise reduction.
Here are four images in the demonstration:
Image #1 - Full scene (downsampled) unprocessed, using 1/20 exposure
Image #2 - Full scene (downsampled), processed, using 1/20 exposure
Image #3 - 100% crop of processed shadows, using 1/20 exposure
Image #4 - 100% crop of processed shadows using 1/10 exposure, simulating a camera with +1EV greater DR
---------------------------------------------
Steve Spencer wrote:
snapsy wrote:
Jack Flesher wrote:
snapsy wrote:
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.
What's your definition of "properly-exposed"?
Fair question. Non-specular highlights not blown. And I consider the Sun or any other direct light source in the frame essentially a specular highlight.
Here's an image from my S1R. The highlights in the window are metered just at raw clipping (ISO 100 f/8 1/20). I have a second exposure that's 1EV brighter (ISO 100 f/8 1/10). These are without noise reduction.
Here are four images in the demonstration:
Image #1 - Full scene (downsampled) unprocessed, using 1/20 exposure
Image #2 - Full scene (downsampled), processed, using 1/20 exposure
Image #3 - 100% crop of processed shadows, using 1/20 exposure
Image #4 - 100% crop of processed shadows using 1/10 exposure, simulating a camera with +1EV greater DR
Thanks for the example. It is helpful, but as an example it demonstrates, to me anyway, just how rarely I would need that extra stop of DR. This is a pretty extreme case and although I like the lower noise at 1/10, I am pretty sure I would be happy blowing out the window a bit with that slower shutter speed. I find the perfectly exposed window with the shadows raised that much to look unnatural like a framed picture rather than a window. Personally, I would rather see the window with at least some blown highlights. So, if I was taking this shot I would would have bracketed exposure and I could have gotten the low shadow noise of your 1/10 of second exposure and I could have gotten a window I think I would prefer and done all of that with one stop less DR and that is even in this extreme example.
I am not saying that an extra stop of DR wouldn't be useful in at least some situations, but this examples suggests to me that it would almost never matter for my shooting.
---------------------------------------------
Jack Flesher wrote:
Steve Spencer wrote:
snapsy wrote:
Jack Flesher wrote:
snapsy wrote:
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.
What's your definition of "properly-exposed"?
Fair question. Non-specular highlights not blown. And I consider the Sun or any other direct light source in the frame essentially a specular highlight.
Here's an image from my S1R. The highlights in the window are metered just at raw clipping (ISO 100 f/8 1/20). I have a second exposure that's 1EV brighter (ISO 100 f/8 1/10). These are without noise reduction.
Here are four images in the demonstration:
Image #1 - Full scene (downsampled) unprocessed, using 1/20 exposure
Image #2 - Full scene (downsampled), processed, using 1/20 exposure
Image #3 - 100% crop of processed shadows, using 1/20 exposure
Image #4 - 100% crop of processed shadows using 1/10 exposure, simulating a camera with +1EV greater DR
Thanks for the example. It is helpful, but as an example it demonstrates, to me anyway, just how rarely I would need that extra stop of DR. This is a pretty extreme case and although I like the lower noise at 1/10, I am pretty sure I would be happy blowing out the window a bit with that slower shutter speed. I find the perfectly exposed window with the shadows raised that much to look unnatural like a framed picture rather than a window. Personally, I would rather see the window with at least some blown highlights. So, if I was taking this shot I would would have bracketed exposure and I could have gotten the low shadow noise of your 1/10 of second exposure and I could have gotten a window I think I would prefer and done all of that with one stop less DR and that is even in this extreme example.
I am not saying that an extra stop of DR wouldn't be useful in at least some situations, but this examples suggests to me that it would almost never matter for my shooting.
I agree with everything Steve said. This particular extreme example sort of proves my point that an extra stop of DR doesn't really make much difference in a real world image. I too would choose bump the window a stop for believability and I'd be more than happy with the final result in that for the most part, totally irrelevant shadow.
Nice examples and demonstration and interesting comments.
The problem I have with the "Full Scene Processed" image is that there is not enough difference between the apparent light outside the room and the apparent light inside the room--there should be more of a difference between the apparent light levels.
I am not sure that blown highlights are really evidence of what is natural--they be more what we are used to seeing in photographs. They are rarely what we actually see with our eyes. But our eyes do sensitively perceive differences in light levels.
I do think the Full Scene image needs to suggest a greater difference between the light inside and outside the room, but without going all the way to blown highlights.
Can you expand on what you mean by "not enough difference between the apparent light outside the room and the apparent light inside the room"?
When I look at the image, it does not appear that it is brighter outside than inside. If I am in a room on a sunny day and I look toward a window, my eyes register brightness greater than what is in the room. In the Full Scene Processed image, it appears as though the exposure value for the interior of the room and for outside the room would be the same, which of course is not the case. (I think this is what makes the window look a bit like a painting hung on the wall.) My preferred way of handling this would be to have the view through the window look maybe a stop or two overexposed so that it reads as brighter to a viewer. Or have the room be a stop or two underexposed, if the view through the window and the relative dimness of the room is of primary interest.
|
