Creating High Dynamic Range Images

[Guillermo Luijk]

These images are fantastic. None of the strange "backlit" character of usual HDR images.

If you could make a tutorial about this, many forum members would be grateful.
[a href=\"index.php?act=findpost&pid=134076\"][{POST_SNAPBACK}][/a]

There is no secret about those images, they are simple so noise-free in the shadows (thanks to a severe overexposition of +6EV) that they can be processed with a curve to strongly lift the shadows without any posterization or visible noise.
The only "trick" was to leave the window (in the second image) out of the curve mask layer so that it remains unaltered and it does not blow.

The blending process was done through simple pixel selection with a routine of mine, but most blending software (no tone mapping) will provide similar results.

[Chris_T]

There seems to be varied opionions on DR in terms of f-stops in this thread. What are yours: for slides, DSLRs and B/W negatives?

Slides capture the least dynamic range of any film type, and most good DSLRs can do better. B&W negative film can capture the greatest subject DR.
[a href=\"index.php?act=findpost&pid=134154\"][{POST_SNAPBACK}][/a]

[Ray]

There seems to be varied opionions on DR in terms of f-stops in this thread. What are yours: for slides, DSLRs and B/W negatives?
[a href=\"index.php?act=findpost&pid=134310\"][{POST_SNAPBACK}][/a]

I've always understood the DR of slides to be 4-6 stops, negative color film 7-9 stops and B&W film 9-11 stops. But it's always problematic in arriving at a consensus on such figures because there does not seem to be a uniform standard defining a cut-off point with regard to image degradation, which inevitably takes place as one approaches the extremes of both the shadows and the highlights.

[bjanes]

I've always understood the DR of slides to be 4-6 stops, negative color film 7-9 stops and B&W film 9-11 stops. But it's always problematic in arriving at a consensus on such figures because there does not seem to be a uniform standard defining a cut-off point with regard to image degradation, which inevitably takes place as one approaches the extremes of both the shadows and the highlights.
[{POST_SNAPBACK}][/a]

[a href=\"http://clarkvision.com/imagedetail/dynamicrange2/]Roger Clark[/url] has done a nice analysis of digital (Canon 1D MII--one of the best 35 mm style digitals) as compared to slide and negative film and came up with the following: "Kodak Gold 200, in this test, showed 7 stops of information, Fujichrome Velvia 5 stops, and the Canon 1D Mark II, over 10 stops of information!". This is consistent with Ray's observations.

Although slide film has a lower dynamic range than print film, the contrast is much higher than that of negative film and the D-max of slide film is considerably higher than that of negative film. That characteristic means that even low range scanners can capture the dynamic range of negative film, whereas it takes a high end scanner to extract all the information from slides. Because of the low contrast of negative film, it can be difficult to obtain good gradation in the scan.

Bill

[Guillermo Luijk]

the Canon 1D Mark II, over 10 stops of information!". This is consistent with Ray's observations.

Bill, don't you think considering a 10 f-stop DR in a 12-bit RAW based camera is a bit too optimistic? if we assume linear behaviour of the sensor into this usable DR, the entire 10th lowest f-stop would be codified in just 4 different tone non-interpolated levels. I think that's too poor even if noise is not present in such deep shadows, isn't it?

    0EV: 2048 levels, 2048..4095
   -1EV: 1024 levels, 1024..2047
   -2EV: 512 levels, 512..1023
   -3EV: 256 levels, 256..511
   -4EV: 128 levels, 128..255
   -5EV: 64 levels, 64..127
   -6EV: 32 levels, 32..63
   -7EV: 16 levels, 16..31
   -8EV: 8 levels, 8..15
   -9EV: 4 levels, 4..7
   -10EV: 2 levels, 2..3
   -11EV: 1 level, 1

The only thing I can think for reaching such a figure (10 f-stops DR) is to consider a non-linear low end of the response curve of the sensor as a usable range that could include more than 4 effective levels to codify degrees of lightness (of course being these levels "stolen" from upper f-stops, -8EV and up). However the dots in the log-log graphic don't seem to show any non-linearity in the input-output response curve:




And in any case he is using a 16-bit reference, with which I don't agree as those new levels appearing in the 16-bit range and filling it, are purely interpolated. Real captured levels are in a 12-bit range, and only 4 different of them represent the whole 10th f-stop.

[bjanes]

Bill, don't you think considering a 10 f-stop DR in a 12-bit RAW based camera is a bit too optimistic? if we assume linear behaviour of the sensor into this usable DR, the entire 10th lowest f-stop would be codified in just 4 different tone non-interpolated levels. I think that's too poor even if noise is not present in such deep shadows, isn't it?

    0EV: 2048 levels, 2048..4095
   -1EV: 1024 levels, 1024..2047
   -2EV: 512 levels, 512..1023
   -3EV: 256 levels, 256..511
   -4EV: 128 levels, 128..255
   -5EV: 64 levels, 64..127
   -6EV: 32 levels, 32..63
   -7EV: 16 levels, 16..31
   -8EV: 8 levels, 8..15
   -9EV: 4 levels, 4..7
   -10EV: 2 levels, 2..3
   -11EV: 1 level, 1

The only thing I can think for reaching such a figure (10 f-stops DR) is to consider a non-linear low end of the response curve of the sensor as a usable range that could include more than 4 effective levels to codify degrees of lightness (of course being these levels "stolen" from upper f-stops, -8EV and up).

And in any case he is using a 16-bit reference, with which I don't agree as those new levels appearing in the 16-bit range and filling it, are purely interpolated. Real captured levels are in a 12-bit range, and only 4 different of them represent the whole 10th f-stop.
[a href=\"index.php?act=findpost&pid=134381\"][{POST_SNAPBACK}][/a]

Guillermo,

As Ray pointed out, calculation of dynamic range of a photograph is not exact, and the floor may be affected by posterization as well as noise. How many levels are needed in the darkest zone is somewhat arbitrary, but 4 levels is probably not sufficient. I think Roger was using a noise floor.

Norman Koren's Imatest allows calculation of dynamic range associated with varying levels of quality. Here is one such test for my Nikon D200 (which has more noise than the Canon 1D M2), and a low quality image with a dynamic range of 10.2 f/stops is obtained. Again, this is using noise as the floor. Norman agrees that the darkest zone should have 8 levels, which would limit DR to 9 stops with a 12 bit image.



I would think that the new Canon 1D MIII with its lower noise and a 14 bit AD converter should be able to get 10 stops of DR easily.

Bill

[One Frame at a Time]

You guys are making me dizzy!  Not that its all that unpleasant, sometimes I drink wine or beer to the same effect  .  

Seriously, very interesting discussion.  I really like the results.  I am blown away by some of those images.  I now realize that it is part of many images I have seen here and elsewhere.   Is there a simple tutorial as to how to use the HDR in CS3??  I went to the tutorial link by Mark Galer.  Its very impressive but not all that easy to execute as part of work flow.  Too many alt-option-del key presses.  If I could remember all that I'd have been an astronaut!

[nemophoto]

I just played with the Photomatix software. Pretty amazing program and far outstrips Photoshop HDR. The only way I could initially play with the program was converting a RAW image three times. Not the best but workable. The program very quickly assembled the images and tone mapping did the rest. I had a very credible image without a lot of work. It will be interesting to see what it does when I do a series of exposures intentionally.

Photoshops's HDR conversion took about three times longer and the result was totally unusable. I'm grateful Tyler brought this slick little program to my attention.