I am working on a digital imaging quality thesis up here in Stockholm and I am currently looking into different raw-conversions. I have up till now used and compared dcraw and Adobe's raw converter.
The results show clearly how big difference different raw conversions influence the end result when it comes to noise as well as sharpness. But it is also very clear that the settings of each converter is very important. Now I realized hat I am not entirely sure about my usage of dcraw and was hoping you could help me out.
The theses starts out with testing some tiff-outputs from a few cameras and is supposed to be followed up with some raw-image comparisons. For this comparison I want to have a raw-conversion that has no noise suppression, no sharpening, only a clean de-mosaic conversion to 8bit with sRGB color space. The conversion should further use the cameras white balance as well as no change of exposure or highlights.
The comparison settings I've used so far has been:
- ARC with default conversion settings, 8bit Tiff and sRGB (only for comparison)
- ARC with default camera settings but all image enhancement by ACR set to zero and using a linear curve, 8bit Tiff and sRGB
- dcraw -v -w -o 1 -q 3 - W -T
Is this the right way to only get raw-conversion without changing any white balance, exposure or highlights and with 8bit sRGB Tiff output?
Mon, 14 Apr 2008 16:21:55 +0000
Your command looks fine to me. But I'm not dcraw's algorithm expert. You might consider writing a message to Dave Coffin, who wrote the different algorithms in dcraw.
One thing you should note is that the dcraw algorithm operates completely differently between cameras. I've noticed that on some cameras, it does exceptionally well. On other cameras, it does a mediocre job. So, in your thesis, you should provide full disclosure to what digital camera(s) you used for the tests.
Mon, 14 Apr 2008 23:03:17 +0000
thanks for your reply. I did write to Dave Coffin and got confirmation on my command line for dcraw.
Interesting to hear about dcraws different results with different cameras. In my thesis, I will of course clearly show records of which cameras, settings, and converters that was used. Good point.
The thesis main subject is to go through the ISO standards and see how they work in practice. Me personally would like to work towards having some possible standard stable enough to compare cameras and possibly converters. The tools are basically available although some newer ideas of assessing imaging quality might need some more work still. There is, to me, a rather strange void and lack of objective test information available for the interested amateur and professional photographer. There are some attempts in press and web but still very very few use the better tools that are specified in the ISO standards.
The main difficulty right now to make a basic setup for this is to choose one or two common paths or workflows for the camera images. This is because although the tools are good and stable they do depend highly on many things such as of course the use of in camera jpeg in comparison to raw-converted images. What converter and what settings also have a big influence. One test path ought to be in-camera jpegs directly tested, this to show the available quality in a basic setting without any need or interest in post-processing. The other ought to be a raw converted workflow, possible using Adobe Raw Converter since it is possibly the most commonly used by more aspiring amateur photographers as well as pro-photographers. What do you think, sounds reasonable? Any input is appreciated :)
Again, thanks for your input!
Tue, 15 Apr 2008 19:39:41 +0000
Did your message disappear? Read the Forums FAQ.
TrackBack only accepted from WebSite-X Suite web sites. Do not submit TrackBacks from other sites.
Send Ping | TrackBack URL | Spam Control
Title: dcraw by example
Weblog: Camera Hacker
Tracked: Mon, 14 Apr 2008 16:23:53 +0000
Spam Control | * indicates required field
Messages, files, and images copyright by respective owners.
229 Users Online