Old Image Set, Atypical Workflow

Here is an oldie but a goodie. I think I have presented this scene merged and toned in several different applications and workflows.

What I have noticed in the last couple of years is that raw and HDR are converging as the usable dynamic range of sensors and raw converters gets better and better. Similarly, with more programmers and enthusiasts getting into raw acquisition and processing, there are a lot of raw converters and utilities for processing and massaging linear, raw data.

Here is the "before" and "after" of the most recent treatment of this familiar scene:

So what is going on here? What kind of whacked out workflow produced this before and after image?

Unlike most HDR workflows that use raws and merge them to a single 32bit file, I used the Zero Noise approach. I used Zero Noise for Linux on my Mac and merged 3 images to produce a 16 bit linear TIFF that has no noise in the shadows, especially up in the dark brown roof structure. The key to the next half of this approach is to treat the raw merge so that the output has no color profile (in dcraw, this would be the option "-o 0"), has a UniWhiteBalance (all white balance multipliers are 1, in dcraw this would be the option "-r 1 1 1 1") and an output gamma of 1, or linear. Why is this important? Because this is essentially raw data - in other words, I have used Zero Noise to create a super-raw file made up of the best exposed pixels from the three original raw images. Raw data with no noise.

Now, to make this into an image, you need a tool that can handle such data and preserve color fidelity, etc. as well as the tonal range so that you can redistribute the pixels into a pleasing final image. This particular tool for the job could be ACR, but ACR can't handle this kind of file effectively. Enter "Color Perfect" by CF Systems - this software is a plug-in for Photoshop that was originally designed to preserve color integrity in the conversion of scanned film negatives into positive images (the "ColorNeg" module). Out of that original concept grew the ColorPos module, which has recently included a major update called "Perfect Raw" - you feed the plug-in a linear, UniWB TIFF made from a raw image and the plug-in performs manipulation using the authors' unique approach to preserving color integrity. For more information, see:

http://www.c-f-systems.com/Plug-ins.html

and the author's treatise on his conceptual framework here:

http://www.c-f-systems.com …ps.html#complet​eintegrity

I've been using this as a viable alternative to straight raw conversion and I have found that color and tonal adjustments are as good, if not better, than most converters and the workflow is much more intuitive in this paradigm. Anyway, I digress.

Because I was able to make a linear, UniWB TIFF in Zero Noise, that is essentially noise free in the shadows, I can feed this to Perfect Raw and use these controls to tone the data. The result is then tweaked in PS to boost local contrast and make final adjustments.

This is sort of a hybrid raw-HDR approach, but one I think is worth pursuing if you want to get natural looking results in a somewhat more familiar workflow than the sliders in HDR applications. This is not something I just cooked up, but rather the result of trying and refining different techniques for about 10 years. If this is way over your head, don't sweat it, I'm still grasping this workflow too!

kirk

Here is the final with a little bit brighter mid and shadow tones.

... and darker tones.


kirk

And here is a 100% crop out of the file processed by Perfect Raw - I have applied a super crazy steep curve to the image to boost the shadow tones (note the super dark areas have funky color clipping due to this crazy adjustment). What you should note is the lack of noise. POW!

kirk