On linux I use entangle, ZereneStacker and a Stackshot rail.
I make a sequence of raw images.
Then I use the Darktable command line interface to create a matching *.tif file for each raw.
mkdir tifs
for file in *nef
do
darktable-cli $file tifs/$file.tif
done
Then I cd into the tifs directory and use ZereneStacker to create a stacked tif.
Then use darktable again (on the stacked tif) and export that to a final tif.
Then i use Gimp on the final tif for minor touchups and then export a jpeg or png. In the first darktable-cli step darktable does some low-level default editing, whether you like it or not.
I can modify the above sequence of events so during the first darktable-cli step I can add manual editing of my own. When I do this I choose one of the orginal raw images, from the middle of the stack, and open that with darktable
darktable capture_0018.nef
And then edit manually, usually increasing contrast with several "modules' and perhaps an exposure adjustment. This makes a darktable log file named capture_0018.nef.xmp, that logs the editing steps. I can now apply all those manual edits from the one file to all.
for file in *nef
do
darktable-cli $file capture_0018.nef.xmp tifs/$file.tif
done
So now my question. Is there any reason to do so?
When I create a series of tif files reflecting manual edits to increase contrast as a first step, does that increased contrast help the subsequent ZereneStacker steps?
I occurred to me it might help, because stacking algorithms (if I have this right) rely on pixel-to-pixel contrast comparisons to choose which pixels to preserve.
I'm trying to devise an experiment to test this but it isn't easy becauase....at some point I have to use darktable to edit a stacked tif. And depending on which stacked tif I start with I have to make manual editing choices that differ from the other scenario. So even when I usually do end up with one image looking a lot better than the ohter I'm never sure if it was because of manual editing early on or near the end that made the difference.
Does that make sense?
Work flow question
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
- Posts: 54
- Joined: Tue Oct 16, 2012 11:02 am
- Location: Bozeman, MT
- Contact:
- rjlittlefield
- Site Admin
- Posts: 23564
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Re: Work flow question
Yes, stacking algorithms do rely on pixel-to-pixel contrast comparisons.pittendrigh wrote:occurred to me it might help, because stacking algorithms (if I have this right) rely on pixel-to-pixel contrast comparisons to choose which pixels to preserve.
However, simply increasing the contrast will not help because the software will end up making the same decisions, just using bigger numbers.
To improve the stacking results you need to improve the signal-to-noise ratio, for example by noise reduction.
This is a general principle. There are some exceptions, for example when curves adjustment is used to increase the brightness and contrast of very dark parts of the subject, then those parts may become less vulnerable to contamination by artifacts from nearby lighter parts.
--Rik
-
- Posts: 54
- Joined: Tue Oct 16, 2012 11:02 am
- Location: Bozeman, MT
- Contact: