Hi all,
I think there's no greater learning experience than seeing some of the professionals work with data I have obtained. I tend to blame my bad tracking/alignment or other setup problems for the lesser quality of my astro photos, I wonder how much of that can be fixed in post? So rather than having a tutorial demonstrating working with the top of the crop acquired image, it would be a little bit of a challenge to improve on what I think for many people without permanent telescope setups is reality: Not always perfectly round stars, drift, only 98% in focus etc...
I also would be very interested to see what some of you come up with creatively and aesthetically, given this stack of 150 x 30 second exposures of a wide field around the Trifid and Lagoon nebulas.
The original FITS file created by DSS can be downloaded here (~130MB):
http://www.inside.net/trila.zipThis is what I have done with it:
http://www.inside.net/trila_2.jpgI processed it as follows:
- AutomaticBackgroundExtractor (division, not subtraction, since I have no flats)
- Resampled to 90% width, to reduce some of the drift (this only works because by pure chance the drift is roughly X oriented). I also tried to do that with a motion blur PSF L-R deconvolution, which, scientifically speaking, would be the proper way to try and fix this. But that didn't do a very good job at all.
- HistogramTransformation, midtones at 0.003
- Then two HDRWaveletTransforms: The first one to accentuate larger features, default settings with a 5x5 B3 spline scaling function, the second one with a 7x7 Peaked (1). Interestingly, the results look much better when those transforms are done AFTER the HistogramScaling, which is somewhat contrary to the way the HDR Transforms should be done, on a linear dataset namely.
If you can document what you've done that would make a very valuable tutorial for us simple DSLR people.
Cheers
- Balt
