Post Reply 
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
how true color with stiff?
08-07-2007, 17:47
Post: #1
how true color with stiff?
Hi,
I have a question:
if I have three scientific fits images taken with three different filters, is it possible to adjust with stiff all cut levels in automatic way in order to take into account of the real color of objects.
Now I must interactively run more and more time stiff working with ds9 by hand and changing sky/min & max levels to produce a pretty "by eye" resulting image.
Is there a standard way to work?
i.e.: can I run sextractor to find background min/max for each band and run stiff with those cuts???
thank you a lot.
Find all posts by this user
Quote this message in a reply
08-07-2007, 18:02
Post: #2
RE: how true color with stiff?
Actually, unless you image has specific features (like more than 1/3rd of all pixels with exactly the same value, etc.), you can generally get pretty nice results simply with a line like stiff r.fits g.fits b.fits because STIFF applies cuts based on quantiles. But this assumes that the image background in the FITS image is already close to 0. If not, you may either use SWarp or SExtractor to subtract it, or add the -SKY_LEVEL option to the STIFF command line to specify the sky background value(s).
- Emmanuel.
Visit this user's website Find all posts by this user
Quote this message in a reply
08-07-2007, 18:14
Post: #3
RE: how true color with stiff?
Thank you for the fast response, i just know the stiff-quantile algorithm and it is a good first approximation, but I'm searching something more "natural".
i.e. I found in Lupton et al, "Preparing Red-Green-Blue (RGB) Images from CCD Data" (arXiv:astro-ph/0312483v1), a good exercise.
He said that there is the possibility to use the same colors to code color-magnitude diagrams, providing a natural `index' to our images. He introduced the use of an asinh stretch, which shows faint objects while simultaneously preserves the structure of brighter objects in the field, such as the spiral arms of large galaxies.
How this can be done using stiff?
Thank you and have a nice day.
Find all posts by this user
Quote this message in a reply
08-07-2007, 18:45 (This post was last modified: 08-07-2007 21:01 by Emmanuel Bertin.)
Post: #4
RE: how true color with stiff?
The asinh luminance compression proposed in Lupton et al. paper is all but "natural" in my opinion (Robert and I have had this discussion but unfortunately I have not been able to convince him so far Smile ). The main reason why most astronomical images look ugly on screen or paper is simply because they are not corrected for the display gamma. The dynamic compression offered by the asinh transfer function certainly helps boosting faint surface brightnesses, but this cannot be a substitute for proper gamma compensation (which Lupton et al. do not apply, leading to colour shifts in dark image parts). The results you get with STIFF (putting aside colour-space conversions, but this is another story) are as close to natural as they can be. If you wish to push faint surface brightnesses in STIFF, you may apply some GAMMA_FAC luminance gamma factor (see doc). An example of result obtained with STIFF lies right in the TERAPIX home-page (the galaxy).
- Emmanuel.
Visit this user's website Find all posts by this user
Quote this message in a reply
08-08-2007, 10:57
Post: #5
RE: how true color with stiff?
thank you very much you will not convince Lupton but sure you will convince me... Smile
I will study with more details the stiff documentation...
Find all posts by this user
Quote this message in a reply
08-08-2007, 11:28
Post: #6
RE: how true color with stiff?
One possible improvement that I am thinking of to bring a more "photographic" look to STIFF images would be to add some soft saturation feature (the current behaviour rather gives a "video" look).
- Emmanuel.
Visit this user's website Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)