no-compresso!

October 3, 2007 12:28 AM

Astronomy image question: Where to find the best image of the hubble deep field? or the deep field south, or the ultra-deep field? I've got jpegs, but imagine I was a real astronomer looking at these things, where do i find the best, uncompressed versions?
posted by Gankmore to Science & Nature (6 answers total)

The hubble site itself has fairly uncompressed images.
posted by so_ at 2:37 AM on October 3, 2007

For instance...

Starting here:
http://hubble.nasa.gov/multimedia/types.php

Stsci:
http://www.stsci.edu/resources/

HDF:
http://www.stsci.edu/ftp/science/hdf/hdf.html

Pretty Pictures:
http://hubblesite.org/newscenter/archive/releases/1996/01

This particular image:
http://hubblesite.org/newscenter/archive/releases/1996/01/image/a/

Highest Quality Download Options:
http://hubblesite.org/newscenter/archive/releases/1996/01/image/a/warn/

You can get a 1.25Mb Jpeg or a 8Mb Tiff...I think that should be big enough for most uses, Tiffs are uncompressed.
posted by so_ at 2:46 AM on October 3, 2007

Beyond that, investigate the:

"Data analysis

Hubble data can be analysed using many different packages, but STScI develops the custom-made STSDAS (Space Telescope Science Data Analysis System) software. The software contains all the programs needed to run pipeline reduction on raw data files, as well as many other astronomical image processing tools, tailored to the requirements of Hubble data. The software runs as a module of IRAF, a popular astronomical data reduction program."

http://en.wikipedia.org/wiki/Hubble_Space_Telescope
posted by so_ at 2:52 AM on October 3, 2007

Yeah, if you were a real astronomer you wouldn't be using tiffs - they're just too short of the metadata and flexibility needed really. You'd want to get hold of something like a FITS file.

I've never had cause to use HST data myself, but it looks like this would be the place to go. That page lead me to here for the UDF. I followed that to some fits files, and I started downloading one but it was 170-odd megabytes.

If you want to actually look at the contents of a fits file I'd recommend something like DS9.

The images for the consumption of the general public go through a lot of image processing of all kinds to get them that pretty - just as a warning.
posted by edd at 4:24 AM on October 3, 2007 [1 favorite]

The upper limit on the resolution you can get for the north and south deep fields is determined by the CCD in WFPC2. The detector is only 1600x1600 pixels maximum (ignoring the stairstep). So the high-resolution tiffs that have already been linked to are really the best you can do. The Ultra-deep field, however, was done with ACS, which has a resolution of 4096x4096. If you want a lot of pixels, that's the way to go.

STSci makes the FITS files for the observations available, but that will give you a fairly ugly picture. You'll get one image per filter, and you'll have to combine them properly to get a nice color image. Then there's all sorts of post processing, like cosmic ray rejection, flat fielding, etc (I'm not sure what the details are for WFPC2). Overall, this is a lot of work, and takes a lot of skill and familiarity with the instrument. If you want to have a look at the raw images, by all means get the data from hubble and look at it in ds9. That is an interesting thing to do. But you won't get a better-looking image than the jpegs or tiffs.
posted by kiltedtaco at 2:05 PM on October 3, 2007

Most of the flat fielding and other standard astronomical corrections are done in the frames that are available (judging by the UDF frame I looked at earlier today). Balancing the curves and the like for human consumption hasn't been.

Quite agree though. The tiffs are the best thing for the human eye.
posted by edd at 4:51 PM on October 3, 2007