The theoretical maximum compression for APOGEE's noise levels and bit depth is approximately 2.6. In practice, APZIP’s compression of ~2.0 is considered as good as can be expected for the specific astrophysical data it handles.
: The resulting multi-extension FITS files are compressed using the FPACK utility and the lossless Rice compression algorithm . 3. Usage and Accessibility : The APZIP code is part of the
: Detector reads are converted into difference images (the first read plus difference images) to reduce dynamic range. : The APZIP code is part of the
: The corresponding program APUNZIP.PRO is used to restore the raw data cubes for reduction. : The APZIP code is part of the
: The APZIP code is part of the apogeereduce software product. While primarily an internal tool for the Sloan Digital Sky Survey (SDSS) , versions of the pipeline are hosted in public repositories like the SDSS-III SVN repository. 4. Comparative Efficiency
: An average difference image is computed and subtracted from all difference images, resulting in "residual" images that are easier to compress.
The APOGEE instrument produces roughly and up to 100 GB per night . APZIP was created to speed up data transfer from the Apache Point Observatory to the Science Archive Server (SAS) and to minimize overall disk usage. It compresses raw data cubes by taking advantage of the fact that successive reads of the detector arrays are highly similar. 2. Technical Compression Algorithm