StarTools

The recommendation when stacking in DSS is to use the autosaved FITS file for use in Star Tools. I've had a quick play with DSS to see what differences there are between the autosave TIFF and FITS files. In this example, the TIFF was 114MB and the FITS 186MB. I had wondered if they were of different bit-depth, but opening them in my trial copy of PixInsight, I see that both are 32-bit floating point 3 channel files. These are considerably different file sizes, so what is different about FITS files, and what makes them preferable for processing in Star Tools (if you can explain in layman's terms ?

Thanks - Ian

Admiral wrote: These are considerably different file sizes, so what is different about FITS files ...
Thanks - Ian

I believe the difference is FITS file metadata (headers) are in a human readable (ASCII) form while TIFF file metadata is not. The ASCII format is not as compact. You can load a FITS file into an editor (e.g., Notepad on Microsoft Windows) and actually read the header information. I'm speculating that your FITS file may also contain more metadata information than your TIFF file.

If unfamiliar with the FITS format, you may wish to consult something like http://www.fileformat.info/format/fits/egff.htm.

Admiral wrote: ... and what makes them preferable for processing in Star Tools (if you can explain in layman's terms ?

I defer to someone more knowledgeable about StarTools.

That's what I am led to believe, having looked at Wiki. But the difference in file size is 72MB, which seems a devil of a big header! Which makes me think that there is something more to this, especially as it is the preferred input file to Star Tools. Perhaps it's just the space allocated to various bits of the file is different, but I can't say I'm conversant with the intricacies of file structures.

Ian

If they have the same sample size but a big size difference, then the TIFF might have (lossless) compression enabled. One way to check it is to zip both and see if they are both compressible (meaning they contain uncompressed data) or only the FITS is.

Thanks all. Certainly the file size seems to imply that, but I had assumed that the DSS autosaved TIFF file wouldn't be compressed, though why I should think that I don't know.

Ian

Some time ago I did DSS of assorted Globulars and processed them in STools We found (Ivo & me) that the 32 bit int Fits showed more dynamic range and definition than the tiff format for these objects.

M3 http://astrob.in/98042/0/