Parts of the SGPro 4.5.X release set will work to make aspects of AutoFocus better. While we certainly have tools and equipment we use for testing, we stand no chance at matching the breadth of equipment used here in the community. As such, we are trying to build up our test database such that it is a bit more diverse.
For a long time now, SGPro has had an AutoFocus option that will instruct focus runs to capture data we can use for auto focus replay, called “FocusPacks”. The original intent was to help us help you tune your AF settings, but they are also wonderful for testing that AF works properly (after we make changes) with more than our limited set of equipment.
So, a request for focus packs… you can enable focus packs as follows:
In AF settings:
Focus packs will be saved to the folder of your choosing. Not looking for any specific settings, just literally whatever you use for focus.
We would very much appreciate any packs you can share. Just zip them up and post them on a file share of your choice and we will add them to our test library.
I looked into one of the focus packs and see a jpg image for the plot curve and a FITS file for the image, but the header info lacks equipment info that I normally see in a light frame FITS file. Is that expected?
The header info in the focus pack FITS shows:
SIMPLE = T / file does conform to FITS standard
BITPIX = 16 / number of bits per data pixel
NAXIS = 2 / number of data axes
NAXIS1 = 6252 / length of data axis 1
NAXIS2 = 4176 / length of data axis 2
BZERO = 32768 / offset and data range to that of unsigned short
BSCALE = 1 / default scaling factor
CRPIX1 = 3126 / reference spectrum pixel coordinate for axis 1
CRPIX2 = 2088 / reference spectrum pixel coordinate for axis 2
CTYPE1 = ‘RA—TAN’ / standard system and projection
CTYPE2 = ‘DEC–TAN’ / standard system and projection
END
I’d have to check, but I’d say it’s normal as we don’t ever really save AF frames and so the amount of metadata we pass to the save function is probably very limited.
Thanks! Yes please. It is directly related to the current dev effort (once and for all automating the “tuning” of autofocus. We can use this data to piecemeal together “tuning runs” for testing beyond what we have on hand.
Just 2 or 3 successful runs is perfect really. Preferably on different types of targets… Galaxy, Nebulous and globs (if you have any).
I wanted to give you ones with different types of targets, but I had about 600 directories and the file names are not sorted in any order, and there’s no quick way to determine what target a focus run was for. The .fit files are too dark to see an object.
Going forward, it would be helpful if the target’s name was in the directory name, e.g., AFPack_M44_2025-04-08-18-21-22_AFID-1_23 for “M44” taken April 8, 2025 at 18:21:22. And unless the AFID is needed, I’d drop it from the name.
On a related topic, I have been recording AF packs for a backspacing analysis software, and I noticed that the FITS header is extremely basic (this has been pointed out in the past threads). Would you mind adding some more useful information ? The minimum would be to add FOCPOS obviously, but other basic parameters such as pixelscale, binning, temperature (FOCTEMP), filter, exposure, etc. would be useful as well in the AF packs.
If you are still interested, I can share my AF packs, but as they were for measuring my optical system, they were taken in very boring areas in the sky! The AF curves are pretty much textbook curves without any particular difficulty.
In addition to the existing headers present now (SIMPLE, BITPIX, NAXIS, NAXIS1, NAXIS2, BZERO, CRPIX1, CRPIX2, CTYPE1, CTYPE2). I have added the following to saved AF Pack FITS files: DATE-LOC, DATE-OBS, CREATOR, XBINNING, YBINNING, CCDXBIN, CCDYBIN, EXPOSURE, FOCPOS, FOCTEMP, FILTER.
In addition, I have added non-standard header HFR just for convenience.
These new items will be present in the next 4.5 beta on Thu / Fri this week.
Feel free to request additional data… it is likely they will be simplistic to add.
Full disclosure… these “spontaneous additions” don’t get a lot of testing other than making sure they don’t bust everything next to them. That said, lemme know if it’s a problem in some way and we’ll fix it.
