I recently had some issues solving with ASTAP, and discovered an issue in recent releases.
If you check the Use Scale box in Align options, ASTAP will be called with a "-fov VALUE" flag, but that value will be populated by a value that's not consistent with what ASTAP expects. ASTAP's doc says the value should be the vertical field of view of the input image, but Ekos sends it something like the max of the width and height's field of views. This inconsistency caused ASTAP to fail solving several images I tried. When I disabled "Use Scale", and thus -fov wasn't part of the ASTAP command, ASTAP worked fine.
My recommendation is: if you are using ASTAP, you should disable "Use Scale" until this issue is resolved. Doing that doesn't seem to slow down ASTAP, and it does make it solve better, at least in the examples I tried.
AP1100 mount, GSO RC10 mount w/RSF focus, ZWO ASI1600 imager & guider, Astronomik Filters, ONAG.
KStars/Ekos/Indi on NUC10 with Ubuntu 22.04
Projects: Greedy Scheduler, Terrain, Polar Align, Analyze, Linear Focuser, SEP MultiStar & GPG Guide, FITS autostretch.
I did have an email conversation with Rob--he's aware of the issue.
I was just letting folks know.
It seems like EKos is sending the max dimension to StellarSolver, so it's somewhere between Ekos and StellarSolver and ASTAP.
Interestingly, ASTAP seemed 5-10% quicker without scale! Was I just lucky? I have disabled scale.
@han.k Also, FWIW, it might be interesting for ASTAP (or all the StellarSolver solvers) to have a "if the search fails, try again without scale/position with timeout" sort of backup possibility. This could be expanded to other parameters as well, if there are more. Perhaps enabled with some flag.
@han.k Finally, as I'm on the topic, does ASTAP have a star-detection interface? StellarSolver has the possibility of using ASTAP to solve location, but perhaps it could also be an alternative star detector as well?
Hy
AP1100 mount, GSO RC10 mount w/RSF focus, ZWO ASI1600 imager & guider, Astronomik Filters, ONAG.
KStars/Ekos/Indi on NUC10 with Ubuntu 22.04
Projects: Greedy Scheduler, Terrain, Polar Align, Analyze, Linear Focuser, SEP MultiStar & GPG Guide, FITS autostretch.
Why there would be a difference in execution speed, I can only guess. Maybe FOV setting is lower. It will still solve if the FOV is 20 to 30 % too low but less reliable and reduces the speed. As you already mentioned, ASTAP requires the image height in degrees. If it is unknown you could execute it with "-fov 0" which is equivalent to auto and it will try all FOV's between 10 degrees and 0.4 degree but I'm not sure if that is possible with EKOS. The program SharpCap uses this option in case solving fails. But I think this option is better reserved for installation (once) or images with no FOV specified.
ASTAP can be used as star extractor using the the option -extract , see www.hnsky.org/astap.htm#astap_command_line
You only have to specify the minimum star SNR For solving I would specify something like -extract 10. The output is a csv text file.
Note ASTAP is now also available as a command line executable version for headless setups.