Please add a Min and Max star size HFD to Asiair for guiding.

Kevin_A

One of the biggest issues I see in Asiair is the lack of multistar guide star HFD pixel amount min and max control and ultimately refining the programs multistar selection. Increasing gain to much higher levels works somewhat but sometimes Asiair still randomly picks guide stars of various saturation intensities anyways, some still bright ones and some dim. Often I will see a flat top on the main star selected HFD image telling me I am guiding on an oversaturated star. If I lower my gain then I loose dim stars. I can ether reduce my exposure or drop the camera gain which requires me to redo darks for a lower gain or speed up my exposures, neither are practical. For multistar guiding to work well all the stars should be of similar intensities to which they are still not. If Asiair had a Min and Max guide star HFD pixels then you could set it a say Min 1.5px so it does not guide on bad pixels and maybe I use Max 20px so it guides better than on an overly saturated star, and all multistars are relatively equal in intensity, which is the intent using multiple stars. Then you will get better guiding due to equal star sizes and best size stars for different setups…with less hassle. Please consider adding this standard multistar feature. I just want to guide on similarly weighted multiple stars already!

w7ay

Kevin_A

Limiting HFD won't help in general, since HFD does not change once a star becomes unsaturated.

The problem with ASIAR multi-star guiding is a combination of these two factors:

1) ASIAIR uses a SNR-weighted centroid average (told to me by the ZWO developer while it was being implemented a couple of years ago). I.e., if you have a star with extremely good SNR relative to the rest, it will carry all the weight, and end up representing the case of no better than one-star guiding.

2) stellar magnitude distribution. Take a look at the HYG catalog, which contains pretty much every star that a guide scope can hope to see.

https://www.astronexus.com/hyg

If you plot the distribution of stars, you will find an exponential type tail. I,e., there are many more Mag 0 stars than there are Mag 1 stars, there are many more Mag 1 stars than there are Mag 2 stars, etc. Because of this effect, even if you get rid of saturated stars (all unsaturated stars should show the same HFD with a good guide scope, independent of brightness below saturation). So, because ASIAIR is using SNR weighted centroid averaging, the brighter stars will still hog the centroid, and you end up with the equivalent of only 2 or 3 stars even if you see 12 stars on the ASIAIR display. This is equivalent to changing from 1 second to 2 to 3 second exposures to average out "seeing" for single star guiding.

You cannot avoid this unless you change the centroid averaging algorithm from a SNR weighted one.

Prior to v1.5 or so, ASIAIR would religiously ignore saturated stars, so you can basically increase the guide camera gain to the point where you can make ASIAIR ignore magnitude 2 or brighter stars.

Each magnitude is a factor of 2.5 in ADU, or about 12 dB between magnitude 0 and magnitude 2 -- so that gives you an idea of how much extra gain needs to be applied to ignore dimmer stars. That had worked very well (when using ASIAIR, I had increased the gain by 10 dB after ASIAIR first sees 12 stars), at the cost of slightly noisier guide images (but, still better than allowing bright stars to hog the centroid averaging).

Unfortunately, the recent ASIAIR (at least since v2 beta) no longer ignores saturated stars. It will even choose Vega. The only way now, as long as you want to keep using ASIAIR, is to move the guide scope so that the FOV is pointed at a region of the sky where there are fewer bright stars. We used to use x-y stages to find a reasonable guide stars in the good old days, now, we need to use the x-y stages to ignore bright stars!

(Another usability bug in v2 is they are now sampling the mount position much slower now (perhaps once every 10 to 15 seconds), so when you slew (or meridian flip the mount), you will not see the mount coordinates change. There are multitudes of lesser bugs. )

I had not fed back beta anomalies in v2 beta because ZWO now requires you to enter an email address, and hell if I am going to give ZWO yet another email address for their marketing department to spam.

Both the bugs above got to their general v2.0 release.

Chen