Hi!

I've been working a lot these past few months to learn how to use autofocus properly, and to optimize the focus function. Now it seems to work nicely (sensitive to collimation on my C8), but still there is one issue that I am not comfortable with, and that is that Ekos focuses after each and every capture. Given that one focusing round might take 2-7 minutes, this means that lose a lot of imaging time. So I wonder how to optimize this. I realize that I need to refocus when temperature changes, but between each and every capture......?

My thinking is that it seems impossible for my subs not to exceed the HFR value automatically given when focusing (even with the marginal that it includes). I normally focus with exposure times between 3 and 5 secs, and my subs are typically taken with 300 secs. Even if everything is perfect (guiding, wind, temperature), wouldn't the HFR exceed the focus value just because it is so much longer exposure?

So: what is the best way to optmize the focusing, so it actually does not focus after every shot, but only when needed? Even focusing after every second shot should give me substantially more imaging time (and with the clouds over here, imaging time is very precsious...).

Any thoughts on this is much appreciated.

Magnus

You could slightly increase the HFR threshold value so it doesn't re-focus when it is slightly above the threshold minimum value.

Hi!

Yes, thanks. I guess I had slightly misunderstood how the HFR throeshold worked - my understanding was that it was measured on the sub taken, but of course that is not how it works - the focus module takes a new image and measures that, and if that exceeds the threshold value, it starts focusing.

So: yes, I can manually increase the HFR threshold - but not when running the scheduler...

So I guess what happens here, is that there is random variations in the HFR value, that makes the "test" capture by the focus module typically exceed the threshold. THis could pehaps be due to bad seeing here, in my suburban area, or something else. So: is there a way to reduce this randomness? I'm thinking in terms of taking longer exposures (for focusing) to average out the randomness, or very short, to avoid seeing problems. What is the typical way the rest you treat this? What exposure values do you use, and why? And Jasem - would it be possible to change the margin added to the HFR, allowing the user to choose a percentage value?

Magnus

I am getting that issue too, maybe capturing a series of focus test imagens, instead of one single image, 5 images for ex. and making a median count on them so It could disperse the error? Just an idea... Increasing the value of HFR is not always the best.

Hi!

Interesting that you recognize these issues! May I ask, what exposure time do you typically use for focusing?

There used to be a possibility to average a number of exposures, but it seems to have been removed, or..?

Another aspect of all this is, of course, to speed up focusing. What takes most time for me, is the download of images. And when it takes some 15-20 subs for a focusing session, that becomes a significant amout of time. Perhaps one could do focusing at the indi server (R-Pi) rather than on client....?

Magnus

Depending on the night, I am using 1 to 1.5 second images @ ISO3200 with my CANON 550D.
The download is a real problem, even a cropped image takes 5-10 seconds to download.
That takes a lot of time focusing.