Hello,
I actually asked this before and got some help.
The advise was to not auto de-bayer so that Ekos would use the gray-level image instead of the red channel for focusing.
It helped a few months ago but my recent sessions again were not good - sometimes the Ha is sharp but Oiii is bloated, sometimes it's the other way around.

I use same focuser settings for quite a while. Linear 1-pass. SEP profile - default.

The problem seems to be worse for shorter focal lengths (or rather the F-number?).
I.e. 280mm is worse than 950mm. Or f/4 vs f/6?

Is there something in the new/improved Focus module which can address this?
Like HFR vs FWHM vs Fourier? SEP?

Thanks!

I experimented with different settings in Focus module - nothing really made any significant difference.
The only thing - I found Fourier measure algorithm to be unreliable (it doesn't always produce a well defined maximum).

So, right now if I do not use auto de-bayer I have well focused Blue/Green and bloated Red.
If I do use auto de-bayer then the situation is reversed - bloated Blue/Green and focused Red.

In the first case the focus point is about 10140 steps.
In the second case the focus point is about 10240 steps.

So the best focus should be somewhere in between, I guess. Like 10190.

Maybe there should be a possibility to add an offset?

?

Hi maxthebuilder,

I can't think of much off the top of my head that would help you in this situation. The issue is that the different bandpasses focus at different points on your system (as you know). So the solution would be:
1. Somehow make Ha and Oiii focus at the same point so the problem goes away. I suppose this would require some sort of optical device in your imaging train, but I don't really know how this could be done.
2. Work out some sort to compromise solution on the focus position that plays off the focus points so that the picture you are intending is in the best focus possible for the picture as a whole, although the individual Ha and Oiii will each be away from their optimum focus positions.

From memory the star detection process looks at the first channel. For the ASI2600 MC Pro looks like the bayer pattern is RGGB so that would tie in with Ha being in focus and Oiii not in focus.

If you don't debayer presumably you get a greyscale composite of (R + G + G + which would put focus towards Oiii and leave Ha more out of focus. Which again, seems to match what you observe.

So I suppose there are a few things that could be done... give the user the control to select which channel to use for Focus (R, G or or use some sort of synthetic channel and give the user the ability to construct the synthetic channel, for example by selecting weightings to use on R, G and B.

Another option might be to allow an offset from focus to be specified (as you mention above). So run Autofocus and then apply an offset that is user specified. The question would be... how stable is that offset, i.e. does it change with temperature or altitude for example.

Would you always want to focus at the same point (average of the channels or user supplied offset) or would you want to change it based on the target? So for example, for detailed Ha targets would you want to have the Ha more in focus and for more detailed Oiii targets would you want to Oiii's focus point to be more in focus?

Hi John,
Thanks much for answering!
Looks like this matter is a bit complicated...

The other night I did try the offset. I focused with/without de-bayer and used the focus point in between.
This worked - both Ha and Oiii were the same but both slightly out of focus, of course.

The good thing for me that this is really an issue only with my FRA400 with a reducer (f/3.9) - but that is so wide-field (280mm) that the difference in focus doesn't really kill the end result.
(In my last sessions I had red halos around stars).

At native f/6 FRA400 doesn't show much difference between Ha and Oiii.
Same with RC6 (either at 950mm or 1380).

But, yes, it would be nice to be able to at least select which channel to focus on - R, G or B.

Thanks!