Sequencer Autofocus

[gcs]

Manually using the multi-star FWHM allows several variables like apply display stretch, sensitivity, noise reduction, etc.
When setting up an autofocus sequence in the editor, there are no such options, just the step information.
Does the sequencer not allow using or changing those, or are they somehow imported from the focus assistant?

[admin]

Hi,

those settings are re-used from the last ones set in the manual mode when you use autofocus in the sequencer, so as long as you have previously set them up in a working fashion and done a manaul scan/focus to test, you should be fine using the sequencer autofocus.

cheers,

Robin

[gcs]

Thanks, that's great.
But I'm confused about using display stretch.
One stretch might be good for the star count, but a higher stretch might be used of image monitoring.
When autofocus runs will it change the current stretch?

[admin]

Hi,

it will use the current one, but that tends not to matter as much as you might think. The display stretch is only used for the first phase of star detection (finding the bright blobs that might be stars). Any reasonable stretch will make the faint stars much more visible and allow them to be picked up when they might not be detected in the unstretched image. Beyond that initial detection, all the measurement of FWHM etc goes on on the unstretched image.

cheers,

Robin

[gcs]

Thanks Robin,
I have been using the autofocus in sequence for some time and it has worked well.
Recently I went from a 0.8 reducer/flattener to a 1.0 flattener only.
I noticed my previous settings didn't focus very consistently now.
The sample points are significantly scattered around the fitted curve, and there seemed to be best focus variation from run to run.
I think it's from the longer focal length and higher F#.
So I've been trying to improve the response.
I've never had to average frames in autofocus before, but now I'll try 2-3 frames to see how that works.
I'll continue using the display stretch even if it results in hundreds of stars.
Thanks again for your help!

[admin]

Hi,

I think the first thing that I would consider trying in that situation is to widen the range of scan for the autofocus - that will put the stars further out of focus at the two ends of the scan, deepening the 'U' shape of the graph and counteracting the increased noise that you are seeing by increasing the signal.

Actually, this is a change that you can pretty much expect from going to a slightly longer focal length - the longer focal length means a narrower light cone, meaning that if you are still focusing to the same extend either side in the scan, the star sizes will be less enlarged at the limits of the scan than they were before...

cheers,

Robin

[gcs]

You're absolutely right Robin!
And I've already done that, but will perhaps increase it some more.
Thanks Again!