×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Perfecting the Scheduler

The following user(s) said Thank You: Eric
5 years 10 months ago #25815

Please Log in or Create an account to join the conversation.

  • Posts: 1185
  • Thank you received: 370
Hi Eric,
may I add one topic to the whish list? Currently, re-focussing is determined on the photography tab, where the single sequence is running. In my setup, a sequence is very simple: LLRGB - and then repeated by the scheduler. In the schedule I defined refocusing after 60min.
In this setup refocusing never happens, since one sequence takes 2*8+3*4min = 28min.

Would it not make more sense having these definitions on the scheduler side rather on the sequence tab?
Cheers
Wolfgang
5 years 9 months ago #26459

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Perfecting the Scheduler

Hello Wolfgang,
I don't think the HFR and periodicity of re-focusing should move to the scheduler. These are concepts related to the capture tab. However, the Scheduler is optimizing imaging time when a job repeats, so that no alignment, no focusing and no guiding calibration will happen. I did a fix recently, forcing realignment when the job is not using guiding, so that the target doesn't slowly move out of the frame. As I understand your wish is similar, I could add an option to force focusing unconditionally at the beginning of a repeated job. That would make three options to control focusing in that situation: forced, HFR-bound and periodically.

-Eric
5 years 9 months ago #26495

Please Log in or Create an account to join the conversation.

  • Posts: 1185
  • Thank you received: 370
Hi Eric,
I think it depends how the capture part is used. If it is used defining all captures for a single target, it makes sense having these features there. If it is used like I do where the capture part contains only one sequence and the scheduler defines how often it is repeated, at least the time-related features like re-focusing every x minutes is less helpful since one sequence is shorter than a typical re-focusing interval.

But maybe it makes sense to step back and think about the characteristics of the capture tab and the sequence tab. What do you think of the following:
  • The capture tab contains all aspects to ensure, that a certain amount of frames are taken with a defined quality. That means, HFR-based re-focusing and guiding-deviations belong to this tab.
  • The scheduler tab contains overall control and all time-related aspects. That means, time interval based re-focusing should belong to this tab, since such an interval is strictly time related and independent of a selected target. If we think this very consequently, I would propose shift meridian flipping here as well.

What do you think about this?

-Wolfgang
p.s.: forcing focusing before each sequence would not be very helpful in my case. One typical LLRGB sequence takes 2x8+3x4=28 min, that means every half an hour a re-focusing will occur. This is too often, so it would not be helpful. As a workaround, I will enlarge a single sequence so that it is slightly longer than the refocusing-interval.
Last edit: 5 years 9 months ago by Wolfgang Reissenberger.
5 years 9 months ago #26508

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Perfecting the Scheduler

Wolfgang, I tried to implement a proof-of-concept on forced autofocusing when a job repeat. But I realize this might not be your use case.

Trying to categorize your use of the scheduler, I think you have only one job that your repeat, perhaps indefinitely. You use the scheduler for the automated target tracking, alignment, focus and capture.

In that situation, the best autofocusing solution is probably the one with the HFR boundary. But I agree it requires trusting the ability of the capture tab to consider refocus is required, hopefully not for all captures.

If you need a fixed periodicity for autofocusing, then you're left with what you did as a workaround, that is, having the procedure trigger after a certain time. As you observed, this requires the sequence job to be long enough.

That is basically why I proposed to implement the forced autofocusing option at the beginning of a batch in the scheduler. This is more difficult than I thought, because the state machine of the scheduler is scattered in multiple functions, so I'll need to rework eventually.

In the current state of code, it's probably difficult to extract the autofocusing period from the sequence and move it to the schedule. However I can check if it is possible to reuse the interval from one sequence to the next inside a repeated job, so that the time origin is the beginning of the schedule job and not the beginning of the first sequence job of this schedule job.

-Eric
5 years 9 months ago #26564

Please Log in or Create an account to join the conversation.

  • Posts: 1185
  • Thank you received: 370
Hi Eric,
reusing the interval would be great, this would resolve my problem.

-Wolfgang
5 years 9 months ago #26632

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301

Replied by Eric on topic Re:Perfecting the Scheduler

I proposed phabricator.kde.org/D13388 for this.

-Eric
5 years 9 months ago #26701

Please Log in or Create an account to join the conversation.

Time to create page: 0.303 seconds