AF on HFR Trigger is not working as expected

[Ryan Haveson]

First sub after previous AF Run
[2021-02-21T03:41:27.8320] [INFO] [Message] Average HFR: 2.6253145047253, HFR σ: 0.308432044631832, Detected Stars 248

AF Triggered but numbers look wrong
[2021-02-21T04:31:48.7352] [DEBUG] [MemberName] ShouldTrigger
[2021-02-21T04:31:48.7352] [DEBUG] [FileName] E:\Projects\nina\NINA\Sequencer\Trigger\Autofocus\AutofocusAfterHFRIncreaseTrigger.cs
[2021-02-21T04:31:48.7352] [DEBUG] [Message] Autofocus condition extrapolated original HFR: 0.033335189213354 extrapolated current HFR: 0.141422034264246

[2021-02-21T04:31:48.7362] [DEBUG] [Message] Autofocus after HFR change has been triggered, as current HFR trend is 76.43% higher compared to threshold of 15%

‌

For this series:

‌

[Ryan Haveson]

This seems off:

Extrapolated goes way off in this time frame.

[2021-02-21T04:21:33.0733] [DEBUG] [Message] Autofocus condition extrapolated original HFR: 2.59080252467999 extrapolated current HFR: 2.59002847123123

[2021-02-21T04:21:33.5331] [INFO] [Message] Average HFR: 2.52305131335049, HFR σ: 0.368369958741734, Detected Stars 291

[2021-02-21T04:21:33.6607] [INFO] [Message] Saving image at C:\NINA\LIGHT\M 51 - Whirlpool Galaxy\2021-02-20\2021-02-21_04-20-31_M 51 - Whirlpool Galaxy_LIGHT_L_-10.00C_121g_60.00s_0241.xisf

[2021-02-21T04:22:34.6651] [DEBUG] [FileName] E:\Projects\nina\NINA\Sequencer\Trigger\Autofocus\AutofocusAfterHFRIncreaseTrigger.cs
[2021-02-21T04:22:34.6651] [DEBUG] [Message] Autofocus condition extrapolated original HFR: 4.87924985856138 extrapolated current HFR: 4.77997269219212

[Ryan Haveson]

The AF before the weird items above was quite a while beforehand

[2021-02-21T03:41:27.8320] [INFO] [Message] Average HFR: 2.6253145047253, HFR σ: 0.308432044631832, Detected Stars 248

[2021-02-21T03:41:27.9776] [INFO] [Message] Saving image at C:\NINA\LIGHT\M 51 - Whirlpool Galaxy\2021-02-20\2021-02-21_03-40-25_M 51 - Whirlpool Galaxy_LIGHT_L_-10.00C_121g_60.00s_0203.xisf

[Ryan Haveson]

I looked at all the HFR values between the events above and all were between 2.43 and 2.78…

[Ryan Haveson]

Went through and manually fed in the subs (from #250 - #260) one by one using the camera driver and a sequence with a 30 sec delay to give time to load the next image. The issue did not repro. Fun… Will have to look closer at the code to see where this could go wrong.

[Stefan B]

Did you capture some snapshot images in the imaging tab in between, in this session? It could be possible that these were incorrectly counted for the trigger as points, as they are also part of the history now and aren’t yet filtered out and could explain the strange calculation.

[Ryan Haveson]

No, this was the middle of the night. I had not touched it since 22:00 or so

‌

[Ryan Haveson]

And this was 40 or so subs after the previous AF so well beyond the sampleSize boundary

[Stefan B]

Yea i think we have isolated the issue and it will be fixed in the next build. Additionally the trigger will then also only consider subs for the current filter when a filter wheel is used

[Stefan B]

Fix is available in Nightly 48.

[Stefan B]

Removing version: 1.11 Nightly (automated comment)