Unanswered questions: AF methods "weighted average" / "best fit"


#1

Ken, Jared,

I asked these questions before (threads: “Old and New Auto Focus Method” and "Shape of a “V-curve”), but did not get an answer yet. I think they can be answered only from the developers:

  1. In SGP’s logfiles after a AF run there are comments like these:

[11.10.17 22:10:00.104][DEBUG] [Camera Thread] New auto focus method calculated focus at: 24907
[11.10.17 22:10:00.104][DEBUG] [Camera Thread] Old auto focus method calculated focus at: 24903
[11.10.17 22:10:00.104][DEBUG] [Camera Thread] Using old auto focus method for focuser position
[11.10.17 22:10:00.104][DEBUG] [Camera Thread] New focus position is at 24903 (@16,90C). Moving focuser…

… and after taking the validation frame:


[11.10.17 22:10:22.058][DEBUG] [Camera Thread] Auto focus is complete (weighted avg method: 24903; HFR 1,28)…

I was not able to find one case where SGP would use the new auto focus method (best fit). In the auto focus plot, the straight lines of the best fit (new) method are shown (which in the above case intersect at 24907) but the focus position value of the weighted average (old) method is indicated and used (in the case above: 24903).

  • What is the reason for this behaviour?

There is no mention of the 2 AF methods in the manual and there seems to be no setup for a choice in SGP, which method to apply.

  • Is the new method still in an experimental stage?
  • How can we make use of the results of the new method?
  1. More questions according to binning of AF shots:
    I am using a DSLR (Canon EOS 600D). When AF binning is set to 1x1, the logfile contains comments like these:

[11.08.17 02:10:33.282][DEBUG] [Main Thread] --> Find stars (normal)
[11.08.17 02:10:33.427][DEBUG] [Main Thread] AF frame was too large… downsample = 0,25…
[11.08.17 02:10:34.188][DEBUG] [Main Thread] Star detection using min star size of 2px…
[11.08.17 02:10:34.188][DEBUG] [Main Thread] Star detection using max star size of 40px…
[11.08.17 02:10:34.210][DEBUG] [Main Thread] Find stars took: 927 ms…
[11.08.17 02:10:34.210][DEBUG] [Main Thread] Star list contains 65 stars…
[11.08.17 02:10:34.210][DEBUG] [Main Thread] Calculating HFR with sample size: 6
[11.08.17 02:10:34.955][DEBUG] [Camera Thread] Auto focus HFR calculated at: 3,64

When AF binning is set to 2x2, the corresponding spot of the logfile looks like this:


[11.10.17 22:10:19.462][DEBUG] [Main Thread] --> Find stars (normal)
[11.10.17 22:10:19.488][DEBUG] [Main Thread] AF frame was too large… downsample = 0,75…
[11.10.17 22:10:20.756][DEBUG] [Main Thread] Star detection using min star size of 4px…
[11.10.17 22:10:20.756][DEBUG] [Main Thread] Star detection using max star size of 80px…
[11.10.17 22:10:20.784][DEBUG] [Main Thread] Find stars took: 1322 ms…
[11.10.17 22:10:20.785][DEBUG] [Main Thread] Star list contains 201 stars…
[11.10.17 22:10:20.785][DEBUG] [Main Thread] Calculating HFR with sample size: 6
[11.10.17 22:10:21.051][DEBUG] [Camera Thread] Auto focus HFR calculated at: 1,29

  • Why is the image still downsampled when AF binning was set to 2x2?
  • With a DSLR, is it better to use 1x1 or 2x2 AF binning in order to get a correct HFR calculation?

Bernd


[Smart] Auto focus help
Temperature Compensation Implementation
#2

Is it really asking for too much to get response to my questions?

Bernd


www.mainsequencesoftware.com