View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0001862||CCdciel||General||public||17-12-09 10:41||18-01-15 14:42|
|Reporter||Thomas Jaeger||Assigned To||Patrick Chevalley|
|Summary||0001862: Autofocus Dynamic FindStarPos : Access violation|
|Description||Yesterday I successfully captured some frames of NGC281 and the Maffei Galaxies. This was my second try with CCDciel, I like it.|
I am quite familiar with autofocus in Sequence Generator Pro, so I decided to start with CCDciel's Dynamic Autofocus which is similar, but nothing happend. Later examinations of the log showed a Access violation. Never got CCDciel to start an curve.
|Steps To Reproduce||1. Focused a star manual with focus aid|
2. Made a short image and select a the star again
3. Press autofocus (dynamic), nothing happend, log showed Autofocus error with Access violation
|Additional Information||Cannot fancy the parameter min.SNR, maybe to complicated for most of the users. Changing values here from -10 to 20 does not solve it.|
|Tags||No tags attached.|
screen capture 2017-12-07 17-41-43.png (45,166 bytes)
screen capture 2017-12-07 17-41-43.png (45,166 bytes)
log-snip.log (2,478 bytes)
The access violation is not the main problem because it is trapped, the important point is it cannot find and measure the star.
This message is probably because the star is too near of the frame border or because the "Star detection window size" is too big relative to the "Focus window size", both option in Preference/Focus.
Beware this two values are in unbinned pixel, so divided by two (to keep the same FOV) when using binning 2x2.
You may want to use at least 400 for "Focus window size" and 100 for "Star detection window size".
Let the min SNR to 0 for now, this disable this checking. You can set a reasonable value for your equipment later because the SNR is show in the log for every measurement.
Also this "Dynamic" method is less tested than the standard Vcurve and it is possible you get some problem. But for me it work with the Indi simulator and for now it not even started as it cannot detect the focus star.
Can you try with the increased window size as mentioned?
Thank you very much for the reply.
My settings were:
star detection window size =200
focus window size=400
OK I will decrese star detection window size to 100
After all, reading all documentation, I do not understand completly how CCDciel chooses a star or stars to determine focus. At my session I selected a star with doubleclick, than I have pressed autofocus.
How autofocus works with an automated sequence is also unclear to my now. I will try it later.
I confirm that 200/400 produce the same failure as you describe the simulator.
1/4 ratio is probably a minimum, I personally use a 1/10 ratio similar to the default value of 20/200.
I will add a check of the values ratio in the preference to avoid this problem.
For autofocus it not use the selected star, it always try to find the brightest star near the center of the image after checking it is not double.
In a sequence it first slew to the nearest star of the magnitude you select at the bottom of the autofocus options. Then return to the target after the autofocus is complete. You must have a working plate-solving setup for this to work correctly.
If you are sure the field you are imaging contain a suitable star for focusing you can avoid slewing to the focus star by checking "Stay in place for autofocus" in the sequence editor.
OK, what should I do? Next clear night I will use 80/400.
Thanks for your explanation in case of autofocus. I always use "stay in place" because, my mount is never connected to my pc and in my eye platesolving is too uncertain using it unattended.
For more than three years I took unattended astrophotos the whole night with great sucess. I try to minimize sources of error. Using PHD-Guiding is one of the success factors.
Yes 80/400 must work.
But beware this will really be 40/200 with 2x2 binning, so you must be sure it work with the highest HFD you can get. If the area is too small you can try 100/600.
It is more easy to use the manual focus to adjust this value as the star is detected the same way as for the autofocus.
There is no problem to "stay in place" for autofocus, principally if you not do narrow band imaging. Just be sure the autofocus exposure time is high enough to get a good star image.
Revision 742 check the focus window is at least four time the star window size.
Version 0.9.22 is available.
It add the possibility to take the mean of 3 or 5 exposure for each measurement point also for the Dynamic method.
I can confirm "access violation" is gone with 9.22. Thanks.
Last night dynamic method failed all the time. But it was my fault, analyzing the logs now I know that "autofocus tolerance" was too high. If it is ok for you, I will post more suggestions for autofocus operation in a separate feature request in the near future.
Issue can be closed.
||Thank you for your report.|
|17-12-09 10:41||Thomas Jaeger||New Issue|
|17-12-09 10:41||Thomas Jaeger||File Added: screen capture 2017-12-07 17-41-43.png|
|17-12-09 10:41||Thomas Jaeger||File Added: log-snip.log|
|17-12-10 15:11||Patrick Chevalley||Assigned To||=> Patrick Chevalley|
|17-12-10 15:11||Patrick Chevalley||Status||new => feedback|
|17-12-10 15:11||Patrick Chevalley||Note Added: 0004342|
|17-12-10 16:21||Thomas Jaeger||File Added: Bildschirmfoto vom 2017-12-10 15-59-04.png|
|17-12-10 16:21||Thomas Jaeger||Note Added: 0004343|
|17-12-10 16:21||Thomas Jaeger||Status||feedback => assigned|
|17-12-10 18:39||Patrick Chevalley||Note Added: 0004345|
|17-12-10 22:35||Thomas Jaeger||Note Added: 0004349|
|17-12-11 22:08||Patrick Chevalley||Note Added: 0004355|
|17-12-12 11:15||Patrick Chevalley||Note Added: 0004357|
|17-12-12 16:17||Patrick Chevalley||Note Added: 0004358|
|17-12-26 09:59||Thomas Jaeger||Note Added: 0004370|
|18-01-15 14:42||Patrick Chevalley||Status||assigned => resolved|
|18-01-15 14:42||Patrick Chevalley||Resolution||open => fixed|
|18-01-15 14:42||Patrick Chevalley||Target Version||=> 1.0|
|18-01-15 14:42||Patrick Chevalley||Note Added: 0004384|