My optical setup is made up of an FSQ106 with Canon Eos 60d CentralDS. On the FSQ106 I mounted a Robofocus. Typically work at f / 5.
Unfortunately, I must realize that the automatic focus routine is almost always wrong.
I tried the focusmax solution with maxim dl and for what I was able to verify, with this same configuration it is fine, with errors of the order of the step. When needed, I exit the SGPro sequence, fire with FocusMax and then resume SGPro.
a question
I find SGPro a beautiful application but the issue of auto focus makes it unusable for me, are you thinking of introducing a similar FocusMax routine shortly?
Can you give details of your auto focus settings. Several users, coming from Maxim DL / FM have problems with SGP - once they understand how SGP works and the required images to determine HFD, they are normally perfectly satisfied. I have FSQ85 and 60Da, no problems.
I’m trying to remember the settings at this time, anyway I can tell you that I’ve tried one of everything …
…auto focus data points: from 5 to 11
…step size: from 4 to 20
…maximum star diameter: from 2 to 15
…crop auto focus frama: from “no” to 50%
…exposure time and ISO: very varied
If I refocus the fire repeatedly the result is always different, the curve is always different.
And then suddenly loses the temperature. (NA in pannel).
What I see during the formation of the curve is that the stars that consider calculating the HFR change, are not the same and so I think that the statistical average does not work well. but these are trivial considerations, I do not know the logic of the program.
What I can say is that if I do not move anything, if I use maximdl and focusmax, I do and refocus the fire … fast and accurate fire with at most one step difference!
I do not know what to say.
SGPro I really like it very much, it is a simple and powerful solution, if I could solve this problem I would be extremely happy.
I think this is a settings issue: Many users have Taks and DSLRs. It is the exposure time and ISO that are critical. The last time I used mine, it was using 9 second exposures at ISO 800. Maxim and FM do focusing completely differently, choosing a bright star and doing very short exposures. You need more generous exposures with SGP, so that many stars are detected.
Try 9 points, with a step size so there is a 3:1 ratio of HFR between max and min (for me, about 20 steps of 4 microns), the crop size at 50% is way too big - try 15%.
The second thing to understand that SGP is not like FM which works out the rough focus position from anywhere. It ideally needs to be within the range. Before doing the first AF run, use frame and focus exposures for 10 seconds and manually step the focuser so that the stars appear roughly focused.
The third thing - I have seen several Taks produce unusual V-curves at the focus point. It looks more like a flat W. This doesn’t happen with my other refractors. This also occurs with all focus algorithms and if you try to only take AF exposures around the very central region of the V-curve, it will get confused. You need some readings on the two ‘wings’, to create an estimate of the central point.
Lastly - just because another algorithm comes up with the same answer, does not necessarily mean it is more accurate. I once worked for an industrial gauging company, using IR spectroscopy. Our competitor used to show how ‘inaccurate’ our system was by pointing out the varying readings…they demonstrated their product with a low gain and high offset to show how stable they were - a very nasty and effective psychological trick on production managers.
There you go - first attempt was good, with clear V-curve. You can see the settings. There is no reason why you cannot achieve the same. In this case, the exposure time is on the limit - and if you go a long way out of focus, the lower star intensity causes the V-curve to fold back. If I use about 10 seconds, the V-curve extends out much further before getting confused by faint, diffuse blobs. [update] = my step size is 6.13 microns, using a Lakeside motor on the Tak.
I did the tests and unfortunately the results are not satisfactory. it works once on 5. At the moment I’m doing this: I do not do autofocus, I use the reminder option. when asked to do autofocus with FocusMax and then resume the sequence. So the fire is fast and good but unfortunately I must always be present :(. I hope I’m not misunderstood, I want to be constructive. If a FocusMax-like Focusing Technique was implemented, perhaps with the automatic search of the stars closest to the target that fit the focus, and then returning to the target after focusing, I would include an adequate review of the cost of the focus SGPro license. A similar technique is implemented in the application called Starkeeper Voyager. This product in my way of seeing is far from SGPro, I tried it and I would not make any judgments, surely I will not use it but have this formidable feature of auto focus. If you happen to try to look at it. thank you very much for the time you have dedicated me. if someday you say you have implemented the function you know that you will see me jumping for joy. meanwhile I will continue to use SGPro. Grazie, Marzio
but unfortunately I must always be present :(. I hope I’m not misunderstood, I want to be constructive. If a FocusMax-like Focusing Technique was implemented, perhaps with the automatic search of the stars closest to the target that fit the focus, and then returning to the target after focusing, I would include an adequate review of the cost of the focus SGPro license. A similar technique is implemented in the application called Starkeeper Voyager. This product in my way of seeing is far from SGPro, I tried it and I would not make any judgments, surely I will not use it but have this formidable feature of auto focus. If you happen to try to look at it. thank you very much for the time you have dedicated me. if someday you say you have implemented the function you know that you will see me jumping for joy. meanwhile I will continue to use SGPro. Grazie, Marzio