Some problems with the Wavefront Estimator script

[AlphaZentauri]

Dear PixInsight Community,
I really would love to use the Wavefront Estimator script, nevertheless I always get the error, that the diameter of the defocused image is too small. I entered the values for my Lens (135mm Zeiss F/2) and decided to go for an exposuretime of 8 secs. So with the parameter "Corrugation resolution" set to 13.5 I get a defocused image diameter of 77.8 pixels and a star with a magnitude of 1.7 because of this I took the star Beta Tauri. Unfortunately I can´t get over 50 Pixels in diameter with my Zeiss lens extrafocal because it is a normal camera lens which doesn´t go much past infinity. Having that in mind i recorded 12 images extra- and intrafocal of the star at 8 seconds and 100 ISO. The star reaches about 4000 ADU per Pixel which is about 25% of the cameras saturation limit (12bit). After converting all NEFs to XISFs and importing them into the script I started the script and it ran for a while but after having loaded all images and after calculating the defocus threshold it just quits during Intra-focal frame registration with the message "The defocused image diameter is too small". I absolutely don´t understand what I am doing wrong and I would be very happy if anyone could help me with my concern.

Best regards,
Othmar

[mschuster]

Hi Othmar,

Sorry for the problem. The script is now configured to accept star diameters between 48 and 320. I am guessing one or more of your images may be just below the lower limit.

There is a chance that the script will work OK with sizes somewhat below 48, but this is something I should test to be sure.

Would it be possible for you to place a zip of your images, along with a txt file giving the script parameters you are using, in Dropbox? If not possible, I can generate small synthetic images for testing purposes. Please let me know.

Thank you,
Mike

[AlphaZentauri]

Yes of course I can thank you really much for your quick answer! I also had a quick look at the images yesterday evening and I found out that the star is 40 Pixels in diameter when the image is in linear state. As soon as i stretch the image the diameter expands to about 50 pixel so i am quite sure that this could be part of the problem. Should I send you the raw files or the already converted XISFs?
Thank you again very very much for such a quick response I am really glad about that.
Best regards,
Othmar

[mschuster]

Converted XISFs please. Also, if not bias-subtracted, a bias master XISF if available.

Thanks,
Mike

[AlphaZentauri]

I think I´ll subtract the bias myself, my masterbias is around 400mb large and I think that would take too much time to send the bias frame to you.
Thank you very much again for your effort.
Best regards,
Othmar

[mschuster]

Just an update, I sent you a post with results. I modified the script to allow smaller stars, and ran a synthetic test to check the sanity of the script's results, similar to the simulations shown in the script's documentation. The lens exhibits large on-axis coma, with a script estimated magnitude likely twice larger than reality due to algorithm approximations. For such large aberrations, this is a reasonable result given the method used on a small image with bayer rgb artifacts. I will release a new version of the script shortly with the lower size limit.

A second update: The twice larger than reality problem was due to a bug in the Bayer RGB detector support code. I found and fixed this problem. Also I added support for all color filter array detectors. Frames must be bias-subtracted and must not be demosaiced.

Thanks,
Mike

[AlphaZentauri]

Thank you very much for your great work Mike! I really appreciate your work.

Best regards,
Othmar