I have not been able to get plate solving to work in Siril V 1.4.0 Beta with Windows 11. I have been using Siril for 3-4 years. I am using a Mallincam DS16C with 0.5 focal reducer on an Orion 8-inch RC 1600mm FL, permanently mounted on an iOptron CEM 60 mount. The attached image shows my try at using Seril Server (I have also tried Astrometry.net) to plate solve my image of M46. The image used is the result.fit file from Siril. I used 2x2 binning. The coordinates agree closely with ones given by TheSkyX. The Plate Solver automatically fills in 1524mm as the FL, which I think is wrong. With 0.5 FR shouldn’t the FL be 800mm? I have tried focal lengths of 1524, 1600, and 800 along with pixel sizes of 3.8, 7.6, and even 15.2 in various combinations all with the same results.
I used the same image online at astronomy.net and it worked perfectly.
When it stops running I get “Plate Solving failed. The image could ot be aligned with the reference stars.” along with:
13:13:33: Findstar: processing for channel 1…
13:13:34: Using 1989 detected stars from image.
13:13:34: Using already downloaded catalogue NOMAD
13:13:34: Fetched 3266 stars from NOMAD catalogue
13:13:34: Plate Solving failed. The image could not be aligned with the reference stars.
I’m not exactly a newbie. I’ve been using this setup for almost 10 years, well, the Orion is new. I was using a Meade LX200 12-in. I’ve only ever gotten Plate Solving to work maybe one or two times in the past. I used to just skjp it but now you have to do it first to be able to use Photometric Color Calibration. Obviously I’m really missing something. Any advice would be greatly appreciated.
Not all acquisition software use the same convention regarding binning, some keep the physical pixel size irresp. of binning, some write the pixel size after binning. Siril can adapt to both conventions, you can adjust the behavior in Preferences:
It was always the case, now we’ve just split that into 2 separate stages, so that if Siril can’t solve the image, you can still solve it with another solver (we also integrate interface to astrometry.net) and go on using the photometric calibration tools.
Now, as said, we need an image to figure out if smthg is wrong.
Thanks, Cecile, I will check out the info about binning. I had no idea.
I had already run the entire thing once again with similar null results. The image used is attached; it is the result.fit from this last run. I also ran it through nova.astrometry.net and it worked. The pixel scale was 0.646 arcsec/pixel.
I just checked and “Update pixel size of binned image” is checked. To ask a dumb question, how does it know the image is binned? Do I still put in the correct camera pixel size?
The acquisition software writes another key, smthg like XBINNING or BINX. Normally, everything should be filled out correctely when reading the image metadata. In any case, values should be 800mm for the FL (1800*0.5) and 7.6 for yhe pixel size (3.8x2 for binning). This gives a sampling of 1.957"/px so i don’t understand how the samping found by astrometry.net relates to the description of your gear.
I can’t see the link to your file though
I tried to back calculate smthg from the sampling found by astrometry, the only combination I find is if the focal reducer ratio is 0.75 instead of 0.5 and no binning…
sampling = 206 x 3.76 / (1600 x 0.75) ~ 0.645"/px
I’m using 3.76 microns instead of 3.8mm as per your spec because, well, all that’s a more common pixel size and I suspect the 3.8 value is just a roundup.
Can you try platesolving with those numbers, 3.76 and 1200mm FL?
Well, I’ll be hornswoggled! That worked, Photometric Color Calibration is no longer grayed out. Apparently, the focal reducer that was sold to me as being 0.5 is really 0.75. Cavet emptor! I also would not have thought that a round up that small would cause a problem. Now I can get on with my image processing. I can thank you enough for your help.
Actually it would solve with 3.8 as well, I was just trying to find the sampling from astrometry while keeping a “classical” reducer factor.
The sensitivity of Siril solver to accuracy of the sampling is also a parameter than can be tuned in the Preferences. I think we have 10% as default. And it can be increased up to 50% if I remember well. However beware, the larger the percentage, the slower the plate-solving (and it may also become prone to false positives…)
