Hi, I’m loving 1.4 and mosaics etc. I’m blown away with SPCC too, especially how fast it is with local downloaded data.
Could anyone help with border feathering errors? I’ve ended up with a lot of random sessions fit files and I get output errors. I imagine its because my fits are not all nicely overlapped.
I’ve noticed the estimate option along with centre of gravity can tell me what fits it doesn’t like but there isn’t much info I can see to work out what to deselect when feather isn’t happy.
I might need to tidy up and only use known even mosaic grids perhaps.
Can you share a log where we can see the errors being thrown? Feathering should not be affected by images having overlaps or not. The only feature that should be affected is if you tick normalize on overlaps (or if there’s a bug of course).
Otherwise, to remove images that may be off, you can use the Plot tab after you’ve registered with sequence platesolve. If you plot Y position vs X position, you should be able to spot the culprit.
Normalization failed. Check image 984 first. It is bad, all black.
So I go to the frame list and deselect that one, but the same error, same image number.
Apologies if I’m missing something I’m on a bit of a crash course
Update: If I clear out my process folder, delete the bad image file and start again it finds another one. I might be able to just do that but it’s slow having to start again.
I have been, although I’ve just now tried leaving debayer on during conversion to see if that makes a difference. It’s entirely possible I’m missing something as I’ve picked all this up in a short space of time.
I wouldnt be surprised if this just worked if I did a nice new neat 3x2 grid also. Might be an edge case with my real mix of questionable data I’m trying to throw to together.
We still have a ticket opened with drizzle producing in rare occasion black images, which then fail when normalizing. Nothing to do with overlaps or feathering. Just a corner case that is hard to capture. But if you don’t drizzle, it should succeed.
Ah I’ll stop using drizzle for now then. I was actually thinking after the excitement of being able to I should not assume it was always better. Thanks, your help much appreciated.
I am using a seestar. But I’m not bothering with mosaics from it as I get more images just stitching myself. You’re right though, this does do a great job, I’ve seen it work great. It’s probably the drizzle issue you mentioned I’ve started falling over on.
If you are able to create a 2-image sequence with one image ok and another black after applying registration, i would be extremely grateful… for now, we have a sequence where this happens after producing more than 200 images using a single thread, making extremely difficult to debug
I’m concerned I’m asking some basic stuff here but I just tried again and got:
Output index mapping failed (2018/2842)
The same 2018 error. I realise now 2018 is the amount of images out of 2842 I was trying to stack. I had put a big wFWHM on it as I haven’t figured out how to do more than 2048 yet.
I’ll try the test you suggested. I’m currently watching it apparently work after unticking feathering. It’s definitely got further atleast.
Update, I’ve just been told I’ve done too many posts for my first day So I’ll try adding in an edit.
Yes it all works without feathering, border “80”. I think I’ve tried lower values and had the same issue.
Another update: I’ve just done it all (without drizzle) in two batches, and it worked and feathering did too. Something to do with the amount I was trying all at once perhaps?
Just to clear this up a bit. The “failed check image xxx” was likely drizzle causing a black image, as you explained. When I got past that, I was then getting the Output index mapping failed (2018/2842) which I seem to have avoided by not throwing as many fit files at it at once.
So I was hitting a couple of different things. I’ll draw a line under this lot and just try again knowing what I do now on, some nicer new grid data when I can. Thanks for your help! Siril is amazing. Love it.
