Solar Mosaic tool/sequencer Output files Won't Stitch?

[chuck666]

I've posted a similar question on Cloudy Nights and not getting many responses. So I'm posting here.

Today, I ran the sequencer in Auto mode as I have all previous times. My gear is and AM5 mount, an ASI 432mm camera, a Astar 80 PHQ scope and a Quark Chromoshere filter. I decided to use 250 pixels of over lap and 150 pixels of edge clearance. 16 Mono and ser video. The sequencer set up 10 panels. I used the default 1000 images. Selected start. Everything progressed normally and I had 10 ser files.

The videos looked fine with the feeling that the overlaps may not be lined up. For reference. The mosaic is 2, 3, 3, 2 shape.

The Sharp Cap manual suggest to stitch with Affinity Photo for example. I own that and that's what I used. First, you have to convert to tif files for Affinity. I ran the ten files thru AutoStakkert and had 10 stacked tif files as expected. Then, inputted into AP, ver 2.5.3 the latest.
After opening and selecting "new panorama", I added the 10 files. NOTE: several people said in CN that that's what they do and it works. Now, I don't know if they used the newish tool or created their own mosaic files. I suspect the manual method, old school.

Well, when AP finished prodessing; I got 3 partial stitched images that didn't use all the files. If I then tried to stitch images 1 & 2; the top pair, it says no panorama found. Other combinations also do that. If I try the images that were used in the original groupings, they stitch?

Sorry for the longish post. I'm stumped and what settings to change?

Chuck

[chuck666]

I forgot to mention that I have a ZWO Tilter in the image train. Could that be my issue?
Chuck

[admin]

Hi Chuck,

sorry to hear that you are stuck at the stitching phase - I think that the most likely causes will be to do with possibly a lack of overlaps or difficulty with the stictching application finding matching features - I don't think things like the tilter will have a significant effect.

I haven't yet had a problem stitching with Affinity (note - V2, not V1 - not sure if I have tried since the latest updates mind you), but that obviously doesn't mean there can't be problems...

I would start with the first row of panels and check that you can see features on the right of each panel also visible on the left of the next one. Similarly, you should see features at the bottom of each panel on row one that are also visible on the top of the panels on row 2 - this gives you confidence that the panels are really overlapping and also a feel for how much overlap you have. You could also try things like setting up a panorama from just the top row to see if that joins, then row 2, etc.

How big is the overlap in terms of % of frame size, rather than in pixels?

cheers,

Robin

[chuck666]

Robin,
Missed your reply; I must not correctly done notifications?

Thanks for the tilter answer........

MY SC ver is last weeks. So, current. AP is also up to date 2.5.3 as of last week.

Sticking: I have tried sticking the first row; panels 1&2.. It says no panarama. The last auto one generated 10 panels. 2 3 3 2 matrix. I created 3 panos. HA.. I took some time and tried to stick various combination and couldn't get many of the adjuacent panels to stitch. And this latest attempt; I created about 30% overlaps in pixels. I think it was 225. Now, I didn't do any inspection of visible features; But, are low contrast captures an issue?

I just noticed a few minutes ago another post from someone having similar problems and the issue of backlash was discussed. I haven't knowingly used that except the default 4as, I think. He said turning it off helped.
I'm in AZ and we have had clouds lately, kind of rare for us. I may be able to test backlash tomorrow. And I have NOT put any values in for backlash on my ASO EAF focuser.

I have only been doing this craft for about 4 years and have always used the ASI AIR as my control software. But, I must say the SCPro and a miniPC are quite fun. Thank you for you assistance and great software.

Chuck

[admin]

Hi Chuck,

glad to hear that you are having fun using SharpCap

Certainly low contrast in the images could be a problem when trying to stitch them together, as the panorama stitching program will be looking for features that it can match up between the right side of one panel and the left of the next (and so on). There needs to be something (granulation, sunspots, etc) for it to work on... Feel free to share some or all of the panel images here if you want as often seeing the image is much more helpful than even a detailed description!

If your mount already tries to eliminate backlash then you might want to turn the backlash correction function off. On the other hand, if you see the image drift for a few seconds after moving to a new panel, that usually indicates backlash in RA being taken up by tracking - in that case consider turning the size of the backlash correction up to make that drift after each movement go away.

cheers,

Robin

[chuck666]

Thanks for the reply.

StIll clouds today. Cannot retry until tomorrow.

EDIT: I did NOT remember that an AM5 mount, that is what I'm using; is a strain wave design and has NO BACKLASH due to that design. And there are NO settings for BL in the setup module! Could this be the cause of the "not stitching"???? And, if so; an alert in the sequencer may be needed?

With a ZWO ASI 432mm camera, there is NO gamma control. I seemed to be missing a way to add contrast of my active image? Suggestions or have I missed a simple setting or parameter adjustment?

Cheeri-O
Chuck

[admin]

Hi Chuck,

having backlash correction on in SharpCap with a zero backlash mount shouldn't be a problem - all that the backlash correction does is make the movement to a new position happen in two stages - an initial move to close to the desired position then a final small move that should take up the backlash. If there is no backlash then this should have no effect on positioning and therefore none on the images either...

ZWO once upon a time had gamma adjustments for their cameras, but removed them some years ago. It's generally considered best to apply that sort of contrast boost after stacking/processing (that way you can remove it, adjust it, apply it again, etc, whereas with a gamma on the camera the adjustment is baked into the saved frames and cannot be undone or cleanly adjusted later).

For display on screen, SharpCap has a display stretch function that can boost contrast, stretch faint stuff, etc, but only affects the image on screen, not the saved files.

cheers,

Robin

[chuck666]

Robin,
Thank you for confirming things. I am starting fresh this morning. The sun just became visible over the horizon. Supposed to be better seeing today and I have moved my setup so I can image sooner. MY backyard location required me to wait for the sun to clear my house. By then, here in AZ it got quite hot.
New setup, so need to update my polar finder and other things. Hopefully, I will be able to ge SC Pro up and running and retry the solar mosaic sequencer again.

Later,
Chuck

[chuck666]

Here are the bottom 2 files jpg save as.

Question: I use Autostakkert 4.0.11 in it's first section on Image stabilization, there is expand or crop. What is your recommended choice?

I have tried both. Seems to not matter; but it should? These 2 files used expand setting. Also, I lessened the overlap on these. About 15% now.

Affinity would not stitch any of the 3 pairs of fileslll "No panorama found?"

Chuck

[admin]

Hi Chuck,

I get the same results in affinity, but being able to see the images gives me some ideas. I tried to spot an alignment between them by eye and I *think* I found one - the two points with the red arrows correspond, I think

Screenshot 2024-07-02 215905.jpg (80.72 KiB) Viewed 9188 times

But, it's hard to be 100% sure as there are significant brightness differences between the areas and I think that is stopping Affinity from spotting the panorama match. Image #5 (left side) looks almost overexposed towards the right hand side - there are bright areas that are almost devoid of detail, like you would see in an over-exposed image - except they are grey not white. The detail is much better defined in #6, although that has a dark patch in the top left corner.

These could be a stacking issue, or they could be coming from the video data directly - worth watching through the videos with no adjustments to see how they look and if they show the same symptoms.

My guess here is that flat correction is going awry somehow and leading to both the over-exposed right hand side and the darkened top-left corner, mainly on the grounds that it is a single thing that could possibly explain both odd effects.

As to Autostakkert - I think I usually use crop, but actually for panoramas I often use PlanetarySystemStacker, as it will batch process all the SER files with the same settings without needing them to be done individually.

cheers,

Robin