Solar Mosaic - Feedback

[Altocumulus]

Cheers Robin, the results of my Mosaics are in the Gallery.

Will be available to help testing, when required and when the weather behaves....

Geoff

[Altocumulus]

I've got large files transferring across my network from an Ha session earlier this morning. Frontal cirrus and cirrostratus now invading.

I didn't make any changes to the backlash settings, but increased the overlap to 400.

Watching the capture I noticed each alignment point was targetted, correctly, by the routine, and then it shifted upwards to take. I suspect I have all my segments shifted by some degree upwards so there's an excess of Space at the top at the expense of the bottom. I took separate frames to compensate (hopefully enough for a stitch).

[admin]

Hi,

There are two things I can think of for the drift here that might explain it

1) If the drift is gradual, then it may be backlash being taken up by the mount's tracking motion in the RA direction. The fix for this would be to increase the backlash allowance in the mosaic setup, so that SharpCap will move the mount far enough in the tracking direction when approaching the panel position to avoid any drift


2) If the movement to the second position is fairly brisk, then you are seeing the two stage goto procedure that is trying to take up the backlash in the mount before each capture. SharpCap works out the position needed for the panel, then moves to it in a two step process - the first step is offset in both RA and Dec to try to ensure that the final approach is in a consistent direction and takes up backlash in RA. The second position here is the one that SharpCap has calculated and is aiming for.

Did you get a chance to try the four buttons to check alignment (goto top/left/bottom/right edges)?

Also, yesterday's update contains some new improvements for mosaic alignment, including

* Auto alignment : backlash takeup on the first measurement point as well as points 2 and 3 to give consistent results

* Auto and guided alignment : improved calculation of the rotation component of the alignment

Anyway - lots of things to try, but I am looking forward to seeing your images!

cheers,

Robin

[Altocumulus]

Cheers Robin.

So am I, 18 frames this time, and that might not include the extra 3 for the bottom pole region.

What confused me is that the alignment points were hit correctly, mount allowed to settle - and then it shifted up before capture. Surely if it hits the points correctly it doesn't need to move again?

[Altocumulus]

Hi Robin,

Needed the extra 3 - so 21 frames in the mosaic and it's not looking good as there's no real coherence in the proms where frames overlap.

I normally stitch then process but it looks like I'll have to try some processing first.

Uncropped mosaic 135mb, some cropping in Affinity Photo 2 brings it down to 95mb - but imppg will not open the mosaic. There's an issue within the disc with uneven tones showing.

[Altocumulus]

Hi Robin,

Looks like an overlap of 400 was too much - way more than needed and seems to be complicating the stitch.

Looking at each three or four horizontal frames that cover the width of the disc, each one to the right is stepped down, rather than horizontal as in the mosaic tool display - see attached.

thirdlayer.jpg (153.34 KiB) Viewed 279 times

[admin]

Hi,

the downward stepping indicates that the measurement of the rotation of the sun's disc relative to the camera didn't come out right - maybe out by about 10 degrees in this case.

When I went back over the code of this last week, I discovered that the rotation angle was only being calculated from one (X) component of the alignment data, whereas it can actually be calculated from both X and Y to give two estimates. The update from Monday uses both estimates and averages them, which should on average reduce the error in the calculation of the rotation. Not sure if you were using the latest update or not though...

cheers,

Robin

[Altocumulus]

I think I may have missed that version - so will pop it onto my pcs when the weather allows the next session.

Cheers

Geoff