3.94
#...
·
 |
|---|
| I also have some fields to add, one with really bad data taken at 40% moon, guess your algo will find it easily.... Will check the sheet and add the data tomorrow, there is still a bit of stacking to do.... |
1.20
#...
·
·
1
like
|
|---|
|
Here's our collective progress so far - 135/1120 imaged fields superimposed onto a map of the entire sky. As the projection used is equal-area it gives a direct impression of the fraction of the sky that we've covered to date. The image resolution is 12960 x 6480 which corresponds to 100 x 100" pixels at the centre ie one tenth of the linear resolution and one hundredth of the area of the final image (gulp!) . If you right-click and select 'open image in new tab' you can see the full resolution version. The mosaicing is very basic at the moment as I haven't used my colour normalisation or background equalisation tools, just a simple equalisation of the median RGB values of each field. The small image scale rules out a more sophisticated approach for now. Brian: we seem to have a small sliver of a gap at about 80 degrees south - I'll investigate in more detail, but if confirmed I'll figure out the coordinates and run an infill image. Construction of the image is now fully automated via a script so it will be easy to update as more fields are added. It's also proved a nice test of the overall mosaicing pipeline. Brian, Michael: what do you think about posting this as a regular AB collaboration image? Maybe we can drum up some more contributions if we can show some progess?  |
16.52
#...
·
|
|---|
James Tickner: Hi James Looking good… Please post. Will focus on the remaining fields at -75 and -70 initially in the next lunation CS Brian |
|
3.94
#...
·
|
|---|
|
Can you add the northern pole area from Todd? Then things do not look so empty up north…. Thank you for your big efforts! Michael |
|
3.94
#...
·
·
1
like
|
|---|
|
I have prepared new data for James, as weather is rough atm there are some holes that now get obvious when I merge the fields. Data is from two different cameras, my hope is that James will be able to create more even result than what I was able to create with Mosaic By Coordinates and then merging the data in photoshop. Michael  |
|
16.52
#...
·
·
1
like
|
|---|
|
Thanks to ongoing clouds, I have looked again at the issue of mosaic making. I am sure @James Tickner is doing a lot better, but I lloked into doing "superfields" with GraXpet, MosaicByCoordinate and Photometric Mosaic. Using 3 x 3 mosaics to generate 18 x 12 degree images on 15 x 10 field centres, I was able to produce the following. Note that this is the first time I have been able to push the post-processing to include HT blackpoint clipping oat 0.5% of data, arcsinh stretch (1.5) and star reduction.   And the annotated versions   |
|
1.20
#...
·
·
1
like
|
|---|
|
Apologies for being offline for a few days. The mosaics are looking really good, Brian! I've been trying to build out more of the machinery for whole-sky mosaicing, but this is proving quite slow going. I think the idea of producing some 3 x 3 or similar 'superfields' similar to the ones you've made is a good idea - it will be an opportunity to test some of my stitching and gradient correction tools on a more practical scale. Let me see what I can get together in the next week or two and I'll post the raw files for you to play with. By the way, with a bunch of fields that Michael has recently completed and a handful that I've managed to add this lunation, we are now comfortably over 25% of sky coverage - something of a milestone! I'll start to put together another whole sky star map to mark the process. |
0.00
#...
·
·
1
like
|
|---|
|
Hey guys, That looks like some awesome regress you are doing, impressive work. At least on my phone display, I can't see any seams or gradients in Brian's images. It will be interesting to see how James routines compare to that. @Brian Boyle a little off topic, but did you hear anything from the PixInsight team regarding the MARS project? I could imagine they have quite an interest in our data. CS Gerrit |
|
16.52
#...
·
|
|---|
Hey guys, Hi Gerrit, Yes, I heard from Juan. And they do have an interest. I propose to make the data on my DropBox available to them - although I am not quite sure when/how they want us to provide the data. CS Brian |
|
16.52
#...
·
|
|---|
James Tickner: Hi James, I am going to hold off on further mosaicing. It is a tedious process with PI, and I await your results with great anticipation. CS Brian |
|
3.94
#...
·
|
|---|
|
By the way, we have also shared our data with the GraXpert team, their lead developer already had a first look and things look promising! So our survey will serve multiple purposes, glad to be part of this journey! |
|
3.94
#...
·
·
1
like
|
|---|
| @James Tickner , would it be possible for you to calculate a sheet for a 40mm lens on a fullframe camera? Background is that I ordered a Sigma Art 40mm yesterday and perhaps Brian will do the same thing (he told me he thought about this for a while) and it would be good to have a plan (and perhaps a sheet) to collect fields. Resolution of this combo is 30.53" per pixel (calculated for a Nikon Z6ii with 5.92" pixel pitch and resolution of 6048 x 4024) From what the GraXpert developer showed me yesterday those even more wide field versions can help to get the gradients right in our 135mm fields, here's an (for me quite impressive) result he did in less than two hours playing with our data where he took a mosaic that he has done with a longer focal length and used our data to correct the gradients: Top Final result, Bottom Original  And here the full overview with his data, the reduced data from our dataset and on the right the gradient maps  |
|
3.94
#...
·
|
|---|
Another result from GraXpert Discord: Top left the original, top right the upscaled data from the survey, then bottom left the final cleaned up image, bottom right the gradient mask. Honestly, I am highly impressed... |
|
1.20
#...
·
|
|---|
Here's a (rather ugly) tiling of the sky for a 40 mm lens + FF camera combination, giving by my calculation a 33.4 x 48.5 degree FOV. I've assumed a 20% overlap between fields, which results in a 48-tile mapping. Attached CSV lists file centre positions. Let me know what you think! field_centres_33x48.csv |
|
3.94
#...
·
|
|---|
|
Thank you, very helpful, much less tiles than what I prepared in a crude way. Now if only snow and clouds would move away…. Michael |
|
3.94
#...
·
|
|---|
|
I was able to get roughly 3 hrs of moonless skies yesterday (but unfortunately not cloudless..) to give the Sigma 40mm another try wide open I like the amount of signal it aquires wide open but the stars are another story, lots of issues in the corners wide open, so I guess I will have to switch to APS-C format (which on the plus side enables me to get more of that Ha as the Z6ii I am currently using is not Astro modified). Here's the first frame, only SPCC, GraXpert and BlurX applied, 1hr of 1min exposures at ISO 8000, center coordinates are RA 03 30 00 and DEC 50 00 00 tracked, not guided on my StarAdventurer GTi  And here's the Barnard's Loop region, Now with the last 20 Minutes not stacked:  As expected, Cygnus is not that good, too many clouds:  |
|
1.20
#...
·
·
2
likes
|
|---|
Here's the current state-of-progress, showing all fields imaged so far on a whole-sky map. I haven't tried to apply any gradient correction, but images should be approximately colour balanced. Together we've covered approximately 26% of the whole sky. |
|
3.94
#...
·
|
|---|
|
Progress is definitely visible, glad that Dan Watt has contributed in the North, it still looks pretty empty and scattered but definitely better than last time. @James Tickner will you also again post the updated image? And another wish, would it be possible to get a csv for 40mm and APS-C, I learned that I need to use an Astrocam and I don’t have fullframe and will likely not get one unless they drop to half of the price…. Michael |
|
1.20
#...
·
·
1
like
|
|---|
Michael Ring: Your wish is my command  A round one hundred 33 x 21 deg fields for the whole sky.field_centres_33x21.csv  |
|
3.94
#...
·
|
|---|
|
Many thanks for the new list! Number of fields has doubled with the switch to APS-C, sniff.... Hopefully next friday and saturday will mark a break of the constant rain/snow/full moon pattern that I have right now. When looking at the southern part of the panorama there seems to be a pretty constant pattern of light pollution repeating in the frames, those could be a perfect area to try if 40mm frames can help fix this, I will also try to run this area through GraXpert to see what it can do... Michael |
|
3.94
#...
·
·
1
like
|
|---|
| @James Tickner@Brian Boyle I read something quite relevant for us in the BlurXTerminator documentation: For best color balance results, it is recommended to apply BlurXTerminator in the Correct Only mode prior to precision color calibration (e.g., SPCC). After calibrating color, apply BlurXTerminator again with any desired sharpening settings. Testing with this workflow during development of AI4 typically resulted in more consistent results from SPCC (i.e., the same or better dispersion in the R/G and B/G fits). see here: https://www.rc-astro.com/blurxterminator-2-0-ai4-release/ There are also detailed instructions for mosaic making Putting all of this together into a workflow, starting from individual mosaic frames, might look like this:
According to this I'd propose that we change documentation for submission, People that have BlurX Ai4 should deliver 2 files, one untouched, no SPCC and a version that has BlurX applied in Correct Only mode (unless we want to apply BlurX) We should definitely put in the docs (if it is not already in) that people should not run SPCC on their data.... Thoughts? |
|
16.52
#...
·
|
|---|
|
Hi @Michael Ring I think @James Tickner is doing this a slightly different way. Has started from OSC data [step 1_ and then doing 4, 2 and 5. No BXT in phase 3 or e.g. GraXpert in phase 2. For my own mosaics I have been doing 2 (GraXpert), 4 [Mosaic by Coordinates} 5 [Photometric mosaic] 6 [image solve] 7 SPCC and 8 BXT. This work well for up to 6 panels. But after that it is simply too tedious to use Photometic mosaic - which only does things pair wise..... Looking forward to @James Tickner results.... CS Brian |
|
16.52
#...
·
|
|---|
|
Hi @Michael Ring I think @James Tickner is doing this a slightly different way. Has started from OSC data [step 1_ and then doing 4, 2 and 5. No BXT in phase 3 or e.g. GraXpert in phase 2. For my own mosaics I have been doing 2 (GraXpert), 4 [Mosaic by Coordinates} 5 [Photometric mosaic] 6 [image solve] 7 SPCC and 8 BXT. This work well for up to 6 panels. But after that it is simply too tedious to use Photometic mosaic - which only does things pair wise..... Looking forward to @James Tickner results.... CS Brian |
|
3.94
#...
·
·
1
like
|
|---|
| @James Tickner I finally had 3 clear nights, ther last night with in fact no clouds at all, so I was able to gather some more data with the 40mm Sigma. I have uploaded the data to my share and have removed the data that was gathered with the Nikon Z6ii. https://www.mycloud.ch/s/S0012A31EE7B15509D2A9051E1B72E54FE47FEC8168 Here's what i have until now, unfortunately the bad weather is back so likely I will have to wait until next year to gather more data.  I mostly followed the workflow as outlined by Russell Croman, I had to do SPCC before merging the tiles because I got very strange results when trying to run SPCC on the pano with that much black space. I was also able to collect some data with the Samyang 135mm lens of the same area, so it should be possible to play with gradient removal based on the 40mm stacks. I am currently stacking this data, may take 1-2 days until this is done. Michael |
|
16.52
#...
·
|
|---|
| Looking really promising @Michael Ring |
