StarTools

A clear and mostly okay seeing night finally popped up over the weekend, albeit maybe not the best transparency and humidity was quite high. I manged to get a few hours on an interesting cluster and stars region. There were a few ways to orient this to frame up various nearby stars, this just happened to be the one I chose.

Being so star-centric, I anticipated maybe some shadow and ringing issues with SVD and thus launched right into 1.9.536 for this; but I'll try the upgraded 1.9 later when I get some time.

First is full FOV processed at 50% bin, then a crop version processed through at 71% bin. The Astrobin link has the higher resolutions.


https://astrob.in/etubaz/0/

Nice one, Mike! I didn't know this one. Everybody is imaging galaxies when pointing the scope to Coma Berenices I especially like the star colors, looks well balanced. And stars are resolved well into the core of the brighter glob

The stars of the cluster in the lower left are a bit oblong. Do you have an idea what's the reason? Haven't seen that in your images until now.

Regards
Stefan

Wonderful! Is that really nice framing on purpose, or just a happy accident?

Nice shot indeed.
The globular close to M53 is a difficult target to spot visually. I tried it a couple of times in my 18 incher at the times but only succeeded to cath a couple of stars and i cound not really see the glob as suuch.
Indeed better photograph it. The difference in brightness between the two is striking.

Stefan B wrote: ↑Thu Apr 20, 2023 7:51 am The stars of the cluster in the lower left are a bit oblong. Do you have an idea what's the reason? Haven't seen that in your images until now.
Regards
Stefan

Thanks, Stefan. Well, I think my images have been riddled with flaws for some time now. And let's not forget I just went through my trianglular stars phase, before hopefully coming up with a fix for that. I'm just too lazy to want to spend much time trying to experiment with tilt and coma corrector spacing. Or, lately, too busy with work. But I really need to spend time on it.

The last couple targets, I think that imaging both sides of transit and stacking helped round off some, but not all, stars. There's also a matter of just how much downsampling is done, both in ST and then later as part of jpg export in Gimp. The more things are kept closer to original scale, like say the astrobin link, the more flaws are apparent. Bit of a trade off, as scale helps one see the mini background galaxies too.

Here, I think I am more eggy towards the lower left, which is also where NGC 5053 is. I'm going to theorize that I have a shifted collimation, as I noticed (in both lights and flats) that the centerpoint of the vignetting field is displaced towards the right side. Thus, I may be mirror-aligned, but that alignment is angled with respect to the OTA? Maybe that means a have a slight rotation in the secondary. Hopefully some work with the sight tube clears that up and centers my field again on the sensor.

admin wrote: ↑Thu Apr 20, 2023 10:40 am Wonderful! Is that really nice framing on purpose, or just a happy accident?

Thanks, Ivo. Despite the aforementioned flaws, there is a nice crispness to the stars here, which I was hoping for. A victory for OptiDev and SVD.

Framing is intentional, I gave up the accidental a while back. Stellarium is used for a FOV and camera rotation plan. I have marks on my focuser drawtube so can start out close. After TPPA and B-mask focusing, I used the NINA plate solving to get camera rotation nailed down. Then the NINA framing wizard, with the offline space survey turned on, lets me drag the FOV rectangle around until I'm happy with it. Hitting a button then sends the center coordinates to the sequencer (simple or legacy sequencer for me), and I flip on the toggle for plate-solved slewing to be the first step once I hit go. I'm not exactly sure what I have my tolerances set to, but I think it's around 50 pixels RA and DEC. Inside of that, NINA will then fire up PHD2, let it settle, and start taking pictures.


Edit to add to framing this particular target: There are a number of good possibilities here that include some nice colored stars of medium magnitude. Just outside my FOV, top of frame I think, is a real nice lineup of stars. I considered various framings and rotations including those, but instead just went for the two globs somewhat offset, and inclusion of the mag 4 Diadem, and a decently strong red star nearby. With Diadem seeming (by some indices) to have a white-blue tint, I thought that would be a good display of varying colors.

Mike in Rancho wrote: ↑Sat Apr 22, 2023 4:56 pm And let's not forget I just went through my trianglular stars phase, before hopefully coming up with a fix for that. I'm just too lazy to want to spend much time trying to experiment with tilt and coma corrector spacing. Or, lately, too busy with work.

I totally get that, Mike. I am not one of the guys being obsessed with star shapes. I just thought it was unfortunate that the region of oblong stars coincided with the cluster.

Mike in Rancho wrote: ↑Sat Apr 22, 2023 4:56 pm Thus, I may be mirror-aligned, but that alignment is angled with respect to the OTA? Maybe that means a have a slight rotation in the secondary. Hopefully some work with the sight tube clears that up and centers my field again on the sensor.

Sounds reasonable...have never considered the possibility. I see that in my flats, too. Some day I should invest in a better collimation/alignment tool beyond the laser...

Anyway and again: cool image!

Regards
Stefan

Stefan B wrote: ↑Sun Apr 23, 2023 7:16 pm
I totally get that, Mike. I am not one of the guys being obsessed with star shapes. I just thought it was unfortunate that the region of oblong stars coincided with the cluster.

Sounds reasonable...have never considered the possibility. I see that in my flats, too. Some day I should invest in a better collimation/alignment tool beyond the laser...

Anyway and again: cool image!

Regards
Stefan

Hey Stefan,

Nothing wrong with noticing a glaring issue, in fact it bugged me too! I did take apart the camera train and changed the spacers. I'm now at 56.7 from the CC to the sensor. Also used the autocollimator a bit more carefully to try to really get things aligned, right before going out to take some test star shots. I think it helped. I'm still not perfect across the entire APS-C field, but it's mostly the long side edges and corners, and really just one that kind of stands out. Hopefully I can keep that up and have reasonably correct stars. At least in the important locations.

Took another shot at this data with the new 549 release. Ran it through at full resolution also to see whether that would help or not with the ringing issues. Not so sure it did, but boy did it reveal my optical flaws. Hopefully my backspace changes have that fixed for the next time I get out.

Crop is slightly different since I just did one click on each slider, and I might have stretched a bit more? Also within HDR I think I kept a little more strength towards the M53 middle core, and seems to give a better look.

I still couldn't avoid all the ringing/shadow artifacts, despite centroid tracking and bumping deringing to 75. That seems to be about as far as I can push that setting. Any more, and it kind of becomes pointless (he he) as the stars actually grow rather than resolve inward. So I added no-iteration deringing in Shrink, which seemed to help clear the remnants up.

I suppose some of it could be dud sampling? There were a fair amount of choices, but not really a ton. Most of the field was widely scattered little white chips that seemed star-free. As to the others, I'm still uncertain what SVD is using for determining the PSF - the same regime as 1.8 where the sampled area is that which is within both the white outline and the blue box? Makes me wonder if I am sampling wrong and cutting off the outer PSF edges.

In any event, once the final is resampled back down to a postable size, things aren't really too bad.

Mike in Rancho wrote: ↑Sun May 07, 2023 8:50 pm Took another shot at this data with the new 549 release. Ran it through at full resolution also to see whether that would help or not with the ringing issues. Not so sure it did, but boy did it reveal my optical flaws. Hopefully my backspace changes have that fixed for the next time I get out.

Crop is slightly different since I just did one click on each slider, and I might have stretched a bit more? Also within HDR I think I kept a little more strength towards the M53 middle core, and seems to give a better look.

I still couldn't avoid all the ringing/shadow artifacts, despite centroid tracking and bumping deringing to 75. That seems to be about as far as I can push that setting. Any more, and it kind of becomes pointless (he he) as the stars actually grow rather than resolve inward. So I added no-iteration deringing in Shrink, which seemed to help clear the remnants up.

I suppose some of it could be dud sampling? There were a fair amount of choices, but not really a ton. Most of the field was widely scattered little white chips that seemed star-free. As to the others, I'm still uncertain what SVD is using for determining the PSF - the same regime as 1.8 where the sampled area is that which is within both the white outline and the blue box? Makes me wonder if I am sampling wrong and cutting off the outer PSF edges.

The sampled area/blob is outline by a boundary. In 1.9, anything that falls within the box that is not part of the outlined blob, will not be part of the sampled area and you can safely ignore that now.

Getting good samples is fairly critical. I'm currently exploring some ways to make those easier to evaluate/see/pick, as well as techniques to avoid ringing earlier in the process. Be mindful that, depending on your dataset, using samples is not always necessary/optimal; a "perfect" synthetic Moffat profile may yield better results if your optics and tracking were otherwise close to ideal.

admin wrote: ↑Mon May 08, 2023 9:35 am The sampled area/blob is outline by a boundary. In 1.9, anything that falls within the box that is not part of the outlined blob, will not be part of the sampled area and you can safely ignore that now.

Getting good samples is fairly critical. I'm currently exploring some ways to make those easier to evaluate/see/pick, as well as techniques to avoid ringing earlier in the process. Be mindful that, depending on your dataset, using samples is not always necessary/optimal; a "perfect" synthetic Moffat profile may yield better results if your optics and tracking were otherwise close to ideal.

Thanks Ivo. I remember it being mentioned as one of the 1.9 improvements, that we could go forward with samples despite holes or protrusions. A kind of interpolation? In any event, I was wondering if the same concept applies to the outer reaches of a stellar PSF. Say you have a sizeable white starfishy outline, but the blue box is small and visibly does not cover that entire stellar profile - SVD's PSF would not just fall off the edge, I imagine?

I do try to avoid synthetic what with SVD being superior. Synthetic usually requires some fiddling with the controls (increasing), yet can show more raised background noise right from the get go. The pinpointing of SVD is great, it's just that pesky ringing or shadow shapes thing.