[Sticky] Colored centering of the stars, align channels

This is with one-shot color or using filters?

@vincent-mod

I only used the Astronomik L2 UV-IR Block 2″.

Mm, difficult one for me, you could try to experiment with the pixel interpolation in tab 6, unfortunately I have not a good understanding of which one would help you in this case so it requires some trial and error. If that doesn't change anything, I'm happy to take a closer look at the data later.

I tested the pixel interpolation all through. But it brought no improvement.

I think that comes from the atmospheric dispersion. If the object is not so high.

Would be cool if APP could optimize that. Since this is currently not possible with any other program. As far as I know.

Hello Daniel, try experimenting with the values in diffraction protection. Start the integration with automatic settings and then read the value in the terminal output. Then stop integration and increase the value manually. If that doesn't help, try adjusting to values between 2 and 5. Depends on the optics used.

Maybe play with rejection algorithms.

I hope this helps.

With best regards Henry

Hello,

I experimenting with diffraction protection. But with no attitude I had an improvement.

But I found a way with APP.

So that it goes quickly, I have only stacked 10 pictures.

In tab 2 "Calibrate" I checked "Align channels" at the bottom.
Then all images were saved as a calibrated FITS file.

After that I reloaded and stacked these FITS in APP.

Without "Align channels"

With "Align channels"

It would of course be great if APP could do this process in one. So you don't have to cache and reload the images.

Ok, that is a really good call there, it seems it's actually meant specifically for this. Sorry I missed that, I've never had this issue and I also don't save my frames in between so I never saw the explanation to that tool. Thanks for experimenting and letting us know! I'll ask if this can be done more in line with the regular workflow.

Unfortunately, this is not possible without intermediate storage.
If the box "Align channel" is checked and the integration is started directly, the effect will not be visible.

Here again with 6.5 hours exposure time reworked.

Without "Align channels"

With "Align channels"

The M104 is very low at my location. (20-25 ° altitude) This shows the effect well.

Yes it's not possible right now. I've spoken with Mabula about this and it's not as easy as it may seem to integrate that function in the workflow. It is on his to-do list though so it should be possible in a future version.

I did the same thing with the galaxy from the first post.

Hello Daniel, just saw your Youtube video on the subject. I hadn't thought about the option Channel Alignment, I have activated it by default. If you want to combine images from different cameras, you can still get colored artifacts around stars, even if you use the Channel Alignment option. I had this problem when I combined DSLR images with images from my ZWO CMOS camera. Then the diffraction protection helped. 😉 

M.f.G.

Hello, 

ok cool. Thank you 🙂

For the others... here is the video. But it is in German.

Hi,

That's very interesting. Thank you for the feedback.

Have you tried the channel alignment with the final integrated image ?

David

Hi,

No, I didn't try that. But think that the quality is better to align the individual RAW images.

I've made this thread into a sticky, great stuff. I'll also go through the other stickies that are outdated and such.

Can this be done for the integrations of each filter? Could I do my normal workflow, where at the end I get an integrated image for each filter, and then reload those files back into APP as lights, and then check "Align Channels" and then Save Calibrated Images, and use those in the RGB combine tool?

Mmm, good question. I never tried that, but my guess would be that you should be able to do that when it's RGB in the end.

I definitely see this shooting mono through filters - I will give it a try and see if this workflow solves my problem as well.

Thanks for this sticky post! I have always ticked the box thinking it would already do that in auto-mode... Good to know, how it really works.

Useful technique. another powerful option in APP that is not as obvious as maybe it could/should be.

I find it almost always necessary using quad band filter with OSC to fix color issue with red and blue sides of stars from those filters. Luckily, my broadband OSC data never had this issue yet. But always with filters and it works quite well, although takes a long time to save all the aligned images. 

Reviving this thread...

I was under the impression never had this issue in APP before, but a recent M31 image stacked with 2.0.0 beta 6had this RGB misalignment. Thanks Daniel for the great video, that workaround resolved this problem. But I cannot confirm this has anything to do with atmospheric refraction. M31 was 60°-80° up in the sky and I had massive misalignment(2-3 px shift between each R,G,B). This issue also showed up in both clear filter and duoband filter stacks, so should have to do with the optics used neither.

However 2 1/2 years later this bug still doesn't   seem to be targeted. Before I bought APP, I used DeepSkyStacker which had an Align channels function if I am not mistaken. I am not sure if there is any reasonable use case where one would want an RGB channel misalignment in stacking. Anyway, it is clearly a bug as the "align channels" function only works when saving&reloading calibrated files. It does NOT work as an option of the usual workflow though, giving the user the incorrect expectation of a channel aligned result.

When the box is ticked, the channels should be aligned in the stack. If system restrictions would require a save/reload to do so, the user should be alerted (and/or system should automatically deroute such path using temp saved calib frames). I feel the worst case option is to have APP simply spit out a faulty result without any user alert or instruction how to circumvent that.

Clear Skies,

Jochen

I agree that it should be clear that align channels does not, in fact, align channels unless you take specific extra steps.

Nearly all of my OSC stacks have channel misalignment even if I’m shooting at fairly high altitudes so there really is a fundamental need here for OSCs.  So it would be a great benefit to me if the default one click “integrate” option allowed align channels to operate.

Perhaps I’m missing something where the default workflow should not attempt to align the color channels.  Possibly it’s an architectural limitation in how the overall system was designed and didn’t include OSC channel alignment as a fundamental basic step.  That would be unfortunate, but understandable at this point.