PIPP / AS!2 Image Distortion

[SnakeyJ]

I've been using PIPP in join mode to merge consecutive 4200 frame SER files of Saturn (grading, re-ordering and saving the best 20% of frames as an SER file for AS2. This seems to work about half the time, but when I subsequently stack in AS!2 I get some odd image distortion as below:

However, I can open the output SER in R6 and stack without issue - so it seems that PIPP is playing nicely. AS!2 works fine with the source files and does work half the time with the PIPP output files - am I perhaps being too agressive with the number/size of APs or stacking options here?

[cgarry]

Hi jake,

I have never seen this before but I wonder if AS!2 is being thrown by the relatively bright background that I think I can see in your frames? When PIPP crops and centres it fills new pixels with true black which looks to be somewhat darker than the actual background of your frames.

If you use your usual options but disable these options:

Processing Options->Object Detection->Centre Object In Each Frame

Processing Options->Cropping->Enable Cropping

Then the planet will bounce about all over the place but AS!2 should (I believe) deal with this. However, you will not have any true black background added because the centring operation is disabled and AS!2 might be happier. If this works I shall look at the possibility of filling the background with a colour that matches the background colour rather than just using black.

Cheers,

Chris

[SnakeyJ]

Hi jake,

I have never seen this before but I wonder if AS!2 is being thrown by the relatively bright background that I think I can see in your frames? When PIPP crops and centres it fills new pixels with true black which looks to be somewhat darker than the actual background of your frames.

If you use your usual options but disable these options:

Processing Options->Object Detection->Centre Object In Each Frame

Processing Options->Cropping->Enable Cropping

Then the planet will bounce about all over the place but AS!2 should (I believe) deal with this. However, you will not have any true black background added because the centring operation is disabled and AS!2 might be happier. If this works I shall look at the possibility of filling the background with a colour that matches the background colour rather than just using black.

Chris,

Absolutely spot on and this now stacks perfectly in AS!2 - very big thanks for this and for your extremely active work on this software (just realised you had another 2 releases out since I last updated to 2.28.

I shall have to watch out for the backgrounds in future, though think I had quite a bright moon perhaps 25-30 deg west of Saturn during this session on 21-22nd May.

Many thanks again - Jake

[cgarry]

Excellent news. I think the combination of a bright moon, hazy skies and the high gain required for Saturn makes bright frame backgrounds inevitable at times.

Obviously AS!2 does a centre of gravity type centring operation and this is being thrown out by PIPP's background filling operation. Come to think of it this could be the cause of a slight misalignment issue I have been occasionally seeing in my own processing using AS!2, well it would be nice if this is the cause of it.

Anyway, I shall update PIPP to calculate an average value for the frame background and use this value for the filling operation instead of just using black. That is nice and easy to do and should solve this issue.

Cheers,

Chris

[SnakeyJ]

Chris,

Thanks again Chris, at times this forum and community just amazes and this is one such occassion.

Though I love imaging the moon, it does seem to intrude and outshine most of the sky when least convenient, which seems to be just about most occassions that the clouds part to reveal the night sky

I shall have another go at combining my channel captures and see if I can get a decent result out of Winjupos tonight!

[cgarry]

You are very welcome Jake and thank you for posting your question. Believe it or not, one of the hardest things about working on a project like PIPP is not the coding itself but actually finding out about problems users are having with it. So it is very useful to me when people like yourself take the time to ask questions about problems they are seeing.

I have to echo your comment about this forum and community, the willingness of the members to help and share knowledge is amazing and it is great to be a part of it.

Good luck with WinJUPOS!

Cheers,

Chris

[stargazerslivingroom]

Great Saturn, Jake and thanks for a really useful thread. Hadn't heard of WinJUPOS...That does look interesting!

Steve

[SnakeyJ]

Thanks to Chris for his fantastic response, work around and an unbelievably quick development release - Although this wasn't actually a bug in PIPP, he's amended the software to take an average colour for the background rather than adding pure black pixels around the centred frames, compensating for the bright backgrounds in my dodgy captures - This now works beautifully, joining 3 successive 4200 frame SER captures for each RGB channel and selecting/ordering the best 3000 frames. The resulting files stacks without issue very quickly in AS!2 and is giving much nicer results (more detail/less noise).

For anyone who hasn't yet had a look at PIPP, I would strongly recommend taking a good look @ https://sites.google.com/site/astropipp/ - I believe its also available on OSX and Linux via Wine!

[SnakeyJ]

Just for completeness, here's the finished red channel image from the original source SERs above, best 3000 frames of 8400 sorted and graded in PIPP ver 2.2.11 development release and then 2000 frames stacked in AS!2 for final result:

No post processing other than AS!2's own sharpen routine.

[cgarry]

Cheers Jake.

I have now uploaded the v2.2.11 release of PIPP which has the changes to fix the misaligned frames in AS!2 problem plus extra options for colour to monochrome conversions that has helped with my lunar and solar image processing.

That is a great result with the red channel Jake, very sharp for just the auto-sharpened image and no wavelets. Hopefully that will be the start of a cracking Saturn image.

Chris

[SnakeyJ]

That is a great result with the red channel Jake, very sharp for just the auto-sharpened image and no wavelets. Hopefully that will be the start of a cracking Saturn image.

Think I could get a great image out, if you can only write a routine to clean up the mush that is my blue channel! Seriously though the blue channel is proving difficult due to intensity and poor seeing - and although I appreciate this is difficult for most, particularly at 8" aperture, I will need to spend some time tweaking the camera/capture settings in firecapture to get optimum results. The FC gain slider control is somewhat odd/non-linear with the last 1% increase seeming to operate a 2x multiplier. Though despite these oddities this is a very sensitive little camera and an absolute joy in comparison to the firefly.

[cgarry]

Ah yes, the joy that is the blue channel! The 'Noise Filter' control on the 'Processing Options' tab in the latest version of PIPP was added in an attempt to get better results from the blue channel. I don't think it has really helped but it is there to try if you want.

The only way to improve the blue channel that I know of is to use an ADC. It gives some improvement to the blue channel and also potentially allows a true L channel to be used which is obviously much brighter and can therefore be captured with a shorter exposure.

Cheers,

Chris

[SnakeyJ]

Chris,

I was only asking tongue in cheek and hadn't previously noticed this function in processing options - but I have tried and it definitely gives an improvement. I've tried it on a range of my blue channel captures and obviously the more noise the more pronounced the effect, but it seems particularly good at removing speckling and background noise from the image, but does not seem after stacking I can't detect any noticeable loss of detail though contrast looks boosted. A lot of my blue channel capture is very marginal, but here's a comparison of one of the better captures processed in PIPP (same options as prev Red Channel, but with Median Noise Filter Disabled and Enabled).

I think that shows a marked improvement and will definitely do some more comparison on future captures. The feature definitely gets my vote so please leave in for the immediate future! Only wish my other blue channel captures were as respectable

I am considering an ADC, after first seeing one come up on ABS 6-8 weeks ago. I've done a little research on this and been keeping an eye on Emilf's output from Craiova, Romaina - He uses an 8" F6 dobsonian, ASH ADC and DBK21AU618as and has produced some stunning results at this aperture, so an ADC is definitely on my wish list - hopefully in time for Jupiter in Dec/Jan! It would certainly improve my luminance channel at these low elevations, which is bright but diffuse. I also use IR-PASS (Astronomik's PP742), which works but my best results seem to come on the Red channel which is a lot brighter and allows higher frame rates - RRGB for me for the time being.

[cgarry]

Wow, the noise filter approach does seem to have worked for your example. Maybe this approach does warrant further investigation after all.

The trouble with the IR pass for luminance approach is the planet does not look the same in IR as it does in visible light so you end up with a fake looking image. Though I do think that there is possibly scope to bring out detail in Saturn's rings using IR as luminance if not the planet itself.

I was talking to Martin when he was selling that ADC on ABS and neither of us could not believe that it did not sell instantly. They really do make a difference yet their general take-up by imagers seems to be quite low. I would look at the Pierro-Astro ADC now as that is cheaper and also works for UV, which you might want to use for Venus.

Cheers,

Chris

[SnakeyJ]

Wow, the noise filter approach does seem to have worked for your example. Maybe this approach does warrant further investigation after all.

The trouble with the IR pass for luminance approach is the planet does not look the same in IR as it does in visible light so you end up with a fake looking image. Though I do think that there is possibly scope to bring out detail in Saturn's rings using IR as luminance if not the planet itself.

I was talking to Martin when he was selling that ADC on ABS and neither of us could not believe that it did not sell instantly. They really do make a difference yet their general take-up by imagers seems to be quite low. I would look at the Pierro-Astro ADC now as that is cheaper and also works for UV, which you might want to use for Venus.

I shall do some more testing with the noise filter on some older data to see how it looks and fingers crossed I should be able to get some fresh data over the w/end.

Thanks for the information on Pierro Astro ADC - this looks very interesting and the price isn't too silly in comparison to many things in this game (as its a small object I might just be able to slip it in under the radar).