HDR combination creates red background and misshapen stars

[ramv]

Hi,
I am a relative newcomer to Pixinsight (have been using it for about 3 months now: a recent convert from pure Photoshop). I have gone through all of Harry's tutorials and processed numerous images with PI with excellent results.

I ran into a problem recently where HDR combining a set of images gives a RED background and colored/black stars. I have used HDR combine in the past with no problems at all. In fact one set of images works fine but another gives the red background and colored stars. The only difference I can think of is that the one that doesn't work is a wide angle (70mm from camera lens). The other is a 530mm telescope and works fine.

I am happy to provide my files as samples if that will help identify the issue.

Thanks for any help in anticipation.

[georg.viehoever]

HDR combine does not depend on having small/wide angle images. I guess that the color tone is due to different filters or color characteristics of the different lenses. You should be able to obtain good colors after doing BackgroundNeutralization and ColorCalibration (in the ColorCalibration section of processes).

Georg

[georg.viehoever]

PS: the misshapen stars that you mention in your subject may be from the wide angle lens. StarAlignment is not very good at compensating the distortions caused by such a lens, see its documentation. Workaround: Use DynamicAlignment.
Georg

[ramv]

Georg, Thanks for your input. I will try this out.
Will Dynamic alignment combine images correctly for HDR, i.e. preserve relative brightness of the images?

[georg.viehoever]

DynamicAlignment only aligns the images. After alignment, you combine them using HDRCombine.
Georg

[ramv]

I have struggled with this  for days now and each time it is looking worse

I managed to get the stars aligned by registering all of the frames against a reference frame while stacking with DSS. PI doesn't make mistakes anymore. Now I am seeing a red cast on the 64 bit image that I am unable to remove with background neutralization.

I have uploaded the tif files for each of the HDR images. I would highly appreciative and grateful it if someone took a crack at it and educated me on how to process this image.

Here are links to the input images.

https://s3.amazonaws.com/Astrophotos/Pixinsight/Barnards+loop+30+secs_dl001.tif
https://s3.amazonaws.com/Astrophotos/Pixinsight/Barnards+loop+3mins_dl001.tif
https://s3.amazonaws.com/Astrophotos/Pixinsight/Barnards+loop+5+secs_dl.tif

Best regards,
-Ram

[georg.viehoever]

I am not really sure where the problem is:
- did HDR composition with default settings using your images
- applied strong HistogramTransform twice
- used STF with decoupled channels to get rid of most of the color problems
Result: See screenshot

My guess is that you are trying to view the result image with STF only. This is beyond the capabilities of STF (that has limited dynamic range due to speed). That is why I apply the HistogramTransform before doing anything else.

Georg

[ramv]

Georg thanks but I am still unable to proceed.
Something is very wrong with my processing/machine. I can get good results if I operate on just one of the TIFs (say the Barnards loop 3 mins) But the moment I try to do anything with the HDR image, it turns to complete junk. DBE just does not work at all.
I tried other things: Converting the 64 bit image to 32 bit and then processing. No luck. The image swings from too much red to too little red. The histogram for red channel is a thin spike whereas for other channels it is dispersed across entire x axis. So I am unable to stretch at all.

I am running this on a somewhat underpowered netbook (but it has 8GB RAM). I can't imagine that would make a difference, would it?

[georg.viehoever]

I did the processing above on an old 4GB notebook... Could you post a screenshot of your problems? I do not see an issue with the result that I got.
Georg