Best way to combine 600sec and 300sec subs/masters (not becuase of desired HDR outcomes)

auroraTDunn

I'm in the middle of reprocessing a bunch of projects and I have a funky one from last Fall where I started imaging (H,S,O) with 600sec subs but after a couple of frustrating events I did the second half of my imaging with only 300sec subs (this was a sanity issue and nothing else really).

Originally I processed with SirilIC and had no issues integrating the two together. But now that I'm back to using WBPP I get two masters for each, obviously.

So my question is what is the best way to combine them into 1 master for each channel?

 - Exposure Compensation in WBPP, but for double the time, is that really smart? For short deviations it makes since but its a huge jump from 300 to 600sec.

 - HDRComposition using the two masters (I've tried this and its OK but logic makes me think I'm loosing out on the benefits of my longer exposures.

 - Doing some pixelmath so manually add them but it would not be a straight 300+600 that would be wrong. So it would have to be, what a combination of some weight for 300 plus some weight for 600 noting the w300 +w600 = 1 (so normalized) but then what use the mean or median? That seems incorrect as I would end up with something similar to HDRCombination. Variance also does not sound correct mathematically. That leaves using a standard deviation  which might be OK??? But how would I calculate the weights between the two to begin with especially since it seems to my thinking that I should have a clean linear fit between the two, at least for the background before I even begin the process.

Note I have a hefty amount of time on this particular project with 2/3 of the total exposure time pe rchannel at 600sec and only 1/3 at 300sec.

BlockHead

Yes, it is smart to combine all of the different exposure times together with one stack. You do need to adjust the exposure tolerance to allow for this (instead of getting multiple masterlights at each exposure time).

The images are weighted according to their signal (quality, depends on the weighting method).

No to the HDRComposition. This is basically reserved for the Orion Nebula.... and I argue even that is lame.

The pixel math is the crude way to weight based on exposure time...but that is the optimal weighting scheme. Please review my material (so many videos!) on the weighting methods. I think you will find the information enlightening now that it actually effects you. That is the issue with my content... people generally do not want to learn things *until* it is actually necessary. :)

WBPP will take care of all of the weighting between exposures. It so happens I (and another) pushed Juan to develop the plethora of options that are now there. 

-the Blockhead

auroraTDunn

Thanks Adam. I was researching this further and found info on how to set this up and have WBPP running right now. It's about half way through so will check on it in the morning.

I know you talk about this in one of your videos but I somehow couldn't remember which one. I need to go back and watch your new WBPP video sometime tomorrow :)

Need to get through this images so I can treat myself to your comet series. I counting that as my reward for finally reprocessing my projects from last year (I used Siril to learn it, mainly because WBPP was crashing endlessly on my old Windoze workstation. Updated PI and workstation and now life is grand again. Not sure which fixed the problem and frankly only care that it works now!

auroraTDunn

Just a follow up just in case anyone reads this for their own solution(s).

I tried the following.

- HDRComposition - worked(ish) but not a decent result considering the unstacked potential.

- Duplicating my 300 and 600 masters and then using all four with ImageIntegration, this also worked but based on this data I preferred the HDRComposition

- A nasty bit of experimenting using Pixel Math to combine them (the worst of these attempts)

- Went back into WBPP and on the "Post-Calibration" tab set the "Exposure tolerance" to 304. Initially this failed completely during the measurement stage. Purged the cache and re-ran and it still failed. In the end I purged and then completely deleted my WBPP output directory and re-ran again. This worked!

Without question reprocessing in WBPP is way and by far the best solution! I'm VERY happy with the results!