PixInsight's new WBPP2.2 Debayer separate RGB Masters for OSC data?

[oymd]

I have been using PI for the past year, and have been processing my OSC data from my ZWO ASI294MC Pro.

I have NEVER processed Mono data before, meaning that I have never worked with R,G & B separate channels before.

With the new WBPP 2.2 which has this new option of Debayer RGB Separation, can some one please give me a broadline how to process my data moving forward?

With my current processing work flow, after the regular WBPP, I get a MASTER RGB file.

I do  Dynamic crop, DBE, PCC, EX Denoise, a tiny bit of Arcsinhstretch followed by the Histogram stretch.  On the NON Linear image, I usually then do Starnet, and work with the starless image: Curves and colour saturation, and possibly some HDRMLT and TGV denoise.

Deconvolution is according to the target.

With the NEW WBPP2.2, I now have 3 separate channel Masters.

Do I follow the same processing workflow on the separate R,G & B Masters? Do I do my regular Dynamic crop, DBE, PCC, EX Denoise, a tiny bit of Arcsinhstretch followed by the Histogram stretch on EACH Master separately?

If so, at what stage do I combine them in RGB combination? When they are in LINEAR or NON LINEAR stage?

Many thanks

Ossi

[scotty38]

Apologies if I misunderstand the intent of your post but I assume you realise that you don't have to take the separate channel outputs, you can just take the RGB as normal, or indeed both as required?

[oymd]

39 minutes ago, scotty38 said:

Apologies if I misunderstand the intent of your post but I assume you realise that you don't have to take the separate channel outputs, you can just take the RGB as normal, or indeed both as required?

yes, of course.

But I wanted to try working on the separate channels, as my latest M33 project has tons of LP, and working on separate RGB channels might help me process better

[scotty38]

ah ok understood

[daz]

I would suggest:

• DynamicCrop
• MureDenoise
• ChannelCombination
• PCC
• ...

DBE on the separate channels is essentially what happens with DBE anyway, so you may as well combine beforehand. There are discussions about PCC before DBE or after - technically PCC does not introduce gradient so it should not make a difference. I have tried both ways and can't  tell if there is any difference between either way!

Remember that if you are using binned colour and unbinned luminance, you'll need to get WBPP to align based on the post-keywords, not on a on a global reference. If not, then PI will resize and resample the colour images based on the luminance reference. Then things like MureDenoise won't work...

 

It's a great tool now - means you can pre-process multi-night data sets easily!!

 

[oymd]

4 hours ago, daz said:

I would suggest:

• DynamicCrop
• MureDenoise
• ChannelCombination
• PCC
• ...

DBE on the separate channels is essentially what happens with DBE anyway, so you may as well combine beforehand. There are discussions about PCC before DBE or after - technically PCC does not introduce gradient so it should not make a difference. I have tried both ways and can't  tell if there is any difference between either way!

Remember that if you are using binned colour and unbinned luminance, you'll need to get WBPP to align based on the post-keywords, not on a on a global reference. If not, then PI will resize and resample the colour images based on the luminance reference. Then things like MureDenoise won't work...

 

It's a great tool now - means you can pre-process multi-night data sets easily!!

 

Many thanks

I have never heard or seen any prior reference to Mure Denoise before?

Where is that? Cannot see it in the processes?

For denoise I initially used MLT with an inverted linear mask when I started using PI last tear. 
 

Recently I have moved to the EZ Processing suite; EZ Denoise. 
 

Is Mure Denoise superior?

Many thanks

Ossi

[daz]

MureDenoise is in the Scripts menu - I'm pretty sure it;'s there by default, I don't think you need to add another repository...

There are two parts to it. First part is to measure the images using two flats and two bias or darks. This gives you values that you can put into MureDenoise (when I get to my PC, I'll do some details and screen shots for you).

It's just enough to knock the fizz off the noise....

 

Although you could just use MurerDenoise on the Lum data, and then do MLT + MureDenoise on the chrominance....