Newbee question

dackerman1

Member
Feb 12, 2021
10
2
Hello all,

I recently started using PI a few weeks ago. Today I processed new data acquired a few days ago. Previous projects came out pretty good but this one... I can't seem to get rid of this purple background.

So far I've applied 4 processes
*DBE
*PCC (including Background Neutralization)
*SCNR
*MLT


Any help would be appreciated.
Thanks,
Darryl
 

Attachments

pfile

PTeam Member
Nov 23, 2009
6,340
210
you'll have to post the unprocessed xisf for anyone to really comment on what to do or what's wrong. you can upload to a file hosting service and then post the link here.

rob
 

dackerman1

Member
Feb 12, 2021
10
2
Here's a link to the master RGB.
m81_m82
It's been processed minimally. I cropped the RGB masters using DC and applied DBE. I did it in this order because I was creating a pseudo luminance. I cannot seem to correct the color later on in processing, even after doing BN and CC. I've also tried skipping these steps and doing PCC instead.
Stretching only turns it a deep purple. I thought maybe it was the way I calibrated, aligned, integrated etc. but I stacked using DSS and the results were nearly identical.

Anyhow... any suggestions are welcome.
 

pfile

PTeam Member
Nov 23, 2009
6,340
210
probably your settings in BN were not right. you need to set the threshold value high enough that it captures the brightest background channel.

in the file you posted, the RGB values in the background are:

Screen Shot 2021-02-26 at 10.07.05 AM.png

so in BN (or PCC's background neutralization area) you should set the upper limit slider to about 0.019. additionally you can define a preview to restrict the tools to an area which is 100% background.

if you do DBE on the combined RGB with defaults set (meaning "normalize" unticked) DBE would have most likely left the background neutral. but since you did the DBE on the channel masters, the channels in RGB are out of whack. that's not to say that doing DBE on the individual masters is wrong, in fact, it can probably be a little more accurate to do it that way.

anyway here is the image after PCC and some stretching and curves:



Screen Shot 2021-02-26 at 10.25.30 AM.jpg

the galaxy colors seem correct but the stars are very blue. most likely this is due to having used a refractor which needs a UV cut filter somewhere in the optical train (if you google 'blue bloat refractor' you'll get a bunch of hits.)

there are definitely still some gradients in the background. i don't know if this is due to bad flattening or light pollution. it's not so much that it can't be hidden by pulling down the background a bit.

rob
 

dackerman1

Member
Feb 12, 2021
10
2
probably your settings in BN were not right. you need to set the threshold value high enough that it captures the brightest background channel.

in the file you posted, the RGB values in the background are:

View attachment 10429

so in BN (or PCC's background neutralization area) you should set the upper limit slider to about 0.019. additionally you can define a preview to restrict the tools to an area which is 100% background.

if you do DBE on the combined RGB with defaults set (meaning "normalize" unticked) DBE would have most likely left the background neutral. but since you did the DBE on the channel masters, the channels in RGB are out of whack. that's not to say that doing DBE on the individual masters is wrong, in fact, it can probably be a little more accurate to do it that way.

anyway here is the image after PCC and some stretching and curves:



View attachment 10430

the galaxy colors seem correct but the stars are very blue. most likely this is due to having used a refractor which needs a UV cut filter somewhere in the optical train (if you google 'blue bloat refractor' you'll get a bunch of hits.)

there are definitely still some gradients in the background. i don't know if this is due to bad flattening or light pollution. it's not so much that it can't be hidden by pulling down the background a bit.

rob
Rob thanks for helping me out. I tried that and it worked out very similar to your results. I'm knew to RGB imaging and PI. I come from a Canon DSLR and Photoshop. Now I'm using the ASI294MM Pro. Great camera but I have huge learning curve ahead of me.

I think most of the gradients are due to light pollution, class 6 bortle, and if I remember the moon was about 18% illuminated and setting about 30 degrees above the horizon, but a lot of it is from vignette. I took flats a couple times using the NINA flats wizard to see if that was it. I was thinking about doing the bias over again too. Oh and I do have a UV/IR cut filter I used on my Canon so I think I'll try and incorporate it into my setup as well.
 
  • Like
Reactions: pfile