PI washed all my color away! (new guy) [SOLVED]

[prefetch]

hi,

i'm a new guy - so please forgive me if it sounds like i'm kind of panicked and if this is something that is common.   

so, i used the batchpreprocessing script and everything seemed to work fine.  here's a shot of the relevant settings:



everything came out kind of green, but i continued doing more processing and i went ahead and used the SCNR process to de-green it, and that seemed fine - but then i started to notice that the color was pretty much all gone!

i had processed these frames in nebulosity previously, and it seemed to handle the debayering just fine. 

here is a frame from the nebulosity debayer (on the left) and the PI 42 frame stack, debayer and SCNR on the right:



so...i've read some documentation on the PI color philosophy, but that didn't really seem to help me.  i just don't get why the default debayer in PI is washing away all my color.  i tried to tweak the colors, but they are basically gone.  there just aren't many pixels left with any color in them.  they are all white-ish, and light colored.  i exported it to photoshop, and it's pretty absent of color information, even though it's technically not greyscaled.  all the color is just washed out.  it's weird.

is this just the way PI works?  am i doing something dumb?

[pfile]

can you post the integrated .fits file somewhere?

only think i can think of is to try unchecking up-bottom fits, but when this is wrong i expect the colors to look wacky, not be gone.

rob

[prefetch]

here is the integrated fits file (63MB):
https://dl.dropboxusercontent.com/u/4404304/washedout_m101.fit

and here is a frame from the original set of light frames (12MB):
https://dl.dropboxusercontent.com/u/4404304/Autosave%20Image%20-021m101_400_.fit

[Alejandro Tombolini]

Try GRBG as debayer pattern.

Saludos, Alejandro.

[prefetch]

i've tried all the patterns - i'm pretty certain the RGGB one is the correct one.

here is GRBG after star alignment and integration:

[pfile]

alejandro is right; GRBG is correct as far as i can tell, for my fits reader orientation (up-bottom).

load that single sub and debayer with GRBG, then open STF and unlink the channels and apply the STF, you'll see that the color is right.

rob

[prefetch]

okay, i must be doing something wrong.  perhaps there is something wrong with how i made the master darks, flats and biases?

again, forgive me for my ignorance - but how do i debayer a single frame?  i'm assuming it's not by using the batch preprocess process (i tried, but it's not happy with a single frame..!) and it's probably best to duplicate what you guys are seeing and then i can start working on finding out where i'm going wrong with the batch preprocess function.

ok, i found the non-batch debayer routine (getting used to how to find things in PI) and you're right - it looks good.

so clearly i'm screwing something up with my darks, biases and flats.  i'll work on that and try and figure things out.

thanks!!!   

[pfile]

okay, open the single sub, then go to Process > Debayer

the Debayer process will open. set it up for GRBG and apply.

there's some chance your fits reader is set backwards from mine, so i think this will affect the debayer pattern. at any rate if you open ScreenTransferFunction and unlink the channels, then apply to the debayered image, you should see something that looks approximately right with a grey background. if not, just go thru all the debayer possibilities and one of them should work. 2 of them will give a weird screen door look in the output and one of them will have a yellow background, and one will be right.

rob

[prefetch]

thanks rob - i found that single debayer function.

so - i got the color figured out (thank you!) - it turned out that something was screwy with my master flats/darks/biases and so i just added the individual frames instead of trying to use a master.  everything turned out okay when i did that.  but now i have another question about this process.

i've spent the last year using nebulosity and photoshop, and i've gotten OK at processing with those tools, but i feel like i'm re-learning everything with PI, so forgive me for these questions!   

i did star alignment and then integration (stacking) and i ended up with these strange streaks in the background.  i've never seen this before.

any ideas of where they come from?  this same image processed in nebulosity does not produce them.

[Alejandro Tombolini]

Hi, this example may help.
Saludos, Alejandro.

[prefetch]

thank you!!

yes, that looks very much like my problem.  i'll work on fixing the calibration.

[prefetch]

and here's what i ended up with.

i have collimation problems (used a hyperstar, which is a pain to collimate) but given the problems with the data, i think this turned out reasonably well for my level of skill:

original single frame:


processed w/ PI:


and a crop:



i'm always open to suggestions or critical feedback.

thanks again for everyone's help.  i'm liking pixinsight more and more.

[pfile]

yeah, collimation, sensor tilt and focus on these fast systems is a serious pain. anyway your image looks pretty good. i would mask the background off and increase the saturation a little bit.

rob

[oldwexi]

Hi Prefetch!
For Hyperstar collimation have a look at my homepage where i describe collimation of Hyperstar.
http://www.werbeagentur.org/oldwexi/hyperstar/tipstest_hyperstar.html

If you follow the 4 major Points, collimation does not take more than 20 Minutes and is in my opinion easy.

You will also find tutorials for PI there:
http://www.werbeagentur.org/oldwexi/tipstest.html

Gerald