Color correction of Astronomics CLS filter

[cdesselles]

With my Canon T2i (550D) DSLR I've been using an Astronomics CLS clip-in filter hoping to reduce the effects of rampant light pollution in my area.  But the final outcome is that the images have a blue cast to them.  This is especially bothersome with planetary photos where the color of features can be critical to a good image.

I'm asking for suggestions on how to deal with this in Pixinsight to retrieve more normal color in my photos. 

(FYI: the in camera white balance is set to Daylight)

[Andres.Pozo]

I think that for planetary photos you wouldn't need a light pollution filter.

For deep space images PI has the processes BackgroundNeutralization and ColorCalibration.

Also, if you are shooting raw images, it doesn't matter the white balance setting in the camera 

[cdesselles]

I think that for planetary photos you wouldn't need a light pollution filter.

I agree and won't be doing so in the future. But I already had the filter installed in the camera and decided to give it a go.  I managed to get several AVI vieos of Saturn that appeared to be among my personal bests for focus and seeing, so I am loathe to just trash them and wish to see what I can do to salvage them.

For deep space images PI has the processes BackgroundNeutralization and ColorCalibration.

Also, if you are shooting raw images, it doesn't matter the white balance setting in the camera 

Unfortunately, no RAW files.  I was using BackyardEOS in Planetary mode for the capture of 1000 to 2000 frame videos.  Backyard EOS does this by taking a rapid series of small JPEGS and then compiling them into an AVI video.  I then stacked the videos using Autostakkert2.  Using PI (I am still very much a novice with this program) I wished to use Restoration Filter to sharpen things up and then make additional adjustments.  But the color correction is eluding me.

[Zocky]

You can select custom WB.
Take an image on a white paper while your CLS is installed, go to custom WB, select image of the white paper, and press select. I use this method with my EOS 350D+astronomik clip CLS.

[cdesselles]

You can select custom WB.
Take an image on a white paper while your CLS is installed, go to custom WB, select image of the white paper, and press select. I use this method with my EOS 350D+astronomik clip CLS.

Is this an "in camera" selection for the DSLR or is it a PI function to be applied to the image?  It does sound like a very proper solution to the issue for future images.  Thanks!   

PS:  Ay suggestions on what I can do for my existing files mentioned above?

[pfile]

it's in-camera but note that all it does is put some metadata in the CR2 file, saying how much the R and B channels (i think) should be multiplied by to restore the white balance. i think it's better to just leave that turned off and do the color calibration using PI's functions.

heck, even ScreenTransferFunction with the channels unlocked does a good job of undoing any color casts.

note that since the in-camera custom white balance is just metadata, the CR2 reader in PI must be set up to pay attention to it ("use camera white balance" in Format Explorer > DSLR_RAW). it's not clear to me that when using the format hint "raw cfa" if the white balance setting is taken as programmed in DSLR_RAW or is overridden to both white balance checkboxes unchecked. BPP uses the "raw cfa" hint.

rob

[cdesselles]

But what about the JPEGs that make up these video captures of planets?  Would in camera custom white balance be useful there also?

[Zocky]

If for some reasone you want to capture JPEGs, than use in camera custom WB as I described you.

[cdesselles]

If for some reasone you want to capture JPEGs, than use in camera custom WB as I described you.

Thanks, Zocky.  It's not that I want to capture JPEGs, but that is the way Backyard EOS v.3.03 does planetary video captures.  It takes a series of JPEGs and compiles them into an AVI.  The normal video mode of my Canon T2i is to produce MOV files.  At any rate, your tip about in camera custom white balance is invaluable.  I'll be sure to set that feature whenever I use my CLS clip-in filter (which won't be for planets anymore.)

[pfile]

i think that the frames that BYE captures from live view are lossless. so not sure they are jpeg.

hard to know if the custom white balance applies to that. but you can ask in guylain's forum, he would know, i'm sure.

rob

[cdesselles]

Rob:

Oh they are definitely JPEGs in BYE's planetary capture.  They are downloaded to a folder (all 1500 of them per capture in my case) named for the target/capture time/iso/etc.  They are then compiled into an AVI by BYE.  I have a seperate thread running on O'telescope concerning some issues with that compilation (you even responded to it if I remember correctly. - I'm listed as PoppaChris on that forum)

Anyway, Probably because the image was compiled from a series of Jpegs - the color correction methods mentioned above didn't work very well.  I real shame as one of the Mars videos stacked out as probalby a personal best for me as far as sharpness, focus, and details were concerned.  Just lousy color that I can't seem to get fixed.  Any other ideas?

(PS:  One thing I have learned is that I will never ever ever use a CLS filter on planetary captures again.)

[pfile]

yeah - i suppose there's probably no reason to use the CLS on a planetary image. but now that it's done, here's what you can do - go ahead and make a CWB with the CLS in the camera, then take a CR2 image with the CWB set, of anything you like. you can use exiftool to dump out the exif information, and in there you'll see the white balance information. you can then try applying those factors (using pixel math) to the stacked RGB image. that should get you into the ballpark, color wise.

rob

[cdesselles]

Thanks for the tip, Rob.  Looks like I'm going to get a crash course in moving out of the "newbie" level of PI use and into a more journeyman level of understanding.  Pixel Math is something I haven't used at all.  Not even sure what I would be doing there.  Never tried to extract EXIF data from one of my pics either for that matter, not to mention obtaining a custom white balance.  A lot to chew on for an old man of my age (63 and feeling every year of it)  But situations like this is one of the things that I love about this hobby.  Always some new challenge.

[pfile]

the pixel math should be pretty simple… just turn off "use single expression" and then assuming the camera has given you R and B factors, make the R expression "$T*<factor>" (no quotes) where <factor> is replaced by 0.8 or whatever. G should be $T, and B should be "$T*<factor>". then configure pixel math to replace the target image. what this means is that the red channel should be <factor> times the target image ($T), the green should simply be the target image with no changes, etc. then drag the triangle to your stacked RGB image and done.

exiftool is probably embedded somewhere in BackyardEos. i'll have to check that out. once you have the exiftool program, you'll probably have to run it in a windows command shell. on the other hand there probably are gui versions of exiftool, or heck, maybe even canon's DPP can show you the exif information in a CR2.

rob