Using fake Luminance with DSLR

[jdonald]

Scott Rosen has a nice video on how to do this with a DSLR and I am wondering if this is possible with PI?

http://www.astronomersdoitinthedark.com/dslr_llrgb_tutorial.php

[Josh Lake]

Sure, I do it all the time, though I wouldn't call it 'fake', just an extracted luminance.

If you have RGB data at any step, you can click the Extract CIE L* Component icon on top or use the Image:Extract menu item. Then you can work separately on the luminance channel and re-combine later with LRGBCombination. This is a very effective technique for gaining both detail (with the work on the L frame) and saturation from the chrominance. The before/after is often significant when the saturation settings are tuned carefully.

[pfile]

problem is, and juan has tried to explain this to me before, L* is not luminance. it's Lightness, which is really a psychovisual thing.

i think at the very least if you want to do this you should set the RGB weights in the original image to 1:1:1 before extracting L*.

i am still not clear on how, if it's even possible, to make the most correct Luminance data from an RGB image. perhaps it's just the sum of the 3 channels. of course, really recovering the L is impossible (i think) because of the gaps between the R/G/B filters. though on a DSLR there is significant overlap between the filters.

[Josh Lake]

I guess in that sense it really is kind of a trick. You're not actually adding anything new to the data, of course... but it does help to increase the saturation and do the chrominance noise reduction step. I didn't even know about this method until seeing it in Vicent's video.

[jdonald]

Fake was a strong word.  I should have chosen a different word.

What a great video!  I am going to try this later on my M78.

Thanks for the replies and the support.  I finally got my first ever batchprocess working! in fact I couldn't get DSS to work at all!  I was surprised that I got PI script to work, never had much luck before.