LRGB SynthLum - best method?
Posted: Wed Apr 01, 2015 12:23 am
Ivo or others, I'd like some feedback on what you consider is the best method for creating a synthetic luminance image in StarTools using LRGB filters (i.e. a LLRGB or HaRGB for instance).
My standard method has been this:
1. Load the R monochromatic channel using "Open", wipe (use temporary autodev to check), then denoise, then save. Repeat for G and B. Keep these as my 'clean' but unstretched chrominance channels, RC, GC and BC.
2. Prepare the monochromatic L channel (dev, wipe, HDR, contrast, deconv, denoise, etc.) and save. Repeat this procedure for each of the R, G and B channels, and save as RL, GL and BL.
3. Use Layer module to do an iterative weighted blend of the L, RL, GL and BL channels. Save SynthLum.
4. Use LRGB to open the SynthLum in the L channel, and RC, GC and BC in the colour channels.
5. Fire up the Colour module, do colour balancing etc.
6. Do final touch-ups, including a wipe check, further HDR if required, plus other stuff like Life or Repair etc.
This seems to work out pretty well. However, I can see that there might be other ways to do it. One that occurred to me was to FIRST use the LRGB module to load up the R, G and B channels only, and then do step 2 from the above on this composite, but including the colour balancing etc. Save this as an RGB-L. Then do a separate process of the Luminance monochrome data. Then, as a final step, do the weighted combination in Layer.
Both methods would preserve the correct colour saturation levels, but the latter strikes me as simpler. But the final Layer step might desaturate the colour. Would it be effective do you think? (I guess I can road-test it!). Or do people have other workflows for LLRGB?
My standard method has been this:
1. Load the R monochromatic channel using "Open", wipe (use temporary autodev to check), then denoise, then save. Repeat for G and B. Keep these as my 'clean' but unstretched chrominance channels, RC, GC and BC.
2. Prepare the monochromatic L channel (dev, wipe, HDR, contrast, deconv, denoise, etc.) and save. Repeat this procedure for each of the R, G and B channels, and save as RL, GL and BL.
3. Use Layer module to do an iterative weighted blend of the L, RL, GL and BL channels. Save SynthLum.
4. Use LRGB to open the SynthLum in the L channel, and RC, GC and BC in the colour channels.
5. Fire up the Colour module, do colour balancing etc.
6. Do final touch-ups, including a wipe check, further HDR if required, plus other stuff like Life or Repair etc.
This seems to work out pretty well. However, I can see that there might be other ways to do it. One that occurred to me was to FIRST use the LRGB module to load up the R, G and B channels only, and then do step 2 from the above on this composite, but including the colour balancing etc. Save this as an RGB-L. Then do a separate process of the Luminance monochrome data. Then, as a final step, do the weighted combination in Layer.
Both methods would preserve the correct colour saturation levels, but the latter strikes me as simpler. But the final Layer step might desaturate the colour. Would it be effective do you think? (I guess I can road-test it!). Or do people have other workflows for LLRGB?