onyXMaster
Filter Forge, Inc.
Posts: 350
|
ShadowHighlight Controlled by Dave C. Schneider http://www.filterforge.com/filters/1176.html
|
Posted: September 19, 2006 5:48 am |
Details
E-Mail
|
onyXMaster
Filter Forge, Inc.
Posts: 350
|
I would like to recommend you to use the Desaturate component with the "Weighted Average" method instead of Extract Lab component for mask generation. The Desaturate is faster and performs better in terms of on-screen perceptual "lightness" than L channel from Lab.
|
Posted: September 19, 2006 5:50 am |
Details
E-Mail
|
Dave C. Schneider
|
Thanks for the hint! Speed is certainly an important point.
There seems to be some disagreement, however, about the best desaturation method. Some people on Photoshop forums forcefully claim that the L channel in Lab is the best desaturation method since Lab was developed to be as "perceptual" as possible. I have no clear opinion about that nor do I know any research.
Anyway, I will try desaturation and post an update to my filter.
David
|
Posted: September 19, 2006 6:45 am |
Details
E-Mail
|
onyXMaster
Filter Forge, Inc.
Posts: 350
|
Well, while I agree that L from Lab is a very good perceptual "lightness" generally, it depends on how you interpret RGB values for RGB<->XYZ transfer (sRGB D65 assumed in Filter Forge) and what is the white point for XYZ<->Lab transfer (D50 assumed in Filter Forge). For this concrete case, I believe that weighted average with gamma correction produces better results than Lab in most user cases, simply because it _ignores_ the white point and performs all operations in RGB device-dependent space.
|
Posted: September 20, 2006 1:05 am |
Details
E-Mail
|