Menu
▾
▴
[E] [ enblend-Bugs-2881272 ] blending uses wrong gamma
|
From: Enblend <enb...@li...> - 2009-10-19 09:01:18
|
Bugs item #2881272, was opened at 2009-10-18 14:34 Message generated for change (Settings changed) made by cspiel You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=696409&aid=2881272&group_id=123407 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Output artifact >Group: v4.0 >Status: Pending Resolution: None >Priority: 5 Private: No Submitted By: Erik Krause (erik_krause) Assigned to: Nobody/Anonymous (nobody) Summary: blending uses wrong gamma Initial Comment: While this is no problem as long as very similar colors are blended it gets strange if blending affects different colors. Background: Blending bright red (RGB 255,0,0) and bright green (RGB 0,255,0) each by 50% results in RGB 128,128,0 which is a far too dark olive color. It should be a yellow of the same brightness as the red and green, which is about RGB 187,187,0 To test this create two test images that contain either color. Alpha mask them such that there is an overlapping region. Feed them to enblend. In Photoshop you can set blending gamma in the color preferences. A value of 1.0 produces nice yellow, no blending gamma (which means the color space one is used) gives olive... The problem is about the same what Helmut Dersch described for interpolation ten years ago: http://www.all-in-one.ee/~dersch/gamma/gamma.html ---------------------------------------------------------------------- >Comment By: Christoph L. Spiel (cspiel) Date: 2009-10-19 11:01 Message: An example (with only two images) attached to this issue report would be helpful. ---------------------------------------------------------------------- Comment By: Andrew Mihal (acmihal) Date: 2009-10-19 07:12 Message: You can pass in tiffs with embedded ICC profiles and use the -c parameter to blend colors with the CIECAM02 color appearance model. It would be straightforward to add a simple gamma-corrected blending model to the software. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=696409&aid=2881272&group_id=123407 |