Don't worry, this one doesn't come with a pull request!

We've done our own link between OCIO and OIIO (as a part of oiiotool).

One thing, and this is connected with the Data Window issue that I posted about 
earlier, is that, if I pass an image with a smaller data window than display 
window through oiiotool, and tell it to convert from linear to Cineon, I end up 
with any black pixels that were inside the original data window converted to a 
Cineon value of approximately 0.093, but anything that was outside the original 
data window ends up black.

What I would propose is to have something in the ImageSpec class that stores 
what the background pixel colour is (ie, what to assume the colour outside the 
data window is). When a colour transform is applied, it is also applied to this 
background pixel colour. And when ImageBufAlgo::crop() is applied to an 
ImageBuf, and new area that it needs to fill in should be filled in this colour.

Shake had this concept of the "bgColor". Nuke uses the nearest valid pixel.

What do people think?

Hugh Macdonald
nvizible – VISUAL EFFECTS

[email protected]
+44(0) 20 3167 3860
+44(0) 7773 764 708

www.nvizible.com

_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org

Reply via email to