Page 1 of 1

RGB to LAB

Posted: 2009-09-30T09:39:25-07:00
by redsmith
Hello. Quick question:

when converting an image to the LAB colour space using GIMP; then doing the same with IM there is quite a difference between the output of the applications in the L channel. Am I misunderstanding something or is it a bug of some kind?

(ImageMagick, left; GIMP, right )
Image Image

Thanks in advance, R.

Re: RGB to LAB

Posted: 2009-10-05T05:59:33-07:00
by redsmith
Bump. Any reply would do - even if it's to point out I'm a numbskull; and to read the docs more closely.

Re: RGB to LAB

Posted: 2009-10-05T12:43:45-07:00
by fmw42
one of the two is using the log (option) of LAB for the L component. do a google search on LAB and see what it is supposed to be and if there are controls for log or not.

Re: RGB to LAB

Posted: 2009-10-08T01:55:39-07:00
by redsmith
Hello, fmw42, and thanks for the response. I haven't got the time to chase technical papers around but taking your mention of 'log ' as a starting point I applied an '-evaluate pow 3' to the separated channel and got something more like the expected result (from GIMP). I think I was mistaken in thinking the GIMP decompose to layers would give something like a pure channel separation. I'll experiment a bit further. Thanks again.