Kraellin
![]() |
i've probably suggested this before, but it's been so long i dont recall. i'd love to see a cmyk assemble and extract.
If wishes were horses... there'd be a whole lot of horse crap to clean up!
Craig |
|||
Posted: April 22, 2007 2:47 pm | ||||
Sjeiti
![]() |
I second that... just needed that today and couldn't find it
|
|||
Posted: January 16, 2008 2:56 am | ||||
StevieJ
![]() |
+3 .....would be handy for pre-print color separations.....
Steve
"Buzzards gotta eat...same as worms..." - Clint :) |
|||
Posted: January 16, 2008 9:17 am | ||||
onyXMaster
Posts: 350 |
I don't think so.
Bad CMYK separation without proper color matching is no good for printing. Good CMYK separation requires proper color matching with a target device profile, along with a proper separation setup (coating at least). While you can get away with the "RGB is almost device-independent, so let's slap on this sRGB label, or hell, just stick with gamma", you can't do the same stuff with CMYK -- it's the #1 device-dependent color space ![]() On the other hand, if Vlad decides that some generic "CMYK" will do, it's very too hard to implement. Please note that my previous statement does not imply that it will be implemented soon even if Vlad agrees that having "CMYK" splitter/combiner is a good idea -- you know what we're working on, right? ![]() |
|||
Posted: January 18, 2008 3:25 pm | ||||
Kraellin
![]() |
making us wait? ![]() ![]() If wishes were horses... there'd be a whole lot of horse crap to clean up!
Craig |
|||
Posted: January 18, 2008 3:31 pm | ||||
Sphinx.
![]() |
Extracting and assembling CMYK channels is really not that hard if you disregard color profile conversions ("Bad CMYK seperation"). But there is little use for it really, partly because of the device dependent nature, but mostly because FF do not support 5 channel composites (ACMYK).
I attached a small snippet that shows one way to extract/assemble CMYK channels - if you simply want to play around with intermediate CMY(K) channels and reassemble to RGB, its a good starting point. The snippet is based on GIMP's internal conversion routines (really simple stuff). There is an alternative assembly cluster that might be a bit more useful - it lets you select the cmyk dye to use when assembling (I included Adobe Working CMYK). I did not include a custom extract version though. Note that the GIMP cmyk -> rgb formula does a poor job converting cmyk input composed with key pullout > 0. RGB _-_ CMYK.ffxml |
|||
Posted: January 19, 2008 4:52 am | ||||
Sphinx.
![]() |
||||
Posted: January 19, 2008 5:10 am | ||||
onyXMaster
Posts: 350 |
When I re-read my own post I found that I was going to say exactly the opposite -- "not too hard to implement" ![]() Yes, generic CMYK splitting is easy stuff, sorry for making a strange typo "very too" instead of "not too". Also, I would like to not that to provide assemble/extract CMYK components, there is no need for us to support 5-channel operations in the UI. Inner workings of FF can work with unlimited number of channels with arbitrary meanings ![]() |
|||
Posted: January 19, 2008 5:17 am | ||||
Sphinx.
![]() |
Cool! Tell me more ![]() I was thinking about the way things seem to work now (ARGB based sample transfer), but perhaps there is more going on under the hood ![]() |
|||
Posted: January 19, 2008 5:26 am | ||||
onyXMaster
Posts: 350 |
The pipeline between components is ARGB, yes (this can be changed, though a bit hard), but internally, FF supports pixel formats with arbitrary precision and number of channels (provided there are conversions set up in the code).
|
|||
Posted: January 19, 2008 5:33 am | ||||
Sphinx.
![]() |
Ok, yeah determining if a pipeline is truly grayscale (with or without alpha) could become a costly check. I'm simply thinking about the optimization possibilites here - you very often see 100% grayscale branches (not even alpha, or that is alpha = 1 constantly), which then at a later stage (e.g. near result) will utilize/modify (A)RGB indepently. Maybe if this should be supported, it could be via some component that forces gray or gray + alpha only processing in the successor chain until a ARGB mode component is introduced. Oh btw Onyx. while I have your attention - did you see this topic: A few alpha channel issues..? Specially the alpha and blendmode issue is something I'd like to know the "status" of (i.e. has it been filed as a bug?) |
|||
Posted: January 19, 2008 5:51 am | ||||
onyXMaster
Posts: 350 |
There is actually a "grayscale" pipeline, you know it as "curves"
![]() Sorry, can't tell about the status of this problem since I'm at home now and don't have access to the issue database. |
|||
Posted: January 19, 2008 8:19 am | ||||
StevieJ
![]() |
Couldn't CMYK assemble/extract components be set up with color separation ranges for user control to match them??? Steve
"Buzzards gotta eat...same as worms..." - Clint :) |
|||
Posted: January 22, 2008 1:46 pm | ||||
saurabhgayali |
Not just CMYK extract and assemble. We also need a color input component with CMYK enabled or only CMYK.
I want to make a filter to know which two colors should be mixed in watercolor to get the color you wish. But the inputs dont have CMYK at all. saurabh.gayali@gmail.com |
|||
Posted: May 21, 2018 11:17 am |
Filter Forge has a thriving, vibrant, knowledgeable user community. Feel free to join us and have fun!
33,711 Registered Users
+18 new in 30 days!
153,531 Posts
+36 new in 30 days!
15,347 Topics
+72 new in year!
21 unregistered users.