Inconsistent image distortions
-
I've encountered a baffling problem with images, and I'm hoping someone who has seen something similar might suggest what's going wrong. I'm working with CMYK bitmaps (not RGB) and generating separate cyan, magenta, yellow, and black bitmaps for printing. I'm adding some markup to the black plane before the bitmaps are generated. The problem is the black markup is appearing on all four planes, and where it does appear, only every 4th column is written. It's almost as if the black markup is being diluted over the four planes. Now for the baffling part: It only happens on SOME machines. It works correctly MOST OF the time. I've only seen this malfunction on three machines. I've tried copying the same .EXE and .DLLs and the results are consistent on each machine, i.e. it appears to be machine-specific, not file-specific. On two machines that have the same service pack (3) of Windows XP, one exhibits the problem, and one doesn't. Has anyone seen any behavior like this before, or have any idea what's causing the inconsistent results? Thanks!
-
I've encountered a baffling problem with images, and I'm hoping someone who has seen something similar might suggest what's going wrong. I'm working with CMYK bitmaps (not RGB) and generating separate cyan, magenta, yellow, and black bitmaps for printing. I'm adding some markup to the black plane before the bitmaps are generated. The problem is the black markup is appearing on all four planes, and where it does appear, only every 4th column is written. It's almost as if the black markup is being diluted over the four planes. Now for the baffling part: It only happens on SOME machines. It works correctly MOST OF the time. I've only seen this malfunction on three machines. I've tried copying the same .EXE and .DLLs and the results are consistent on each machine, i.e. it appears to be machine-specific, not file-specific. On two machines that have the same service pack (3) of Windows XP, one exhibits the problem, and one doesn't. Has anyone seen any behavior like this before, or have any idea what's causing the inconsistent results? Thanks!
Hi Alan, never heard such a thing, I find it quite intriguing indeed. First thought was about undercolor removal, but then you said it is black you added. Some questions: - how do you observe this? by watching a color component bitmap on screen? or only by actually printing/plate making? - the added text is pure black? - what is the background color? - what is the ImageFormat at the input, anything with lossy compression (e.g. JPEG)? - if the black adition is H pixels high (or W pixels wide), are the color shadows also H pixels high (or W pixels wide, give or take 3 due to your "every 4th column" remark)? - how solid/saturated are the colored shadows of your black additions? or is it mostly an outline effect? (that would suggest your text got added with anti-aliasing, resulting in some gray pixels at the edge, and then maybe your font rendering logic differs on some machines.) - assuming your input images are not CMYK, what are they? RGB? how many bits? and what code do you use for CMYK conversion? if that is part of a library, please check for version differences here. - any chance you have dithering/screening code running *before* you do color separation? - any chance I could see some of the images (in, out) myself? :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that.
[The QA section does it automatically now, I hope we soon get it on regular forums as well]
-
Hi Alan, never heard such a thing, I find it quite intriguing indeed. First thought was about undercolor removal, but then you said it is black you added. Some questions: - how do you observe this? by watching a color component bitmap on screen? or only by actually printing/plate making? - the added text is pure black? - what is the background color? - what is the ImageFormat at the input, anything with lossy compression (e.g. JPEG)? - if the black adition is H pixels high (or W pixels wide), are the color shadows also H pixels high (or W pixels wide, give or take 3 due to your "every 4th column" remark)? - how solid/saturated are the colored shadows of your black additions? or is it mostly an outline effect? (that would suggest your text got added with anti-aliasing, resulting in some gray pixels at the edge, and then maybe your font rendering logic differs on some machines.) - assuming your input images are not CMYK, what are they? RGB? how many bits? and what code do you use for CMYK conversion? if that is part of a library, please check for version differences here. - any chance you have dithering/screening code running *before* you do color separation? - any chance I could see some of the images (in, out) myself? :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that.
[The QA section does it automatically now, I hope we soon get it on regular forums as well]
Hi Luc, The problem has been solved: The images that showed this phenomenon had a 5th, alpha channel, which was being used for White, so the pixel format was CMYKW. (Research on the use of extra channels is being done concurrently, and the three machines that showed the problem just happened to have images with extra channels.) Thanks for your help! Alan
-
Hi Luc, The problem has been solved: The images that showed this phenomenon had a 5th, alpha channel, which was being used for White, so the pixel format was CMYKW. (Research on the use of extra channels is being done concurrently, and the three machines that showed the problem just happened to have images with extra channels.) Thanks for your help! Alan
No problem. Glad you managed to solve it. :)
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that.
[The QA section does it automatically now, I hope we soon get it on regular forums as well]