Bug 689494 - Differences in tiffsep and tiff32nc
Summary: Differences in tiffsep and tiff32nc
Status: NOTIFIED FIXED
Alias: None
Product: Ghostscript
Classification: Unclassified
Component: Graphics Library (show other bugs)
Version: 8.60
Hardware: Macintosh MacOS X
: P4 normal
Assignee: Ray Johnston
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-10-03 10:43 UTC by Marcos H. Woehrmann
Modified: 2008-12-19 08:31 UTC (History)
0 users

See Also:
Customer: 190
Word Size: ---

Attachments
screenshot.png (165.08 KB, image/png)
2007-10-03 10:48 UTC, Marcos H. Woehrmann 		Details
Simple_Rect_Trans.pdf (635.01 KB, application/pdf)
2007-10-03 10:49 UTC, Marcos H. Woehrmann 		Details
Cyan.tif (489.57 KB, image/tiff)
2007-10-03 13:30 UTC, Ray Johnston 		Details
Yellow.tif (489.57 KB, image/tiff)
2007-10-03 13:31 UTC, Ray Johnston 		Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Marcos H. Woehrmann 2007-10-03 10:43:51 UTC 
The customer reports and I've verified that there are significant differences in the output of Ghostscript 
8.60 using the tiffsep and tiff32nc devices (see screenshot.png, attached); gshead (r8268) behaves the 
same way.  Note that 32nc.tiff and sep.tiff are both CMYK files, so there shouldn't be any question of 
CMYK -> RGB conversion causing the difference.

The command lines I used for testing:

  bin/gs -sDEVICE=tiff32nc -sOutputFile=32nc.tiff ./Simple_Rect_Trans.pdf
  bin/gs -sDEVICE=tiffsep -sOutputFile=sep.tiff ./Simple_Rect_Trans.pdf
Comment 1 Marcos H. Woehrmann 2007-10-03 10:48:20 UTC 
Created attachment 3434 [details]
screenshot.png
Comment 2 Marcos H. Woehrmann 2007-10-03 10:49:03 UTC 
Created attachment 3435 [details]
Simple_Rect_Trans.pdf
Comment 3 Ray Johnston 2007-10-03 13:28:57 UTC 
The Cyan and Yellow separations show the problem. It looks like the opacity
of 0.5 is being ignored and the original C=0.7 Y=0.2 is being used.

This is lower-right rectangle is written by:
%Resolving: [11 0]
<<
/OPM 1 /BM /Normal /CA 0.5 /OP false
/SMask /None /ca 0.5 /AIS false
/op false
/Type /ExtGState /SA true
>>
endobj
/Fm0 Do
%Resolving: [10 0]
%Resolving: [8 0]
<<
/I false
/K false
/S /Transparency /Type /Group >>
endobj
0.7 0 0.2 0 k
0 i
/RelativeColorimetric ri
/GS0 gs
%Resolving: [9 0]
462.197 256.749 -294.366 319.718 re
f
Comment 4 Ray Johnston 2007-10-03 13:30:12 UTC 
Created attachment 3436 [details]
Cyan.tif

Cyan separation
Comment 5 Ray Johnston 2007-10-03 13:31:02 UTC 
Created attachment 3437 [details]
Yellow.tif

Yellow Separation
Comment 6 Ray Johnston 2007-10-11 09:37:58 UTC 
This was caused by inverted blending logic in pdf14_cmykspot_put_image.

Fixed by rev 8287 (but the fixes for 689493 are needed as well for testing).
Those fixes will be committed next.