1994.01.28 15:32 "TIFF Group4 Files and Photometric interpretation", by Timothy Platt
To whom it may concern:
My apology for bothering you. I work for Xerox Imaging Systems I am having a problem with a customer of ours for ScanWorX API (An OCR software package.) The issue concerns TIFF files with Group4 Compression with the Photometric (tag 262) set to 1 (min-is-black). The customer insists that when we render some of his files into our memory for OCR we are getting the sense wrong and hence OCR is failing. He claims that for Group4 decoding the initial-refference line starts at 0xFF..... for the min-is-black case decoding proceeds from there and then the image is inverted if black needs to be one for processing purposes. I have "Facsimile Coding Schemes... Group 4 Facsimile Apparatus (Malaga-Torremolinos, 1984)" as refference. It states that "The refference line for the first coding line in a page is an imaginary white line" which is open to interpretation and perhaps we get it wrong.
Could you shed any light on the problem? Also if you have a copy of the above document but the CCITT, Geneva: 1988 Revision I would be ever so appreciative. My feeling is that however we deal with the issue there will be files that will render improperly however, I want our software to be in line with the TIFF 6.0 specification AND reality as defined with the world of existing TIFF files wherever possible.
Again, my apology for bothering you.
Yours - Timothy J. Platt firstname.lastname@example.org (508)977-2068 Fax: 2433