2011.06.07 17:06 "Re: [Tiff] Questions on TIFF LZW compression", by Bob Friesenhahn
While I am not an expert at TIFF compression issues, I suspect that you are correct that the predictor has little value (and may cause harm) at depths other than 8 (and perhaps 16) bits since the compressor works on an octet stream.
Question rather is, does it predict octects, or pixel values modulo 2^N?
LZW works with streams of bytes, not pixel values. TIFF LZW is similar to standard LZW except that I understand that it includes an implementation defect (which must be supported).
I'm not so much concerned about the compression performance - as you say, there are clearly better codecs available - but rather about correctness. I really wonder how specs as sketchy as the TIFF specs got accepted.
The TIFF specification is an Adobe publication which was formulated according to the internal rules they followed at the time. It is not like an ISO standard.
I understand. Is there a newer spec than rev.6 available? I don't see much of this in the specs I have.
It seems exceedingly unlikely that Adobe will ever release a new major version of the TIFF specification. Over the years they have produced a number of documents describing changes and additions. Cris Cox from Adobe has been quite helpful, but Adobe's focus is not on legacy formats like TIFF so it is difficult for him to obtain corporate support for issuing new specifications.
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/