AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2014.12.21 20:12 "[Tiff] PNG in TIFF ?", by Ulf Zibis
2014.12.21 20:20 "Re: [Tiff] PNG in TIFF ?", by Toby Thain
2014.12.21 20:49 "Re: [Tiff] PNG in TIFF ?", by Joris Van Damme
2014.12.22 01:54 "Re: [Tiff] PNG in TIFF ?", by Tom Lane
2014.12.22 02:29 "Re: [Tiff] PNG in TIFF ?", by Bob Friesenhahn
2014.12.22 02:55 "Re: [Tiff] PNG in TIFF ?", by Toby Thain
2014.12.22 03:19 "Re: [Tiff] PNG in TIFF ?", by Bob Friesenhahn
2014.12.22 08:51 "Re: [Tiff] PNG in TIFF ?", by Joris Van Damme
2014.12.22 14:59 "Re: [Tiff] PNG in TIFF ?", by Tom Lane
2014.12.22 15:30 "Re: [Tiff] PNG in TIFF ?", by Joris Van Damme
2014.12.22 23:02 "Re: [Tiff] PNG in TIFF ?", by Ulf Zibis
2014.12.27 22:24 "Re: [Tiff] PNG in TIFF ?", by Ulf Zibis

2014.12.22 15:30 "Re: [Tiff] PNG in TIFF ?", by Joris Van Damme

Tom,

> IIRC, the only allowed predictor in TIFF is previous-pixel (horizontal differencing)

Yes, YRC.

I don't know what compatibility hazards you face if you use it. (Of course, that caveat can be applied to depressingly many things in TIFF ...)

You said it! The same applies for a multitude of things, a great many of which we use routinely.

In my experience, support for predictor and flate compression is actually very good in all main stream stuff. Even stuff that has historically persisted in the category "rather limited", or "braindead", or even "plain wrong", up to and including the build-in Windows TIFF codec, has advanced greatly in the last half a decade or so, and supports all this without any problem. The vast majority of things, being based on LibTiff, have been supporting it for decades.

Thanks for your thoughts on predictor efficiency.

Best regards,

Joris