AWARE [SYSTEMS] Imaging expertise for the Delphi developer
AWare Systems, Imaging expertise for the Delphi developer, Home TIFF and LibTiff Mailing List Archive

LibTiff Mailing List

TIFF and LibTiff Mailing List Archive
March 2002

Previous Thread
Next Thread

Previous by Thread
Next by Thread

Previous by Date
Next by Date

Contact

The TIFF Mailing List Homepage
This list is run by Frank Warmerdam
Archive maintained by AWare Systems



Valid HTML 4.01!



Thread

2002.03.20 16:37 "Changing Orientation Tag", by Erika Nesse
2002.03.22 05:56 "Re: Changing Orientation Tag", by Jon Saxton
2002.03.22 21:13 "Re: Changing Orientation Tag", by Peter Smith
2002.03.23 05:38 "Re: Changing Orientation Tag", by Tom Lane
2002.03.25 14:05 "Re: Changing Orientation Tag", by Erika Nesse

2002.03.23 05:38 "Re: Changing Orientation Tag", by Tom Lane

Peter Smith <psmith@creo.com> writes:
> I would lay the blame for this inconsistency in viewer behaviour on the 
> Tiff 6.0 spec itself, which states:
>
> "[The Orientation tag] is recommended for private (non-interchange) use 
> only.  Most images are scanned and stored in the Orientation = 1 format, 
> and most TIFF readers can only handle this case."
>
> It makes me wonder why this tag was even included in the spec.

TIFF has always been designed on the principle that the writer can write
anything he damn pleases, and perhaps the reader will be able to make
sense of it.  Look through the spec: surely 75% of it describes features
that a run-of-the-mill reader would be unable to cope with.  (Show me
a reader that will do something sensible with YCbCr, associated-alpha,
32-bit-float-sample images ... to take just one example.  If you say
your code can handle that nicely, I'll start asking about tiles,
padding, downsampling, and planar configuration.)

I think the spec is being quite honest here to tell you what part of
the described feature set is actually likely to be portable.

If this approach to compatibility doesn't thrill you, perhaps you
should consider another file format.  For instance, PNG, which was
designed in large part on the principle that TIFF blew this issue
badly.  The PNG spec expects that every reader should be able to
do something sensible with every PNG file, which is certainly not
the design assumption of TIFF.  On the other hand, you can't stick
a YCbCr, associated-alpha, 32-bit-float-sample image into a PNG.
You pays your money and you takes your choice...

			regards, tom lane
			organizer, Independent JPEG Group
			member, PNG development group
			onetime member, TIFF advisory committee