AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2005.03.19 15:35 "[Tiff] Floating Point Sample Format", by Charles Auer
2005.03.19 16:23 "Re: [Tiff] Floating Point Sample Format", by Frank Warmerdam
2005.03.21 19:51 "Re: [Tiff] Floating Point Sample Format", by Chris Cox
2005.03.21 20:56 "Re: [Tiff] Floating Point Sample Format", by Frank Warmerdam
2005.03.21 21:12 "Re: [Tiff] Floating Point Sample Format", by Chris Cox
2005.03.21 23:26 "Re: [Tiff] Floating Point Sample Format", by Joris

2005.03.21 23:26 "Re: [Tiff] Floating Point Sample Format", by Joris

I did notice the comment in the Spec that the SMINSAMPLEVALUE and SMAXSAMPLEVALUE aren't supposed to be used for scaling.

Part of the rationale might possibly have been:

  1. If the spec said these values are supposed to be good for scaling, then that's one more rule that potentially makes a TIFF file 'invalid', and there's plenty of those already. On the other hand, if it doesn't say that, the values aren't any less useful because any reader that has the ability to scan and derive proper values before any rendering or otherwise necessary scaling, will do so anyway, for robustness, and any reader that doesn't have that ability because of single-pass/speed needs, will use those values anyway, because it's the only indication and thus best possible guess.
  2. Scaling may not even be a trivial notion, in some applications. Some data may find itself with e.g. 2% in the 0-145 range, 96% of the data in the 145-155 range, and 2% of the data in the 155-300 range. What scaling values are appropriate then? Some data may not make any linieair sense even.

Joris Van Damme
info@awaresystems.be
http://www.awaresystems.be/
Download your free TIFF tag viewer for windows here:
http://www.awaresystems.be/imaging/tiff/astifftagviewer.html