2001.03.28 18:52 "RE: Bit depth", by Max Martinez

I don't know about that, Frank. I think that will tend to exacerbate a problem that already exists with the min/max sample values.

In our experience TIFF min and max sample values are, in general, so unreliable that we simply ignore them in our software. We would like to use them as (and perhaps this is Andrew's intention as well) an indication of an appropriate contrast stretch for initial display of the image but because of the combination of the public domain libraries inability to write per sample min/max combined with TIFF writers putting useless information in there (like simply the range of the bit depth), we find that we are better off not using the information at all.

If Andrew attempts to indicate the dynamic range of the sensor through these values, I suppose he will get better results than not putting anything at all, but he won't get appropriate results for images whose actual data does not really span the range of the original sensor.