Blake Winton wrote:

> Logan Shaw wrote:
> 
>> Now, we can assume that rowbytes is never smaller than width * pixelsize.
>> If it were smaller, then the data would not fit!

> Can we?  What about compressed bitmaps?

I could be wrong, of course, but I assume that rowbytes isn't defined
for compressed images.

Regardless, I was actually talking about the subset of bitmaps which
are not compressed.  Comparing compressed bitmaps is a whole different
game.  For example, in theory, you could have two bitmaps with different
pixel data and different compression types, but the compressed data
could be the same.  (Or, at least I have not seen proof that this
cannot happen.)  Or you could have bitmaps with identical pixel data
but different compression formats and different compressed data.

Or you could even have two compressed bitmaps that use the same
compression type and have the same pixel data but have different
compressed data!  This type of thing can happen when a newer version
of an algorithm is tweaked to use the same format but increase the
compression algorithm slightly, or when a different time/space
tradeoff parameter is chosen for the same algorithm.  In the real
world, this really happens with the deflate algorithm.  I don't
know details of the compression algorithms that Palm OS uses, but
I haven't seen any guarantee that there can't be variations in
the compressed output from the same uncompressed input even with
a single algorithm.

So, the obvious way to compare two compressed bitmaps would be
to decompress them, then compare.  It would be very difficult
to do anything else and be sure it's a correct implementation.

  - Logan

-- 
For information on using the Palm Developer Forums, or to unsubscribe, please 
see http://www.palmos.com/dev/support/forums/

Reply via email to