2004.10.28 12:03 "[Tiff] why does tiffcp::cpSeparateBufToContigBuf takes BYTES_per_sample as parameter?", by Sachin Garg

2004.10.28 17:50 "Re: [Tiff] why does tiffcp::cpSeparateBufToContigBuf takesBYTES_per_sample as parameter?", by Sachin Garg

But I am confused (as mentioned in my original post) how tiffcp routines (cpSeparateBufToContigBuf and cpContigBufToSeparateBuf) handle all bit-depths, they seem to be byte oriented.

As far as I can see, they are indeed byte-oriented. This doesn't make any difference when PlanarConfiguration=contiguous, but I guess it will cause problems when PlanarConfig=separate AND bits per sample is not a multiple of eight. If that combination is really important to you, you should write your own code (that's going to be a lot of bit-shifting).

I had written a version (with lots of bit-shifting) but its dead slow, effecting the overall application performance. So just decided to check how libtiff does similar stuff.

Sachin Garg [India]
http://www.geocities.com/schngrg