2005.09.23 21:11 "[Tiff] Additional Lossless Compression Schemes", by Frank Warmerdam

2005.09.25 01:09 "Re: [Tiff] Additional Lossless Compression Schemes", by Joris Van Damme

But it doesn't have quite the flexibility that TIFF needs.

In so much as tiled support or something else?

Let's not take the OJPEG route, for obvious reasons. Every strip/tile has to be a totally independently compressed block, i.e. an image all by itself, whether compressed with JPEG or JPEG2000 or whatever. Thus, we don't need JPEG2000 to support tiles, we need TIFF to support tiles as it does, and we need JPEG2000 to support only image compression (applied to individual strips or tiles, but that should not matter on the JPEG2000 compression level). So, it doesn't matter if JPEG2000 doesn't support tiling, we are not supposed to use that that anyway.

It also doesn't matter if it doesn't support some particular bitdepth and can be applied only in a limited number of TIFF colorspaces and bitdepths (just like JPEG, G3 and G4), and it doesn't matter if it cannot be applied to predicted data (just like JPEG), etc.

Thus, basically anything that can compress an image is a valid option to include in TIFF, I think. I don't actually know JPEG2000, but with a) the lack of need for compression schemes to support the image subdivision that we impose on the TIFF level, on the one hand, and b) the lack of need for compression schemes to support all color spaces and bitdepths and prediction and such on the other hand, I cannot imagine why any compression scheme at all would not be a suitable compression scheme in TIFF, and I am curious too what 'flexibility' TIFF needs and JPEG2000 cannot support.

Joris Van Damme
Download your free TIFF tag viewer for windows here: