In JPEG 2000 standard, all the code blocks within the same tile of a component need to have the same width and height according to the JPEG 2000 standard. The width and height information of a code block are specified in either the coding style default (COD) and/or coding style component (COC) marker segments. However, the constraint of fixed code block size is due to the fact that COD and/or COC marker segments in a JPEG 2000 compressed file only can carry one set of code block width and height information for an image or a tile or a component. We have studied the implication of using variable code block sizes at different DWT resolution levels instead of the fixed size code blocks prescribed by JPEG 2000 standard. The simulation results show that there is no significant impact on compression efficiency if one uses different code block sizes at different DWT levels compared to the fixed size code block size.
There is no clear answer to the question of which block size is best. 64x64 gives the maximum coding efficiency (minimum distortion for given coded size) almost always with 32x32 some few percent less efficient. 64x64 usually also results in slightly faster processing since there are slightly less processing overheads.
The main reason we have always left 64x64 as the default is that it reduces the risk that others might blindly use 32x32 in benchmarking the compression efficiency of JPEG2000.
On the other hand, 32x32 blocks given finer granularity which leads to more efficient service of small windows of interest via JPIP. Also 32x32 blocks end up requiring only half the processing memory during encoding or decoding.