Hello. Is there a software which checks differences between audiofiles? The software should find largest common segments from both audiofiles.
I also need to compare two audiofiles in two separate computers: instead of transferring the whole audiofile from computer to another, the software should compute checksums for large blocks and transfer only those checksums. Different offsets should be accounted -- by sending a few samples from beginning of each block, for example. An alternative would be to search the block for location which has most energy or most number of different sample values. Any ideas for developing those software? The software could work by thinking the first audiofile as reference and the other as possibly corrupted file. Then the reference file is divided to blocks, and list of entries containing <32 bytes of block> and <block checksum> generated (and transferred to other computer). All that was trivial, but it is not that trivial how to search the equivalent blocks at the other computer. Sure we know in what order the blocks should be, and that could be used in the algorithm. The problem seems to be how to start the block matching: take first block with enough energy and search all matches (not just the first one)? Then it would be easy to search both forward and backward. The enough energy here means that we don't match silences --- they give wrong matches easily. Best regards, Juhana
