I downloaded 3 very large fake files, bitprints below. Each of these files is created in such a manner that my file sharing software identifies it as an MP3 file.

http://bitzi.com/lookup/OQFCMMODCLUJQT2YR64Q4XBDW75Z2OD4

http://bitzi.com/lookup/KZJH35GOPMDI35VFUY5PPCHZAPBAGPUL

http://bitzi.com/lookup/SEQWBAV6N7D64K2TIN676POGLHAHS7BR

Bitcollider attempts to extract tag data from these files where none exists. The time and CPU load it requires to accomplish this is considerable. ie: Choose any of these files with the file name set as MP3, and Bitcollider will take hours to hash.

It would be good if Bitcollider limited it's field of data from which to extract MP3 tag data, or time out if it cannot locate expected/required data within a reasonable amount of time for a lowly 386DX system, or just add a feature that can determine if a file is or is not an MP3 ignoring it's file extension (changing the extension works around the issue).