March 2006 Archives
In the middle of this week, my mp3 drive decided to corrupt itself. There appears to be nothing wrong at the hardware level, but I had significant directory corruption.
Attempting to repair this using the utility that usually fixes all - Disk Warrior, made things significantly worse.
I turned to a piece of software that has saved my data (and that of others) on many occasions - Data Rescue X II. Took all night but I got back about 125GB of data - which was something approximating the amount of data that was on the original disk.
However, many recovered files were no longer helpfully named. Some were corrupted and many were lacking id3 tags which would have made re-discovering the names easier.
MP3 Rage came to the rescue with both detecting the thankfully small amount of files that were seriously corrupted and renaming those that still had ID3 tags.
Finally I thought I was left with approx 2000 - un-named files, enter ieatbrainz - a tool which sits between itunes and the musicbrainz audio fingerprint database.
This has allowed me to identify many files that would have otherwise gone into the "unknown" bucket.
I thought i could see an end in sight with this, 2000 odd songs to sort through in iEatBrainz and then merge this with the already tagged and named files.
However I discovered a bit of a problem - two in fact. Some music files were split into two by the recovery tool, this is very hard to identify unless you can hear the files are related. Another tool mp3 trimmer allowed joining of mp3 files and in some cases once joined ieatbrainz managed to identify them properly.
Secondly, and this more concerning - I noticed a lot of 100MB mp3s, this in itself was suspicious - as it's a pretty exact/large size for a music file. So I dragged them on mp3 trimmer thinking it was just a case of 1 track with 90 odd megs of padding due to the file recovery process. Not so, these 100MB files seem to contain 10+ tracks each, so in a russian dolls like process, I end up with more tracks to sort through.
Also, the 100MB tracks don't just appear in the "unknown names" bucket, many of the files that appeared to have proper id3 tags and were renamed appropriately - upon closer inspection some are also 100MB in size and will need to be split/analysed.
Fun Fun Fun. lotsa work on my part to verify/identify all of these files - this could take months.