I was wondering if you were ever planning on building some sort of audio track recognition, i.e. somthing like Shazam in the way that you could provide a list of audio tracks that your mp3 tag app would play each, for say 10-15 seconds to recognise the song, then pull the appropriate artist, album, track details, art work from the online database etc.
The reason for asking is that I recently 'allowed' (but never again) iTunes to manage my music library but for some unknown reason the process managed to complete screw up the file names / mp3 tag information. I've now a few thousands tracks sat in an UNKNOWN folder with just a track number as a file name and the mp3 tag details are all blank, hence I'm currently doing the above describe functionality manually which is becoming a tedious task in the little spare time I have.
My own experience with Picard is very bad. The 'recognized' songs are more then 80% wrong or will be allocated to the wrong album.
If you have a high quality in the song tags, Picard can help you. If the tags are empty, you will have a lot of manual work with your songs.
Actually, they discuss the introduction of AcoustID at MusicBrainz. But until they collect and support this fingerprints officially, it is just another 'nice try'.
Please let us know, if you find a way to improve the detection rate!!
Hence my recommendation to either replicate or team up with Shazam. I'm sure there are a few million people like me that are looking for such a 'all in one wonder' application that nails audio track identification and pulls the correct details etc.
Not to worry, I'll persevere with both mp3 tag and MusicBrianz Picard (MBP), what MBP doesn't identify correctly I'll use Shazam.