Files can be messy, and music files, doubly so. When ripping CDs into MP3 files, not everyone takes care to rename the files accurately--and even if they do, different people may have different naming and filing schemes. When that happens, MusicBrainz Picard (free) can help.
Fourteen years ago, I ripped several Faith No More CDs to MP3. At the time, I used a naming scheme that made lots of sense to me--I called the files "01.mp3," "02.mp3," and so on. I felt quite clever at the time, because by adding a leading zero I could get Windows Explorer to sort the files correctly.
Today, that naming scheme doesn't really work for me. I carefully file my music collection according to artist and album, with each filename containing the complete track name preceded by its number ("05 Tremendous Dynamite.mp3"). Converting my ripped MP3s to this scheme manually would take a long time, and could be a frustrating experience.
One step above completely manual conversion would be tagging the files using the CDDB database or FreeDB.org. These are both old, established projects (the latter based on the former) that work to recognize music according to its "digital fingerprint." However, they're both focused on complete CDs rather than individual tracks. In contrast, the MusicBrainz database (which Picard queries) is built around tracks, with each individual track getting its own fingerprint.
Another issue is that both systems employ user-generated content with no filtering. That means you can get ten listings for the same CD, or listings full of typos and strange capitalization. MusicBrainz lets users edit content, but other users must vote on it before it gets entered onto the database and served up in reply to queries. This makes for better results and less duplication.
Getting up to speed with MusicBrainz Picard can take a few moments. The first step is to "throw" your files into the app--you can literally drag and drop them in, or do it with a built-in folder tree. You then need to "Cluster" them, which divides them into sensible groups (by album, usually).
Once you've clustered the files, you can either click "Lookup" or "Scan." The difference is that a lookup is done on an entire cluster at once, and uses whatever existing metadata is already in the files to query the database. It's quite fast.
Sometimes Picard's lookup doesn't work, and then you need to scan the files. Scanning is done on a file-by-file process, and uses each file's audio fingerprint. It takes much longer, but it works well.
If you already know what files you're trying to tag (as in the case of my Faith No More album), you can also just manually query the server. Just type "Faith no More" into Picard's search box, and you'll be taken to a page in your default browser listing all of their releases. It takes just a click to pull the info back into Picard.
Once you have the correct album information, it's just a matter of dragging the files onto the album. Picard usually matches each file with its metadata correctly, but you can make corrections if needed.
Picard uses the metadata to update each file's tags, but you can also have it rename the files and even place them in folders according to any naming scheme you see fit. All in all, MusicBrainz Picard is a very efficient, high-quality way to bring some order into your vast, sprawling music library.