Tagging IV

In my previous mail (now archived at http://www.zunester.com/tag1.htm, with picts inline), I’ve shown how important media tags are to the smooth operation of media players and devices. This time I will discuss some of the problems tags cause when things go wrong, and some solutions for those problems.

PROBLEM: Missing/Incorrect data

Given the importance of tags for successfully using your media library, files that include incorrect or missing tag information can be a huge problem. Genres aren't very useful if they don't match your listening habits, and artists aren't very useful if they're mis-spelled, or listed as "unknown". Traditionally, this is a key reason users didn’t use media libraries or devices – they just used Windows Explorer to sort through lists of files. As media libraries have grown, this method has failed to scale. So users have had to spend a lot of time using specialized and arcane "tag editors" to look up and/or manually enter information.

As a result, the contents of your tags can become more valuable than the music they contain. I personally am religious about backing up my music collection, not to protect the music, but to protect my tags. Re-ripping my music would take time, but re-catogorizing 10,000 songs by genre would be horrible.

Of course, many users have files that come from random sources, and those users must not only redo genres, but must fix up artist, album, and track number tags to get their media working. This takes a lot of time.

SOLUTIONS for Missing/Incorrect data.

"Bulk" tag editing tools, where you could specify the "artist" tag for 200 files, or the "genre" tag for 1000 files in one step. Windows Media Player 9 Series introduced a new bulk tag editor that enables exactly this scenario, and it works for mixed libraries of WMA and MP3 files. Another favorite tool of mine is Tag&Rename, which includes extremely powerful bulk editing tools. In addition, it has the ability to pull tag information out of filenames so if you've got 50 files named "artist-album-songtitle-track#.mp3", it can move this information into tags where it's more useful. Numerous other such tools exist.

But such bulk editors still require knowledge and time on the part of the user. When we studied users, we found that most don't even *know* tags exist - they just knew that the media library is a tree view of artists, albums, and genres. For this reason, we enable drag-and-drop within the media library pane. Users can drag songs or whole albums from one genre or artist to another in the tree, and the player takes care of the tagging in the background.

Another way to clean up your tag data has recently been added to some media players. Windows Media Player, Tag&Rename, and MusicMatch are now capable of getting information about files from internet services like AMG and CDDB. For example, in Windows Media Player, the user simply indicates a file to be tagged, selects "Find Album Info", and the player attempts to identify the song based on the filename and tags. If it's not sure, it offers several options for the user to choose from, and then fills in the tag information for you.

A unique feature of Windows Media Player called "Auto Info" feature will *automatically* supplement missing tags for songs it has identified. WMP users need never even know that their data was missing it'll just be fixed up in the background for them.

PROBLEM: Multiple tags

While missing information in tags can be vexing, at least it’s understandable. The Multiple tags problem is both frustrating and confusing.

One unfortunate consequence of the many versions of the ID3 standard is that an MP3 file can have *several different* tags in it. In practice, the file typically looks like this:

---ID3v2 --------------- Sound -------------ID3v1--

But what if the two sets of tags contain different – and contradictory data? Then you've got two types of contradictory tags. This happens when a user's library of MP3 files was created with some tool that wrote data to v1 tags. Now they are using another program that only knows about ID3v2 tags – but their *device* displays the v1 data. So no matter now many times they edit the v2 tags, the device will always show the wrong spelling of "Grateful".

The problem is further exacerbated by proprietary "extensions" to tags that some programs add to mp3 files. For example, one program extended the numeric id3v1 standard to include new genres, but many devices don't recognize those codes.

As more users are trying to bring mp3 files onto devices, they're encountering these problems with increasing frequency.

SOLUTIONS to multiple tags

A number of vendors have provided tools for debugging and fixing this problem. In particular, I recommend Tag&Rename for its fine-tuned ability to view and edit each type of tag individually. It even has a "tag synchronization wizard" for solving these problems.

The WMP engineers also recognized this problem, and we've done a number of things to help users behind the scenes. First and foremost, when you edit the tags in a file, we look to see if there are multiple tags and replicate all changes into all tags. So the user above would have modified *all* of his tags to be "Grateful" and the change would be reflected on *all* of his devices.

What's more, the Auto Info feature will not only supplement data in files, it will automatically *fix* file tags by synchronizing the v1 and v2 tags in the background. So again, just by using the media player, the quality of mp3 tags improves.