Digital music 2.0

Digital music is mainstream, that’s for sure. But we’re only now seeing the true power of what having one’s collection digitized can do — beyond the obvious portability of it all.

In the vinyl and CD eras, one navigated a music collection by album. There was no other way to do it. You selected the record/disc and then maybe the track and that was that. Digital music libraries with filterable metadata, smart playlists, and all manner of apps for organization make navigating music a lot more flexible, if not significantly easier than the old days.

CoverBuddy gives you an iTunes-like interface that represents albums as cover art thumbnails. CoverFlow takes this concept one step further and presents 3D cover art that you can flip through as though through a booklet of CDs.

Fun for sure, but it is online music services that truly open up possibilities.

Most talked-about these days is Pandora from the Music Genome Project. Basically a streaming radio station set atop a massive database of style data for thousands of songs and artists, Pandora delivers tunes based on the internal characteristics of a single song (or more) that you like. Once you start listening you can further hone your tastes — er, genetic composition.

Pre-dating Pandora is Last.fm and their Audioscrobbler service. This too delivers customized recommendations and a personal radio station, but it is based on what like-listening users have played rather than a close (human) analysis of styles. (It also powers the playlog of Ascent Stage.)

MusicBrainz offers a different slant. Think of it as a wikipedia for musical meta-tags. CDDB on steroids. MusicBrainz offers downloadable applications to help you properly tag your music in a way that reflects the user-contributed info in their vast database.

But this presents a problem — at least to me. Certainly artist and album information can be somewhat standardized, but much of tagging is subjective. For example, I find that I almost always start listening by heading into a genre first and then to an artist and then (maybe) to an album. The genre category is my front door. But it is also the most subjective and least standardized. One person’s Ambient is another’s New Age, Heavy Metal another’s Hard Rock; Dance another’s Techno. But that’s a good thing. There’s opportunity for personalization, to make the categories your own. Here are mine.

1980’s If it was released in this decade and has that new wavy feel (i.e., not classic rock) then it goes here. Obviously a problem category since it is the only chronological one.
Alternative If it was ever played on mainstream radio and is not 1980’s or classic rock, it goes here.
Ambient Mostly electronic. Not New Age.
Audiobook Including spoken word.
Children’s Obvious, though certain bands like They Might Be Giants have kid albums that might as well be in other genres.
Christmas Obvious.
Classic Rock There’s certainly a cutoff date for this in my mind, but I have no idea what it is.
Classical Obvious.
Country For my wife. Please disregard.
Electronica Most everything, but increasingly difficult even to know what part of an album constitutes electronic.
Halloween Obvious. (I love Halloween.)
Jazz Obvious, though there’s much overlap with certain sub-genres of Electronica.
Mashup My newest genre. For categorizing music whose reason for being is to mess with generic labels.
New Age Gotta put Ottmar Liebert somewhere.
Oldies I suppose this is chronologically-bounded too. Classic rock and roll, pre-1970.
Original My own music.
Pop Not 1980’s, not rock, not alternative. Prince, for instance.
Soundtrack Both scores and soundtracks, actually.
Surround Sound There’s no confusion on this one. Pure sonic muscle-flexing.
World Global styles.

There are more here than I would like, but this is the smallest number that adequately divides. My feeling is that keeping the number of these doors few is key. Too-fine generic subdivision makes a top-level category useless. I have a friend who sub-divides using the Grouping tag religiously. (There’s even a guy out there who hacked iTunes to let him more easily categorize classical music.) Yet, to me, that way insanity lies. You can always further describe something, but how much is enough? Is genre a function of chronology, musical style, popularity?

I’m disgressing. The point is that there is no answer to these questions and that is a good thing. Genre is personal. I’m the first to admit that my categories make no good sense and overlap horribly. I’m all for data standards, but not in this case.

Which isn’t to say that digital music depersonalizes the experience. If anything it has multiplied the possibilities of expressing oneself. Collaborative, themed mixes are all the rage these days. And just recently Jason Freeman released the iTunes Signature Maker, a stunningly cool app that scours your music collection and creates a unique sonic “signature” of your musical taste — a kind of schizophrenic flashback through what matters most to you. The output is uncanny. Here’s mine (2:12 minutes, 3.1 MB, MP3).

2 Responses to “Digital music 2.0”

  1. cbjg says :

    where to begin…with my experience in knowledge management, I could never adopt some of your genres (e.g., soundtrack, 1980s, alternative) as they do not specifically refer to any type of music (imho). to me genre should be MECE (to use a nice consulting acronym – mutually exclusive, collectively exhaustive) – hence the religious use of the various sub-genres in the grouping field. Allmusic.com is a great resource.
    I use smart playlists for everything and the better the metadata in your tags, the more powerful these become. My only complaint in taking this approach is that the grouping field is free-text and I do not know why apple has not yet incorporated “keyword” functionality (like in iphoto) so you could easily add, modify, delete single terms without have to redo the entire field across multiple songs.
    As for the itunes signature (what a great little app!), after some experimentation, I found that selecting based on my ratings, no restrictions on artist/album combos and only overlapping 2 songs at once produced a nice signature where you could identify many of the songs.

  2. cbjg says :

    more on the itunes signature app – we need something that will analyze, compare and contrast multiple signatures and identify the connections between musical tastes – the DNA of itunes! while iscrobbler sort or gets to this, it doesnt go far enough.