Back in the early Zee's (actually, before then, but saying Zee is more annoying), i had a Creative Labs Juke Box, with a mighty 6GB storage for all my tunes. Loading the beast with music meant using some semi-horrid interface that stuffed your tracks into an /artist/album/track type directory. It was slow, but worked well enough and various versions of whatever didn't choke when they hit directories containing several hundred entries.
Later i had a Zen, which had a bit more horsepower tucked into it's shiny plastic being, but it still used the same sort of setup as it's predecessor. This was further enforced because all the Music Rental Sites of the time continued to drive home the point that you needed to divvy things up before it could make any sort of sense out of things. Perhaps this was a relic of using MTP (microsoft's method of inflicting DRM on the masses) which also relegated tunage to subdirectories. So, for various reasons, i continued the practice. i'd rip albums and stuff the tracks into appropriate directories, then later copy them over to devices.
There was a problem with this practice, however. i had duplicate tracks show up.
Perhaps it was a track that i pulled earlier, but forgot to label, or maybe i just dragged it into the wrong directory ("Various" instead of "artist" or i got the album wrong for whatever reason.) To be honest, a lot of the tracks i was buying were winding up in the "various" directory because i'd only buy one or two tracks. Having Artist/Album/Track.mp3 was just a bit wasteful to me (and to the disk).
That's when it hit me, What the hell am i doing?
In databases, there's the concept of normalization (segregating data into tables to reduce the overall size) and warehousing (size be damned, we're going for speed.) Normalization forbids duplication, warehousing embraces it. Every MP3 player worth it's weight in lithium-ion batteries has a database built in. Said database automatically builds indexes for artist and album, plus genre and any other bit of meta data you care to toss into said MP3, WMA, FLAC, OGG or AAC file you like. What's more, Linux, Macs and Windows all have directory and search tools that do the same damn thing.
So i flattened my music.
Now i have one directory that contains all my tracks, all 1118 of them. Each track is titled uniquely, mostly "Artist - Title.type", but a few are just "Title.type" and that serves pretty much as a simple indicator of "hey, this is a track" Honestly, i could give them a name based off of the UTC that i acquired the track for all i care since i use the tracks meta data to identify the actual content.
In case you're curious, i had one conflict, and i renamed and recovered the file. i also had 48 duplicates that i was able to prune out. (and yeah, that explains why i was hearing some songs more often than others)
So, how are you keeping your music?
It does, but Songbird still fails. i can't right click duplicate tracks to delete files, it doesn't understand MMS streams (even though VLC has zero problem with them), and unless i'm scraping a music blog, i really don't see what's so all-fired impressive about it.
The search continues.
Artist/Album/Track Number - Song Name.mp3
I've got a *lot* of mp3's (something like 12,000 or so) so its easier to find them on my disk(s).
I switched to letting iTunes handle it all for me, mostly so that I didn't have to update everything myself anymore. I'm at around 15K songs, and when I get a new album, I just import it into iTunes. I also buy through iTunes, but when I do that, I burn a CD, and rip it to high bit-rate MP3 for storage. I know I lose some quality, but if iTunes ever dies, I'll have everything in non-DRM'ed MP3 format. It makes me feel safer about my music.
As far as the duplicate issue goes, the biggest problem I have is when I have older MP3s of an album, at a lower bit-rate. I end up with two files. I've made it through a lot of my albums, but sometimes there's ID3 differences that I made to one and not the other, which makes it difficult to match up duplicates in an interface.
My first player was a D-Link with something like 256Mb (upgradeable to 512!). It cost $40 and had a screen like a solar-powered calculator but it played for 16 hours on a AAA.
Then I got a 20GB iPod back when they were still black and white, but haven't used it outside the house for over a year (partically because it's the size of an old-school Gameboy, but mostly because I don't like iTunes' "one-way" path for music storage).
So now I have an 4GB Sansa with Rockbox loaded and it's absolutely the best MP3 player I've ever used. Drag-n-drop MP3s like a jump drive, switch to FM radio and The best part is that it cost less than my archaic D-Link abomination.
Save This Page
Sadly, this is a topic I think about A LOT.
Originally, I was in the well-named (Artist - Song.mp3) flat directory camp. Nobody had enough space or time to practically rip whole albums. I also used Windows Explorer to cue into WinAMP.
I dreamt of the OS being able to handle this stuff natively, casting directories dynamically out of the ID3 data.
Then came iPods and iTunes, and they both screwed my world over. In their directory-driven, album-oriented world, everything became nestled and the interface was database driven.
I hate it. It sucks. iTunes offers a lot of nice features, but I do not want it to run off its database - the data needs to come from the file system (ideally in real-time). It's how I roll. My editing of ID3 tags and filenames (largely via MP3TAG) needs to not interfere with the player's operation.
Nowadays, however, space is less of an issue, so a file system that is organized by artist/album/artist - song.mp3 is Ok with me. I sorta like albums, although the duplication problem you mention is quite prevalent as a result.
However, when I go to listen to music these days, I usually use the OS search for *.mp3 (or a subset) and use that result to feed VLC player. The non-intrinsic meta-data is for display, but my navigation whenever possible is based on the file system.
Having data repeated (in the directory/file name and in the ID3 tag) bothers me in a technical way quite a bit, but not as much as my application (iTunes or whatever) duplicating the ID3 data within itself. An unnecessary synchronization problem. I hope future consumer file systems free us of this - people will do far more interesting things with data.
Anyway, in answer to your question: In Artist/Album/Artist - Song.mp3 format. And it's always mp3; for better or worse, I maximize compatibility.