I have a non-technical, fun question for each of you - how do you
listen to music? :)
I had an interesting chat with a friend yesterday, and he listens to
music in a *much* different way than I do. This got me thinking -
since people listen to music in different ways, is there any way we
could use that to improve IMMS? Could we make IMMS more useful? Usable
by more people? Useful more quickly? Maybe if we understand how people
go about organizing and listening to their music it would help.
So, if any of you feel like chiming in, tell me how you listen to
music. Here are some questions to get you started:
- How do you acquire new music?
- How do you store/organize your music?
- How and where do you listen to music?
As an example, here is my answer.
I usually acquire music by purchasing CDs and then ripping them to my
computer (maybe 10% of the time), or else buying a song online (90% of
the time). I constantly hear songs that I like when walking around
stores, restaurants, clubs, lounges, etc. and I'm always bugging
people/waitresses/store clerks to tell me what song is playing :P
Once I have music on my computer I dump it into one large folder,
import it to one big playlist, and hit play.
I *hate* having to organize my music, make playlists, or tell the
player what is good and bad. I want the computer to figure it out for
me, and luckily, IMMS does. I never want to think about it.
I listen to music in three main places: home computer, work computer,
and ipod. My home computer dual-boots windows and linux, so right now
I can only use IMMS about half the time, and my work computer is
permanently stuck on windows.
From my own description I have three goals for IMMS:
- make IMMS run on Windows (working on it)
- make IMMS run on my ipod (current best bet, probably rockbox)
- make some kind of utility that can merge my db play records and keep
IMMS data in sync across multiple locations (I added a ticket to the
IMMS tracker for this -
http://code.google.com/p/imms/issues/detail?id=3 , and thanks to list
member Brian for discussions about this off-list)
Note I am not specifically trying to get feature requests or anything
from you, though if you have requests that's cool too. But I am
interested in how you store and listen to music, to see if we can get
a broader picture of what people do.
Feel free to reply off-list if you're not comfortable broadcasting. I
can collect and share anonymous results if people are cool with that.
Dave
> - How do you acquire new music?
When I was in college (starting roughly 5 years ago) I began by
ripping a few CDs, then amassed a somewhat substantial collection from the
college LAN. After moving back home, I ripped yet more CDs from my
family's collection. Other than that, my two biggest sources are Youtube
and Web archives of free music like "Hear the Choirs Sing." I don't
*acquire* new music much at all nowadays, actually. Lately, I'm starting
to get back into last.fm, which I used briefly in 2007. But disk space is
tight enough that I can't yet afford to get much more music at the moment.
> - How do you store/organize your music?
I store my music on my hard disk, in a directory tree that's kept
synchronized between my computers. The organization is a unique structure
that makes sense to me, in which the three primary criteria were
keeping directory sizes small enough to be manageable, keeping full-path
filenames relatively short, and making sure that I could remember what a
file was from its path and filename without having to look at the tags. I
have various "favorites" directories (at first single directories, but now
full trees) to which I hard-link my favorite music. (Actually, most
recently I just make a full duplicate directory tree full of hard links
and remove any track I listen to that isn't a "favorite.") The idea here
is that I can pick a track at random from the "favorites" directory and
not have any non-favorite music, or pick a track at random from the whole
thing and have an increased chance of picking a favorite track.
> - How and where do you listen to music?
I usually listen to music at my computer (which, whether it's my laptop or
desktop, is almost always running Gentoo Linux). Because I tend to run
CPU-, memory-, and I/O-intensive programs at levels slightly beyond what
my computer was designed for, particularly nowadays, a heavyweight media
player like Audacious is usually too bloated for my needs, so I use
MPlayer. (When my system isn't heavily loaded and I think of it, I use
Audacious with IMMS, as I have since XMMS was dropped from Gentoo. I
occasionally try out other media players, but they never quite measure up
to IMMS.) Since there's no way I know of to use MPlayer with IMMS, I have
a script that plays five random files from my current main "favorites"
directory and optionally asks me after each one whether to delete it from
that directory. If I want specific music, I have another script to play
the files and directories I specify in a random order (but an order
determined beforehand, so "play the previous track again" does what I
want).
On the rare occasions I'm on Windows, I mount the Linux partition as a
drive and use the Windows version of MPlayer, using those same scripts
under Cygwin.
--
Jonathan Lovelace
kingj...@gmail.com
http://shinecycle.wordpress.com
I strongly support RFC 1855.
1. I buy CDS and rip them to ogg/vorbis using abcde.
2. (much less frequently) I buy mp3s online
3. I scour a few resources for free mp3s by hand, mostly:
http://www.rcrdlbl.com
http://www.daytrotter.com/
and amazon
4. I wrote a script that spiders a list of mp3blogs and downloads any
mp3s there automatically:
https://launchpad.net/barbipes
> - How do you store/organize your music?
The stuff I rip/download manually is meticulously tagged (using
musicbrainz picard) and placed in a directory structure like:
/artist/album/tracknumber-title.ogg
Most importantly I delete a lot of tracks that I end up not liking,
which is necessary, because 4. fills up any disk space I purchase pretty
quickly. This means a lot of albums are missing songs, even the ripped
ones, and people who care about albums as an indivisible unit of the
artist's intentions think I'm a barbarian. ;)
> - How and where do you listen to music?
On my computer, almost exclusively. I work from home, but when I didn't
I used to synchronize my entire collection between my work and home
machines. (using rsync.)
I listen using quodlibet http://code.google.com/p/quodlibet/, the best
music player I've yet found, and very easy to extend in python.
I wrote my own plugin (which has some overlap with imms) "autoqueue"
https://launchpad.net/autoqueue that queues similar tracks, to create a
more or less consistent but not predictable listening experience from my
huge (~1.5 TB) and very diverse library.
It looks up similar tracks in 4 ways, in order:
1. acoustic similarity using mirage http://hop.at/mirage/
2. similar tracks from last.fm
3. similar artists from last.fm
4. songs with the same folksonomy tags locally (I use the id3 'grouping'
tag for this, and synchronize the tags with last.fm.)
Any of these can be turned on or off.
The plugin is very configurable, and there are settings for:
- Number of days between plays of songs by a certain artist/performer.
- Number of days between plays of the same file. (Scaled by rating)
- Length of queue to keep generating. (This is useful if you want to
generate a CD worth's)
- Restrictions: an arbitrary quodlibet search that restricts what songs
can be queued, so that I can listen to music released in 2010, or music
with specific words in the title, etc. A very helpful tool for
generating themed mix cds.
Hope this is of any use.
--
eric casteleijn
https://code.launchpad.net/~thisfred
http://thisfred.posterous.com/
>I store my music on my hard disk ... I have various "favorites" directories to which I hard-link my favorite music.
>The idea is that I can pick a track at random from the "favorites" directory and not have any non-favorite music, or pick a track at random from the whole thing and have an increased chance of picking a favorite track.
Interesting! So is this folder structure a work-around because IMMS
doesn't really do what you want? Or maybe you just like to be able to
make quick/good playlists? If IMMS had a feature that let you generate
a playlist of x songs or x minutes of music would help you at all? Or
would you keep the folder structure regardless?
>Since there's no way I know of to use MPlayer with IMMS, I have a script that plays five random files from my current main "favorites" directory and optionally asks me after each one whether to delete it from that directory. If I want specific music, I have another script to play the files and directories I specify in a random order (but an order determined beforehand, so "play the previous track again" does what I want).
Another list member - Brian - mentioned MPlayer in another thread. If
IMMS worked for MPlayer would that help you? Wikipedia claims that
MPlayer supports IMMS plugins; I wonder if we could use the existing
IMMS plugin and get it working on MPlayer. If you are running immsd
under cygwin the same plugin might even work on the Windows version of
MPlayer...
I created a ticket for IMMS on MPlayer -
http://code.google.com/p/imms/issues/detail?id=15 - and I'll look into
it in the near future.
Excellent information! Thanks for sharing :)
Dave
I am not familiar with musicbrainz or mirage, so thanks for the tips
on those. So did you write autoqueue because you were using IMMS but
it didn't do what you want? Or did you write autoqueue first? Would it
help you if IMMS had some of the features you mention from autoqueue,
or are you much happier with your existing script now?
It is very cool seeing people whose habits differ greatly from my own.
> Hey Jonathan, thanks for the reply.
>
>
>> I store my music on my hard disk ... I have various "favorites" directories to which I hard-link my favorite music.
>> The idea is that I can pick a track at random from the "favorites"
>> directory and not have any non-favorite music, or pick a track at
>> random from the whole thing and have an increased chance of picking a favorite track.
> Interesting! So is this folder structure a work-around because IMMS
> doesn't really do what you want? Or maybe you just like to be able to
> make quick/good playlists?
This folder structure predates my usage of IMMS (or any graphical media
player); it's primarily so that I can easily find a file I want using the
command line (with tab-completion) but not have the terminal screen fill
up if I want, say, ten songs from different subdirectories.
> If IMMS had a feature that let you generate
> a playlist of x songs or x minutes of music would help you at all? Or
> would you keep the folder structure regardless?
Yes, a playlist generator would be very helpful. (An option to generate a
playlist of songs taking up a certain amount of disk space would also be
useful; I have an MP3 player that I use very occasionally.)
>> Since there's no way I know of to use MPlayer with IMMS, I have a
>> script that plays five random files from my current main "favorites"
>> directory and optionally asks me after each one whether to delete it
>> from that directory. If I want specific music, I have another script
>> to play the files and directories I specify in a random order (but an
>> order determined beforehand, so "play the previous track again" does
>> what I want).
> Another list member - Brian - mentioned MPlayer in another thread. If
> IMMS worked for MPlayer would that help you?
Yes, it would help if there were some way to tell IMMS about my listening
habits when I use MPlayer (which tracks I quit midway through and which
tracks I listened to all the way through), and to get adaptive playlists
that I can give to MPlayer.
> Wikipedia claims that
> MPlayer supports IMMS plugins; I wonder if we could use the existing
> IMMS plugin and get it working on MPlayer.
The MPlayer docs
<http://www.mplayerhq.hu/DOCS/HTML-single/en/MPlayer.html> mention XMMS
*input* plugins, not any other kind.
> If you are running immsd
> under cygwin the same plugin might even work on the Windows version of
> MPlayer...
I haven't tried compiling IMMS under cygwin.