Better Jukebox Software for Bigger Libraries?

jimjenkins1975 asks: "I recently ripped and encoded my entire CD and Vinyl library, as well as merged my home and work computer's libraries (I work at a music company so my work library is very very large). It resulted in well over 750 GB of MP3's. I was hoping to get away with using iTunes to manage this, however the XML database file has grown very large, and the application itself is non-responsive or very sluggish at best, once it has loaded up (a process that takes several minutes itself). Is there another application (preferably for Mac, but I do have a PC) with similar features out there that can handle a library of this size with aplomb?"

Comments

Amarok in Linux

[rhythmx]

As a GNU/Linux user, even though I refuse to run KDE, I have had the best luck with Amarok. My archive (only about 150 GB) is nearly entirely rips of my albums. It has just about the best interface I have seen for dealing with a large (and sorted) archive. The features I like most are album cover manager, last.fm integration, ipod-style (artist->album->track) menus, the wikipedia info and lyrics based on context, and the random-album play mode.

There is a gnome equivalent but it is not quite as stable. I can't speak for the MacOSX crowd, but when in Win32 (rare these days) I reluctantly choose to use Winamp.

Some tips from my experience:

• Be an ID3 tag-nazi - No player can compensate for 750 GB of badly named media. MP3Tag is your friend for batch editing ID3 tags.
• Sort all your files using a resonable naming system. I use '/path/to/archive/%Artist%/[%Year] %Album%/%02Track% - %Title%.%Ext%'. This comes in real handy for writing scripts to deal with an archive to large to manage by hand.
• Backup. Backup. Backup.

Re:Amarok in Linux

[Curien]

Question asked and answered. Discussion closed. Next!

Re:Amarok in Linux

[willpall]

Don't forget that Amarok can use PostgreSQL or (I believe) any other standard database for handling it's database. I'm sure that will make dealing with 750 GB of music MUCH more efficient.

Re:Amarok in Linux

[bhima]

You know the exact same can be said of iTunes. I have about 350GB and have noticed a slowdown myself but when it was 150GB I saw no speed problems at all. 750GB is a lot of data, no matter what kind of data it is... I'd hate to use my old G4 cube to deal with an music library that large.

So I think the real answer lies along the lines of what would a database pro do if his application got so big it started to slow down on his existing hardware?

Answer: Buy more, better, and faster hardware.

Re:Amarok in Linux

[spagetti_code]

I'd like to amplify the comments about being an MP3 tag nazi
(does that count as Godwinning??).

The beauty of spending all your time getting the MP3 tags right and the
album art and lyrics *embedded* into the tags is that
someday, guaranteed, you are going to have to move
from one media organiser to another. Amarok wont always be
the killer app and some other smart organiser will take its place.

If you're tag data is good, that switch will be trivial. If, like
itunes, it stores some data in a DB rather than tags, you get locked in.

Tagrunner (commercial s/w) is my pick.

Re:Amarok in Linux

[KDan]

*Dzing* Incorrect answer. Thank you for playing

A "database pro" would look for unoptimized queries and missing indexes and speed up the database on the current hardware. As a matter of fact, most "database pros" work in environments where commissioning new hardware takes 6 months or so and goes through some ridiculous approval processes, so that's actually the hard option.

Daniel

Re:Amarok in Linux

[ajs318]

Wait ..... did I just hear "MySQL" and "proper database" in the same sentence?

Oh, you were comparing it to SQLite. Sorry. Point taken .....

Re:Amarok in Linux

[phoenixwade]

I'm sure we'd have heard a lot more about it since it was first suggested in 2002. We might if we could still hear......

Re:Amarok in Linux

[twistedsymphony]

From:

So I think the real answer lies along the lines of what would a database pro do if his application got so big it started to slow down on his existing hardware?

To:

Your explanation is completely contrary to my experience in dealing with DBA's. When dealing with code they did not design or implement...

Nice switch there...

As someone who designs lots and lots of database application, managing nearly 30 years worth of engineering data, as my full time job... I'd be looking for a new job pretty quick if my solution to every performance slowdown was to just throw more hardware at it.

If your saying: How can I get iTunes to run faster then yes, throwing more hardware might be the answer but that's only because you don't have access to the actual code that runs it... there isn't much of anything you can do about it. Chances are when friends and family ask for help getting things to run faster "buy faster hardware" is the only thing a DBA would have to offer because they didn't build any of the applications in question.

Though, you posed the question as if someone was building an new application from scratch, or at least had the ability to modify an existing application, in which case any DBA worth hiring would look for ways to optimize their queries. Nevermind the fact that iTunes uses XML to manage it's data as opposed to a more conventional database.

Sometimes better hardware is the answer, but it's never the first one you go to. More often then not in the business world you're required to build your application within the constrains of existing hardware.

Re:Amarok in Linux

[LWATCDR]

"750GB is a lot of data,"
It is but itunes should be dealing with a tiny fraction of this.
An MP3 might be three or for megabytes is size. The tag information should be far less than one kilobyte and probably a quarter of that. So let's say that he has 100,000 songs. So at one kilobyte per song runs to a grand total of 100 megabytes of data and 100,000 records in the database. That isn't a big database at all. I have one that I run on an old 300 Mhz P2 using Postgres that has over 400,000 records in a single table. Look uptime for any record is under .1 second using an indexed search of course.
SQLite would handle it with NO problem. Heck a good Xbase system could deal with it.
The problem with Itunes sounds like XML addiction. I swear people think of XML as the perfect tool for every problem. An XML flat file sucks as a database. Itunes is probably reading it all into memory to parse the file which will most likely cause a thrash unless you have a lot of memory free.
It is a fine solution for small datasets but as people here have found it really starts fall apart once the data set gets too big.
I haven't gotten into yet but SQLite looks like a great solution for many programs that need some type of data storage but don't need the complexity of a full blown database.

Mediamonkey

[afidel]

mediamonkey claims to handle 50K+ files without slowing down. It's amazing what you can find in seconds with google =) The search was mp3 media manager.

Re:Mediamonkey

[bigmouth_strikes]

> It's amazing what you can find in seconds with google =)

I'm sure most /. users know about Google. Asking a question on ask./. is about getting opinions, not so much about getting information.

Media Monkey

[complete loony]

I've used Media Monkey on windows for a 45GB archive. I went looking for a replacement when winamp stopped being useful.
The only other thing I can suggest is just using the filesystem to organise your music.

Quod Libet, of course

[keitosama]

Quod Libet is a fantastic GTK+-based music player designed for gigantic libraries. There are so many ways to search in it (for instance, you could search for &(genre=pop, genre=rock, #(lastplayed > 30 days)) to find every pop rock song you haven't been listening to for the last month, if you've got the tags right), so finding the tracks you're looking for shouldn't ever be a problem either.

amaroK

[wall0159]

amaroK works really well for me on ~14000 tracks (80gb). It uses either a mySQL or SQLite database for indexing, so I would expect it to scale pretty well. It supports mp3, ogg, aac, wma, ipods, irivers, ... it's the best and most flexible music player that I've seen.

Amarok again, but with some tips

[aitikin]

Amarok is by far my favorite "jukebox" program. There are only two things it doesn't have that iTunes does and those are the jukebox look (coverflow) and the APE (air port express) integration. Now, you mentioned OS X. Amarok is a great program, and when it's finally ported to Qt4, I will no longer use iTunes unless I have to. Here is a guide for getting Amarok running in OS X, and here is one to get it running "natively". There's a bit of a conversation as to an .app package for it.

mpg123

[Zork the Almighty]

I use mpg123 and the file system hierarchy to organize and play my mp3s. I have no idea what the hell you kids are talking about.

Re:mpg123

[ajs318]

That's the way I did it too -- all directories are named for groups, within which are subdirectories for albums. I even built a CGI-based front end so I could play tracks with a web-based interface. It turned out to be less useful in practice just because I was running the machine without a display. My TV set has RGB inputs (and I know how to generate csync from hsync and vsync) but I'm not sure the machine's onboard graphics will run down to 15kHz.

www.foobar2000org

[doofusclam]

Foobar2000 is my choice for *managing* my collection, which is currently about 55000 tracks split equally between aac and musepack files, totalling about 320gb.

The tagging and conversion features are unsurpassed and it's still nimble even with a collection that size. I don't use it for actual playback, for that I use mpd on my linux box.

hth

well over 750 GB of MP3's

[Anonymous Coward]

Torrent, please?!

Mmm, pop rocks...

[Chris Pimlott]

for instance, you could search for &(genre=pop, genre=rock, #(lastplayed > 30 days)) to find every pop rock song you haven't been listening to for the last month

Mmm, pop rocks...

Amarok - Potential Problems

[penguinchris]

I started using Amarok to do this recently, but I've found that indexing only 40-50 Gb of my collection it chokes up my computer for ten minutes or more. It has completely frozen my computer several times doing this as well. Not just the first time, either, but obviously whenever it needs to re-build the database. It unnecessarily does this if you add an additional folder to the database, which is annoying. It recursively scans your folders if you want it to, and it adds things to the database quickly that way, so why does it have to rebuild the entire database if you add a folder?

Some annoying problems often occur when transferring music to my mp3 player (jetaudio X5L), which it does in a customizable and convenient way so I would like to keep doing it in the future. For now, though, I've gone back to just copying files from konquerer directly. Sometimes Amarok doesn't name the files correctly and album tracks end up out of order on the player, sometimes it claims to have transferred files but they are nowhere to be seen when I'm out somewhere and really want to listen to that new music I thought I just put on the player, and things like that.

Your mileage may vary of course. It's a great program, and I highly recommend it over anything else that I've tried. It's very possible that it's just my particular setup that causes problems with it, considering that the other comments indicate that it works well with much more data than I am using it with.

Of course you're asking about Mac software specifically; maybe this wasn't the best place to do that.

Slimserver supposedly scales that high

[martinde]

Slimserver, while traditionally used to drive a Squeezebox, can stream to any player that can stream MP3 format. (And probably FLAC, AIFF, or WAV, I've never tried it though.) The latest version uses mysql as a backend and I've seen people talk about very big collections like yours on the mailing list. FWIW, I have a squeezebox (rev. 1) and I love it.

At work I have done the other thing people mention, which is attempted to rigorously organize the directory structure my MP3s are stored in, and then used good old xmms to play directly from the filesystem. I see other people talking about amarok but every time I have attempted to use it it's very unstable for me. (My collection is about 80G and it never seems to make it through scanning it.) Is the secret to backend it into mysql instead of letting it do sqllite? Or maybe it's artsd that is problematic? Would anyone like to share their Amarok best practices?

MusikCube

[fozzmeister]

Is a better iTunes in some ways, worse in others, but its built on an SQLLite backend which is semi-exposed and is _super_ quick on my 120GB collection

Re:External HD

[Corwn of Amber]

• Be an ID3 tag-nazi - No player can compensate for 750 GB of badly named media. MP3Tag [mp3tag.de] is your friend for batch editing ID3 tags.

• Sort all your files using a reasonable naming system. I use '/path/to/archive/%Artist%/%Year %Album%/%02Track% - %Title%.%Ext%'. This comes in real handy for writing scripts to deal with an archive to large to manage by hand.

• Backup. Backup. Backup.
• format the whole thing in FAT32

... and the story ends.

foobar2000

[Fweeky]

fb2k is known for being very effecient, even in the face of crazily huge libraries. I dare say you'll hate the default interface/config, but it's not difficult to bend it to your will (though it's not exactly iTunes; more like vim/mutt for music).

Windows only unfortunately, though it is supposed to work well in Wine. Significant chunks of it are BSD licensed.

MySQL and mpg321

[ajs318]

What's wrong with MySQL and MPG321? As in

$ for SONG in `echo "SELECT full_path FROM songs WHERE group LIKE 'kylie%' | mysql music"`; do mpg321 $SONG; done

Same for video?

[LordSnooty]

Has anyone come across something similar for video? I've seen videodb but that seems more geared to DVD. I want something that will catalogue all my downloaded xvid so I can tell straight away if I had a particular title, instead of hunting through a stack of discs. Hashing or fingerprinting them files in some way would also be good, so I can start to share my collection with other people.

Imagine this: set up a torrent tracker, get your members to catalogue their video collection, combine that into one list of all available video, then if someone wants a particular file, the tracker will be able to ask all members with that file to start seeding.

iTunes FYI

[dave420]

iTunes doesn't use the XML to store its library - the XML is there purely to be used by *other* applications. iTunes keeps its library in its own proprietary format, similar to the format of the iTunesDB file on iPods, which is completely binary in nature, and muuuch smaller than the XML spat out :)

I like iTunes because of the COM object, mainly. I wrote a script that uses MusicBrainz to tag my music in iTunes automatically, getting Amazon artwork for that missed by the iTunes Music Store (and embedding downloaded artwork for those with only the downloaded variety, which iTunes doesn't like putting in MP3s on its own).

If I could find an application that allowed media management just as good as iTunes, with the playback features, artwork shits, etc. then I'd jump ship in a second. Especially if it had a SQL back-end. dirty. :)

Re:WinAMP

[Risen888]

Assuming these are mp3, at an average size of 3.5 MB per file, we're talking somewhere in the neighborhood of 70 GB. Correct me if my assumptions are unfounded. I'd call that "medium-big," but not huge.

Anyway, if you're looking for a good GTK jukebox (and yes, Rhythmbox totally blows), check out gmusicbrowser. Excellent browser, can use gstreamer or mpg321/ogg123/flac123 or mplayer as a backend, very adaptable interface, snappy as hell with my 80GB or so (mixed bag of flac/ogg/mp3), I can't speak for how well it'd hold up under a heavier load. It's initial scan with that much data takes a goddamn long time (start it before you go to bed), but once the files are scanned, you're golden. You can set it up to scan for new/deleted/changed files on startup, which is much faster and less processor-intensive than the initial scan, but even that can be turned off or done manually from the prefs menu. Enjoy!

XMMS2 + MusicBrainz Picard

[cparker15]

I've had pretty decent experiences with XMMS2 for playing music from my library and MusicBrainz Picard for organizing it.

One of my requirements is the ability to add an SMB share directly to the media player's library, as my entire music collection is stored on a media server (Maxtor MSS Plus) and accessible via an SMB share. Amarok is unable to add an SMB share directly to its Collection, and requiring root access to mount an SMB share is just stupid, IMHO. Rhythmbox is capable of using GNOME's solution to the problem, the "Network Places" shortcuts, which are GNOME-specific connnection configuration settings saved in GConf and represented as "shortcuts" on the desktop and within Nautilus (and applications that use Nautilus in them). However, as we're all aware, Rhythmbox totally blows chunks.

XMMS2 even runs efficiently on low-end hardware. I turned an old Dell OptiPlex GXM 5166 I dug out of storage (specs, picture) into a headless XMMS2 box. I control it using TurboX2, which is also installed on the old OptiPlex. Playback is perfect, even with a 166 MHz!! clock speed.

I have a little over a month's worth of MP3, Ogg Vorbis, FLAC, and MOD tracks in my music library right now, and I'm adding more on an almost weekly basis (I <3 Used CDs).

Winamp 5x

[casualsax3]

I have around 20,000 files in my collection totaling 90GB and the Winamp Media Library handles it admirably. If you can break things down using the smart bookmarks you'll be even happier with the performance - here's how I have mine setup: (this picture isn't from the computer with my whole collection, but you get the idea)

http://mr2.phpwerx.net/Photos/Sully/stuff/full/w inamp.png

I can go to "Audio" which shows me all of my tracks, or I can go to "Classic Rock" or "Rock" which contain smaller amounts of music, and load a bit faster. Also plays nice with my iPod, including album art.

Do you like all that music?

[teridon]

Honest question here, because I am puzzled. Do you actually like all that music? I have about 40GB of music, but I only listen to about 12 GB of that with any regularity. All the other music I have just isn't that good, and I haven't gotten around to deleting it from my hard drive.

I very rarely find new music that I actually like -- so I'm puzzled when I hear that someone has a 750GB music collection!

Am I just too picky? :)