#musicbrainz

How odd. http://images.google.co.uk/images?um=1&hl=e...

looking at a java api

input appears to be a trm identifier

how do i get that trm for a given piece of sound ?

trms are no longer used

you might have an old library

old as haystack java api yes

you want to look for PUIDs

how

well, that's the technology that's used for song identification, so you'll need to find a library that lets you do that

I'm not sure about the state of the java apis

can i use any of the other fingerprinting systems ?

other than musicip's

no

no because i need a puid

i fold in trying to understand a solution

what libraries am i supposed to use to implement a tagger

how do i turn sound into this proprietary identifier required to index your musicbrainz database

once called musicdns

this is BULLSHIT software archeology work and i dont think its leading anywhere

i'll entertain replies, but after being so hostile none are really necessary

took me some time to get http://code.google.com/p/musicip-libofa/ out of apt-cache depends picard-tagger but there it is

its really a slap in the face amplifind has 300' landed so many links

Depends: libofa0

ok well then bye for now

http://musicbrainz.org/release/d0bdc687-8553-45... <- This CD lists artists as "Brenda Stubbert, fiddle; Brian Doyle, guitar; Richard Wood, piano". Should the release then have something like "Brenda Stubbert, Brian Doyle, Richard Wood" as the artist?

I have what appears to be an earlier release of http://musicbrainz.org/release/e2e3cdd1-e643-4a... which gives all text in English, German and French. Since the original work is from a German composer, should the information in MusicBrainz be in German?

That is, "Wassermusik, Suite Nr. 1 in F-Dur: I. Ouverture (Largo - Allegro)"?

Or Suite No. 1, even.

Leftmost: the water music was actually written for king george the 1st, so you can be damn sure it was named in english :)

and well, it's pretty much your choice what language you choose when the release is multilingual. just stay consistent

Well, George I was German-born. I'm tempted to modify it all. Just worried someone will hit me on the head.:)

he was the king of england ffs...

Doesn't mean he didn't speak German.:)

....

and for your earlier question: http://wiki.musicbrainz.org/Advanced_Relationships

you're looking for http://wiki.musicbrainz.org/Performance_Relatio...

Yeah, I figure it should be possible to get fine-grained with what they did with AR, but I've gone ahead and changed the artists since they're given equal billing.

well err http://wiki.musicbrainz.org/Instrument_Relation...

you don't do that unless they're a group of artists that have multiple releases in the same formula

as in, "real albums", not a compilation...

i.e. only something like this warrants a new combined artist: http://musicbrainz.org/show/artist/?artistid=17...

Hmm. I never saw anything of the sort in the standard. That doesn't make much sense, though, as we're still not able to do multiple actual artists for a release or a track. (Though I understand that's in the new version?)

are those artists of yours the composers, or just the performers?

So the data either ends up having superfluous extra artists (which can be dealt with when the multiple artists thing comes up) or is just incorrect.

Performers. Composers are listed in the liner notes additionally, but I'm only doing it based on performer.

well, not sure about that kind of music, but i'd go classical style on it

composer as track artist, performers as performers, and the release as various artists unless those folks banded together for this specific release...

Hmm. I'm not sure I agree. Performer is usually who's credited for any track, especially since composer is often unknown.

on the other hand i might be utterly wrong :)

Leftmost, can you find the exact same piece of music performed by someone else?

So much of Celtic music is traditional and/or derived from other songs that composer is rarely a good measure. Composer is only listed for a handful of pieces and usually it's only for one piece in the medley.

.. even if the composer is [unknown]

Often not, since they're so often medleys.

I mean, a couple of individual pieces on there are performed by more than one of the groups, but not a full medley.

anyway, looking at this http://www.amazon.com/dp/B00005Y1U0 .. if you add a collaboration artist for each of those, you'll get spanked hard :P

those performers should have instrument ARs for each track. what you'll do about the track artist is a different matter

xlotlu, it actually breaks down to just six collaboration artists.

"Brenda Stubbert, Brian Doyle, and Richard Wood", "Wendy MacIsaac and Jackie Dunn MacIsaac", "Kinnon Beaton, Betty Beaton, and Owen D. Gillis", "Jerry Holland and Allan Dewar", "Buddy MacMaster and Joey Beaton", and "Kinnon and Betty Beaton".

well, you go ahead and do that if you feel the urge... :)

It just makes the most sense to me until the ability to add multiple full artists is available.

except they're performers..

Right, but I think the composer situation would be worse. Track 1 would be "Brenda Stubbert / [unknown] / [unknown]"

Track 2 "[unknown] / [unknown] / [unknown] / [unknown]"

Track three has four known composers and two unknown. Et cetera.

a mess is a mess is a mess

It's a lot less messy with performers, especially as particular groups of performers tend to perform together often.

why the hell does picard have a path length limit for renaming.....?

Even the long filename stuff on FAT32 is still limited, I believe.

there's no limit

There's some common enough filesystem that supports long filenames but is still limited. Can't remember which.

most filesystems are limited to a 255 chars filename, but there's really no limit to the path length

well, 32,767 for ntfs...

Ahh, actual path length? Hmm.

err. ugh. windows..

windows == pure evil

ntfs is capable of storing whatever, windoze limits it to 260

yes yes

Explorer, you mean.

In the Windows API (with some exceptions discussed in the following paragraphs), the maximum length for a path is MAX_PATH, which is defined as 260 characters.

winapi..

Hmm. If I'm doing classical in German, should it be Suite Nr. 1 or Suite No. 1?

xlotlu: It means the shell api, the actual underlying api supports the \\?\ disambigiouation syntax with allows arbitary length paths.

Which.

Actually, I think it's still capped at 32k, but that's pretty arbitary. ¬_¬

FauxFaux: and in practice that means... ?

xlotlu: Applications /can/ use long file names easily, but nobody bothers. Including the Explorer team.

haha

(It makes me so sad)

most people using windows make me sad :P

I bet people who complain without understanding all the facts make mesader. ¬_¬

So much lag on this connection, it's not my fault I can't type, I swear.

For example! I have an untainted debian kernel (not even any firmware), and the supported, free intel driver keeps locking up. Who's to blame now?!

obviously faith!

alternate answer: you for not submitting a bug report, in which you should blame the developers, and they'll get pissed and blame you for the bug report being useless, and then you submit a proper bug report after 10 hours of figuring out how to do it...

yes | reportbug xserver-xorg-video-intel -m "NOT WORKING"

... and blame the developers back, and then they'll figure out it's definitely not their fault, and blame the hardware, and code a special case for that hardware that must have the words fuck and shit at least a few times in the comments

guess i'll patch picard with an option to skip those annoying short paths

done

damn that file renaming annoyed me for too long

wee: http://bugs.musicbrainz.org/ticket/5638

how about making picard not submit a puid if the file's duration varies widely from the MB track?

where "widely" remains to be defined..

widely.. wildly..

xlotlu: I'd be happier if it still submitted, but http://bugs.musicbrainz.org/ticket/4810

brianfreud_: yeah, that makes sense. but you'd get spammed with (proper) 1-second differences

in the long run that's the way it should be. but i'd welcome a quick client-side hack to keep dumb puid submissions away

better to have some noise in the data than no data at all to work with... :)

s/welcome/code/ :)

argh. "some" ..

if it weren't such a major PITA to remove that noise, compared to the dumb press of a button, maybe..

funny. i can see in audacity the difference between a flac and a mp3 that generate different puids

bass on osteroratorium