#musicbrainz

nikki: Wow, that change to the 'edit medium' edits is damn sexy.

isn't it just :D

nikki: Thanks

that's ok :)

ListMyCDs: (Lueskelin eiliset keskustelut). Nyt ehkä ymmärrän paremmin Sibeliuksen ja muiden sävellyksien eri versioiden merkityksen. Mun mielestä sun tyyli tehdä noita "töitä" on ihan hyvä. En kai ole kauheasti väärin linkannut?

ListMyCDs: Finlandian voisit tarkistella jossain vaiheessa. Siellä on ainakin yksi "heviversio" linkattuna (Tikanmäki), mut en tienny mitään parempaa sille silloin...

Jozo: Joo ei siell� kummoisia mokia ole ollut. Listaan tuonne kaikki Sibeliuksen s�vellykset ja sovitukset. Joskus n�it� on lipsahtanut v��riin versioihin kun kaikki versiot eiv�t olleet saatavilla.

ListMyCDs: http://musicbrainz.org/release/5452a736-8a75-30... - Otinpa ton levyn kätöseen, ku se löyty tosta pinosta. Mun kansikuvassa ole tota sinistä jutskaa, sulla ilmeisesti on? Varmaa pari eri painosta tosta..

Jozo: Tuo sininen on muovikanteen lis�tty tarra joka on lis�tty julkaisuihin levyn voitettua Gramophone palkinnon. Julkaisu on muuten identtinen.

nikki: Is the diffing any better than it used to be (regarding the medium edits)

Oh my. Those changes ARE sexy.

yeah.

Freso: http://musicbrainz.org/edit/20949528 (100x100 example)

Leftmost: Might be slightly even better if the tracklist just showed once with diff highlighting, but this is still a HUGE improvement.

If two people performed as a duo for four years and released two albums, credited in different orders, would you split or use artist credits on a duo artist?

hawke, it would be difficult to show sections that had changed instead of been removed or added.

hi there. I came here for something but I forgot what…

Who adding Various Artist release is so difficult... I managed to add release with Various Artist performed various different names (last week)

s/Who/Why/

Release editor is terrible

Leftmost: that about the Altan one? ;-)

hawke_1, yes.

My general instinct is to split, so I'm wondering what consensus is.

Leftmost: I would prefer to split it, but I think others might disagree.

And I can see why we would want it to show up under Altan discography

I'm sure I would split in cases where others wouldn't.

Well, the decision seems to be to have it credited to the artists properly, but monxton wants the switch to be to the Mairéad Ní Mhaonaigh & Frankie Kennedy artist, which I would like to split.

I'm a bit of a crazy splitter, though. I'd split Neil Young & Crazy Horse if I thought I could get away with it.

Leftmost: yeah, I tried that too IIRC. Or something related to that. — ah, it was Crosby, Stills, Nash, & Young I think.

Yeah, I'd want to split that.

It seems we don't have any good guidelines for collaboration artists. At least not that I can find. That might be a good thing to hash out on -style.

Leftmost: I think the problem is more that we need more server support in some cases, but we don’t really know how to solve it.

How so?

good guidelines would help us decide, but it won’t make people happy when their pet artist gets the “wrong” thing done to it.

True enough.

So for example it would be nice to have a way to have collaboration artists split correctly, but still have an artist page.

Or maybe have their releases/RGs show up under some different artist

or…who knows, that’s what I mean by “server support, but we don’t know what to do”

Makes sense.

I like split many "collaborations", but our search can't find them if I split

http://musicbrainz.org/artist/1fb517fd-25e4-4f7... some examples

artist & band is so common. Some are well-known that name, some are not.

Leftmost: http://goo.gl/dh3Vr

Jozo: I think we kind of need that anyway…or something like it.

Jozo: think about the “artist credits” section of an artist’s aliases tab.

hawke_1: Yeh, I like to see them on search results....

hawke_1: I like split all sucks "artist & band" (where artist is often vocalist). Are you against?

Jozo: search results, and a way to see “where is this artist credit used?”

Jozo: Depends on artist and band…

Jozo: e.g. I wouldn’t want to split http://beta.musicbrainz.org/artist/ee55e4e8-807...

hawke_1: What about http://musicbrainz.org/artist/8fa7b39a-6faa-49a... and "Topi Sorsakoski & Agents", "Agents & Jorma Kääriäinen"? (I selected this for example, cos it has some english references in wikipedia)

Jozo: I don’t know…I would lean towards splitting, but I don’t know the artist well enough to say.

this is where having good guidelines would be helpful (as Leftmost suggested)

hmm. in that case, i'd be tempted to merge 'Topi Sorsakoski & Agents' into the group, and just make it an artist credit.

hawke_1: yes. it handles inserting tracks better now, e.g. https://musicbrainz.org/edit/19198346 and https://beta.musicbrainz.org/edit/19198346

nikki: beautiful.

kepstin-work: +1

that's more of the case of a band which includes the lead vocalist in its name, which renamed itself on occasions when the lead vocalist changed or they didn't have one

kepstin-work: That's good rule. I would go still further... If artist or band have did some work outside "collaboration" and are well-known outside this "collaboration"... then I like split also

http://musicbrainz.org/artist/80acb463-cfe4-44e... - I like split

http://musicbrainz.org/artist/ba2a3bb1-23a4-4c1... Timo Rautiainen & Trio Niskalaukaus - I don't like split (still don't know why... maybe 'cos they are so well-known this name)

I do splitting-day at some point... Then you can vote my splits

Eww, this release has 000000000000 for all of its ISRCs.

what? not my isrcsubmit.py then (would just show no ISRCs)

JonnyJD: When you update it using new apis?

JonnyJD: (I dont want answer guestions sucks "How many tracks are on the previous (actual) discs altogether?")

hmm if you look at http://musicbrainz.org/artist/80acb463-cfe4-44e... it's wikipedia page is a bit

"Elonkerjuu, formerly known as Lauri Tähkä & Elonkerjuu"

so we should rename it maybe

I'm using isrcsubmit-cdrdao.py.

?

ZaphodBeeblebrox: Lauri Tähkä was vocalist... "In March 2011 the singer Lauri Tähkä announced that the band's career is over.[1]"

ZaphodBeeblebrox: After then band has continued their career and Lauri Tähkä has also...

that's just another case of a band whose name previously contained the lead singer renaming themselves when the singer leaves

jozo: isrcsubmit2 is in the pipeline and will depend on python-musicbrainz-ngs and python-discid (and won't need backends anymore)

jozo: and I want to fix that duplicate ISRC slide down bug (in libdiscid) before that

JonnyJD: I have followed it... icedax gives me duplicates also

Not yet tried new libdiscid...

Is libdiscid python api uptodate for getting ISRCs?

How is libdiscid maintainer?

http://packages.qa.debian.org/libd/libdiscid.html --- maintained: Debian QA Group

Someone might take that over....

libdiscid is maintained again on debain (sebastian sth. I think)

and I maintain libdiscid itself (now)

libdiscid can currently get ISRCs, but since that duplicate ISRC bug seems to be at firmware level (or something) we need to really do raw checks

JonnyJD: Debian QA Group is emergency maintenance of orphaned packages and so... http://qa.debian.org/

these raw checks are implemented for linux in a branch, but getting that to work for windows seems to be easy and for os x I currently don't now

rozo: http://packages.qa.debian.org/libd/libdiscid.html Sebastian Ramnacher is the new maintainer

*ramacher

isrc duplicate bug probably is just an issue with the drive's subchannel reads not being properly aligned with track positions or something :/

I worried for nothing... http://bugs.debian.org/cgi-bin/bugreport.cgi?bu... Sebastian Ramacher <sramacher [at debian.org]> adopted package

jozo: that's why I said ;-)

I think the ISRC data in the subchannel should be repeated over the track, so you might be able to workaround the issue by just reading from slightly after the track starts

because that bug surfaces on different OSes when issuing direct SCSI commands, and only on certain drives -> probably firmware

no, repeating is not necessary, actually

I can test it tomorrow...

Is EAC open source?

unfortunately not

but it wouldn't be hard to get a trace of the scsi commands it sends

I looked into the code of cdrdao and cdda2wav/icedax and both use the same SCSI command (which leads to duplicates, but only on certain (but many) drives)

there is one scsi command to read the q-subchannel data -> that has problems on some drives

there is a scsi command to READ cds where you can extract q-subchannel data -> that one seems to work on all drives (but is only optional in the spec)

from eac's behaviour, it seems like it uses a subchannel extraction command and reads over a range looking for encoded isrcs

interestingly, the READ command also only ready once, but reads fine. Using that q-channel command it actually helps issuing the command multiple times helps for some (but not all) cases

that would be READ command

mostly in how it's much faster when it finds an isrc; it seems to retry multiple times and take longer if the tracks don't have isrcs

Jozo: yes so therefore we should rename the thing Elonkerjuu, right? if it is called that now

you have to read a couple of sectors until you find a sector with ISRC information, but the first such sector always gave the correct ISRC for me

yeah, that seems to match the eac behaviour i've seen.

kepstin-work: my implementation currently reads at most 150 sectors (2 seconds), Per spec there should be at least one ISRC in 100 sectors.

I still have to implement some kind of check if the READ command is implemented, since it is optional in the spec (and do the broken command otherwise)

Freso: pong

ZaphodBeeblebrox: Or split Lauri Tähkä & Elonkerjuu to Lauri Tähkä and Elonkerjuu (what is now impossible due relations on Lauri Tähkä & Elonkerjuu)... I don't know how I can manage this well and easily

JonnyJD: you should probably warn the user in that case, too. I suspect more drives implement it than not, tho.

Jozo: I would not split that. "Lauri Tähkä & Elonkerjuu" is the full band name.

(or it used to be, before they renamed to just "Elonkerjuu")

well, both of my drives on my dev machine implement it (and one drive gives duplicates with that other implementation)

kepstin-work: my thoughts exactly

JonnyJD: hmm, do you think scanning the track until you find the same isrc twice would help?

kepstin-work: Merging Lauri Tähkä and Elonkerjuu to Elonkerjuu is evil also.

Jozo: why? it's the same group

just with the name changed after one member left

the broken command does only allow to specifiy a track, no sector. And the working command doesn't seem to need to find more than one ISRC

we would add both names as aliases with dates set, and artist credits would be used to say the group's name at time of album release

kepstin-work: Same group, but it's nothing same after lead singer left... Lead singer was all what they was

kepstin-work: I will announce a test release when I implemented the new method also for windows (currently only for linux)

kepstin-work: It's bad give credits only to band when there was band AND lead singer

Jozo: the lead singer is a member of the band...

kepstin-work: It's still well-known "Lauri Tähkä" and some band...