#musicbrainz

awaygoove is now known as catgroove

catgroove is now known as catmooose

one o to many

today sucked bigtme

there seems to be an extremely long wait time in between puid lookups

what's it doing?

Jugdish: Looking up the pweeds?

:)

hmm

i don't see why it should take a full minute to look up one puid

It doesn't, in my experience, maybe the server is busy?

could be

luks: a question about how puid-lookup is done...once a release is pulled from the server and brought into picard, does it bring in all the puids for that release? and does this mean the server does not need to be accessed again while matching up the individual tracks in that release?

puid lookup = decode the file, calculate fingerprint, query musicdns.org, query musicbrainz.org

it does query both servers for each file

i'll admit i don't fully understand the details of puid's but once you have a local copy a release's info including all its puid's, does the server really need to be queried again if any of those puid's matches a file?

nope, this is a good idea

but still, that will eliminate only the last step

i'm just trying to think of a way of decreasing the number of accesses to the servers

i'll make it download all puids when downloading a release and first compare to already loaded releases

:)

luks, would you consider it smart to install the alpha picard over the earlier dev versions?

it shouldn't cause any problems

excellent

how about installing over the old 072?

there will be problem with uninstalling 0.7.2 if you want to have them both

overwriting should be ok

i've noticed that when you delete a release on the right-hand pane it always jumps up to the top of the list, which is kinda weird

there is no need for me to keep both, so hooray

rpedro__ is now known as rpedro

is it possible to tell picard not to write to some tagfield at all, for example, albumartist?

Kilu: you can unset it using tagger script

enable scripting and put $unset(albumartist) there

hooray

I suck at scripting, so I didn't even dare to try

:D

:)

hmmm

whatup

looks like the unset also disables the tag from the rename strings

ah

that will need some tricks then, '$set(_albumartist,%albumartist%) $unset(albumartist)' and use %_albumartist% in the naming format

(I hope that will work)

x.x

looks scary

zout is now known as zout_away

but seems to work okay :)

sorry for spamming the wiki, lukas :)

no problem :)

Jugdi5h is now known as Jugdish

set(unset:set)

:P

sorry?

I was just jokinmg

booooooooooooooooooooom

I don't know any code

:p

luks?

her is the other bug: error in QTCORE4.dll

lukas, I can't find the info for how tagger script works; I rememeber I saw a page once, but I can only find the PicardQt/Scripting. When is it applied? I tried to define $if(%date%,$set(year,$left(%date%,4))) and use it in the dir name as %year% but it doesn't work.

I think they are applied when the album is loaded from the database. if you change the script after, you have to refresh the album. if that's what you meant :D

well, it doesn't work in the "test tags" example, at least

hmm, let me see

$if(%date%,$set(year,$left(%date%,4))) is not very safe if you want it only for file renaming

this way it will write tag YEAR to ogg or flac files

neetds a %performer% variable for cases like http://musicbrainz.org/album/7b7e58ff-bd9d-4fc9... where I want the trackartist to be the performer

hmm, I wonder if it actually works without any instrument or vocal type

how do I tell it to put the %artist% into the composer id-field *if* there is a 'perforer* ar?

$if(set something)

if not erh.. something

I'm afraid this is not possible right now

oh

erh.. shouldn't it be?

I mean with the code there...

because perfomers can be %performer%, %performer:orchestra%, ...

no it's just a 'perfrmer' type ar

I know, but in picard it uses also instruments/vocals

aaah

there is a separate AR type for simple 'performed by'

yes there is

:)

I didn't know

heh

I think picard doesn't even use that

luks: that's what I mean about the tagger script, I'm not sure how it works. Can I have variables that I use as sub-expressions? (But that are not automatically used for something else, as years seems to be.) And can I use newlines and indentations? I have some big lisp-like expressions for file naming...

nope, I tried it

and my biggest pet-peeve: can I find out the year of the first release, for remasters/re-releases that have the necessary AR?

bogdanb: you can use variables that starts with _

also with that inplace, I'd want the script to put the MBartist in the composer-id tagfield (in the file) IF there is a "performed" ar

but it's usually better to unset them when you don't need them

k, I can do that

and the script is whitespace sensitive, but in most places it doesn't matter

so then all my mozart files have 'mozart' in its 'composer' field, but thetrack and tag says 'whatever orchestra' or somesuch

well *i* don't actually wan it.. I think, but I am certain others wantr that

I need to figure out how to handle all the performer types better

okidoki

could do something like, %performer:vocal:all% or (%performer:vocal%) then %performer:vocal:soprano%

hmm..

foodtime

moin! are there any tagger developers in here?

im looking for a good way to generate PUIDS with libofa and without using lame or some other external programms. Any experiences?

you need something to decode the audio files

so the options depend on your target OS, whether it's open-source or not, etc.

yes I played around with libmad but decodings seems to be slow on mp3

Linux- GPL

you need only the first 123 seconds from the file

picardqt on linux uses ffmpeg

it has quite simple API and codecs for most of formats

err, 135 seconds

didn the time you need to decode depends on the content?

well, yes

but there isn't many songs with long leading silence

are 135 seconds the minimum you need?

130

(I think)

To handle the time you need in the tagger is quite unergonomic.

unergonomic?

then with TRM generation, the lib decided when there was enough data to calc the print.

The ofa lib did not I think

no, TRM uses only 30 seconds, nothing else

even if it's 30 seconds of silence

Shure? For me, it looks like the TRM algorithm take as much data as it needs for the fingerprint

yes, I'm sure