Re: [Jajuk-developers] Beta version of the new Java client for the MusicBrainz XML Web Service available

[Nicolas S. <nic...@st...>]

I agree that the possibility to use the  Musicbrainz database would be =20
a great feature. I have used the service in the past with an external =20
app called iEatBrainz to update the metadata of my tracks and it =20
worked fairly well. Having such a feature built right into the player =20
instead of using an external app would be a nice time-saver.

Since many songs start similar (and therefore give multiple results =20
from the database) each track should  be checked separately for =20
correctness before applying any changes. IMO Jajuk should use CDDB =20
when importing tracks and MusicBrainz for updating/fixing the metadata =20
later on.



Quoting Nebu Pookins <ne...@gt...>:

> =09From what I understand, with cddb or freedb, you provide (a hash of)
> the duration of all the tracks on a CD, and cddb can return a list of like=
ly
> CD track listings. You are unable to, for example, identify the track
> information for a single track (since you don't necessarily know the
> duration of all the other tracks on the same album).
>
> =09In contrast, MusicBrainz uses an acoustic fingerprint on the audio
> data itself, so in theory it should be able to recognize and identify an m=
p3
> file, for example, even if some modifications were done since the file was
> ripped from a CD (for example, if the volume was normalized, or if the
> leading and trailing silence were removed).
>
> =09In practice, I find MusicBrainz isn't really all that good at
> identifying music using the acoustic fingerprint method. Still, it's bette=
r
> than CDDB which only uses track duration, and only works if you have the
> whole album.