almost merged those last week and someone thankfully caught it
2006-02-20 05133, 2006
BGreeNZ
If I tell ppl not to merge remasters, then no albums will be merged *ever*, because they *might* be remasters :(
2006-02-20 05151, 2006
wolfsong
the merge portion of the ui would benefit from a side-by-side view
2006-02-20 05120, 2006
wolfsong
tell them not to merge ones marked as remasters in the annotations
2006-02-20 05131, 2006
Schika
yep scrolling up and down is hard to find differences - side by side would be better
2006-02-20 05156, 2006
BGreeNZ
I always compare them by loading each merge candidate into a separate browser tab. Seeing them side by side would be better, though. Even if it meant scrolling sideways (usually a big no-no in web page design)
However, such a large difference on CDs would be a clear indicator of different song versions (e.g. studio vs. live, or album vs. radio edit)
2006-02-20 05143, 2006
HairMetalAddict
In-A-Gadda-Da-Vida (17:05) vs. In-A-Gadda-Da-Vida (2:56) <--- Methinks one is an edit. ;-)
2006-02-20 05156, 2006
BGreeNZ
intgr: Do you know for a fact that the UUIDs are kept? I'm not sure they are
2006-02-20 05119, 2006
BGreeNZ
lol @ HMA
2006-02-20 05108, 2006
intgr
BGreeNZ: It wouldn't make any sense *not* to keep them.
2006-02-20 05124, 2006
intgr
Because the UUIDs are the preferred way of linking to items.
2006-02-20 05124, 2006
yalaforge
UUIDs on merges or deletes?
2006-02-20 05130, 2006
intgr
Merges.
2006-02-20 05145, 2006
BGreeNZ
Are UUIDs the same thing as permanent links?
2006-02-20 05147, 2006
yalaforge
unfortunately, they aren't kept :-(
2006-02-20 05152, 2006
intgr
No?
2006-02-20 05104, 2006
Schika
the problem is not the analogue media - it's the digitalizing (the binary system). But yeah you have to remove the "burning fire noise" on some old vinyls which affect also the whole thing
2006-02-20 05121, 2006
yalaforge
BGreeNZ: the UUIDs are the weird ascii characters at the end of a permanent link. 36 characters.
2006-02-20 05133, 2006
BGreeNZ
Oh.
2006-02-20 05144, 2006
yalaforge
intgr: no, but they should be. there's been a ticket for quite a while
2006-02-20 05154, 2006
BGreeNZ is ever-so-slightly less confused
2006-02-20 05118, 2006
yalaforge
in fact, using the old moderations it could be possible to recover the UUIDs, I'm not entirely sure
The '69158f97-4c07-4c4e-baf8-4e4ab1ed666e' part is the UUID
2006-02-20 05140, 2006
intgr
yalaforge: I hope so.
2006-02-20 05148, 2006
yalaforge
yup, in its 36 character ascii representation. in fact it's a 128 bit random value
2006-02-20 05102, 2006
intgr
I would qualify this as a 'flaw'.
2006-02-20 05129, 2006
yalaforge
wouldn't be difficult to implement that. go ahead.
2006-02-20 05156, 2006
intgr
I would if I knew Perl at all.
2006-02-20 05103, 2006
BGreeNZ
Schika: How does digitizing an LP affect its length?
2006-02-20 05109, 2006
yalaforge
it has nothing to do with perl, that's a DB issue
2006-02-20 05118, 2006
intgr
yalaforge: Oh, hmm...
2006-02-20 05125, 2006
BGreeNZ
It would be no different to playing it, would it?
2006-02-20 05140, 2006
intgr writes it down onto a 3-meter long TODO queue.
2006-02-20 05143, 2006
yalaforge
tables have to be created, transitional scripts written
2006-02-20 05111, 2006
yalaforge
it's on my 3-meter long TODO as well. and not just on mine. see?
2006-02-20 05124, 2006
intgr
;)
2006-02-20 05112, 2006
BGreeNZ wishes he had the resources and knowhow to work on the database.
2006-02-20 05105, 2006
yalaforge
BGreeNZ: it's difficult. better start on the client side to get into MB programming
2006-02-20 05135, 2006
intgr
I can't say I'm exactly overflowing with experience either. :)
2006-02-20 05158, 2006
yalaforge
well, that's something you need on the server side
2006-02-20 05136, 2006
yalaforge
a small mistake and replication clients (mirrors) get out of sync
2006-02-20 05113, 2006
Schika
BGreeNZ, when you have to correct the de-noize and such ... the amp level will manipulated and if the track is fading out earlier as it is - when you would pump up the volume on your Hi-Fi system where's the vinyl playing
2006-02-20 05104, 2006
BGreeNZ
Schika: When you rip a vinyl, do you digitize it to a WAV first, or attempt to encode it directly to MP3 etc?
but the end of the track is now at 9:55 (in the digital version) instead of 10:00 (in the alalogue version)
2006-02-20 05152, 2006
Schika
another point is that digitalisation means: getting something into a 1 or 0 scheme = reduction of information. You can't have 0.999999999999999999999 or 0.111111111111111111 ... however I guess I killed the chat ... and runns away
2006-02-20 05142, 2006
juhae
Schika: Actually, I've been following with interest what you've been talking about.
2006-02-20 05104, 2006
juhae
Don't have any fancy equipment, but I have a crapload of vinyls which could use some digitalization at some point.
2006-02-20 05143, 2006
BGreeNZ
Sorry, Schika, I was cleaning up my browser tabs. That's why I didn't acknowledge you
2006-02-20 05127, 2006
Schika
No prob. However, I did digitalizing some old vinyls for my dad - which was a shitload of efford to get them in back in a listenable fashion
2006-02-20 05120, 2006
BGreeNZ
So, you're saying that the original digital recording is 10:00, but after you're done editing it it's 9:55?
2006-02-20 05111, 2006
BGreeNZ
What happened to the other five seconds, or perhaps more accurately, how did you process it?
2006-02-20 05116, 2006
Schika
on some old records I thought they are recorded besides a campfire
2006-02-20 05145, 2006
BGreeNZ
Also, at CD quality (44.1kHz, 16-bits per sample per channel) or greater, the quantization noise should be far from audible
2006-02-20 05113, 2006
BGreeNZ
You said you were working at 48-bits per sample, so that's over 281 *trillion* possible digital values per sample. You wouldn't be able to hear the difference there if you tried...
2006-02-20 05111, 2006
BGreeNZ
Schika, juhae: I hope I didn't scare you away
2006-02-20 05128, 2006
luks joined the channel
2006-02-20 05141, 2006
Schika
sorry I was making coffee ... however I worked sometimes with 48-bits not always ... however the whole process of digitalisation / filtering and such could summarize up to that mentioned 5 seconds
2006-02-20 05138, 2006
BGreeNZ
At any rate, when sampling at CD quality or better, quantization noise should be the least of your worries :)
2006-02-20 05123, 2006
BGreeNZ
So, would I be correct to say that the original analogue recording is 10:00, the original (unfiltered) digital recording is 10:00, and you silence out 5 seconds while editing?
2006-02-20 05143, 2006
Schika
yep
2006-02-20 05147, 2006
BGreeNZ
or, are you changing the speed, by resampling?
2006-02-20 05153, 2006
Schika
not that I would know
2006-02-20 05112, 2006
BGreeNZ
Are you deliberately chopping off 5 seconds, or are you using an interpolate / resample function to "average out" the noise?
2006-02-20 05149, 2006
Schika
If the source quality is good then nothing will/should be chopped off
2006-02-20 05113, 2006
BGreeNZ
So you should end up with a post-edited file the same (play)length as the master file in that case, shouldn't you?
2006-02-20 05130, 2006
Schika
sure
2006-02-20 05120, 2006
BGreeNZ
Where is the problem, then?
2006-02-20 05103, 2006
Schika
I was just explaining how it could came to a track duration difference you asked for - or not?
2006-02-20 05134, 2006
_thom_ is STILL compiling crap - plugins for libtunepimp now
2006-02-20 05123, 2006
yalaforge
crap, ah
2006-02-20 05149, 2006
BGreeNZ
Ah, but if you've deliberately chopped 5 seconds of noise off each end of the track, then you've chosen to edit the final playlength of the track. The original version (on the LP) is still more or less the same length as it would be on a CD, right?
Say, Schika, maybe you'd like to write a HowToDigitizeLP ? :)
2006-02-20 05147, 2006
Schika
BGreeNZ, yeah the vinyl and the CD track durations would be the same
2006-02-20 05137, 2006
BGreeNZ
So the entries in MusicBrainz *should* be the same, and they should be merged, correct?
2006-02-20 05100, 2006
Schika
yeah if they are simular
2006-02-20 05133, 2006
Schika
especially for electronic music the releases often differ in track order or tracks itself
2006-02-20 05139, 2006
BGreeNZ
Schika: If you can find me an example of a CD and LP version of the same album, where the track times differ by *more* than 5 seconds, then there might be grounds to change that tolerance. However, 5 or 10 seconds is a pretty big discrepancy to account for.
2006-02-20 05121, 2006
Schika
Like the DJ-Kicks mix series, the CD release is mixed and the vinyl release is unmixed - the vinyl release has different track durations
2006-02-20 05100, 2006
BGreeNZ
Yes, but in that case, the music isn't identical, is it?
2006-02-20 05133, 2006
Schika
same artist / track titles on both releases
2006-02-20 05143, 2006
HairMetalAddict
Fleetwood Mac's Tusk album (original CD release, not the remaster). The CD ended up with an edited version of "Sara" because it was 2 LPs on 1 CD. But it was marked on the CD as an edit, so that may not count. ;-)
2006-02-20 05153, 2006
BGreeNZ
Yes, but do they *sound* the same?
2006-02-20 05146, 2006
Schika
no - they can't cause the CD is DJ-mixed with playing 2 or more tracks the same time
2006-02-20 05119, 2006
Schika
* at some points
2006-02-20 05138, 2006
BGreeNZ
Then they shouldn't be merged anyway, should they?
2006-02-20 05159, 2006
Schika
no
2006-02-20 05121, 2006
BGreeNZ
The reason for the 5 second tolerance, as far as I am aware, is to account for different pressings of the same CD or slightly different playlength of an analogue media, due to wow/flutter etc
