SongKong Jaikoz

SongKong and Jaikoz Music Tagger Community Forum

New user trying to figure out how to process 200+ GB of music

#1

I tried Jaikkoz on a few hundred files and it seem to do a good job but I got more than 200 GB of music and then using the UI becomes reaally slow/hard.

Especially the fingerprint process is “deadly” slow and requires massive amount of time to do (understandably).

Is there a way to run the fingerprint process via commandline and have jaikoz skip the files that already have been processed thus I can restart it again and again without having to wait for jaikoz to import all the files again ?

Another issue I got is how to “sync” up the Edit and View pane.
i.e. is there a way to click an element in the view pane and see it selected in Edit ? With thousands of lines it become tedious to scroll to find the matching row to see what were previously stored in a “purple” field.

And thanks for me finally having something that actually seem to be able to fix tags properly :wink:

#2

I’m considering writing a small introduction to how to identify bottlenecks and make Jaikoz work faster.

If you could help out a bit with verifying some of the issues in that regard, it would be helpfull.

What is your OS, java, and hardware specifications?

#3

[quote=MAzE5h1p69wB]I’m considering writing a small introduction to how to identify bottlenecks and make Jaikoz work faster.
[/quote]

The problem is not really so much that it takes time, but that I can’t stop and start the process easily in case I have to close/stop Jaikoz between reboots/other work etc.

[quote]
If you could help out a bit with verifying some of the issues in that regard, it would be helpfull.

What is your OS, java, and hardware specifications?[/quote]

Mac OS X: 10.6.4
Hardware: Macbook Pro 13", 2.53 GHz Intel Core 2 Duo, 4 GB 1067 MHz DDR3, SSD Drive
Java:
java version “1.6.0_20”
Java™ SE Runtime Environment (build 1.6.0_20-b02-279-10M3065)
Java HotSpot™ 64-Bit Server VM (build 16.3-b01-279, mixed mode)

#4

in case of the fingerprinting process you can use MusicIPMixer from Amplified Music Services http://www.amplifindmusicservices.com/what/downloads.php

#5

And jaikoz will Pick that up ?

I also noticed that jaikoz still trues looking up ids/fingerprints for Songs it has already processed and thus trying again is waste of time. Any way to let jaikoz only ask again if explixilty asked for it for songs it already processed ?

#6

I thought it didn’t try and generate an accustic ID if it already had a unique amplifind ID.

And I would assume that would be picked up, yes.

You can easily check that yourself, by loading some files in Jaikoz, chose to show the IDs, find a file without an ID and then move it to a different directory.

Next use MusicIPMixer to get the ID, then load it in Jaikoz again.

#7

Yes it does seem to not check again if I got a fingerprint but there is alot it cannot find fingerprints for so when running again it’s spending alot of time checking files that I know it won’t need to,but ihavent found a way to filter it.

#8

Running Magic Mixer doesn’t not really help because although it does a lookup from the same AmpliFIND server as Jaikoz it doesnt store Amplifind ID (Puid) for the tracks it finds (because Puids were added by Musicbrainz after Magic Mixer was released).

Jaikoz will not check again for a song that already has an AmpliFIND Id but it will check again for songs that it has already checked but that do not have an Id because if Jaikoz cannot find a match for a song it then does a full fingerprint analysis and submits the analysis to AmpliFIND, then over a 24 hour period AmpliFIND will create a Puid, and Jaikoz needs to check again to retrieve the Puid.

#9

MusicIPMixer doesn’t get the PUID, that’s right. But it does the fingerprint and analyze and that’s what takes time. After that the song is in the Amplifind database. To get then the PUID (after 24hours wait time) is only a moment of time

#10

I tried amplfind mixer …it took close to 20 hrs to load the library set and it ran for 10 hrs processing; but not its stuck.

So yeah, this won’t work - don’t want to wait 20 hrs more to just see it fail again ;(

#11

I think you just need to work in batches of 10,000 songs at a time.

#12

yes, that could be a solution but how do I tell jaikoz to load the same 10.000 songs between restarts?

#13

yes, that could be a solution but how do I tell jaikoz to load the same 10.000 songs between restarts?

[/quote]

or even better, how do I get it to skip files it already processed ?
or how do I identify files that are already processed and load it via commandline with those left?

#14

yes, that could be a solution but how do I tell jaikoz to load the same 10.000 songs between restarts?

[/quote]

or even better, how do I get it to skip files it already processed ?
or how do I identify files that are already processed and load it via commandline with those left?

#15

Jaikoz will skip files that have already been matched, but will reattempt to match ones that didnt match first time if you reprocess them, but the memory restrictions come into play on loading the files not matching the files so you dont want to reload matched files. You could of course close them again and that will free the memory but I think you should have a PROCESSED and UNPROCESSED folder, when you have matched songs use File and Folder Correct or Save and Move to put them in your processed folder, that way you can work through them.

There is no command line option

#16

This is an old topic, but I’ve been doing some large processing myself, and figured I’d mention just how much RAM Jaikoz uses with larger collections.

With 26 thousand songs at a time (about 150GB), Jaikoz is using 4GB of RAM (private working set, 9GB total commit size).

#17

The way Java works is if you allocate it 4GB gb it wont make that much effort to release memory if there is plenty available, I’m sure you can process 26,000 files with much less memory than that, I would say apromixately 1GB, what does it say in the bottom right hand corner of the Jaikoz status bar.

#18

what does it say in the bottom right hand corner of the Jaikoz status bar.

3.63 GB of 7.95GB

#19

Oh I see so 8GB has been allocated, yet it never uses more than half. If you needed to you could allocate alot less and it would still work