SongKong Jaikoz

SongKong and Jaikoz Music Tagger Community Forum

MUSICDNS Duplicates + Memory Issue

#1

Hello. Looks like a good product. I am currently testing it out before I use it on my 50k song library.

1st question:

Looking back at this message http://www.jthink.net/jaikozforum/posts/list/148.page I’m trying to get a sense of what will happen when I throw multiple songs at the library. How does the app know which album a duplicate song is from? if the metadata suggests one album over another will it choose that or overwrite it all? what if there is no metadata?

At a worst case scenario, i’d rather not lose the (maybe correct) metadata already in the songs. Also with 50k songs, i’m not really going to be able to look at individual cases…

2nd Question is wondering if the issue described here http://www.jthink.net/jaikozforum/posts/list/504.page has been resolved. I have my support file ready to send if necessary per the discussion in that thread. Is there a file i need to edit? the total mem on my computer never got above 50% usage (I have 2gb RAM).

#2

[quote=theruffian]How does the app know which album a duplicate song is from? if the metadata suggests one album over another will it choose that or overwrite it all? what if there is no metadata?

At a worst case scenario, I’d rather not lose the (maybe correct) metadata already in the songs. Also with 50k songs, i’m not really going to be able to look at individual cases…
[/quote]
Jaikoz isn’t going to delete any songs unless you run the Delete Duplicates, if you do run it you can configure it to only identify multiple songs as duplicates if:
They have the same Acoustic Id
They have the same Musicbrainz Id
They have the same Musicbrainz Id and Acoustic Id

by setting the Preferences/Local Correct/Delete Duplicates/File is a Duplicate Option.

So if you want to be ultra careful set it to they have the same Musicbrainz Id and Acoustic Id, this would mean the audio would have to sound the same, and the metadata would be the same because both ttracks have been identified as the same track in Musicbrainz (and a track can only be linked to one album).

[quote=theruffian]
2nd Question is wondering if the issue described here http://www.jthink.net/jaikozforum/posts/list/504.page has been resolved. I have my support file ready to send if necessary per the discussion in that thread. Is there a file I need to edit? the total mem on my computer never got above 50% usage (I have 2gb RAM).[/quote]
There isn’t a problem with setting the memory it is explained in the help file but it is OS specfic e.g for OSX you just need to modify Info.plist. There was another problem that caused OutOfMemory to be reported when it hadn’ really run out this has also been fixed.

If you are having problems setting your memory then send me your support files.

#3

Thanks for the detailed expalanation.

On Q1, Is the delete function able to determine the duplicate’s bit rate and delete the file with the lower bit rate?

What i’m actually worried about is less the deletion than having for example a best of X and then also their regular CD. I would hate for the app to rename a song in the best of as a song on the regular cd because when i view this in itunes i’d like to see the song in both places.

I assume that if the metadata (even if wrong) says something like best of the app is able to differentiate and knows they aren’t the same.

Basically i’m wondering how the app differentiates between having the same Music ID and MB number. When it comes up with a music ID how does it know which MB number (when there are multiple) to choose from?

On Q2, I’ll look at the OS X help files. Thanks Again.

By the way, i’m really impressed that your app isn’t more well known (because it’s the only app around with this capability and functionality) I’ve tried a lot of other apps that just correct locally, etc, but aren’t all that great. This is what i’ve been looking for for a while. Finally a way to automatically get all the metadata available without even having the right song, artist or album name. I imagine this also helps remove all those extra blank albums when you view music in coverflow.

#4

[quote=theruffian]Is the delete function able to determine the duplicate’s bit rate and delete the file with the lower bit rate?
[/quote]

Yes, it will keep lossless audio formats (i.e Flac) over lossful formats (i.e Mp3) . If files are the same audio format the delete duplicates action deletes the file with the lowest bitrate, or if they are the same deletes either one.

[quote=theruffian]What i’m actually worried about is less the deletion than having for example a best of X and then also their regular CD. I would hate for the app to rename a song in the best of as a song on the regular cd because when i view this in itunes i’d like to see the song in both places.
[/quote]
If both songs have been correctly matched in Musicbrainz, they will have different Musicbrainz ids because they are tracks of different albums and hence wont be regarded as duplicates.

The matching algorithm is quite complex but basically it matches the database using both acoustic ids and metadata to find the best matches, the best match is determined by giving each match a score out of 100 weighted for different elements such as artist name match, album name match, no of tracks match, length of song match ecetera. You can se the score that are calculated by using Remote Correct/Manual Tag from Musicbrainz

#5

Super slick.

I’ve been using iDupe to delete duplicates (which has been touch and go), but am psyched to be able to use Jaikoz, as it is one less step and also it seems like you really put in a great deal amount of features and sophistication in the software.

Thanks for the quick responses.