Advanced MetaData Handling / Tagging Tool

Guest · Post by **Guest** » Sat May 12, 2007 9:29 am

Originally proposed:
http://forum.dbpoweramp.com/showthread.php?t=12424

however there is no evidence the idea will be used.

Below is a quick mock up of a proposed worksheet to work with multiple meta data services at once and choose the best meta-data for tags. I choose the following album because its a good example of the differences between the databases and the fact that none of them are perfect. The meta-data in the final column most closely represents the data from the album. AMG is actually very good and only differs in this case a bit for meta-data style. Though there are plenty of examples were AMG is just plain wrong.

Meta-Data from AMG, GD3 (not shown) or MusicBrainz is given 2x the weight of FreeDB, tracktype, or MusicDNS data since they are more carefully (but perfectly groomed). If multiple sources agree on the meta-data is is automatically moved to the final column. If sources do not agree then one of the sources must be selected or the data can be typed manually. Green next to the meta-data means the sources are in agreement. Red means there is a consensus the meta-data is wrong, or that there is no agreement (ie only 2 sources but both with different data and same source confidence). Greyed out meta-data has been manually selected. If there is only one source for the meta-data it is trusted.

Hopefully we would also then be able to submit the corrected meta-data to all the sources. And even maybe build another database.

-AMG: would require a contract with AMG. They have good meta-data and its worth it. In addition, they offer song fingerprinting, making the tool useful for tagging existing tracks, not just when ripping from CDs

-GetDigitalData: unlimited access for <$5/year, with no commitment for numbers. They offer song fingerprinting, making the tool useful for tagging existing tracks, not just when ripping from CDs
# Artist Field
# Album Field
# Track Field
# Album Artist
# Genre (narrowed to approx. 32 specific genre fields)
# Cover Art (400x400 images)(as opposed to 250x250 from AMG)
# Year published
# Publisher
# Composer (classical genre)
# Conductor (classical genre)
# Compilation Information (Individual track artists for compilation albums)
# UPC & ISRC if required

-MusicBrainz: seems to be free, as would incorporation of their PUID generator. They offer song fingerprinting, making the tool useful for tagging existing tracks, not just when ripping from CDs

-Tracktype: free

-freeDB: free

-MusicDNS official support would likely not be free, but it would be easy enough to add a framework so end users could access it for free. They offer song fingerprinting, making the tool useful for tagging existing tracks, not just when ripping from CDs

Anyone else have any ideas/feedback?

Eli · Post by **Eli** » Sat May 12, 2007 9:30 am

sorry, didnt realize I wasnt logged in.

Post by **Peke** » Sat May 12, 2007 6:30 pm

I have read your post in dbPowerAmp and it certainly rise really good questions.

Tagging problem become more and more problematic.
Back in 90's I had one Ripper (Xing Audio catalyst) Winamp as Tagger and CD-s as sources allong with CDDB

BTW Audio catalyst has lifetime Gracenote subscriptions

but gets only CDDBv1

I do not want to talk off topic but as soon as someone call Gracenote I remember FreeDB msg that CD-s I posted are Copyrighted by Gracenote and are rejected to enter the database. After sending logs of my original posting to original CDDB and prove that I have been poster of those CD info they were entered in FreeDB. Basically I neded to repost several hundreds of CDs again due to CDDB

Standardized world wide acceptance for accurate info is must needed (take a look at Last.FM) which will mostly benefit all of us even RIAA and copyright owners can profit from that without playing with consumers. Here in Serbia there was good idea about that which never made to real life unfortunately.

Steegy · Post by **Steegy** » Sun May 13, 2007 4:43 am

Some things I feel different about:
- You probably forget that some albums have less tracks or tracks extra with the same album name (e.g. bonus tracks, special editions with same album name) and so it's not easy for a program to compare these albums (because you might have the right cd, but are missing hidden tracks, and multiple sources also have multiple album matches), and it should also compare album art for the different sources or the different album results per source.
- It doesn't mean that something is right because most sources are saying it. E.g. artists starting with The on one source are almost always more correct than the artist without The in front (even if it's 1 against 3)
- Some people like a lot of track title detail (e.g. detailed remix specification), others don't. You seem to choose the way in between, but a program can't do it like that.

Just wanting to say that your suggestion is an ideal. We probably need some AI that doesn't exist yet to solve these kinds of problems. And implementing everything like you say would be lots of work.

Eli · Post by **Eli** » Sun May 13, 2007 9:44 am

Albums with different numbers of tracks will not match when ripping an original CD with the schemes used by any of the databases. This is not a problem. One nice solution available in dBpoweramp and WMP is to store the CD TOC in metadata when ripping a CD. This makes it easier to look up the CD in a database later. Hopefully more codecs will add support for TOC storage soon.

As for album art, I completely agree. A similar interface could show album art and allow you to choose the best one. For archival purposes I like to have the highest resolution. Off topic, but one thing I would really like is the ability to resize album art for adding to metadata when transcoding, ie take a 500x500 folder.jpg file and have it resized to 100x100 for storage on my ipod.

I realize that many sources agreeing doesnt mean that the data is correct. That is one of the beauties of the system. You can see multiple sources of data in one view, and with a simple click of a mouse choose the correct one if one is not chosen. Thats much easier and faster then typing. Also, not all sources are considered evenly. Groomed databases such as AMG, musicbrainz, and GD3 would have a higher weight then something like FreeDB when the system is choosing which resources are probably considered correct.

As for the level of detail people want, ect. Again if what you want is available from one of the sources, its a simple mouse click. If not, you can simply type it in. But the goal should be to have data that most accurately reflects the data on the album, IMHO. But the system lets everyone choose what they want with much greater ease then any other system available.

Eli · Post by **Eli** » Sat Aug 25, 2007 9:21 am

It seems that BOTH ripstation and dbpoweramp will be adding / developing variations on this them. Would be great to see this tool in MediaMonkey!

trixmoto · Post by **trixmoto** » Tue Aug 28, 2007 10:15 am

Can't use AMG as a source because they've kicked off about this. Otherwise a great idea though. Maybe this should be moved to the scripts section though, as the "Autotag from Amazon" has now become "Autotag from Web Source" with scripts to control the lookup?

Mizery_Made · Post by **Mizery_Made** » Tue Aug 28, 2007 11:25 am

trixmoto wrote:Can't use AMG as a source because they've kicked off about this.

Unless you pay the licensing fees for their Database access, or whatever.

Guest · Post by **Guest** » Sun Sep 16, 2007 9:03 pm

I know people tend to forget it, but yes, you can actually pay for the service!

trixmoto · Post by **trixmoto** » Mon Sep 17, 2007 5:06 am

No scripter is going to pay for a license so that everyone can use his script though.

Eli · Post by **Eli** » Fri Nov 16, 2007 11:16 am

trixmoto wrote:No scripter is going to pay for a license so that everyone can use his script though.

Im not talking about scripting. Im talking about official support for the tool.