Consolidating tags of the same tracks

Does anyone have an efficient workflow for dealing with inconsistent tags in the same track?

I have gotten carried away with 'tag sources', and ended up with a huge amount of tracks that kinda look like this:

Album Artist Artist Title Genre Year
The Who The Who Baba O'Riley Oldies 1971
Various Artists The Who Baba O'Riley 70s rock 2007
Various Artists The Who Baba O'Riley Hard Rock; Electronic Rock 2010
Various Artists Who, The Baba O'Riley unclassified 1990
Various Artists The Who Baba O'Riley rock 2008

So, I want them all to have the year from the first result, and the genre from the third, and I need to replace the track artist for the fourth.

I also need to flood the genre across all tracks that are the same artist and title but appended with " [Live]".

I have been SLOWLY working through these in mp3tag, but the fact that mp3tag doesn't list out the genres of the tracks in the dropdown is making this very fiddly.

I'm familiar with export-import and Excel, but that would also be quite arduous. but is there a tool out there that assists with consolidating the tags of 'duplicate' tracks, rather than simply deleting them?

The easies way to assign the same data to several files is to use the tag panel.
I don't see any other way in MP3tag.
So, if you have a (play)list of alleged duplicates, you would still have to select the grouped files manually (or with a filter) and then assign the data.

The next problem: what qualifies a duplicate?
Here is a lengthy discussion

and perhaps it is quite allright if the data in the tag fields stays as it is.

E.g. I know a number of files that have the same data in TITLE but are actually different as one is the normal pop version and the other the Salsa version... So, unifying the genre would then be a mistake actually.

The only real issues are with the genres and the years.
My titles are quite meticulously tagged with stuff like "[Demo]", "[Live]", "[Acoustic Version]", "[Remix]" etc.
If there's a 3-minute version and a 6-minute version of the same song, then I'm okay to assume that their year and genre are the same. If I haven't, for example, added [Live] by mistake, I can live with going back and making that amendment afterwards. The priority now is to efficiently pick which year and genre to populate tracks with the same title and genre.
It's baffles me that none of the duplicate-sweeping tools includes the functionality to do this prior to deletion.

You have to enable
File->Options->Genres->Prefer Genres in List at Tag Panel
to see only the Genres which are in the files of your selection.

I'm just curious:
Why do you assume that this 2 songs share the same year and genre? One could be made years before the other one and the genre could have changed from "House" to "Techno".
Or is this something that you can exclude for your collection?

No recognition tool can ever tell you for sure that 2 similar songs are similar enough to make any decision. It can calculate if songs are identical or how many percent difference between two files exists. But I doubt that such a tool will ever tell you "Song 1 is the House-Version with Vocal parts" and "Song 2 is the Techno-Live-Version with additional 3 minutes extended drum part".

Even if you see this pragmatical and it is completely your choice - the tagfiled year ist normally taken from the release. So the year of the original album, the year of a single or the year of a compiliation should not be the same as long as you care at all about the album-tag.

To decide if something is the same track you should at least consider the length of the piece as a minimum criterion.

I suppose it depends on what genres of music are in your library. I guess that applies quite a bit to electronic music, but not the majority of my library. Invariably, if there are two different track lengths of a 'same artist-same title' track, one is a 'single edit'.

I know of a sampler "Summertime (40 Versions)" which features the song "Summertime" in 40 versions (not surpfisingly). Some are in Jazz, there is Pop, Soul, Easy Listening, Swing, etc.
"Clocks", each time performed by Coldplay, exists in a Pop version and a Salsa version.
So this is more or less a plea to have a thorough look whether the data really is the same for all the files that look similar.
I completely agree that "The Who" and "Who, The" should become the same.

I too have many duplicates and meticulously tagged fields. The best way to do this, IMO, is to use the Tag Panel as ohrenkino suggested. Sort your columns first by Time, then by Artist, the by Title. This will help you sort covers and remixes from the original and also the different versions by the Artist. When you are sure you have a batch that are the same, highlight them all, and use the drop-downs on the fields to choose the correct information, then hit save. It is tedious but it is the only was to keep things straight in my experience.

Thanks. Yes, I think you're right. I don't think there's a demand for tools like I'm looking for because most wouldn't consider keeping multiple copies of the same track. I did that once and deeply regretted it. I like listening to my library on shuffle and occasionally going "hang on! This is different to the version on the album!". An obvious example of this is 'Rock Me Amadeus" by Falco - there are several 'single versions', all quite different.
I'll persevere with mp3tag, and try using sorting and filtering.

1 Like

I am in the process of removing all duplicates myself. I debated keeping multiple copies where was in the greatest hits, but decided against it. My playlist will serve as my greatest hits. It is tedious, but stick with it and your library will be clean. As one with thousands of live Grateful Dead music, I know a thing or two about having multiple copies of a song. :exploding_head:

The disadvantage of playlists is that they only work with stable file locations.
As soon as you modify anything in the filename, the link is broken - you will not find that file without a rather extensive search in your player (which does not rely on filenames but on metadata).
I would not tear apart compilations and other albums, esp. not complete ones for the sake of a couple saved cents of storage space. The time to restore the original but now broken playlist is much too valuable.

I agree with this. Way too much trouble to verify the differences between release versions versus just keeping the complete albums intact.

Besides, doesn't this give you bragging rights to the size of your music library? :smile:

Indeed, it would.
I once used the iTunes function to show duplicates - and then it showed less then 5% of the total number of files in my collection. So I think that is tolerable.

You both make good points and I have waffled on both sides. It is good to have one's thinking challenged and reassess prior strategies to determine if they are valid.

Ohrenkino's thoughts on creating and using sorting fields really got me thinking of whether it is worth the trouble (for now I think it is, but my conviction is not as strong as it once was). Here are my thoughts:

I know I have a lot more duplicates than 5%, I have already taken out 10% of my library in duplicates and I am only at Marley. It is more about the space on my phone than on my computer, but there are ways around that as well. I do use ,flac and have some 24 Bit 96 kHz+ music so the size can be quite large.

I have debated myself keeping greatest hits just to have another opinion of what the greatest hits are, or a more universal greatest hits, that is not determined by me. I may be changing my stance on this, and just keep a couple of greatest hits albums and just not source any of my playlist from them but rather from the versions in individual albums. This has caveats though, as some of the best masters for songs might come from an MFSL mastered greatest hits.

As far as changing file names, I almost never do that, and when I do I only change filenames within my players drive with MP3Tag. I do not do it outside and I make sure to update changes on iTunes using deduper script from turingtest2 (great guy) in the iTunes for Windows forum. This allows me to repopulate all my playlists with the proper version if I find a better mastering. I am very meticulous with my music, as I am sure you both are in your own way. That which I have sorted through is a copy of an original folder, which I keep untouched.

I actually brag more about how organized my library is, and all the information contain therein, as far as involved people, composers, dates of composition, periods, a proper classical music naming and complex genre scheme, than the size of it. Still it is very large at over 80k (ahem :slightly_smiling_face:), and that is after removing 8-9k duplicates (I had imported different masterings of records, now I just determined the best master and deleted the rest). I am sure that is not that large compared to some people here.

Anyways, thanks for your input. I do appreciate it.

The problem of the playlists and the broken links could be overcome if the name of the playlist gets saved in a user-defined field at creation time of the playlist, preferably together with the position number.
Filtering for that name and sorting by the field would restore the list in MP3tag and the playlist could be updated, again will all the files in it.

A very good suggestion that I have been pondering myself. Why not create playlist fields where we could save our playlist. I have never ordered a playlist, though I have thought about it. If I did, I would like that to be saved as well. Thanks for the input!