Delimiters, sub-delimiters (!) and flattening/normalising

For some songs (beyond my normal sphere) that I tagged (over the weekend), Discogs (via Mp3tag's Search feature) returned multiple genre tags, separated by ", " (comma-space). It was first time I experienced this "in the wild" (previously I only expected to make my own tags that way).

I assume that only the comma is delimiter, the space should be regarded as "invisible". So it's not the case that comma-space is a two-character delimiter). Correct so far? (uncertainty over this has been bugging me).

Now the real issue...

Some of the so-delimited tags from Discogs contained what look like sub-tag delimiters, such as "/" or "&".

For example "Stage & Screen" (whatever that means here) in...

  • Madonna - Express Yourself (1989): [Electronic, Pop, Stage & Screen]

Also "Funk / Soul" in...

  • Hall & Oates - Private Eyes: [Rock, Funk / Soul, Pop]

My first reaction was Huh?. Apart from the fact that I don't regard this (Hall & Oates) song as Funk or Soul (was someone having a joke?), why didn't they just use "Funk, Soul" ? What they have essentially done is to create a second level, a sub level, of multiple-genre tagging.

So what to do with sub-tags, when encountered? My first reaction would be to flatten the two levels by simply including "/" and "&" along with "," all as equivalent delimiters. Then one could programatically just replace "/" and "&" by ",", when translating as-received tags into (let's call them) normalised tags. Motive: to simplify rotation rules creation in my auto-scheduling radio station playout app (RadioDJ).

Or maybe - based on a sample of one (!) "/" means "not sure", in which case any tags with "/" inside them should be regarded as questionable, unreliable. Partly based on the "sample of one" and partly from implication: not having used the (de facto) standard tag delimiter implies lack of thoughtfulness generally (or am I being harsh?, e.g. such tags could have been "inherited" from earlier days).

Would it be helpful for Mp3tag's Search feature to offer (an option) to normalise (hence split) such sub-level compound tags?

Any thoughts? Anyone "been there" already?

I am a little puzzled that you have problems with genres and subgenres after I participated in this thread:

Also, I think that a discussion about the discogs data quality would be a thing for the discogs forums. MP3tag is here only the messenger.

If you are not satisfied with the retrieved data and you want to adapt it to your needs, there are a number of action types that will help you to replace strings and split fields. I doubt that a ready-made function will be implemented for that as I bet that there are just as many users who want to keep the data as it is.

As clarification...

In my previous post (which you reference) the overall concept was "how can I best design my own tag system?". Whereas in my current question-post, I raise two main points:

  1. The simultaneous occurrence of two levels of separator/delimiter and thus of tags in single results from an online music database (here being Discogs, but I wonder if this happens more broadly - anyone know?) and how such cases are best interpreted, leading into the question...
  2. Could Mp3tag automate that interpretation in some way? (right now or in principle, e.g. as a user option) - say to allow multiple simultaneous separators/delimiters to be defined.

BTW for anyone interested:

  • the Madonna Express Yourself example can also be seen directly on Discogs here
  • ...while the Hall & Oates example is here.
  • For the latter, the only even vaguely Soul and/or Funk feel track is "I Can't Go For That". But yes, I should raise that on the Discogs forum - I merely mentioned it here in passing.

You only have to create a fitting action group.
The post-processing IMHO is nothing that any automatism should do but that should be tailored according to the individual needs. And for this purpose MP3tag features actions.
The web source should get the data as accurately as possible, that is what I think.


This is precisely why I have not used web sources as the only solution for my tags. Even if most users find them reliable, there is often room for some interpretation for genre. So I allow those sources to guide me, but ultimately it is what i choose to use in my library that matters to me.

Same goes for any other fields where I have found things like song titles or even artists have been mis-typed or even in some cases where tracks are listed out of order.

This is the nature of online data sources. Somebody else had to put it there, and there are no guarantees about it’s accuracy.


This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.