Audiotool board archive

Tag based genres

Leinholz · started 2019-04-13 22:05 · updated 2019-09-26 23:42

I'm not so happy about the genre selection/situation that is implemented at the moment.
I think there would be more artistic freedom if you could choose a genre from a tag list and especially have the possibility to create a genre tag if your desired tag doesn't exist.
You could still keep the concept of promoting the top genres. For example you could feature the top 10 used genre tags or something like that.

Comments (14)

2019-04-14 10:30 · 2019-04-14

Sorry, but I don't understand your suggestion. One can choose an "official" genre (which is displayed as a tag) and some user defined tags.
Only the genre-tag will be technically taken into account for the association with the genre.

2019-04-14 19:51 · 2019-04-14

Sorry, I was a little unclear. Basically I am trying to suggest that the genres on the "Music" page could be flexible. I sometimes find it difficult to categorize tracks in those couple of available "official" genres. Therefore it would be cool if you could suggest/create genres and the top so and so many would be displayed on the "Music" page.

2019-04-15 15:43 · 2019-04-15

I saw this appear in my newsfeed awhile ago, and I was meaning to comment on it, because I really like the idea, in theory.
@chordofdestruction In attempt to clarify...
I didn't read this in depth, but to my understanding, the audiotool community would decide on what the official genres are (I guess, of course with moderative assistance). So, the most frequently used tags could potentially become official genres, if used often enough for viable tracks (e.g. not WIPs/experimental, memes, etc.)

Apollo · reply
2019-04-15 15:54 · 2019-04-15

Yeah, I see what you're saying. I really like this idea because genres in general are a flexible thing. Confining them to what's currently available is not only artistically limiting, but developmentally, it doesn't future-proof audiotool. As an extreme example, there may popularly be new genres in the future we don't even know the names of yet. But also, there are popular genres not listed in the current implementation that I would love to follow.
I'd actually go as far to argue that every community created genre should at least have its own page (yes, even memes). That way, you can easily keep up with what you're truly interested in and its completely futureproof. (I would love to listen to Jazz and Complextro, but those are legitimate genres that aren't supported by audiotool)

2019-04-15 18:37 · 2019-04-15

@apollo Thanks for elaborating. That's exactly what I was thinking :D

Apollo · reply
2019-04-15 18:44 · 2019-04-15

No problem! Hopefully its helpful to others as well
But I'm fully on board with this, would definitely like to see something like this implemented.

2019-04-15 21:30 · 2019-04-15

That's a difficult topic. We have to avoid that we end up with single-song-genres.
As an example: I'd like to distinguish between: Hardcore, Gabber, Jumpstyle, Hardstyle, Schranz, Rave - but currently we don't even have enough tracks to get the existing Harcore genre filled with enough content (the situation has improved in the recent weeks, though).
More genres will make it harder to develop any chart-system.
Another problem is that lots of songs obviously have the wrong genre tag.
But I'm totally on you side that we have to improve something. I suggested something like weighted genre associations - 20 % Techno, 70 % Goa, 10 % Death Metal. This way we can list tracks for niche genres as long as some users see a relation. We'd have to take the relevance somehow into account.
I still believe it would be best if users can suggest genres for each track.

Leinholz · reply
2019-05-02 19:10 · 2019-05-02

Yeah, I see the problem of single track genres. Maybe some kind of threshold could be implemented to avoid that. For example a genre only gets a genre page when a certain amount of tracks have been submitted to it.
About the problem of currently not filled genre charts: Maybe the wrong genres are listed ;)
I like the idea with the percentage system. Not sure how the chart listing could work with that though. Would tracks get listed in every genre chart page that you select then?

Leinholz · reply
2019-05-02 19:13 · 2019-05-02

I think such a flexible system would level off after a while. After all the artists want to get recognition for their tracks. Therefore they are more likely to submit / make tracks in genres that are more popular.

Leinholz · reply
2019-05-02 19:16 · 2019-05-02

Or the devs could moderate when a genre gets a page..

Leinholz · reply
2019-05-02 19:17 · 2019-05-02

On the other hand there's not much point submitting a track to a genre that doesn't even have a page.

2019-09-04 00:50 · 2019-09-04

We've been working on this topic in the recent weeks. I've decided against implementing user suggested genres. Assume someone decides to publish a track with a genre of UK Drill (which is a sub-sub-genre of Trap and thus a sub-sub-sub-genre of Hip Hop). It would be legit to offer this track to a user who's interested in Hip Hop. It would also be legit to let that track participate in some Trap charts. But because that genre is unknown to us, we cannot deduce any relations and the track might get lost in the voids.
Here's the deal: We offer a new set of hand-picked genres (including some new, based on the most popular tags) and equip them with relations between them. Additionally we allow to categorize the tracks even further. The aforementioned track might be declared as follows:

  • Genre: Trap
  • Mood: Aggressive
  • User Tags: #UK Drill
    This is similar to what we have now, but the mood part allows to filter the track without introducing more and more sub-sub-sub-genres. We will also make sure that once a user tag becomes too popular to ignore, we add it as another genre along with all relations to the existing set.
Known As I · reply
2019-09-26 23:42 · 2019-09-26

I just recognized that after some iterations we're pretty close to what @beatbob suggested :D
Internally that implementation became really complex.