Stations: Algorithm doesn't account for artists that shift genres
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Can we boost this? The algorithm currently in use does not account for artist that shift genres. not only that, it tends to apply artist to genre's that aren't appropriate. I made a station from the Artist "ODDEEO" which is a late 80's bar music style sound that uses vocoloid for vocals. They have multiple collaborations with other vocoloid artists, however, the station will rarely play "ODDEEO" and gets lost in Scandinavian Pop like Grimes. It has deviated far from what the Music Genome Projected intended, which was to build upon the sound of a song and play other songs that match the fundamentals of the music, and instead took the short route of grouping similar artists and play their top hits. This station sadly is continuously suggesting EDM and Techno (i.e. Rezz and DeadMau5) is a world apart from the sound of ODDEEO - a massive deviation from the intent of the music genome project.
Even a station based on G3, Joe Satriani's touring group, deviates from the guitar virtuoso/technical genre and start playing base metal that is more mainstream like AC/DC. It got too broad far too fast and isn't helping bring the truly similar artists the proper exposure. you used to find independent artists on Pandora, but now it's on par with normal radio - 9 out of 10 songs are overplayed hits played on random for 1 song to actually be what you want to listen to.
Moderator Edit: Edited title for clarity
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@dasgones I moved your post over to the My Collection space.
Sorry to hear that your stations aren't playing the way you want them to.
Pandora stations adapt with every song to try to find the right music for you. It does this by using a combination of the detailed information in the Music Genome Project, the station seeds and Thumbs feedback you have given along with feedback given by other listeners under similar conditions.
There are a number of variables that come into play when selecting the next song:
-
The complete range of musical attributes for all of the songs we have by the artist you used to create the station. For example, for a station based on Brian Eno, you might hear rock songs and ambient music. You might try creating a station from one of the artist's songs instead (e.g. Lucky Star by Madonna). This should make the station much more specific.
- The type of feedback you've given on the station (all the Thumbs Up and Thumbs Down ratings) and any additional artist or song seeds you've added to the station. Pandora lumps all this together and tries to determine what the best musical match is based on our musical analysis.
To improve your stations, here are a couple of things to try:
- Thumb down tracks that do not fit. This will prevent that song from playing again, and Pandora will learn more about what you don't want to hear.
- Make sure each of the thumbs-up you've given fits in with the type of station you are trying to create. If not, then remove that thumbs-up rating from the station page.
- When listening from a computer, a mobile device or the Sonos app, you can access Pandora Modes to customize your station experience. You can pick different modes like Crowd Faves, Discovery and Deep Cuts.
Of course, Pandora Premium subscribers can also create custom playlists in order to have complete control of what they hear on Pandora. If you want more information about that, let me know.
Thanks so much for listening!
Adam | Community Moderator
New to the Community? Introduce yourself here
New music alert! Picks of the Week (1.6.25)