Another one here. Vocal techno is a pain in my ears. There are few exceptions like in Bonzai Records classic tracks like “Let there be house.” or Emmanuel Top. These male voices are more aggressive and are helping to inject some additional energy into the songs, while female voices seem to drain it.

Goth industrial stuff from like the female singer from L’Âme Immortelle was a nice exception with “Bitterkeit”.

Last.fm and scrobling was way more accurate 20 years ago than anything available today.

I would bet that you can train Pandora's algo to respect this preference by thumbing down every vocal techno track. The music genome annotators will have a boolean for vocals in the analysis. You'd likely start using recommendation strategies that use the annotations more than normal. Pandora is still the best for lean back listening.

Have you tried brain.fm? It’s closer to the Lofi genre, has zero vocal which I really appreciate and no quick abrupt, distracting changes. It’s very rhythmic.

It happens to me that I keep skipping certain song but no matter what, it gets again into autoplay.

What playlist creation tool are you using ?

There's many auto generating dynamic ones. Like the Monday and Criday ones. Plus the DJ new AI thing and then the better stuff is often in all the human curated Spotify lists.

There's also a way on desktop to take an existing g playlist and "create similar" where it tries to make something similar too one of your already made playlists. You have a lot more control using this

Spotify was never a good product IMO and for techno it has always been obviously a bad product as many many songs and variants never make it to their platform.

Even YouTube music works better for me most of the time, but SoundCloud clearly takes the lead

Spotify especially likes to push its own AI generated slop tracks because then it gets to pay the royalty to itself

Yes, just because an algorithm knows what someone wants doesn't mean the algorithm will give it to them.

In fact, the algorithm is there to make money by changing the choice a person would make otherwise - the subtlety I assume is figuring when someone will accept a "good enough" choice handed to them. But just limiting choice to what you want someone to make might be a simpler approach.

Similar here: I occasionally listen to music on Youtube late in the evening, and always watch a math or physics video in the morning. Simple, but it appears impossible for Youtube.

Sure would be nice if we enforced Payola laws. It's hard to believe that at one point the US Government sought to regulate fairness on something as low-stakes as radio. Today corporations do much worse and our politicians mostly sit by and watch. Maybe a small fine.

> When listening to personalized playlists on Spotify and YouTube Music I skip every single electronic track with vocals. I listen to a lot of techno and related genres, and the moment someone starts singing I skip the track. I've been doing it for years. None of these advanced algorithms have ever noticed this simple pattern.

Holy shit, I’m the exact opposite. I’m so sorry. I am the reason you have to listen to all that vocal trance.

I actively thumbs down the 99 of every 100 tracks I don’t like, the services still don’t get it.

I really wish I could use an LLM to select the recommendations

There's just no incentivization to do better at this, or they would in a heartbeat

> I think the songs get pushed by who pays to push, and most of those songs have vocals.

They get pushed also by what's the most profitable to play out, which is why you get so many bootleggy "Live Version" and shitty remixes.

It's often hard to get Spotify that actually all you want is the unremixed unedited unremastered album version of a track.

you're on wrong platform, soundcloud

Hey just let them merge a few hundred thousand more PRs of slop and maybe it will work?

Spotify is chock full of payola. Every single playlist l they generate (and auto shuffle mode) has paid ads in it. It’s a terrible product.