Understanding the necessity of spotdl

[therepanic]

In the current implementation, we use spotdl, which in turn uses ytdlp under the hood. Now we poll tracks in real time, get their meta information, including track names. However, then through spotdl, we again get these tracks under the hood to use ytdlp. In our case, it would be better to use ytdlp directly.

We need to figure this out correctly, or if I understand everything correctly, then upload the fix.