I understand that stem generation is a new feature and a work in progress. However, what I do not understand are some of the current limitations:
While the stem generation process runs in the background, it is possible to select another song and generate stems for it as well. The subsequent stem generation tasks are queued up automatically. This means that batch stem generation is effectively implemented, but the user experience is cumbersome and unintuitive.
Is there a reason for this design choice? It seems to unnecessarily complicate the process for users.