A set of models from the Ultimate Vocal Remover program, which are based on the old VR architecture. Most of the models are vocal, but there are also special models for karaoke, piano, removing reverberation effects, etc. Overall the models have average separation quality, but may be useful in some scenarios.