In this page, we introduce a few audio examples drawn from the test set of MUSDB18. Each audio example have been downsampled to 16kHz (Anchor), then processed successively by the Apollo model, the AudioSR model, and our model HP-codecX.