AI Music Generation from Text

MusicLM is an innovative artificial intelligence model developed by Google Research that generates high-fidelity music from text descriptions. For example, given “a calming violin melody backed by a distorted guitar riff” as input, the AI would generate music that matches this description. The model uses a hierarchical sequence-to-sequence modeling task to generate the music at 24 kHz, ensuring consistency over several minutes. It outperforms previous systems in both audio quality and adherence to the text description. Furthermore, MusicLM can be conditioned on both text and a melody, allowing it to transform whistled and hummed melodies according to the style described in a text caption.