MusicGen

About

MusicGen, by Meta AI, which is part of the AudioCraft suite of generative AI models. MusicGen is a controllable model for music generation that can generate music conditioned on textual or melodic features (influence audio), allowing better control over the generated output. Meta AI envisions MusicGen as a new type of instrument, similar to synthesizers when they were first introduced

Features

• Text-to-Music Generation: MusicGen is a text-to-music model capable of generating high-quality music samples conditioned on text descriptions or audio prompts
• Controllable Music Generation: MusicGen allows for better control over the generated output by being conditioned on textual description or melodic features (influence audio)

Limitations

• Lack of Emotional Depth: AI-generated music, including MusicGen, may struggle to create music with genuine emotional depth, as it lacks the emotional understanding and connection that human composers bring to their work.
• 60 seconds: We only support max of 60 seconds of audio generation at a time.

Usage Tips

• Use Stereo for more Depth: Use the Stereo model for more depth in the generated audio.
• Use Short clips for Influence Audio: Use Influence audio to better control the generated audio. Influence audio can be a short audio clip or a melody.