Meta releases framework for creating sounds and music to open source

Meta, formerly Facebook, has introduced Audiocraft, a framework that utilizes AI to generate high-quality, realistic audio and music from text prompts. This isn’t the first time Meta has ventured into the field of audio generation; however, the company claims considerable improvements. With three generative AI models – MusicGen, AudioGen, and EnCodec – Audiocraft is designed to simplify the use of sound and music generators. Ethical and legal issues have been raised regarding MusicGen’s ability to learn from existing music and produce similar effects. Meta has stated that the model was trained with Meta-owned and specifically licensed music, and that vocals have been removed from the training data to prevent replication of artists’ voices. Despite this, the company doesn’t outright prohibit commercial applications.