Meta will allow users to generate music from text using ‘AudioCraft family of models’
Facebook's parent company Meta has introduced a new generative AI-based tool called AudioCraft which can help users high-quality, realistic audio and music using text prompts. According to a release by Meta, the company is releasing a set of three audio models namely, MusicGen, AudioGen and EnCodec. The company says MusicGen has been trained on Meta owned specifically licensed music and can generate music using text prompts.
Meanwhile, Audiogen has been trained on public sound effects and can generate audio using text prompts. Meanwhile, the EnCodec decoder allows users to experience high-quality music generation with ‘fewer artifacts’. Meta has also announced new pre-trained AudioGen models which can help generate environmental sounds resulting in sound effects like dog barking, cars honking, footsteps or wooden floors.
The company also informed that it is also releasing all of the AudioCraft model weights and codes. "With AudioCraft, we simplify the overall design of generative models for audio compared to prior work in the field — giving people the full recipe to play with the existing models that Meta has been developing over the past several years while also empowering them to push the limits and develop their own models." Meta said in the release on Thursday. Artists have earlier raised concerns about artificial intelligence tools affecting their livelihood.
However, Meta says that the AudioCraft family of models will act as tools for musicians and sound designers helping them to provide inspiration and help people quickly brainstorm and iterate on their compositions in new ways. Earlier this year, Google introduced its experimental AI model called MusicLM which could help turn written prompts into music. The
. Read more on livemint.com
