Meta has recently announced the launch of ‘AudioCraft’ as its new generative-AI tool for crafting audio and music from text – simplified and made easy for users.
AudioCraft has the capability to generate sound, music, and compression all in the same place while producing high-quality and realistic audio and music with long-term consistency.
Users can easily generate audio and music from a simple text prompt, so that even “professional musicians can explore new compositions without having to play an instrument”.
Sounds are made “easy to build on and reuse” so users who want to build better sound generators or music generators can do it all in the same platform and “build on top of what others have done”.
Consequently, content creators and small business owners can add soundtracks to their videos using the new AI tool. With a simple text input describing what the users want to hear, AudioCraft will turn it into a sound or music output.
According to Meta, their generative AI tool for audio and music consists of three open-sourced models, namely, MusicGen, AudioGen and EnCodec, which has their own functions that make their tool advanced in the field of AI-generated audio.
To elaborate, Meta explained these three models as;
- MusicGen was trained with Meta-owned licensed music which generates music from text prompts.
- AudioGen was trained on public sound effects which generate audio from text prompts.
- EnCodec, which was stated to be an improved version of their decoder, allows high quality music generation.
In addition to this, they have also released their pre-trained AudioGen models where it enables users to generate environmental sound effects like birds chirping, vehicles honking, and footsteps on concrete flooring.
“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 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 a release.
Meta also sees this as a tool that could benefit sound designers and musicians by giving ideas and inspiration in a new yet innovative way to help with their composition.
The power of AI has now reached the ability to generate and produce high quality sounds and music. The integration of AI to sound-generation will change the way in how sounds and music are produced as well as how people will listen to them in the generations to come.