New Delhi: Meta’s new AI tool AudioCraft is a fascinating innovation that allows users to create music and audio from text prompts. It is based on three generative AI models that can produce sound effects, melodies, and even orchestral compositions from simple descriptions. Here are some of the features and benefits of AudioCraft:
- AudioGen can generate realistic and diverse sound effects from text inputs, such as “a dog barking” or “a car horn honking”. This can be useful for creating immersive soundscapes for games, movies, podcasts, or other media projects.
- MusicGen can create music from text inputs that specify the genre, mood, instruments, tempo, and other aspects of the desired song. For example, “a pop dance track with catchy melodies, tropical percussions, and upbeat rhythms, perfect for the beach”. MusicGen was trained on over 20,000 hours of music that is either owned by Meta or licensed for this specific purpose.
- EnCodec is a neural network-based audio compression codec that can reduce the size of audio files without compromising the quality. It can also help generate higher-quality music with fewer artifacts, preventing audio manipulation from causing distortion.
AudioCraft is an open-source tool that Meta has released under the MIT License. This means that anyone can access, modify, and use the tool for their own purposes. Meta hopes that AudioCraft will contribute to the advancement of generative audio technology and inspire more creativity and experimentation among researchers and practitioners.
If you are interested in learning more about AudioCraft, you can visit Meta’s website or read their blog post. You can also listen to some of the audio samples they have provided on their website¹. They are quite impressive and demonstrate the potential of AudioCraft.
