Meta, the technology company known for its advancements in artificial intelligence (AI), has unveiled a new set of AI tools that will revolutionize the music industry. These tools, collectively called AudioCraft, consist of three generative AI models: MusicGen, AudioGen, and EnCodec. With these models, users can create high-quality and realistic music simply by inputting text.
MusicGen and AudioGen are responsible for generating sound from text. While MusicGen focuses on creating music, AudioGen produces specific audio and sound effects. To demonstrate the capabilities of MusicGen, Meta provides an example prompt: “An 80s driving pop song with heavy drums and synth pads in the background.” Users can visit MusicGen on HuggingFace and experiment with the demo, describing any type of music they would like to hear from any era.
EnCodec, on the other hand, is an audio codec that utilizes neural networks to compress audio and reconstruct the input signal. Meta has released an improved version of EnCodec that enables higher-quality music generation with fewer artifacts. This enhancement ensures that the music produced by AudioCraft is of the utmost quality.
In addition to the generative AI models, Meta has also made the pre-trained AudioGen models available to users. These models allow users to generate environmental sounds and sound effects, such as a dog barking or a floor creaking. By providing access to these models, Meta aims to empower users to create a wide range of audio experiences.
Furthermore, Meta has demonstrated its commitment to collaboration and innovation by sharing the weights and code for all three open-source models. This move enables researchers and practitioners to leverage the models and train other AI models based on them. Meta believes that AudioCraft has the potential to become a new standard instrument, akin to the impact that synthesizers had when they were first introduced.
Meta’s CEO expressed excitement about the future of MusicGen, stating, “With even more controls, we think MusicGen can turn into a new type of instrument — just like synthesizers when they first appeared.” This sentiment reflects the company’s belief in the transformative power of AI in the music industry.
While Meta’s AudioCraft is groundbreaking, it is not the only generative AI model of its kind. Google released its own model called MusicLM earlier this year, which also transforms text into music. Furthermore, a recent research paper revealed that Google is utilizing AI to reconstruct music from human brain activity. These developments highlight the increasing integration of AI in the music landscape.
The introduction of AudioCraft by Meta signifies a significant step forward in the use of AI in music creation. With its user-friendly interface and powerful generative AI models, AudioCraft has the potential to revolutionize the way music is composed and produced. As the technology continues to evolve, it is clear that AI will play an increasingly prominent role in shaping the future of the music industry.
