Explore the Possibilities of AI-Generated Music with Google’s MusicLM
Google researchers have developed a groundbreaking artificial intelligence (AI) system called MusicLM that can generate minutes-long musical pieces from text prompts. It can even transform a whistled or hummed melody into other instruments, similar to how systems like DALL-E generate images from written prompts. Google has released samples of the AI’s music, showcasing its impressive capabilities.
Listen to the AI’s compositions and explore the possibilities of this technology with MusicLM. From 30-second snippets of songs created from paragraph-long descriptions to five-minute pieces generated from one or two words, MusicLM is capable of producing a wide range of music. You can even hear its interpretation of phrases like “futuristic club” and “accordion death metal.”
The AI can also simulate human vocals, though the sound is grainy or staticky. To demonstrate this, Google has released an example of “story mode”, where the model is given a script to morph between prompts. For example, the prompt “electronic song played in a videogame (0:00-0:15) meditation song played next to a river (0:15-0:30) fire (0:30-0:45) fireworks (0:45-0:60)” resulted in the audio you can listen to here.
Google has released a research paper explaining the technology in detail, and is also publicly releasing a dataset with around 5,500 music-text pairs to help when training and evaluating other musical AIs. However, the company is being cautious about releasing the model due to potential risks of “misappropriation of creative content” and cultural appropriation.
AI-generated music has a long history dating back decades, and MusicLM is yet another example of how far this technology has come. To hear the possibilities of this technology, check out the samples produced by MusicLM.