Microsoft has submitted a patent named “ARTIFICIAL INTELLIGENCE MODELS FOR COMPOSING AUDIO SCORES” on the WIPO IP Portal. Microsoft is developing an intelligent audio composition engine that will allow it to create noises, music, and other audio components for a wide range of media, including movies, TV programmes, games, and even live recordings. The patent references dynamic moments in games, implying that it might provide scores that alter based on the activities of the player. According to the patent abstract, parameters may be defined using visual, audio, and textual elements and prompts (together referred to as ‘Dataset’) to train a variety of AI models on how to create audio scores.
The recent introduction of AI has been groundbreaking, crossing numerous areas of art and media. Although there have previously been a number of AI tools for audio synthesis available, Microsoft’s recent patent seems to indicate that their own AI model ecosystem will be the most extensive and complex system of machine aided music composition to date.
