NVIDIA announces an experimental, automatically generative artificial intelligence modelFugatto, officially known as "Foundational Generative Audio Transformer Opus 1", is mainly used to create audio content, or modify existing music, voice or sound details, and is advertised as being able to handle content in multiple languages and accents.
This model has been described as a "sound all-purpose knife." It primarily uses artificial intelligence to understand content and process sound details. For example, it can quickly create a prototype of a song through artificial intelligence, and then derive different styles, performance methods, and dubbing content.
Users can use custom sounds as training material for generated content, while game developers can leverage existing sound assets to create more application resources or adjust the sound effects in games to meet the needs of players. Furthermore, this model can generate sounds that vary over time, such as the sound of wind crashing across land as a storm passes over land, and can also be trained for specific sounds.
It is currently unclear whether NVIDIA plans to open this model to the public. It is possible that it will still be used in specific fields for academic research.
