Nvidia said its new model can generate any mix of music, voices and sounds described with prompts using both text and audio ...