audiocraft-audio-generation

Text-to-sound generation creates sound effects and ambience from text descriptions using AudioGen. It applies to sound design workflows like Foley generation, allowing rapid prototyping of environmental audio from descriptive prompts.

Melody conditioning enables style transfer by guiding audio generation with an existing melody across multiple model variants. This allows you to condition MusicGen outputs to match specific melodic structures while altering the surrounding musical style.

Environment setup for reproducible audio generation involves specifying dependencies and configuring generation parameters. The Skill defines recommended workflows including model variant selection and specific configurations to ensure consistent, reproducible audio outputs.

AudioCraft supports both music composition and sound effect generation through its MusicGen and AudioGen models. It leverages EnCodec to process audio, enabling text-to-music creation alongside text-to-sound effects for comprehensive audio prototyping.