Pre-Publication V3. The application of machine learning algorithms and neural networks intended to alter existing manipulated sound audio sources and/or generate derivative synthetic audio such as music audio compositions, speech synthesis, and sound effects from human generated initiated inputs.
Deliberation Summary:
(1) Refine the audio terminology for precision, preferring sound to audio in describing source material, replacing a compositional phrase that is not in industry use with a direct reference to music, and describing the inputs as human initiated rather than human generated. (2) Replace “manipulate” with “alter” and “prompt” with “inputs”; the phrase “existing sounds” was adopted to distinguish source material from generated audio without the ambiguity of alternative wording.
Pre-Publication V2. AI Sound Synthesis
The application of machine learning algorithms and neural networks intended to manipulate audio and generate create derivative new audio compositions, speech synthesis, and sound effects from human generated text prompts.
Deliberation Summary:
(1) Consider whether this process will always originate from a text prompt, suggest broadening how the input is described, including replacing the reference to text with human generated input. (2) Wording changes framing the process as intended to generate rather than to create.
Pre-Publication V1. Sound Synthesis
The application of machine learning algorithms and neural networks to manipulate audio and create new audio compositions, speech synthesis, and sound effects from text prompts.