/Patent Tracker
GooglePatent applicationAudio & voiceInterest 6/10

Google patents faster audio generation using fixed-step neural networks

Google has filed a patent for a method that creates audio signals using a neural network that runs a fixed number of steps, regardless of how long the sound is. The system processes chunks of data simultaneously using attention mechanisms to predict the final audio tokens, rather than generating them one by one.

Why it mattersThis approach could make computer-generated sounds, like speech or music, produce much faster by avoiding the slow, step-by-step process used in current audio models.

In the filing’s own words

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating an output audio signal. In some implementations, the output audio signal can be generated by using a generative neural network in a number of iterations that is independent of the length of the output audio signal.

What exactly they claim (claim 1)

1. A method performed by one or more computers, the method comprising: receiving a masked representation of an audio signal as a sequence of tokens, wherein the sequence of tokens comprises one or more masked tokens; processing the masked representation of the audio signal using a neural network, comprising: processing the masked representation of the audio signal using an embedding layer of the neural network to generate a sequence of embeddings that comprises a respective embedding of each token in the sequence of tokens; processing the sequence of embeddings using a pooling layer of the neural network to generate a sequence of pooled embeddings, wherein each pooled embedding in the sequence of pooled embeddings results from pooling a respective plurality of embeddings from the sequence of embeddings; processing the sequence of pooled embeddings using an attention block of the neural network to update the sequence of pooled embeddings, comprising applying one or more self-attention operations to the sequence of pooled embeddings; and processing the updated sequence of pooled embeddings to generate, for each of one or more positions in the sequence of tokens, a respective prediction characterizing a token that should occupy the position in the sequence of tokens.

Facts

Company
Google (Google LLC)
Official title
GENERATING AUDIO USING NON-AUTOREGRESSIVE DECODING
Publication number
US20260290306A1
Application number
19477525
Filed
Apr 29, 2024
Published
Sep 24, 2026
Inventors
2
Claims
21

Classified under

G10L 13/027 Physics & computing
Read the full filing (official PDF) ↗ Open in the Patent Tracker

More from Google in Audio & voice

Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.