Mixing decision controlling decode decision
AssigneeGoogle Inc.
InventorLars Henrik ELLNER
Life of this patent
Estimating the term…
Abstract
Methods, systems, and apparatus are provided for combining (e.g., mixing) audio signals received from a plurality of participants communicating with each other during a communication session (e.g., an audio conference) based on voice-activity-detection (VAD) data contained in extended headers of Real-time Transport Protocol (RTP) packets. An audio mixing apparatus receives RTP packets from connected clients and extracts the VAD data included in the extended RTP headers to render a mixing decision. Once a mixing decision has been made, the audio frames to be mixed are decoded while other received frames are discarded, thereby preventing processing resources from being wasted to decode frames that are never used.
Claims18
A method for mixing audio signals comprising:
- receiving, at an audio mixing apparatus, audio packets from a plurality of clients in communication with the audio mixing apparatus;
- retrieving audio level data contained in an extended packet header of each of the received audio packets;
- selecting audio frames to be mixed based on the audio level data retrieved from the extended packet headers of the audio packets;
- decoding the audio frames selected to be mixed;
- and
- generating a mixed audio stream by mixing the decoded audio frames.
An audio mixing apparatus configured to perform operations comprising:
- receiving audio packets from a plurality of clients in communication with the audio mixing apparatus;
- retrieving audio level data contained in an extended packet header of each of the received audio packets;
- selecting audio frames to be mixed based on the audio level data retrieved from the extended packet headers of the audio packets;
- decoding the audio frames selected to be mixed;
- and
- generating a mixed audio stream by mixing the decoded audio frames.
Dependencies are inferred from the claim text.