Early transmission in packetized speech
AssigneeMicrosoft Technology Licensing, LLC
InventorsKarsten Vandborg Sørensen, Sriram Srinivasan, Koen Bernard Vos
Life of this patent
Estimating the term…
Abstract
Network communication speech handling systems are provided herein. In one example, a method of processing audio signals by a network communications handling node is provided. The method includes processing an audio signal to determine a pitch cycle property associated with the audio signal, determining transfer times for encoded segments of the audio signal based at least in part on the pitch cycle property, and transferring packets comprising one or more encoded segments for delivery to a target node in accordance with the transfer time.
Claims18
A method of processing audio signals by a network communications handling node, the method comprising:
- processing an audio signal to determine a pitch cycle property associated with the audio signal;
- determining transfer times for encoded segments of the audio signal based at least in part on the pitch cycle property by
- establishing a packet interval for periodically transferring packets associated with encoded segments, wherein a packet may include one or more encoded segments;
- and
- determining a transfer time for a respective packet that is earlier than a packet interval associated with the respective packet based on the pitch cycle property predicting that a delay to include a further encoded segment would exceed the packet interval associated with the respective packet;
- and
- periodically transferring packets comprising one or more encoded segments for delivery to a target node in accordance with the transfer time.
A computing apparatus comprising:
- one or more computer readable storage media;
- a processing system operatively coupled with the one or more computer readable storage media;
- and
- program instructions stored on the one or more computer readable storage media, that when executed by the processing system, direct the processing system to at least:
- process an audio signal to determine a pitch cycle property associated with the audio signal;
- determine transfer times for encoded segments of the audio signal based at least in part on the pitch cycle property, the instructions to determine the transfer times include instructions to direct the processing system to:
- establish a packet interval for periodically transferring packets associated with encoded segments, wherein a packet may include one or more encoded segments;
- and
- determine a transfer time for a respective packet that is earlier than a packet interval associated with the respective packet based on the pitch cycle property predicting that a delay to include a further encoded segment would exceed the packet interval associated with the respective packet;
- and
- periodically transfer packets comprising one or more encoded segments for delivery to a target node in accordance with the transfer time.
A network telephony node, comprising:
- an audio input element configured to convert input audio into digital audio and deliver to an encoder service at a periodic rate;
- a speech cycle processor configured to process the digital audio to determine pitch cycle durations associated with the input audio;
- the encoder service configured to encode portions of the digital audio into encoded segments for periodic packet delivery of integer groupings of the encoded segments in packets over a packet network link;
- the encoder service configured to establish a packet interval for periodically transferring the packets;
- the encoder service configured to determine when encoding a current portion of the digital audio is predicted to span into a subsequent packet interval based in part on a current pitch cycle duration predicting that encoding and adding the encoded segment into a current packet will span into the subsequent packet interval, and trigger early delivery of the current packet without waiting for the current portion of the digital audio to be encoded;
- and
- a network interface configured to transfer the encoded segments over the packet network link.
Dependencies are inferred from the claim text.