Jitter buffer adaptation based on audio content
AssigneeBroadcom Corporation
InventorsMickael Jougit, Laurent Pilati
Life of this patent
Estimating the term…
Abstract
A system and method for discarding or inserting audio frames in a jitter buffer is described. The system and method provides improved audio quality as compared to conventional jitter buffer management systems. In one embodiment, buffer control logic determines whether to discard audio frames to be stored in a jitter buffer or to insert audio frames among audio frames to be output from a jitter buffer based not only on the number of audio frames currently stored in the jitter buffer but also based on the power of the current audio frame to be stored in or output from the jitter buffer. The system and method is generally applicable to any wireless or wired communication system in which audio signals are transmitted between entities operating in different clock domains.
Claims20
A method for generating an output audio signal in a communication system, comprising:
- (a) receiving a segment of an audio signal, wherein the received segment is one of a series of segments of an audio signal transmitted over a communication link;
- (b) obtaining an estimate of the power of the received segment;
- (c) comparing the estimated power of the received segment to a power threshold that varies in a manner that depends on a current number of segments stored in a jitter buffer;
- (d) storing the received segment in the jitter buffer if the estimated power of the received segment exceeds the power threshold and discarding the received segment if the estimated power of the received segment does not exceed the power threshold;
- and
- (e) generating an output audio signal from segments stored in the jitter buffer.
A system for generating an output audio signal, comprising:
- a jitter buffer;
- buffer control logic configured to receive a plurality of segments of an audio signal transmitted over a communication link, to obtain an estimate of the power of each of the received segments, to store a received segment in the jitter buffer if the estimated power of the received segment exceeds a power threshold, and to discard a received segment if the estimated power of the received segment does not exceed the power threshold, wherein the power threshold varies in a manner that depends on a current number of segments stored in the jitter buffer;
- and
- audio output logic configured to generate an output audio signal from segments stored in the jitter buffer.
A system for generating an output audio signal, comprising:
- a jitter buffer;
- buffer control logic configured to receive a plurality of segments of an audio signal transmitted over a communication link, to obtain an estimate of the power of each of the received segments, to store a received segment in the jitter buffer if the estimated power of the received segment exceeds a power threshold, and to discard a received segment if the estimated power of the received segment does not exceed the power threshold, wherein the power threshold varies in a manner that depends on a current number of segments stored in the jitter buffer;
- and
- packet loss concealment logic configured to predict an output audio signal based on segments stored in the jitter buffer.
Dependencies are inferred from the claim text.