| US 7,590,047 B2 | ||
| Memory optimization packet loss concealment in a voice over packet network | ||
| John T. Dowdal, Germantown, Md. (US); Sachin Adlakha, Germantown, Md. (US); and Dunling Li, Rockville, Md. (US) | ||
| Assigned to Texas Instruments Incorporated, Dallas, Tex. (US) | ||
| Filed on Feb. 14, 2005, as Appl. No. 11/57,638. | ||
| Prior Publication US 2006/0182086 A1, Aug. 17, 2006 | ||
| Int. Cl. H04L 12/26 (2006.01) | ||
| U.S. Cl. 370—216 [714/701; 714/746; 714/785] | 4 Claims |
| 1. A method for packet loss concealment, comprising:
receiving voice data packets containing voice signals into a voice playout unit;
detecting, in the voice playout unit, a loss of at least a first and a second consecutive voice packet during said receiving;
providing a history buffer in packet loss concealment unit to retain said received voice signals;
saving, in the history buffer, said voice data signals which arrived in the packet loss concealment unit prior to the lost
first and second consecutive voice packets but are not yet played out, only when the voice playout unit detects the lost first
and second consecutive voice packets;
using, in the packet loss concealment unit said buffered voice data signals instead of at least part of the history buffer
of the packet loss concealment unit as a sample history to generate synthesized voice data signals when the packet loss concealment
unit is ready to replace the lost voice data signal;
copying said received voice data signals to a scratch buffer;
using the scratch buffer as a working buffer to generate synthesized voice signals to replace voice data signals lost with
said first and second consecutive voice packets; and
extending said history buffer to include said received voice signals and said synthesized voice signals.
|