초록 ▼

The mobile telephone detects bad data packets by decoding speech parameters represented by the packets and comparing those speech parameters against acceptable ranges of speech parameters. If the speech parameters do not lie within the acceptable ranges, the packets are discarded. This speech parame

The mobile telephone detects bad data packets by decoding speech parameters represented by the packets and comparing those speech parameters against acceptable ranges of speech parameters. If the speech parameters do not lie within the acceptable ranges, the packets are discarded. This speech parameter-based detection of bad data packets is particularly useful in mobile telephones receiving variable rate data packets wherein an error in determining the rate of a packet results in the entire packet being decoded at an incorrect rate, thereby likely causing an annoying audible artifact in the decode speech signal. Such rate detection errors can occur even though a CRC or other conventional error detection check fails to detect the bad packet. In an exemplary implementation, the mobile telephone receives signals encoded with TIA/EIA/IS-95-A standards. Linear speech parameter frequencies and codebook gain parameters determined using linear predictive coding techniques are tested against acceptable ranges to detect bad data packets.

대표청구항 ▼

[I claim:] [1.]a receiver that outputs a digitized signal containing received speech parameters representative of speech;a decoder, responsive to a bit rate detection unit, that generates processed speech parameters from the received speech parameters, the processed speech parameters being based at

[I claim:] [1.]a receiver that outputs a digitized signal containing received speech parameters representative of speech;a decoder, responsive to a bit rate detection unit, that generates processed speech parameters from the received speech parameters, the processed speech parameters being based at least in part on a detected bit rate of the received speech parameters;a comparator that compares the processed speech parameters with at least one predetermined range of acceptable speech parameters to identify and discard atypical processed speech parameters outside of said predetermined ranges, the comparing being based at least in part on a rate of change of the processed speech parameters; anda speaker that outputs an audio signal associated with processed speech parameters lying within said predetermined ranges.