| 摘要 |
A system and method are provided for improving speech quality for signals divided into a frame structure. Speech information in a signal is detected, and a lost or corrupted transmitted frame is detected. The lost or corrupted frame is replaced by a suitable frame if a number of frames represented e.g., by a counter value, exceeds a predetermined value. The counter value may be changed, depending on whether the system is in a comfort noise generation state or in a muting period. The frame may be replaced by a frame representing mainly background noise, generated at the transmitting end during speech pauses or at the receiving end, or such a frame and a correctly received frame. The predetermined value may be the length of a muting period or may be the number of lost or corrupted frames preceding a speech frame. A first of the correctly received speech frames that follows a number of lost or corrupted frames may also be replaced. The output frames gradually approach pure speech frames. |
