| 摘要 |
PROBLEM TO BE SOLVED: To provide a voice encoding method accompanied by such packet repair processing that S/N and subjective quality are good and the voice in a consonant section is articulate. SOLUTION: Multiple interpolation repair processes are prepared on a transmission side. Assuming that frames to be transmitted are lost on the transmission side, all the interpolation repair processes are tried, frame by frame. Then the waveform interpolated and repaired through the repair processes is compared with a reproduced waveform locally decoded from the packet. Then the index number of an interpolation repair processing system which can obtain an interpolated and repaired waveform closest to the locally decoded reproduced waveform is sent to a reception side together with the packet. On the reception side, multiple interpolation repair processes are prepared as well as the transmission side and if the lost of a packet is detected, a interpolation repair system is selected according to the index number of the interpolation repair system sent together with the frame and the interpolation repair process is carried out. Consequently, when no packet is lost, the interpolated and repaired waveform which is closest to the decoded reproduced waveform is obtained. |
