| 摘要 |
PURPOSE:To provide high quality voice by performing classification into plural classes based on error sensitivity, detecting errors for the respective classes and performing different bad frame masking processings composed of the interpolation of parameters and muting. CONSTITUTION:Coding parameters outputted from a voice coding device 100 are sent to a classifying circuit 110 and classified into the data of 74 bits as important class 1 data and the data of 60 bits as class 2 data with low importance. At the classifying circuit 120, the data of 10 bits of a more important A class and the data of 34 bits of a second important B class are taken out. The CRC code of 7 bits is calculated at a 7-bit CRC calculation circuit 130, an error detection function is added, 5-bit CRC is similarly calculated at a 5-bit CRC calculation circuit and they are inputted to a convolution coding circuit 150. A 2-slot interleaving circuit 160 is given the data of 164 bits from the circuit 150 and the data of 60 bits from the circuit 110 and performs an interleaving processing. |
