摘要 |
The present invention is a low speed speech encoding method based on the network protocol (IP), including: determining the attribute parameter of the input speech during a period of time (TN); hereby determining the best frame length (T) of the following speech processing; encoding the speech data in every T; finally outputting after packing up by TCP or UDP protocol, IP protocol. Using the low speed speech compression encoding method of single frame, alter frame length, intra-frame bit self-adapting, is of advantage to increase the ability of encoding compression and efficiency of transmission. Because the IP datagram partition uses the partition of the best frame length speech encoding frame as a unit, it extremely increases the quality of source speech encoding and decoding. Informal sound test indicates that the MOS value can be increased 0.1 - 0.2.
|