| 摘要 |
A duplex speech transmission method is disclosed wherein a limited bandwidth transmission channel is divided in a plurality of successive time slots, alternate time slots being available for each of two parties to call in Speech signals from either of the parties are divided into time segments with the time segments associated with one of said two parties being offset in time with the time segments associated with the other of said two parties. Segments of speech when present are digitised and stored in digital form. A higher rate of quantization is used when speech is present in the segments of one party and not in those of the other party, said higher rate quantized signal being transmitted in the other party's time slot and the immediately following time slot of said one party. A lower rate of quantization is used if both parties have speech in their respective segments, the lower rate quantized signals being time compressed and transmitted in their associated time slots. By the method being digited throughout the end of a segment coincides substantially with the end of the time slot associated with said segment and thus a delay not exceeding half a segment is incurred. By making the segments have a duration of 2 ms then the method complies with PSTN requirements.
|
