The implementation and optimization of AMR speech codec on DSP

The adaptive multi-rate (AMR) is the mandatory speech codec for GSM system, ranging from 12.2 kbps down to 4.75 kbps. To satisfy the requirement of need for both audio and video simultaneously, a kind of hardware and software of AMR speech encoder based on high-performance chips TMS320C64X DSP is proposed. Considering the limited resource of DSP chips, the optimization scheme is referred to later. The result shows that the CPU-load is reduced from 38 percent to 4 percent. And the time of processing each frame is also reduced greatly after optimizing. It is verified with all the test vectors provided by 3 GPP, and the stable operation on the real-time testing board was also confirmed. So, more resource is reserved for video coding on DSP chips. This is helpful for implementing both speech coder and video coder on one DSP chip.