MELP算法1.2kbps,800bps和600bps的权威论文，共四篇，很有参考价值。-MELP algorithm, 1.2 KBPS，800bps and 600bps authority thesis, altogether four papers, very valuable.

2.训练结果 最终误差：0 迭代次数：8 最终权值：2.6416 1.3022 最终阈值：3.1000 误差收敛曲线 -ganzhiqi

function exp85 clear all p=[0:0.1:1.1] t=[22.4570 26.6012 12.6416 5.9367 6.9265 28.2432 31.5068 37.0166 7.8947 1.0398 12.7095] net=newff([0 1],[5 1],{ tansig purelin }, traingdx , learngdm ) net.trainParam.epochs=2500 net.trai

code c DSP 6416(TMSC6416) for CCS3