摘要
文章针对10阶LSF参数求解,提出一种基于牛顿迭代法的有效算法,并通过实验比较了该算法和经典的Kabal搜索算法的运算量、精度。该算法避免了多项式降次,计算量又远远小于经典的搜索算法,适于用大规模集成电路实现,而且可以推广应用于求解更高阶LSF参数。
Focusing on the characteristics of 10th-order LSF （line spectral frequencies）, this paper proposed an efficient method based on Newton Raphson. Experiment and analysis was given to compare the precision and computational quantity between our method and the traditional Kabal＇s method. Our algorithm not only avoids polynomial deflation but also is more efficient than traditional rootfinding algorithm and it is fit for being realized by VLSI. Furthermore, our method could be deduced to high-order LSF application.
出处
《微电子学与计算机》
CSCD
北大核心
2006年第4期149-152,共4页
Microelectronics & Computer
