The high-speed LSF transformation algorithm in CELP vocoder

We propose a computation reduction method of the real root method that is mainly used in the CELP (code excited linear prediction) vocoder. The real root method is that if polynomial equations have real roots, we are able to find those and transform them into the LSF. However, this method takes much time to compute, because the root searching is processed sequentially in the frequency region. But, the important characteristic of the LSF is that most of the coefficients occur in a specific frequency region. So, the searching frequency region is ordered by each coefficient´s distribution, and the coefficients are searched in the ordered frequency region. The transformation time can be reduced further by this method than by the sequential searching method in the frequency region. When we compare this proposed method with the conventional real root method, the experimental result is that the searching time was reduced by about 46% in average