zoukankan      html  css  js  c++  java
  • Digital biquad filter

    Direct Form 1

    The most straightforward implementation is the Direct Form 1, which has the following difference equation:

     y[n] = frac{1}{a_0} left ( b_0x[n] + b_1x[n-1] + b_2x[n-2] - a_1y[n-1] - a_2y[n-2] 
ight )

    or, if normalized:

     y[n] = b_0x[n] + b_1x[n-1] + b_2x[n-2] - a_1y[n-1] - a_2y[n-2]

    Here the b_0, b_1 and b_2 coefficients determine zeros, and a_1, a_2 determine the position of the poles.

    Flow graph of biquad filter in Direct Form 1:

    Biquad filter DF-I.svg

    Direct Form 2

    The Direct Form 1 implementation requires four delay registers. An equivalent circuit is the Direct Form 2 implementation, which requires only two delay registers:

    Biquad filter DF-II.svg

    The Direct Form 2 implementation is called the canonical form, because it uses the minimal amount of delays, adders and multipliers, yielding in the same transfer function as the Direct Form 1 implementation. The difference equations for DF2 are:

     y[n]=b_0 w[n]+b_1 w[n-1]+b_2 w[n-2],

    where

     w[n]=x[n]-a_1 w[n-1]-a_2 w[n-2].

     

    //ASM example

    /////////////////////////////////////////////////////////////
    //                                                                     //
    //     Process the audio stream                            //
    //                                                                    //
    /////////////////////////////////////////////////////////////

    #include <def21262.h>
    #define SECTIONS 3  
    /* Number of second-order sections (biquads) */

    .section /pm seg_pmco;
    .global _Cascaded_IIR_Filter_SIMD;
    .extern inbuf;
    .extern outbuf;
    .extern delaybuf;
    .extern coefficients;


    _Cascaded_IIR_Filter_SIMD:

    /*****************************************************************************
    The algorithm:

    IIR Second order sections - The cannonic second-order section implemented as
    "Direct Form II" biquads.  Note that the SIMD architecture of the 2126x SHARC
    family enables the two parallel execution units to filter the left and right
    channel simultaneously.  All register moves and memory reads implicitly apply
    to the shadow processing element (PEy) as well as the primary computational
    unit (PEx).
    *****************************************************************************

      Given the most general biquadratic (second order rational polynomial)

                    b0 + b1'*z^-1 + b2'*z^-2
            H(z) = -------------------------- ,
                    a0 + a1'*z^-1 + a2'*z^-2

      we may factor out the gain of the transfer function,

                    b0     (b1'/a0)*z^-1 + (b2'/a0)*z^-2
            H(z) = ---- * -------------------------------
                    a0     (a1'/b0)*z^-1 + (a2'/b0)*z^-2

      and normalize the coefficients, such that

                        a1*z^-1 + a2*z^-2
            H(z) = A * -------------------
                        b1*z^-1 + b2*z^-2

      where A = gain = b1'/a0

            a1 = a1'/b0,  a2 = a2'/b0,  b1 = b1'/a0,  b2 = b2'/a0

      This leaves only four true filter coefficients.  The gain values from
      all of the sections may be combined into a single channel gain applied
      apart from the inner computational loop.  With the simplified coefficients,
      the cannonic direct form II may be written as a pair of difference
      equations:

              w[n] = x[n] + a1*w[n-1] + a2*w[n-2]
              y[n] = w[n] + b1*w[n-1] + b2*w[n-2]

    which leads to the following pseudocode:

    read(x[n])
    f12=0,                                                                     f2=w[n-1],           read(a1)
    --- Loop --------------------------------------------------------------------
    f12=a1*w[n-1], f8=f8 + f12,                                      f3=w[n-2],           read(a2)
    f12=a2*w[n-2], f8=x[n] + a1*w[n-2],                        w[n-1] -> w[n-2]', read(b1)
    f12=b1*w[n-2], w[n]=x[n] + a1*w[n-2] + a2*w[n-1],  f2=w[n-1],           read(b2)
    f12=b2*w[n-1], f8=w[n] + b1*w[n-2],                        w[n] -> w[n-1]',   read(a1)
    -----------------------------------------------------------------------------
                    y[n]=f8 + f12

    **************************************************************************/

      /*  Subroutine that implements the pseudocode above */

    cascaded_biquad:
      bit set mode1 CBUFEN | PEYEN ;                    // Enable SIMD mode
       
      b0 = delaybuf;
      b1 = b0;
      b3 = inbuf;
      b4 = outbuf;
      b9 = coefficients;
      r0 = SECTIONS;

        f8=dm(i3,m1);                                          // read inbuf
        r12=r12 xor r12,  f2=dm(i0,m1),  f4=pm(i8,m8);
      lcntr=r0, do quads until lce;
              f12=f2*f4,      f8=f8+f12,      f3=dm(i0,m1),     f4=pm(i8,m8);
              f12=f3*f4,      f8=f8+f12,      dm(i1,m1)=f3,     f4=pm(i8,m8);
              f12=f2*f4,      f8=f8+f12,      f2=dm(i0,m1),     f4=pm(i8,m8);
    quads:f12=f3*f4,      f8=f8+f12,      dm(i1,m1)=f8,     f4=pm(i8,m8);

      f8=f8+f12;
      rts (db);
      dm(i4,m1)=f8;
      bit clr mode1 CBUFEN | PEYEN;               
      // disable SIMD mode

    _Cascaded_IIR_Filter_SIMD.end:


    //--------------------------------------------

     

  • 相关阅读:
    KLSudoku的数独题目生成方法和难度控制说明
    对XChain和ForcingChain的实现解说
    开源数独游戏软件KLSudoku发布第一个Release版本
    每个 Vim 用户都应该阅读的文章
    自己常用的几个gvim的vimrc设置
    KLSudoku数独游戏软件1.1预览版发布
    KLSudoku数独游戏软件1.1正式版发布
    字符串
    .NET面试大全
    IIS是如何处理ASP.NET请求的
  • 原文地址:https://www.cnblogs.com/ldjrl2013/p/3910368.html
Copyright © 2011-2022 走看看