Sign in

Not a member? | Forgot your Password?

Search code

Search tips

Free PDF Downloads

A Quadrature Signals Tutorial: Complex, But Not Complicated

Understanding the 'Phasing Method' of Single Sideband Demodulation

Complex Digital Signal Processing in Telecommunications

Introduction to Sound Processing

C++ Tutorial

Introduction of C Programming for DSP Applications

Fixed-Point Arithmetic: An Introduction

Cascaded Integrator-Comb (CIC) Filter Introduction

IIR Filter Design Software

See Also

Embedded SystemsFPGA

DSP Code Sharing > Fixed Point FIR Filter

Fixed Point FIR Filter

Language: C

Processor: Not Relevant

Submitted by Shawn Stevenson on Dec 3 2010

Licensed under a Creative Commons Attribution 3.0 Unported License

Fixed Point FIR Filter


 

This code implements a fixed point FIR filter using fixed point arithmetic. The data and coefficients are stored in 16 bits as Q15 numbers (2's complement signed numbers with 15 fractional bits). The code is designed so that several different FIR filters can be applied to the same input data. The code is quite general, so there is plenty of room for optimization for specific cases. (Coding in assembly language is often best if speed is a concern.)

There are four functions defined for using the filter. The firFixedInit function is run once at startup to clear out the input data array. The firStoreNewSamples function is called each time there is a new set of input samples to process. The number of samples can be anything between 1 and MAX_INPUT_LEN. The function returns a pointer to the sample at which the filter should be applied. The firFixed function performs the filtering operation, and produces new samples in an output array. It should be called once for every filter that is to be applied. And finally the firMoveProcSamples function shifts the input samples to prepare for the next block of input samples. It should be called after all of the filter operations have been completed.

The code will saturate on overflow for large inputs, provided that the passband gain of the filter is close to one (can be a bit larger than one, as in my test program).

 

 

 
#include <stdio.h>
#include <stdint.h>

//////////////////////////////////////////////////////////////
//  Filter Code Definitions
//////////////////////////////////////////////////////////////

// maximum number of inputs that can be handled
// in one function call
#define MAX_INPUT_LEN   80
// maximum length of filter than can be handled
#define MAX_FLT_LEN     63
// buffer to hold all of the input samples
#define BUFFER_LEN      (MAX_FLT_LEN - 1 + MAX_INPUT_LEN)

// array to hold input samples
static int16_t insamp[ BUFFER_LEN ];

// FIR init
void firFixedInit( void )
{
    memset( insamp, 0, sizeof( insamp ) );
}

// store new input samples
int16_t *firStoreNewSamples( int16_t *inp, int length )
{
    // put the new samples at the high end of the buffer
    memcpy( &insamp[MAX_FLT_LEN - 1], inp,
            length * sizeof(int16_t) );
    // return the location at which to apply the filtering
    return &insamp[MAX_FLT_LEN - 1];
}

// move processed samples
void firMoveProcSamples( int length )
{
    // shift input samples back in time for next time
    memmove( &insamp[0], &insamp[length],
            (MAX_FLT_LEN - 1) * sizeof(int16_t) );
}

// the FIR filter function
void firFixed( int16_t *coeffs, int16_t *input, int16_t *output,
       int length, int filterLength )
{
    int32_t acc;     // accumulator for MACs
    int16_t *coeffp; // pointer to coefficients
    int16_t *inputp; // pointer to input samples
    int n;
    int k;

    // apply the filter to each input sample
    for ( n = 0; n < length; n++ ) {
        // calculate output n
        coeffp = coeffs;
        inputp = &input[n];
        // load rounding constant
        acc = 1 << 14;
        // perform the multiply-accumulate
        for ( k = 0; k < filterLength; k++ ) {
            acc += (int32_t)(*coeffp++) * (int32_t)(*inputp--);
        }
        // saturate the result
        if ( acc > 0x3fffffff ) {
            acc = 0x3fffffff;
        } else if ( acc < -0x40000000 ) {
            acc = -0x40000000;
        }
        // convert from Q30 to Q15
        output[n] = (int16_t)(acc >> 15);
    }
}

//////////////////////////////////////////////////////////////
//  Test program
//////////////////////////////////////////////////////////////

// bandpass filter centred around 1000 Hz
// sampling rate = 8000 Hz
// gain at 1000 Hz is about 1.13

#define FILTER_LEN  63
int16_t coeffs[ FILTER_LEN ] =
{
 -1468, 1058,   594,   287,    186,  284,   485,   613,
   495,   90,  -435,  -762,   -615,   21,   821,  1269,
   982,    9, -1132, -1721,  -1296,    1,  1445,  2136,
  1570,    0, -1666, -2413,  -1735,   -2,  1770,  2512,
  1770,   -2, -1735, -2413,  -1666,    0,  1570,  2136,
  1445,    1, -1296, -1721,  -1132,    9,   982,  1269,
   821,   21,  -615,  -762,   -435,   90,   495,   613,
   485,  284,   186,   287,    594, 1058, -1468
};

// Moving average (lowpass) filter of length 8
// There is a null in the spectrum at 1000 Hz

#define FILTER_LEN_MA   8
int16_t coeffsMa[ FILTER_LEN_MA ] =
{
    32768/8, 32768/8, 32768/8, 32768/8,
    32768/8, 32768/8, 32768/8, 32768/8
};

// number of samples to read per loop
#define SAMPLES   80

int main( void )
{
    int size;
    int16_t input[SAMPLES];
    int16_t output[SAMPLES];
    int16_t *inp;
    FILE   *in_fid;
    FILE   *out_fid;
    FILE   *out_fid2;

    // open the input waveform file
    in_fid = fopen( "input.pcm", "rb" );
    if ( in_fid == 0 ) {
        printf("couldn't open input.pcm");
        return;
    }

    // open the output waveform files
    out_fid = fopen( "outputFixed.pcm", "wb" );
    if ( out_fid == 0 ) {
        printf("couldn't open outputFixed.pcm");
        return;
    }
    out_fid2 = fopen( "outputFixedMa.pcm", "wb" );
    if ( out_fid == 0 ) {
        printf("couldn't open outputFixedMa.pcm");
        return;
    }

    // initialize the filter
    firFixedInit();

    // process all of the samples
    do {
        // read samples from file
        size = fread( input, sizeof(int16_t), SAMPLES, in_fid );
        // store new samples in working array
        inp = firStoreNewSamples( input, size );

        // apply each filter
        firFixed( coeffs, inp, output, size, FILTER_LEN );
        fwrite( output, sizeof(int16_t), size, out_fid );

        firFixed( coeffsMa, inp, output, size, FILTER_LEN_MA );
        fwrite( output, sizeof(int16_t), size, out_fid2 );

        // move processed samples
        firMoveProcSamples( size );
    } while ( size != 0 );

    fclose( in_fid );
    fclose( out_fid );
    fclose( out_fid2 );

    return 0;
}

 
 
Rate this code snippet:
4
Rating: 4 | Votes: 1
 
   
 
posted by Shawn Stevenson
Shawn Stevenson is a DSP software engineer based in Vancouver, BC. He has over a decade of experience implementing signal processing algorithms for computer telephony and VoIP at HotHaus Technologies and Broadcom. His specialty is the efficient implementation of fixed point algorithms for telephony, such as echo cancellers and speech compression codecs. He is currently working at a Vancouver start-up called Malaspina Labs, where he is working on the implementation of speech algorithms for digital hearing aids. Shawn has BASc and MEng degrees from Simon Fraser University. He has posted some tutorials on DSP at http://sestevenson.wordpress.com/


Comments


No comments yet for this code


Add a Comment
You need to login before you can post a comment (best way to prevent spam). ( Not a member? )