UM0585

16-bit IIR filters

Doc ID 14988 Rev 2

17/25

To use the iirarma_stm32 function, the following conditions must be satisfied:

●

The moving-average coefficient vector, h2, must have a length of 5 shorts:

–

short h2[5];

●

The auto-regressive coefficient vector, h1, must have a length of 5 shorts. The h1[0]

value is not used, so 4 coefficients remain:

–

short h1[5];

●

The number of output samples, ny, must be a multiple of 4 greater than or equal to 8.

●

Input and output vectors must have a length of (ny+4) shorts.

●

The first four elements of the output vector must have the previous outputs.

Example

#define NY 32/*number of outputs, must be a multiple of 4 and >= 8*/

/* Coefficients for the ARMA IIR filter */

short h2[5] = { 0x09c2,

0x270a,

0x3a8f,

0x270a,

0x09c2 };

short h1[5] = { 0x7fff,

0xd24a,

0x72ca,

0xcf4e,

0x1ad4 };

/* Input and output vectors */

short x[NY+4],y[NY+4];

/* Fill the input vector x */

...

/* Fill the 4 previous outputs */

y[0] = y0;

y[1] = y1;

y[2] = y2;

y[3] = y3;

/* Improve the filtering of NY samples */

iirarma_stm32(y, x, h2, h1, NY);

5.2.2

iir_biquad_stm32

function

Table 10 describes the iir_biquad_stm32 function.

Table 10.

iir_biquad_stm32

function

Function name

iir_biquad_stm32

Prototype

void iir_biquad_stm32(uint16_t *y, uint16_t *x, int16_t

*IIRCoeff, uint16_t ny)

Behavior description

Input parameter

– y: output array

– x: input array

– IIRCoeff: IIR filter coefficients

– ny: number of output samples

Output parameter

None

Return parameter

None