Linear convolution

Hi,

Thanks for your post.

The above mentioned DSP library routine does the full length convolution of real vectors x(n) and h(n) which is nothing but the linear convolution of x(n) with h(n) in time domain and typically the input data (x) are padded with consecutive zeros at the beginning and end, so obviously the inputs arrays x and h are required to be aligned on a double-word boundary and accordingly the ouput will also be aligned which would automatically pad the zeros.

Please check the algorithm involved:

void DSPF_sp_convol(float *x, float *h, float *r, short nh,

short nr)

{

short ocntr, icntr;

float acc ;

for (ocntr = nr ; ocntr > 0 ; ocntr−−)

{

acc = 0 ;

for (icntr = nh ; icntr > 0 ; icntr−−)

{

acc += x[nr−ocntr+nh−icntr]*h[(icntr−1)];

}

r[nr−ocntr] = acc;

}

}

Thanks & regards,

Sivaraj K

-------------------------------------------------------------------------------------------------------

Please click the Verify Answer button on this post if it answers your question

-------------------------------------------------------------------------------------------------------