fir filter buffer issue in C

[hkmixxa]

Hi I've been trying to write a function to perform fir filtering with a buffer in C and am having trouble with the output. My filter coefficients were generated in MATLAB with 29 taps and I'm trying to multiply and accumulate with an input signal of 1400 samples. After the 29th calculation, the results are inaccurate and I'm not sure how to alleivate the issue.

My programming skills in C are very poor and any assistance would be greatly appreciated. Below is my source code for the function:

*Note iIndex is the # of rows in the signal = 1400
float filterdata(int iIndex, float signal[], float filtercoeffs[])
{
int m;
float accum;

/* calc FIR */
accum = 0;
for (m = 0; m < iIndex;m++) {
accum += filtercoeffs[m] * signal[m];
}

/* shift delay line */
for (m = iIndex - 2; m >= 0; m--) {
signal[m + 1] = signal[m];
}
return accum;
}

[weaknessforcats]

What do you mean by inaccurate?? A float has only 6 significant digits so you may be encountering round-off error.

Also, this code:

accum += filtercoeffs[m] * signal[m];

assumes there are 1400 filtercoeffs. Is that true?? If not, your are suffering from an array out-of-bounds error.