Why does the "goertzel" function not accept arguments in [N, N + 1) as the FFT index for signal inputs of length N in MATLAB R2024a and earlier?

Why does the "goertzel" function not accept arguments in [N, N + 1) as the FFT index for signal inputs of length N in MATLAB R2024a and earlier?

Specifically, I get a "Maximum Index exceeds the dimensions of X." error when running the following code:

>> Fs = 8000;
>> N = 205;
>> lo = sin(2*pi*697*(0:N-1)/Fs);
>> hi = sin(2*pi*1209*(0:N-1)/Fs);
>> data = lo + hi;
>> goertzel(data,205.5)
Error using goertzel (line 143)
Maximum Index exceeds the dimensions of X.

Error in goertzel (line 100)
        Y1 = goertzel(reshape(X,coder.internal.indexInt(size(X,2)),ONE),INDVEC,1);
 

Answer 1

[MathWorks Support Team]

In MATLAB R2024a and earlier, the "goertzel" function only accepts arguments for the FFT index number over the interval [1,N] for a given signal input of length N.

Starting with MATLAB R2024b, the "goertzel" function now also accepts fractional inputs in [N, N + 1) as the frequency index. The workflow above should no longer error for versions of R2024b and later.