Correlation of sine waves over non-integral number of cycles

[Mike Page]

Hi,

I know this isn't strictly a Scilab question, but I wonder if anyone out
there can help me?  I remember from somewhere that there is a formula for
finding the real and imaginary parts of a sine wave that doesn't have an
integral number of cycles over the set of samples.

Obviously with a whole number of cycles, I can just multiply by a sine and a
cosine and take the sums.  When there are not a whole number of cycles,
there is another more complicated formula (or maybe it is an
approximation?).

Does anyone know what it is, or where I can find it?  I seem to remember it
involved sums of sines squared and cosines squared and maybe sines times
cosines, but I can't remember the detail.

Thanks,
Mike.