Is it better to use a larger FFT or to decimate when trying to resolve the frequency between two closely spaced tones?
Increasing the frequency bin resolution by using a larger FFT size, by decimating with the down mixer, and by using a lower sample rate are all equivalent operations. They can even be combined for even more signal extraction gain. The differences between methods are: • A larger FFT size is not as efficient and requires more CPU power but it has the most flexibility. • Decimating with the down mixer reduces the sample rate but it adds filter skirts (roll-off) to the left and right of the spectrum. • A lower sample rate reduces the sample rate by doing the resampling with the audio cards ADC. This doesn’t utilize a CPU performance hit but it can create a serious artifact problem with some sound cards. The current bin resolution can be found in the Drift Integrators pop down section.
