1
$\begingroup$

I am currently trying to perform an experiment given an audio signal. I am therefore sampling the audio frequency range (0 to 22kHz) with a 48k sampling rate.

When performing a FFT on my signal, I had the following question :

  • Is it equivalent to have a very small FFT size (for example 256), producing a 172 Hz resolution on my signal, in comparison with having a large FFT size (in a way more bars..) and performing a mean on my resulting frequencies in order to bring the number of bars down?
$\endgroup$
  • $\begingroup$ The bin width of the fft is Sampling Frequency/number of samples (Fs/N). Larger fft will help you resolve more. $\endgroup$ – Vaibhav Garg Feb 28 at 9:17
1
$\begingroup$

The two are not equivalent. Taking a larger FFT means you are using more time domain samples (unless you zero pad), in which case you have introduced new information compared to a smaller FFT size

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.