FFT Bin Resolution Calculator

Bin width, frequency resolution and capture time for a given FFT size and sample rate.

Example: A 1024-point FFT at 48 kHz gives 46.88 Hz bins from 21.33 ms of signal, across 512 usable bins up to 24 kHz.

Formula

bin width = f_s / N    capture time = N / f_s = 1 / bin width

Worked example

A 1024-point FFT at 48 kHz gives 46.88 Hz bins from 21.33 ms of signal, across 512 usable bins up to 24 kHz.

  1. bin width = f_s / N

    48 kHz / 1024

    46.88 Hz

  2. capture time = N / f_s

    1024 / 48 kHz

    21.33 ms

    Which is exactly 1 / bin width — resolution and observation time are the same constraint.

  3. effective resolution = bin width × ENBW

    46.88 Hz × 1.5

    70.31 Hz

Frequently asked questions

How do I improve FFT frequency resolution?

Capture for longer. Bin width is the sample rate divided by the FFT size, which equals one over the capture time — so resolving 1 Hz requires a full second of signal regardless of sample rate. A larger FFT at the same sample rate is exactly a longer capture.

Does sampling faster improve resolution?

No — it makes it worse for a fixed FFT size, because the same number of bins now spans a wider frequency range. Sampling faster extends the frequency range you can see, not the detail within it.

Why does my tone read lower than it should?

Scallop loss. Unless the signal lands exactly on a bin centre, its energy splits between neighbouring bins and the peak reads low — up to 3.9 dB with no window. A flat-top window reduces this to essentially zero, at the cost of much wider bins.

Which window should I use?

Hann for general work. Flat top when you need accurate amplitude readings. Rectangular only when the signal is exactly periodic in the capture window, which in practice means a synthesised test signal rather than anything measured.

Related tools