FFT Size & Bin Resolution Calculator
Calculate FFT bin width, window duration, and required size for a target frequency resolution at a given sample rate.
How to Use
- Enter sample rate and FFT size (or desired bin width).
- FFT size must be a power of 2 for typical radix-2 implementations.
- Bin width = fs / N. Smaller bins = finer resolution but longer window.
- Window duration = N / fs.
Worked Example
A 1024-point FFT of audio sampled at 48 kHz. Bin width = 48,000 ÷ 1024 = 46.875 Hz, and collecting the 1024 samples takes 1024 ÷ 48,000 = 21.3 ms.
Separating 50 Hz and 60 Hz hum (10 Hz apart) needs bins a few times narrower than 10 Hz once a Hann window widens each peak by about 1.5 bins. 16,384 points give 2.93 Hz bins, at the cost of a 341 ms window.
The common mistake: zero-padding to get resolution. Padding the 1024 samples out to 16,384 draws a smoother curve with 2.93 Hz spacing, but the two tones still merge, because resolving power comes from the 21.3 ms of real signal, not the padded length.
Show Work
Formulas
History of the FFT
The discrete Fourier transform (DFT) had been understood since Joseph Fourier's 1807 work, but naive computation is O(N²) — impractical for large N. James Cooley and John Tukey rediscovered and popularized a fast algorithm in their landmark 1965 paper "An Algorithm for the Machine Calculation of Complex Fourier Series", reducing complexity to O(N log N) — a factor of ~100× speedup at N = 1024, enough to make real-time digital signal processing practical.
The Cooley-Tukey algorithm's power-of-2 efficiency preference shaped decades of DSP hardware. Single-chip DSPs of the early 1980s, such as Texas Instruments' TMS32010, and most DSP families since include features aimed at the FFT, such as bit-reversed addressing and fast multiply-accumulate for its butterfly operations. Modern FFTW library (1997, Frigo and Johnson) achieves near-optimal performance across any FFT size via machine-tuned code paths.
Most modern signal processing — spectrum analyzers, real-time audio plugins, SDR waterfalls, vibration monitoring, MRI reconstruction — runs on FFT. The bin-width / window-duration / sample-rate trade-off in this calculator is identical to the one every DSP engineer has analyzed since 1965.
About This Calculator
Enter a sample rate and FFT size N, or specify a desired bin width and let the tool solve for the required N. Returned values: bin width (frequency spacing), window duration (time span), number of usable bins (N/2 + 1 for real inputs), and the nearest power-of-2 FFT size for radix-2 efficiency.
Remember that bin width is not the same as true frequency resolution. Spectral leakage from real-world windowing (Hann, Hamming, Blackman) smears each tone across 1.5–3 bins. Zero-padding adds interpolated bins but doesn't improve real resolution. Everything runs client-side; no values leave your browser.
Frequently Asked Questions
What is FFT bin width?
The frequency spacing of output bins. Bin width = fs / N. For 48 kHz sample rate with 1024-point FFT, bin width = 46.875 Hz. Signals within the same bin can't be distinguished.
Why power of 2?
Radix-2 FFT algorithms (Cooley-Tukey) are O(N log N) and run fastest when N is a power of 2. Modern libraries (FFTW, CMSIS-DSP) handle other sizes efficiently but power-of-2 remains most common.
Bin width vs actual resolution?
Spectral leakage (from windowing) spreads a single tone across multiple bins. Actual frequency resolution depends on window choice. Hann window: ~1.5× bin width. Blackman: about 1.7× bin width. Rectangular: 1× but with severe side-lobes.
How do I pick FFT size?
Trade off: larger N = better frequency resolution but longer collection window and more CPU. Real-time audio: 512-4096 typical. Precision scientific: 16k-1M. Limited by buffer RAM and latency tolerance.
What about zero-padding?
Zero-padding interpolates between bins (smoother plot) but doesn't improve actual resolution. The resolving power of an FFT is still set by the signal window length, not the total padded size.
How do I use the FFT Size & Bin Resolution Calculator?
Just type your numbers. The answer shows up right away — there is no button to press. Change anything and it updates by itself.
Does it cost anything or need an account?
No. The tool is completely free, there is no account to create, and it keeps working offline after the page first loads.
Is anything I type uploaded?
No. The tool works entirely on your device, so the values you enter never leave your browser.
Common Use Cases
Audio Spectrum Analyzer
1024-pt FFT at 48 kHz: 46.9 Hz bins, 21.3 ms window — typical real-time audio DSP.
Precision Frequency Measurement
65536-pt FFT at 100 kHz: 1.53 Hz bins. 655 ms window needed to resolve ±1 Hz tones.
Vibration Analysis
Machinery monitoring: 2048-pt FFT at 25.6 kHz gives 12.5 Hz bins up to ~12.8 kHz.
RF Spectrum
SDR with 10 MHz bandwidth and 4096-pt FFT: 2.4 kHz bins — fine enough for wideband surveys.
Musical Note Detection
To resolve adjacent piano notes near A440 (~26 Hz apart), need bins < 13 Hz. At 44.1 kHz, that's N ≥ 4096.
Last updated: