ToolJoltTools

Quantization Noise Calculator

q/√12 noise, SQNR and the effective bits gained by oversampling — with the dither caveat the formula hides.

Quantisation noise (RMS)
LSB size q
SQNR (FS sine)
Oversampling gain
Effective resolution with OSR
e(rms) = q/√12 ; SQNR = 6.02n+1.76 dB ; ΔSNR = 10log₁₀(OSR)
References: Bennett (1948), Spectra of Quantized Signals, BSTJ · Kester MT-001/MT-022 (quantisation noise, oversampling)

The q/√12 model assumes the error is uniform and uncorrelated — true for busy signals, FALSE for small periodic ones, where quantisation becomes harmonic distortion instead of noise (cue dither: add ~1 LSB of noise on purpose). Oversampling only helps if real thermal noise exceeds 1 LSB to randomise the error. Σ-Δ converters push this idea to the limit with noise shaping.

Quantization Noise Calculator computes the noise floor an ideal converter's step size creates, and the oversampling remedy — free, instant and private in your browser. Mixed-signal designers squeezing resolution from modest ADCs use it to skip the datasheet algebra: type your numbers, read the answer with the substituted formula shown step by step, and share an exact permalink of the calculation.

About Quantization Noise Calculator

Quantization Noise Calculator computes the noise floor an ideal converter's step size creates, and the oversampling remedy using the standard engineering relation: e(rms) = q/√12; SQNR = 6.02n + 1.76 dB; +10log₁₀(OSR) dB from oversampling+filtering. Worked live: a 10-bit, 3.3 V converter has a 930 µV rms quantisation floor; 16× oversampling buys 12 dB ≈ 2 extra bits. The result recalculates on every keystroke, the worked-example panel shows your numbers substituted into the formula, and the Copy permalink button encodes the inputs in the URL so a colleague opens exactly your calculation. Everything runs client-side — nothing you type leaves your device.

How to use Quantization Noise Calculator

  1. 1Enter your values — the tool starts with realistic defaults for this exact use case, so the worked example is meaningful immediately.
  2. 2Read the live result and the worked-example panel, which substitutes your numbers into the formula step by step.
  3. 3Adjust any input to compare scenarios, then use Copy result or Copy permalink to share the calculation.

Why use Quantization Noise Calculator?

  • Implements the real formula — e(rms) = q/√12 — with the substitution shown, not a black box
  • Built for mixed-signal designers squeezing resolution from modest ADCs
  • Copy result and permalink buttons — share the exact calculation in a README, forum answer or design review
  • 100% free, no sign-up, runs entirely in your browser (works offline once loaded)

Frequently asked questions

How do you calculate quantization noise?+

The noise floor an ideal converter's step size creates, and the oversampling remedy follows e(rms) = q/√12; SQNR = 6.02n + 1.76 dB; +10log₁₀(OSR) dB from oversampling+filtering. For example, a 10-bit, 3.3 V converter has a 930 µV rms quantisation floor; 16× oversampling buys 12 dB ≈ 2 extra bits. The calculator applies the same relation and shows the substituted arithmetic so you can verify every step.

Does oversampling always add effective bits?+

Only when ≥ ~1 LSB of genuine noise (thermal or added dither) randomises the quantisation error; otherwise you average the same wrong code forever. With that condition met, 4× oversampling + decimation = +1 bit, 256× = +4 bits — the principle Σ-Δ converters industrialise.

When is quantisation 'noise' actually distortion?+

For small, periodic signals: the error repeats with the signal and shows up as harmonics, not a flat floor — audibly ugly in audio fade-outs. One LSB of dither decorrelates it back into benign hiss; every serious audio chain dithers before truncating bits.

Is the Quantization Noise Calculator free and private?+

Yes — completely free with no sign-up or usage limits, and it runs entirely in your browser: the values you enter are never uploaded or stored on a server.

Related tools

Related Electronics tools

Sponsored