What is quantization noise?

Quantization noise is the error introduced when a continuous signal is rounded to the nearest discrete level during analog-to-digital conversion. For a uniform quantizer the error is modeled as white noise uniformly distributed in the range [−LSB/2, +LSB/2].

SQNR (signal-to-quantization-noise ratio) is the ratio of signal power to quantization noise power, expressed in dB. For a full-scale sinusoid with an N-bit converter the theoretical SQNR is approximately 6.02N + 1.76 dB.

ENOB (effective number of bits) converts a measured or calculated SQNR back into equivalent bit depth: ENOB = (SQNR_dB − 1.76) / 6.02. A real converter always has ENOB slightly below its nominal bit depth due to thermal noise and distortion.

Why does signal amplitude affect SQNR?

SQNR rises as signal amplitude increases because signal power grows while quantization noise power stays constant. A signal at 50% of full scale loses roughly 6 dB of SQNR compared to a full-scale signal.

Does waveform type matter?

Yes. A sinusoid has a crest factor of √2 and a square wave has a crest factor of 1; a sawtooth has the highest crest factor of the three (√3 ≈ 1.73), above the sinusoid (√2 ≈ 1.41) and square wave (1). The waveform type changes the signal power calculation and therefore the SQNR.

What are the units of quantization noise power here?

Noise power is expressed in V² (mean-square voltage). You can convert to dBV² or dBW by taking 10·log10 of the value.

Quantization Noise Calculator

Enter ADC bit depth and signal amplitude to compute quantization noise power, SQNR, and ENOB.

Converter Parameters

Bit Depth (N)

bits

Nominal resolution of the ADC/DAC (1–32 bits).

Full-Scale Range (V_FS)

V_pp

Peak-to-peak input range of the converter.

Signal Amplitude (fraction of full scale)

× V_FS

1.0 = full scale, 0.5 = half scale, etc.

Signal Waveform

Affects signal power and therefore SQNR.

Results

LSB Step Size (Δ)

V_FS / 2^N

—

Quantization Noise Power

Δ² / 12

—

V²

Noise RMS Voltage

√(Δ² / 12)

—

V_rms

SQNR

Signal power / noise power

—

ENOB

(SQNR − 1.76) / 6.02

—

bits

Theoretical Max SQNR (full scale)

6.02N + 1.76 dB

—

Common ADC Bit Depths — Theoretical SQNR

Bit Depth	Levels (2^N)	SQNR (dB)	Typical Application
8 bit	256	49.9 dB	Old PC audio, telephony
10 bit	1,024	62.0 dB	MCU ADCs, basic sensors
12 bit	4,096	74.0 dB	Industrial sensors, oscilloscopes
16 bit	65,536	98.1 dB	CD audio, pro audio recording
20 bit	1,048,576	122.2 dB	High-res audio, precision instruments
24 bit	16,777,216	146.2 dB	Studio recording, medical imaging

Summary

Enter ADC bit depth and signal amplitude to compute quantization noise power, SQNR, and ENOB.

How it works

Enter the converter's bit depth (resolution in bits, e.g. 16 for CD audio).
Select the signal waveform type (sinusoid, sawtooth, or square) to set the correct crest-factor correction.
Enter the full-scale range of the converter in volts (peak-to-peak).
Enter the actual signal amplitude as a fraction of full scale (0 to 1).
The calculator computes LSB step size, quantization noise power, SQNR, and ENOB.
Adjust any input and results update instantly without page reload.

Use cases

Verify ADC selection meets the noise floor budget for an audio or instrumentation design.
Compare 12-bit vs 16-bit vs 24-bit converters for a given SNR requirement.
Calculate the effective resolution loss when a signal does not use the full ADC range.
Teach or study the theoretical limits of PCM digital audio.
Estimate ENOB for an ideal converter before adding real-world noise sources.
Check whether dithering is needed to push noise below the quantization floor.

Frequently Asked Questions

Last updated: 2026-09-19 · Reviewed by Nham Vu