Quantization Noise and Dynamic Range Calculator
Find an ADC's quantization step size, noise floor, and dynamic range from its bit depth and full-scale voltage range.
📏 What is Quantization Noise and Dynamic Range?
Quantization noise is the small error introduced every time an analog-to-digital converter (ADC) rounds a continuous voltage to the nearest available digital level. Because an ADC can only represent a finite number of discrete steps, every sample is rounded by up to half a quantization step, and this rounding error behaves like a small amount of added noise. Dynamic range is the ratio, in decibels, between the largest signal a converter can represent (full scale) and this quantization noise floor.
Audio engineers use dynamic range to judge how quiet a recording ADC's noise floor is relative to its loudest possible signal, directly affecting how much a mix can be turned down before noise becomes audible. Instrumentation engineers use it to decide whether a data acquisition system can resolve a small sensor signal riding on top of a much larger full-scale range. RF engineers budget dynamic range across an entire receiver chain, since the ADC is frequently the weakest link limiting how quiet a signal can be detected.
A common misconception is that a higher full-scale voltage range improves dynamic range. It does not. Because both the signal ceiling and the quantization noise floor scale up together with full-scale range, their ratio, and therefore dynamic range in dB, depends only on the ADC's bit depth, not on the voltage range chosen.
This calculator computes the quantization step size (LSB), the quantization noise RMS voltage, and the resulting dynamic range in dB from a converter's bit depth and full-scale voltage range.
📐 Formula
📖 How to Use This Calculator
Steps
💡 Example Calculations
Example 1 — 12-Bit Industrial ADC
12-bit ADC, 5 V full-scale range
Example 2 — 16-Bit Audio ADC
16-bit ADC, 2 V full-scale range
Example 3 — 24-Bit High-Resolution Data Acquisition
24-bit ADC, 5 V full-scale range
❓ Frequently Asked Questions
🔗 Related Calculators
How do you calculate quantization noise?
Quantization noise (RMS) equals the quantization step size (LSB) divided by the square root of 12: noise_rms = LSB / sqrt(12). This comes from modeling quantization error as uniformly distributed across one LSB step, whose RMS value is step_size / sqrt(12) for a uniform distribution.
What is the quantization step size (LSB)?
The quantization step, or least significant bit (LSB) value, is the full-scale voltage range divided by the number of quantization levels: LSB = V_FS / 2^N, where N is the ADC's bit depth. It represents the smallest voltage change the converter can distinguish.
How do you calculate dynamic range from bit depth?
Dynamic range in dB equals 20 x log10(full-scale range / quantization noise RMS), which simplifies to the closed form DR = 6.0206 x N + 10.7918 dB, where N is the bit depth. Notice the full-scale range cancels out entirely, so dynamic range depends only on bit depth.
Why doesn't dynamic range depend on the full-scale voltage?
Both the full-scale ceiling and the quantization noise floor scale up or down together by the same factor when full-scale range changes, since the noise floor is a fixed fraction (LSB/sqrt(12)) of the full-scale range divided by 2^N. Their ratio, which defines dynamic range, is therefore unaffected by the absolute voltage scale.
What is the difference between dynamic range and SNR?
Dynamic range compares the full-scale ceiling to the noise floor (DR = 6.0206N + 10.7918 dB). SNR, as typically quoted for ADCs, compares a full-scale SINE WAVE's RMS amplitude (which is lower than full-scale by a factor of 2*sqrt(2)) to the same noise floor (SNR = 6.02N + 1.76 dB). Dynamic range is always the larger of the two numbers for the same bit depth.
Why is quantization noise modeled as uniformly distributed?
For a busy, non-periodic signal that spans many quantization levels, the rounding error introduced by each sample lands anywhere within plus or minus half an LSB with roughly equal probability, matching a uniform distribution. This is the standard assumption used to derive the LSB/sqrt(12) RMS noise formula.
How many quantization levels does an N-bit ADC have?
An N-bit ADC has exactly 2^N distinct quantization levels. A 12-bit ADC has 4,096 levels, a 16-bit ADC has 65,536 levels, and a 24-bit ADC has 16,777,216 levels, each dividing the full-scale voltage range into that many equal steps.
Does a smaller full-scale range improve resolution?
A smaller full-scale range gives a smaller absolute LSB voltage (finer resolution in volts) for the same bit depth, but the dynamic range in dB stays exactly the same, since both the signal ceiling and noise floor shrink proportionally. Matching the full-scale range tightly to your actual signal amplitude is still good practice, since it uses all available quantization levels.
What real-world factors make dynamic range worse than the ideal formula predicts?
Thermal (Johnson) noise in the analog front end, reference voltage noise and drift, clock jitter, and converter nonlinearity all add noise beyond pure quantization error. A datasheet's measured dynamic range or ENOB is always the more honest real-world number compared to this calculator's ideal theoretical ceiling.
How much does one extra bit improve dynamic range?
Each additional bit adds approximately 6.02 dB of dynamic range (the coefficient of N in the DR = 6.0206N + 10.7918 formula), because doubling the number of quantization levels halves the quantization step and roughly quadruples the signal-to-noise power ratio.