Quantization noise is proportional to which of the following?

• A. Temperature
• B. Step size of quantization; higher resolution reduces error
• C. Input signal amplitude
• D. Sampling rate

Answer: (B)

Quantization error comes from representing a continuous signal with a finite set of levels, so its size is governed by how large the steps between those levels are. In a uniform quantizer, this step size Δ sets the maximum possible error (roughly ±Δ/2) and the average noise power scales with Δ^2. That means, as you increase resolution and make the steps smaller, the quantization noise drops accordingly. Temperature, input signal amplitude, and sampling rate don’t determine the fundamental size of this error in the basic model: temperature and other physical noise sources can add extra noise, amplitude affects whether you stay within the available dynamic range, and sampling rate doesn’t change the per-sample quantization error (though techniques like oversampling can reshape how noise is distributed). So, the best description is that quantization noise is tied to the step size of quantization, and higher resolution reduces the error.