Adaptive Quantization Step

Problem

Choose the uniform quantization step that achieves a target SNR for a signal with the given variance. With target SNR_linear = 10^(target_snr_db / 10) and mean squared error step^2 / 12, solve for the step:

step = sqrt(12 * signal_var / SNR_linear)

Hint

Louder signals or lower targets allow coarser steps.

Test cases

case 1

in: [1,0]

expected: 3.4641016151377544

case 2

in: [1,10]

expected: 1.0954451150103321

case 3

in: [4,20]

expected: 0.6928203230275509

case 4

in: [9,6.02]

expected: 5.196511320715222

Starter codePython

import math
def adaptive_quant_step(signal_var, target_snr_db):
    # Your code here
    pass

4 test cases · run your code against them in the editor