Signal-to-noise ratio (SNR) in the output of a zero-memory nonlinearity is defined and found when the input is a complex Gaussian process. The results are applied to an odd

th law device and to an analog-to-digital (A/D) converter exhibiting both saturation and quantization.