We demonstrate a simple machine-learning-based method to estimate the signal-to-noise ratio of an optical signal based solely on phase noise metrics that are readily extracted from a conventional carrier phase recovery algorithm. The method is then expanded to other impairments as well as BER estimation.
