Algorithm Selection for Short-Delay Self-Heterodyne Laser Linewidth Measurement Under Varying Signal-to-Noise Ratios

To evaluate the impact of signal-to-noise ratio (SNR) on short-delay self-heterodyne laser linewidth measurement, we systematically compare the peak–valley method (PVM) and the power spectral equalization (PSE) method across a 20–80 dB SNR range through time-domain Monte Carlo simulations. A theoretical model incorporating white frequency noise and 1/f flicker noise is established, with 100 independent trials per SNR point. The results reveal complementary SNR dependence. PVM remains the relatively preferable method below 56 dB owing to its reliance on macroscopic spectral envelope features, although both methods exhibit large errors and frequent flagged outputs over much of this regime, while PSE achieves higher accuracy above 60 dB by accurately recovering the white noise floor from the demodulated phase noise spectrum. Above 70 dB, both methods converge to the true linewidth. The effects of fiber length, 1/f noise coefficient, and true linewidth on the critical SNR (defined as the PSE–PVM error crossover) are systematically investigated. Increasing the true linewidth from 30 to 90 kHz lowers the PVM reliability threshold to approximately 40 dB and eliminates the crossover, whereas at 1 kHz neither method remains viable over the scanned range. Based on these findings, an algorithm selection strategy with 56–60 dB as the transition band is proposed, providing a practical criterion for linewidth measurement system optimization.