In a previous paper (Varadi et al.), random-lag singular spectrum analysis
was introduced for finding oscillations in very noisy and long time series.
This work presents a generalization of the technique to search for common
oscillations in two or more time series.