Comments on “A linear prediction method for parameter estimation of damped sinusoids”

For the parameter estimation of damped sinusoids, the linear prediction (LP) was recently applied to a forward–backward expanded-data matrix by Kannan and Kundu [Estimating parameters in the damped exponential model, Signal Processing 81 (2001) 2343–2351]. This version of LP will be called the expanded linear prediction (ELP). A key claim of that paper is that ELP offers a comparable performance as LP. The purpose of this letter is to prove that this claim does not hold true particularly for weakly damped sinusoids.