This contribution deals with phase noise estimation from pilot symbols. The phase noise process is approximated by an expansion of DCT basis functions containing only a few terms. We propose an algorithm that estimates the DCT coefficients without requiring detailed knowledge about the phase noise statistics. We demonstrate that the resulting (linearized) mean-square estimation error consists of two contributions: a contribution from the additive noise, that equals the Cramer-Rao lower bound, and a noise-independent contribution that results from the phase noise modeling error. Performance can be optimized by a proper selection of the symbol block length and of the number of DCT coefficients to be estimated. For large block sizes, considerable performance improvement is found as compared to the case where only the time-average of the carrier phase is estimated.