A post-processing technique that corrects phase misalignment in synthesized speech to reduce artifacts like buzzing.