We propose a new method to mitigate so called “pre-echo artifacts” that occur with transform-based audio codecs owing to their comparatively large block lengths. Our algorithm operates entirely in the time domain. In contrast to previous time domain approaches, a frequency selective pre-echo control (PEC) is achieved by applying the concept of a filterbank equalizer (FBE) with non-uniform frequency resolution. As an example application, the new PEC approach has been used to enhance the ITU-T G.722.1C super-wideband codec. The performance of the algorithm has been objectively measured and optimized w.r.t. various FBE parameters. Considerable quality gains could be obtained which are confirmed by the subjective listening impression. The proposed algorithm can be easily extended to act as an adaptive spectro-temporal pre- and deemphasis filter.