In this paper we describe a speaker normalization method based on spectral shifts in the auditory filter domain. This method is characterized by using an estimated vocal tract length as a criterion to determine the spectral shift value. Certain constraints are found to be necessary for the shift in the auditory filter domain, and two techniques based on these constraints, the One-Bark shift and the refined Bark-scale shift, are introduced. When tested in vowel classification experiments, significant performance improvement was obtained for both techniques. We maintain that the method is useful for speaker normalization in speaker-independent speech recognition.