Simple IIR or FIR f i l m have bem widely used in isolated orword recognition tasks to filter the time sequence of speech spectral parametas, since, despite their simplicity, they significantly improve recognition performance. Those f i l m , when applied to continuous speech recognition, when phonema sized modelling units w used, induce spectral transition spnading and a cross-boundary effect In this work, we show how the use of contextdependent units reduces the side effects of the filters and may result in improved recognition performance. When dynamic parameten are not used. filtering scans to be especially useful, even for clean speech, and when they are, f i l m do well under unmatched training and testing conditions.