Cepstral normalisation in automatic speech recognition is investigated in the context of robustness to additive noise. It is argued that such normalisation leads naturally to a speech feature based on signal to noise ratio rather than absolute energy (or power). Explicit calculation of this {\em SNR-cepstrum} by means of a noise estimate is shown to have theoretical and practical advantages over the usual (energy based) cepstrum. The SNR-cepstrum is shown to be almost identical to the articulation index known in psycho-acoustics. Combination of the SNR-cepstrum with the well known perceptual linear prediction method is shown to be beneficial in noisy environments.