We study vibrato as the more ubiquitous manifestation of a non-stationary tone that can evoke a single overall pitch. Some recent results using nonsymmetrical vibrato tones suggest that the perceived pitch could be governed by some stability-sensitive mechanism. For non-stationary sounds the adequate tools are time–frequency representations (TFRs). We show that a recently proposed TFR could be the simplest framework to explain this hypothetical stability-sensitive mechanism. We propose a one-parameter model within this framework that is able to predict previously reported results and we present new results obtained from psychophysical experiments performed in our laboratory.