The evaluation of the signal frequency and Rate of Change of Frequency (RoCoF) from voltage or current waveforms is used for critical grid control, monitoring and protection applications. However, when step changes in the amplitude or phase of the signal occur, conventional frequency and RoCoF estimation methods, typically based on phasor models, are highly unreliable and can yield large frequency errors. To address this issue, this paper proposes a technique that uses dictionaries based on common signal dynamic models to identify and track amplitude and/or phase steps in AC signals. Additional signal dynamics including amplitude modulations, frequency ramps and harmonic tones are also characterized. Distinct from a previous iteration of this method developed by the authors, the proposed Step Change Detection (SCD) technique separates the envelope and angle of the signal's analytic form for independent analysis of these components. For numerical validation the method is applied to synthetically generated signals with challenging combinations of signal dynamics. The technique is shown to greatly improve frequency and RoCoF approximations as compared to state-of-the-art phasor estimation methods.