Envelope Following via Cascaded Exponential Smoothers For Low-Distortion Peak Limiting and Maximisation
Dario Sanfilippo · Zenodo (CERN European Organization for Nuclear Research) · 2022
In this paper, we discuss the implementation of look-ahead peak limiters using exponential envelope followers and their deployment in multi-band loudness maximisation. Particularly, the limiter’s amplitude profiling algorithm uses cascaded one-pole filters to improve derivative continuity and reduce convolution artefacts when multiplying the input signal by the attenuation gain as shown in a total harmonic distortion analysis. The maximiser algorithm allows for two independent maximisation modalities that can be combined: an RMS-based dynamic normalisation of each band with controllable depth to reach equal power, and a global maximisation gain that boosts the individual bands by the same amount. An open-source FAUST implementation of these algorithms is available on Github at the links: https://github.com/dariosanfilippo/limiterStereo and https://github.com/dariosanfilippo/maximiser.