A Corpus-driven Parametric Modal Reverberator

A parametric modal reverberator is presented in which synthesis pa- rameters are derived from a large, curated corpus of room impulse responses (IRs). The collected responses are subjected to modal de- composition, yielding per-mode frequencies, damping coefficients, and residue amplitudes, together with a short early-reflection finite impulse response (FIR) filter. From the decomposed data, a feature table is constructed per IR comprising standard acoustic indices, per-band damping and density statistics, amplitude distributions, and FIR descriptors—50 variables in total. Six acoustically mean- ingful user controls are selected; since these exhibit substantial pairwise correlations across the corpus, they are orthogonalised via principal component analysis (PCA) prior to regression. Per-band damping and modal density are predicted by robust linear models in log space; residue amplitudes follow the diffuse-field equipar- tition relation, and early-reflection energy is set directly from the clarity index definition. The resulting method maps a six-number perceptual specification onto thousands of modal parameters driv- ing a bank of second-order resonators, making modal reverberators easier to operate for use in audio engineering. Regression diag- nostics and corpus-distribution analysis confirm that the generated impulse responses are acoustically plausible and span the parameter space of the training data, which, in turn, should be sufficiently representative of room IRs used in audio engineering and music production.