Lossless Reveberator Prototypes

Reverberator design is often factored into separate designs of the early reflections and the late ``diffuse'' reverberation. Specific early reflections are easily realized using FIR filter taps. To develop a high quality late reverberation, it is good to begin with a lossless prototype reverberator so that the structure of its late time response can be clearly heard. Nominally, a lossless prototype reverberator is judged by the quality of white noise it generates in response to an impulse signal. For smooth late reverberation, the white noise should sound uniform in every respect. Subsequent introduction of sparse lowpass filters into the prototype network serves to set the desired reverberation time vs. frequency. In other words, starting with lossless networks allows the decoupling of reverberation time from structural aspects of the reverberator. Any ``passive'' arithmetic scheme, such as magnitude truncation, can be used at certain multiplier outputs to eliminate the possibility of limit cycles and overflow oscillations [27].

Since FDNs and DWNs appear to present very different approaches for constructing lossless reverberator prototypes, it is natural to ask what connections may exist between them, and whether there may be unique advantages of one over the other.