Algorithmic Reverb Through History

This project explores the recreation of three influential algorithmic reverbs using PlugData, a Pure Data-based interpreter integrated with JUCE, a robust C++ framework for digital signal processing. By reconstructing these iconic algorithms, the study investigates their sound and historical significance in shaping modern reverb designs. This work provides insight into the evolution of algorithmic reverbs and their influence on contemporary audio processing practices.

Hello, my name is Gage Jones and I’m here to report my findings on algorithmic reverbs through history. Algorithmic reverb are reverbs that mimic a space by a mathematical equation and they could be in real or non real spaces and it’s incredibly controllable. A famous example is a Lexicon 224 which is heard in many famous recordings from the late 70s and early 80s. Different examples of algorithms. First is the algorithm by Manfred Schroeder that started it all, and it sounds like a plate reverb due to its diffuse reverb, tail and diffuse means it’s equal over space and time. It also has no early reflections, so it has a particular sound that really sticks out compared to modern reverbs. So to start off with. I want to show a dry signal, then the Schroder reverb signal. So here’s an example of the dry signal. And here is an example with the Schroder Reverb. So as you can tell it is reverb, but it’s also very glassy and has a very particular phonic to it. So James Moore basically took Schroder’s idea and added a couple of things. He added early reflections he adds early and late reflections being controlled by the same system, and also that the late reverb tail is created by using a network of comb and all pass filters for smoother diffusion. Unlike Schroders Reverb, which just used one of each. Now what does this mean in practice? Well, you’ll notice that that one particular phonic is actually not present in the more reverb due to more smoothing. Lastly, I went for John Dattorro’s reverb, which uses a feedback delay network and also features nonlinear processing which caused a smooth diffusion of reflections. Now what this essentially means for me is I kind of visualize it as a fish tank, so it’s essentially like each reflection is a fish and each time that fish sinks further to the bottom, it gets further away. It’ll take that fish and feed it back up to the top, and then it’ll keep everything that it stored inside of the fish as it transfers it back to the top. So essentially it’s a very rich and it builds very fast into an extremely dense echo. And it also features spatial encoding, which means that each of the reflections are actually localized to wherever they are in the fish tank. If you want to call it that way, and it’s also the algorithm that’s featured in the famous lexicon reverb units, and also still used to this day. I also want to include that all of these examples were created inside of plug data, which is a software that infuses both Max MSP and Pure Data objects inside of JUCE C++ repository. So this is the Schroeder reverb. This is the Moore reverb and this is the Dattorro reverb, so hopefully you found that as cool as I did and thank you for watching.