Creative Audio Processing
09 Reverbs

 

 

 

 

A reverb is many things, depending on the point of view. Originally, a reverb is an acoustic space simulator. From a signal processing point of view, a reverb is a sum of filters. For a mixer, a reverb is a versatile tool than can either add or remove clarity, or radically change a sound, or "place" the sound inside a mix.

 

1. Reverb basics A standard model for reverbs : direct sound, first reflections, diffuse field. Direct sound : the original sound, to be processed First reflections : the original sound is being reflected by surrounding walls ; more obvious in small spaces Diffuse field : the numerous reflections are merged into each other ; often referred to as the reverb itself ; more obvious in large spaces In a studio environment : "small rooms", "ambiances", are mostly first reflections - "large halls", "churches", are mostly diffuse field.

 

2. Different kinds of reverbs Using different algorithms or settings can provide radically different results. For instance : a 480 Large Hall doesn't have anything to do with the TC VSS3 closet presets, though they are indeed both reverbs. In one case, the effect lasts like 3 or 4 seconds. In the second case, it almost doesn't even have any length. And : in the first case, it doesn't seem to transform the original source, it sounds more like something added to it. In the second case, it seems to transform the source itself, very much like an EQ would do. Another example : a Plate reverb is radically different from a standard Lexicon Small Room reverb. A Plate Reverb doesn't evoke any acoustics, whereas a Small Room is all about acoustics and realism.

 

3. Knowing what to do with a reverb Given that reverbs can be radically different from each other, it's useful to make a list of examples that demonstrate a few of the many uses reverbs are good for. - acoustic emulation I have recorded someone speaking in a studio, I want to give the impression this person is actually speaking inside a garage or a wardrobe or a concert hall : a reverb will do it - lead vocals in a mix I want a lead vocal part to stand out from a mix ; a large hall will a reasonable pre delay will help doing it - drum liveliness I want my drums to sound more live ; a medium room / ambience reverb will do that - in fact, if I can, I'd be better off recording those drums *with* the recording room's natural reverberation. - electric guitar "vintage-ization" I have a 70s like electric guitar part which I want to sound even more vintage : an old-fashioned echo chamber emulation will help doing that - harmonic textures I want to modify the timbre of a string part so it sounds more personal : a short reverb with rich first reflections will do the trick ... and so on. Many different possible uses. This means that before switching on a reverb unit, it's not a bad idea to have a precise idea of what you are actually going to do. This also means that when I want to add "style" on a recording, reverbs are the way to go.

 

4. Reverb in post prod vs reverb during the recording   In multi-track production, it's common to close-mike everything, and then to recreate acoustic depth, timbre, realism in some cases etc. in post production. From a strictly artistic point of view, this is not a good idea. True, it's cheaper, and without risks, but the results can be way more interesting when instruments are recorded with the recording ambience. This is especially true in music production. Why is that : close mike recording simplifies timbre, and kind of "normalizes" it. In post production, much work must be devoted to creating or re creating a richness that was lost during the recording. A simple experience can be made to prove the point : take an instrument, take a room. Record the instrument with the ambience. Now sample the room acoustics using a convolution processor, and try to recreate the original sound using close miking + EQ + convolution. It will sound OK, but much more sterile than the original recording. Why is that ? The main reason is probably related to the existence of *acoustic background noise*. When close miking an instrument, the acoustic background noise disappears. When sampling a room with a convolution processor, same phenomenon. In the process, we lose the tiny creaks, rumbles... that take part in making an ambience interesting. Another reason is the *randomness* of a real reverb. Reverb processors, including convolution processors, are deterministic machines. They'll always give the exact same result when they're fed the same audio file. [a notable exception : there are Lexicon presets which include time variants, but is that sufficient ?] On the other hand, a real room will always give different results, no exceptions. ... and that's spectacular : a reverb processor will always give the same result, a real room will never give the same result. A third reason : let's say I record a saxophone player, this player is inside a rather reverberant room, and he's standing not far from a reflective wall. Now, if the saxophone player walks away from the wall by only 1 meter, the recording will be different. If the saxophone player faces the other way, the recording will be drastically different. This also means that when the guy moves while playing (and everyone does that), there will be important variations in the recording. How can I achieve such a richness in post production ? Generally I just can't. Conclusion : whenever you can afford it, and when it suits the project at hand, record *with* the surrounding acoustics. The result will be much more personal.