admin: First posted on 2007 07 20
My last post was about reverb, but I managed to cover only the properties of the natural reverb and to talk a bit about the digital reverbs that we chose. Here is some information about the "unnatural" – i.e., simulated – reverb.
You can simulate a reverb by playing a sound in a room and placing microphones around the room to pick up the natural reverb. A room set up for this purpose would be a chamber reverb unit.
The chamber reverb concept sounds simple and it is. It can, however, have many variations. You could use various rooms, various room treatments, different type and number of microphones, different microphone placement, and different sound origin placement. Chamber reverbs were probably the original way engineers tried to simulate reverb, but they are still used and are a viable option.
While the chamber reverb concept is simple, applying it in practice is not. One big problem is obviously transport. You can set up everything in a studio once, but it is difficult to move this setup to a different place. In a home studio, chamber reverbs are even a bigger problem. If I was to try this at home I would get too much inside and outside noise, my walls do not really sound the way I want and I cannot control that, and I do not really have good microphones for this purpose. So I am ignoring the chamber reverb for now. It really does not work in my home studio and it probably will not work on in most home studios.
A plate reverb is the first "unnatural" way to simulate a "natural" reverb. You take a metal plate and make it vibrate with sound. You then pick up the metal plate vibrations and create reverberated sound. The beauty of the metal plate is that if you can make it vibrate with sound it will produce some residual vibrations. If you knock on the plate with your finger, for example, it will not move only once and stop after your finger is gone. It will continue to vibrate as an "echo" of your original knock. These residual vibrations are the artificially created "plate" reverb.
To figure out how this works we must first define the concept of "transducers". A "transducer" is a piece of equipment that transforms one type of energy into another. Take the speaker driver for example. It transforms electrical energy into mechanical energy as running a signal through the speaker driver makes it move back and forth. A speaker driver is therefore a transducer. Your guitar pickups are another type of transducer. They work in the opposite way, by transforming the mechanical energy of your guitar string into an electrical signal. Microphones are also transducers.
To get a plate reverb, suppose now that you take a metal plate and place a transducer similar to a speaker driver on it. The metal plate will vibrate. You can then place another transducer – one that works as a pickup – to pick up the vibrations of the plate. The second transducer will pick up not only the original plate vibrations but also the residual plate vibrations creating in this way a "reverberated" sound.
You do not actually need the first transducer. You can put the plate in front of a speaker and pick up its vibrations. Even better yet, you can place your electric guitar in front of the speaker and then let the guitar pickups do the rest of the work. The only problem is that, because of its design, the guitar will exaggerate certain frequencies more than other and that you have no way of controlling the length of the reverb.
Plate reverbs were used early (say 1950s), but they are still very relevant now. Many digital reverbs would actually simulate various plate reverbs. What I find interesting about digital simulators of plate reverbs those is that they are smooth (i.e., they have large diffusion) and that they have a specific coloration, which accentuates highs over 2 KHz. If you recall my post about equalization this means that this reverb accentuates frequencies characteristic for a metal clank, for sibilance, and for shimmer. The smoothness of the plate reverb is something that I have always liked. Its coloration is something that for me either "makes it or breaks it". Sometimes it is nice and sometimes it is annoying. Because of its specific coloration I have so far used the plate reverb only on cymbals and on certain vocals. We are stuck with the SM57 on the vocals and one of our vocalists always ends up with well recorded lows and missing highs, so these vocals sometimes lack liveliness and the plate reverb works well.
Of course, if you are working with digital simulators of the plate reverb chances are you can control diffusion and coloration, but then we are really not talking anymore about plate reverb, so more on that later.
Spring reverbs are actually very common. You can kick your Fender amp to hear the springs inside resonate. The spring reverb works similarly to the plate reverb. You can use a transducer to move a spring (or springs) and you can use another transducer to pick up the spring vibrations. The residual spring vibrations simulate reverb.
While spring reverbs work similarly to plate reverbs, they have different qualities. Spring reverbs that I have seen tend to be less smooth and have a different coloration. They accentuate the mid-range and do not create similar sibilance and shimmer. It seems to me that spring reverbs also have more pronounced early reflections. Nowadays I slap digitally simulated spring reverb on everything (I mentioned in my previous post on reverb that my cheap options are the Fender Deluxe Spring and King Spring reverb coming out of the Line 6 PODxt). At this point my complaint about those is that they make my songs too warm and loose, but that is to be expected.
The digital reverb is a completely different animal. There is nothing physical in the digital reverb that vibrates or resonates. A digital reverb is simply a bunch of computations done on a digitally recorded signal. If you read about how digital reverbs are designed you will usually see things like "comb filters" and "all pass filters", but this is not important. What is important about the digital reverb is that, if you have a signal stored digitally, you can repeat and equalize it as many times as you want and you can do so in any way you want. For example, you can repeat it loudly and distinctly and with accentuated highs to simulate a small bright empty room. You can stretch it out, quiet it down and accentuate mids and lows to simulate a bar. You can stretch it out further, bring back some highs, take down some lows, smooth it out, and make it very pronounced to simulate a cathedral.
Hence, a digital reverb can only mathematically simulate what a naturally resonating room, plate, or spring will do. A digital reverb can have so much flexibility, however, that at the end of the day it is the easiest one to make "naturally sounding". I used to experiment by taking some digital reverb that simulates a plate or a spring reverb and by adjusting the qualities that I did not like. I would normally take a digital "plate" reverb and bring its highs down or I would take a digital "spring" reverb and take its highs up. You have to be careful though, as too much movement will have one of two effects: it may destroy or over-exaggerate the reverb. Or it may make it completely unnatural.
Which brings me to the topic of my next post: There is so much you can change in a digital reverb, that you can get qualities that are completely unnatural and sometimes even non-intuitive.