funny stuff about reverb

note: 

wet/dry.

when you send a sound to a reverb machine, one "knob" lets you adjust the percentage of original signal versus "with reverb" signal, or simply "How much reverb" you want added to the end signal, the "finished product" of the reverb machine. in most cases, these are called DRY (no reverb at all)and WET(only the reverb sound). whatever machine you own, there has to be a control that does this or something very close.

also note that "reverb" refers only about the sound that comes back to the source from bouncing off the walls - not the original signal. most applications call for a mix of the two.

difficult-to understand differences between artificial reverb and real reverb.

first of all, obviously real reverb (RR) originates from sound bouncing off surfaces, and in a plain square room these might nto be exactly as parallel as you would imagine; you ought to know that stuff that is *exactly parallel* tends to do weird stuff in the amazing world of audio physics (phase cancellation, anyone ?), while minor deviations affect these maths drastically.

However this is a minor point, granted, as in most modern habitat situations walls do indeed tend to be fairly parallel, and the reverb that reverb machines simulate is based  on parallel walls; mind you, i am not talking about walls that cause standing waves, but rather a room constructed with the wall shape specular on both sides, so, if the left wall is twenty degrees bent to the left, the right wall is bent twenty degrees to the right. Most homes, offices and other non- audio related buildings are rectangular in shape, after all. It's entirely possible to have digital reverb units emulating a cavern, or a church, with large spaces and indistinctive shape; similarly, it is possible to have simulations of trully horrid spaces, such as pipes, tunnels and .. well, even most basic "club" settings are horrid, to my ears.

Artificial Reverb (AR) machines are quite stupid: whatever the room characteristics, the "walls" involved are matematically flat - the formulas around the parameters  do what they are asked to do - modify the sound as if it bounced back from that surface.

I'm not saying that digital reverb units are bad, I'm saying that there are other things that natural ambients do to sound, that a AR unit just do not do; Absorption, for one. And refraction.

AR units create spaces. They go as far as deciding how reflective the "walls" are, but the other two elements are simply not there; they *could* be, in the future, if someone begins working them in to the machines, but as of today ( as far as i know, that is) if you want realistic reverbs from a machine, you have to cheat.

Absorption is a portion of the spectrum disappearing inside a substance - the spectrum of RR is never going to be equal to that of the originating sound; also, different densities and hardness of materials reflect different portions of the spectrum differently, with those few millimetres of plaster on your cellar wall reflecting differently (and at a different time) than the underlying brick. Wave amplitude per frequency is also a factor, granting different wavelenghts different penetration and thus, different reverb times. Low rumbling sounds do not bounce back the instant they hit a surface, they bounce when they encounter sufficient resistance. Shrill reverb sound is typical of digital reverb units, making your bass sound like it's a high hat in a glass room, but sped down. Also, harder materials such as steel beams, concrete and the likes set in motion when stimulated, at different speeds (we're talking both molecular vibration and   actual physical vibration), and stuff that vibates reflects energy differently than stuff that doesn't.

Refraction is essentially what happens when your surface isn't flat down to molecular level, that is, always. Anechoic chambers are, afer all, walls covered with a bunch of spikes. Big spikes granted, but microscopic spikes (such as bad plaster) can cause their own refraction, expecially in ultrahigh frequencies; which do exist, and do modulate other frequencies, wheter we hear it or not. Air pockets, cavities, and other proprieties of the reflecting materials collaborate to create the big differencebetween RR and AR.

So how do we get a good AR sound in a small studio ?

Well, "real" reverb machines, such as plate reverb, or tunnel, do get "real reverb", even though it does not sound like a hall, but they do have some physical components which have their own mechanical acoustic proprieties; if you can use one of those, you are almost there. But if you are broke/have a DSP, then you need to cheat.

First, you should always process your reverb separately. Send the feed to the machine, then have it bounce back a 100% wet mix on a separate channel. Do not record the reverb on the main track (you know this), and also do not record your master with the reverb unedited.

You will need to make some changes to your rever sound to account for real world coloration.

Now, i can suggest some way to make your wet AR sound a little more real, but in the end, it's up to you what sort of coloration you want to add; i doubt that even with a lot of effort and money you will ever be able to recreate the acoustics of a granite castle, or marble church.  

Equalizier on the wet :

if you have a good listening room, less reverb and more low frequencies makes for a better sound, as long as you are prepared to painfully tweak your EQ to get a great response over the whole song; natural reflections change their equalizing with their natural resonance, and although it *is* possible in a MIDI application to select every individual pitch and give it its own EQ, i realize this is impossibly complicated for anyone but the insane.. simply set multiple EQs each to a different frequency, and have each lower one boost more the signal - a lot more, if you take in consideration the F/M curves; make sure that the boosted frequencies do not overlap. If you want to make this perfect, also have multiple wets to feed to each different EQ; i know this sounds like a million tracks, but after all we have ProTools, right ?

Cutting high frequencies from the reverb is also a nice trick; you could always add a resonant shrillness by boosting a frequency in the wet and dampening it in the dry, such as a room with windows, or vice versa a natural environment (moss, trees) would absorb in the wet and boost in the dry. You can chose multiple frequencies to boost and cut, and though this might seem random on paper, you are simulating a very complex event in nature - as i have said before, great natural reverbs are a matter of drastic coloration. It's entirely possible to drastically cut high frequencies from a reverb wet while keeping them untouched in the dry - you might think that it's going to sound muffled and boomy, but it's what real reverb does. When we are en situ, we have a almost "magical" talent for picking up reverb from our surroundings, something that disappears when listening to "2D" sound. Of course, altering reverb thus means having less of it on the track; if the end product is not what you want, simply have less EQ boost/cut, and more reverb overall.

You can also create unusual ambients by filtering a copy of the dry and feeding the result to your reverb unit, which again you record separate. For example, track 1 dry, track 2 dry with filter cutting all but 200-300Hz low mids and hi-pass everything above 3k, feed track 2 into the reverb and get a 100% wet track 3; record tracks 1 and 3. This will sound like a empty ambient, that only refelects those particulair frequencies. For example, a basketball court next to a  glass-walled building.

You could also make that glass building sound further away from the court pavement ..

Find the number of reflective objects you want; then, EQ out all the frequencies you do not want bounced by that object, and feed each EQ to a delay unit. Send both to the reverb unit, and you have your ambient with different distances. And for absorbition ..

Multiple copies of the same reverb, with dramatic EQ cuts, and delays to slow the return of the lower frequencies (they bounce off solid substances after penetrating them, thus their impact point is further and therefore their travel path longer). The lower the frequency, the higher the delay. The effect can be exaggerated to create a vivid impression of massive reflecting structures, and all your friends will think you are awesome when they hear that great RR sound. Again, having at least five wets with different delays is ideal; should you experience something horrid like cancellation, simply skew the delay parameters of one a bit until it disappears. Start with 2ms, and going up - depending on what settings your delay allows you - try to avoid multiple increments.. the Fibonacci series works perfect. That's (1/1)2/3/5/8/15/23/48 etc.. 

Delay can center a sound anywhere in a room - split the signal, feed it to the reverb twice, give each reverb a different delay time - the higher the delay, the lower the wet mix. then, assign each to Left Pan or Right Pan.

There are no "good" ways to simulate the other acoustical proprieties of reverb-creating reflecting materials, such as air pockets, but in the end, the approach is always the same - send to the AR a copy of the track you want to have reverb, and put your processors on it, then mix the result, 100% wet, with the original dry. Exciters, choruses, distortors and flangers (sheet metal room??), whatever effects you can think of, just make sure you do not abuse them - if you find the result is too dramatic, diminish it and couple it with yet another copy of the reverb track, but without the effect; unfortunately, reverb is just like a compressor - it works best when you can't hear it working. High-volume listening helps immensly in getting a really pure sound.

You might be wondering why on earth someone would want to go through all this effort to get a reverb sound, when there are excellent machines from Lexicon that do it for you .. well here is my answer:

Reverb bores me to death. I hate to listen to a song with reverb, after a few listenings i get really tired, the artificialness really begins to stand out. I have always hated that, and too much reverb (which for me is any reverb at all) means i can't listen to that song, ever again.

Older AR units tend to have better sounding reverb, mostly because they were mechanical in nature, and their plates, or tunnels, had their own micro-audio proprieties, and i can tolerate them more. But artificial reverb units, even a 'Lex, is something i positively do not want in my mix.

Having fought over the years on this and now having accepted that people simply will not live without reverb, if you are like me, try these tricks and you might get an all together better result; failing that, at least you put some effort in, and that's it own satisfaction.