Reverb is the sound that remains in an environment when the direct signal has died out. To get a better picture lets take a macroscopic example: a gunshot in a cathedral. It is admittedly a rather rare event to hear a gunshot in a cathedral, but nevertheless it is an ideal example in terms of sound. After the gunshot (the sound coming from the gun is a short blast) the sound echoes for a few seconds, slowly dying out. This fading echo is in fact the original sound reflecting off the surfaces it meets. We have already seen how, in reflection, part of the energy is absorbed and the rest is reflected. So, at every reflection the wave loses part of its energy until it dies out completely.
Reverb follows very specific rules, summed up in the following diagram:
The diagram shows the amplitude of the various reflections and the instants in time in which they take place. The first sound that reaches the listener is still the direct signal, the latter being the shortest path travelled by the sound. After a brief pause called pre-delay, the early reflections arrive which are the reflections that have met one surface only on their journey to the listener's ear. Finally the late reflections arrive which are those that have met more than one surface. They arrive overlapping each other, thus generating a continuous sound. Since these late reflections are very close to one another, they are called reverb cluster.
To reproduce such behaviour, very ingenious techniques were implemented in the past. We will now look at these techniques, that are still simulated on modern digital reverb modules.
Spring reverb: in order to simulate the reverb effect a spring inside a cavity is employed. Two transducers are applied to its far ends which, in turn, transfer the sound signal to the spring. A microphone is used to pick up the sound that has been generated. It's not a particularly realistic simulation; the spring reverb is used in some guitar amplifiers, whereas it is rarely used with vocals. A digital module which uses this kind of reverb may use it to give a guitar a 60s sound.
Plate reverb: two transducers are applied to a vertically suspended plate, one transmitting an input signal, and the other that picks up the vibrations that are generated. In order to vary the effects' characteristics the two transducers can be placed in different spots. This technique offers a more realistic effect than that created by the spring reverb, especially at high frequencies.
Chamber reverb: the sound source is placed inside a room whose walls have particular reflective qualities. Here too, the room's sound is picked up by a microphone. Unfortunately this solution is somewhat limited due to trasportation issues...
The last generation's reverb equipment is practically all digital and employ highly sophisticated simulation algorithms (that became more and more accessible thanks to an ever-increasing processing power available). Here is a list of the main controls available on a digital reverb unit:
Pre-delay: allows pre-delay time to be controlled.
Early reflections: length of the early reflections.
Decay: duration of the reverb cluster's decay.
Mix: the percentage levels of dry and wet (reverberated) signals.
Room dimensions: often values are referred to the shape and dimensions of environments (hall, room, chamber, cathedral, spring/plate).
HF Ratio: high frequencies are the first to be attenuated during reflection. This control allows us to simulate the absorbing capacities of surfaces.
Stereo width: widens or tightens the stereo image of the reverb.
Some units allow different decay times to be defined for different frequencies.
The following is the sound of a dry instrument followed by the same instrument to which reverb has been added in different ways:
We will now add different kinds of reverb to the same sound: