Paul White explains why a great reverb doesn't always make a mix sound better.
Most studio musicians appreciate the importance of reverb in recorded music, but a large number of demos are still spoiled either through over‑use or inappropriate use of this crucial effect. Pop music is rarely recorded in a natural acoustic environment, unless you have access to one of the top studios with a really good live room, so what tends to happen is that sounds are recorded in a fairly dry room, then treated with artificial reverberation to make them sound natural. The same is true of electronically generated sounds that have no natural ambience of their own — they need some added reverberation to make them sound believable. Problems arise, however, when the type and amount of reverberation are wrong.
Take any solo'd track and add reverb to it, and the chances are that it'll sound bigger (in a spatial sense) and more impressive. That's because a solo'd track usually has plenty of space between the sounds. But when the whole mix is playing, there's a danger that reverb can fill all the important spaces that let the detail within the recording shine through. By its very nature, reverb occurs after the sound that caused it, so the effect of too much reverb is to 'smear' percussive events, reducing the contrast between beats and the spaces between those beats. Though the average signal level of a reverb‑processed sound may be higher than the untreated signal, the chances are that it'll actually sound less loud with the reverb added, because one of the ways in which we perceive loudness is to subconsciously register the difference in level between peaks and the lower‑level sounds that come between them. The less contrast there is, the less loud the peaks sound. (Incidentally, it is possible to make peaks sound even louder by extending their duration slightly, which is why gated reverb sounds so powerful. The high level of reverb stretches percussive sounds, but then it suddenly stops, leaving plenty of contrast with the following quieter sounds.)
The effect of too much reverb is to 'smear' percussive events, reducing the contrast between beats and the spaces between those beats.
Gated reverb has become something of a cliché — though you shouldn't let that put you off using it altogether — but similar results can be obtained by using ambience settings. These are characterised by strong early reflections but very little dense reverb afterwards. Dedicated ambience settings tend to appear only on the better reverb devices, but, providing you have a unit where the relative level of early reflections can be adjusted, you can approximate the effect well enough by setting the ER level to maximum. Pick a bright reverb algorithm with well‑defined early reflections, then set a short decay time so that there's minimal 'ring' after the initial sound. These strong early reflections will help strengthen and widen sounds without smearing them excessively; even difficult sounds such as bass drums and bass guitar can sound good with just a touch of ambience added.
Perhaps the biggest dilemma with reverb is that vocals sound great with lots of it, but as soon as you add it, the vocal loses the 'up‑front' quality that we expect from a contemporary pop mix. This problem is related to psychoacoustics — the nearer we are to a sound, the greater the proportion of direct sound we hear. In most environments, nearby sounds tend to seem fairly dry, whereas those at a distance may be more reverberant. Even in a reverberant environment, the perceived level of reverb will be lower for those sounds that are closest, as a greater proportion of direct sound reaches our ears. Distance also tends to dull sound, so for a vocal to sound 'up‑front' it needs to be bright and dry — and these conditions aren't always flattering to a voice.
Fortunately, there are ways to use reverb without losing the required sense of intimacy. One of these is to use a fairly short decay time, again with a high early reflections content. The other way is to place a pre‑delay of several tens of milliseconds before the reverb, to provide some separation between the initial dry sound and the ensuing reverberation. Combining these two techniques can work effectively, but you'll still need to keep the reverb level under control. Listen to a selection of good contemporary music mixes and you'll find that many use so little reverb that you don't actually notice it unless you specifically listen for it.
Because low frequencies take up so much headroom, it's better to remove the low end before the signal gets to the reverb unit input — which should result in a better signal‑to‑noise performance.
The quality of the reverb processor being used also makes a huge difference, so save your best processors for vocals and drums, where the differences show up most. With the better processors, adding a lot of reverb doesn't seem to bury the sound in the same way that lots of reverb from a cheap unit does, and you can often get away with a much higher reverb‑to‑dry ratio before the sound becomes unnaturally muffled. Good reverb quality is particularly evident at short decay times, where lesser units may start to ring or sound unduly coloured. If you have access to a really good reverb unit, you might find that a vocal line sounds more effective treated with a higher level of fairly short reverb than it does with a lower level of a longer reverb.
A potential problem with using a bright‑sounding reverb on vocals is that any sibilance in the original performance will already have been exaggerated by any compression that's been used, and once you add a bright reverb it may reach an annoying level. De‑essing vocals often results in a lispy quality to the sound, so a kinder solution may be to de‑ess the input to the reverb unit and leave the dry vocals as they were. Figure 2 (below) shows how this might be done.
From what's been said so far, you can probably deduce that sparing amounts of bright reverb, or reverb rich in early reflections, will help give a sound presence, width and interest without pushing it to the back of the mix. However, if bright reverbs are chosen for everything in the mix, the contrast element once again gets lost. Classical instruments tend to benefit from more natural reverb treatments, and in most cases that means using quite a lot of HF damping and HF rolloff to simulate a concert hall type of environment. Even in rooms with very hard surfaces, such as stone cathedrals, the reverb can be much less bright than you might expect, mainly due to the air absorption of high frequencies in the large distances between walls.
Classical instruments tend to benefit from more natural reverb treatments, and in most cases that means using quite a lot of HF damping and HF rolloff to simulate a concert hall type of environment.
Using lots of early reflections can also be a bad idea when processing classical instruments, because when a large ensemble plays in a real concert hall, the early reflections have a tendency to disappear. The reason is quite simple: the pattern of early reflections depends on the position of the performer relative to the nearby walls and other boundaries, but if each performer is in a different position on stage, each will generate a slightly different set of early reflections. When these are all added together, individual reflections tend to become masked by the increased complexity of the reflected sound.
Reverberation is a very powerful effect, without which no studio would be complete, but there are dangers associated with its use. In terms of perspective, excessive use of reverb pushes sounds to the back of the mix, while adding more than the merest hint of reverb to bass sounds clutters up the low end alarmingly. There are occasions when long reverb settings work, but these generally require musical arrangements that leave a lot of space for them to work in.
Adding much in the way of reverb to sustained pad sounds seldom works, as the sustain of the pad hides the reverb, which means that you have to add a lot to make the effect noticeable. As a rule, smoother sounds benefit more from coarse treatments with widely spaced early reflections while percussive sounds need a higher density of reverb — otherwise the early reflections sound like somebody ripping cloth! Once you've picked an appropriate reverb sound, you then have to decide whether there's enough space in the music to let you use it as an obvious effect, or whether you should add as little as you can to create a convincing sense of space.
Reverb units can tend to emphasise anything that's bassy or muddy in the material being processed, and since most of the energy in a typical pop music mix resides at the low end of the audio spectrum, perhaps this isn't surprising. The problem can be reduced by EQ'ing some of the low end out of the reverb, and the easiest way to do this is to feed the returns through a couple of mixer channels rather than aux returns, so that you can use the channel EQ to apply bass cut. However, this isn't actually the best way to do the job. Because low frequencies take up so much headroom, it's better to remove the low end before the signal gets to the reverb unit input — which should result in a better signal‑to‑noise performance. Any type of equaliser patched before the reverb input will do the job, but the high‑pass side‑chain filters on a noise gate such as the Drawmer DS201 (set to Key Listen mode) are particularly good for this purpose because of their steep 12dB/octave slope. Using these, it's possible to almost surgically remove the low end without changing the mid and high frequencies in any obvious way. Figure 1 shows a suitable patch for accomplishing this.