When it comes to treating vocals, we generally think first of compression and reverb, but there's a wealth of other vocal treatments that may not be quite so obvious. Not all are applicable to all styles of music, but it's still worth trying them at least once, just so that you know what they sound like. There isn't space here to explore every effect in detail, but you should find something to inspire a bit of vocal creativity!
Vocal EQ has been covered in great depth in previous articles, so I won't revisit it here. Suffice it to say that choosing a mic type and position that gives the right sound at source is better than using EQ to knock a disappointing recording into shape. If you need to add an airy gloss after the fact, the merest broad boost at 12kHz usually does the trick or, if you're lucky enough to have a TC Powercore system, the included Character plug-in or the optional Specific Vocal Enhancer plug-in from Noveltech do the job very simply and effectively.
The first thing most engineers do to a vocal after (or even during) recording is to add compression. As I'm sure you're aware, compressors restrict the dynamic range of sound, making the level more even. In a pop music context, where the backing track levels are also tightly controlled, this can be important in getting the vocal to 'sit' correctly in the mix. Without compression, the vocal level may drift between being too loud and too quiet. Having said all that, using compression alone to tame level differences isn't usually the best way forward. Some singers have great control over their dynamics, while others will get noticeably louder when they sing in certain registers and if you use enough compression to deal with this, the vocal can end up sounding seriously squashed. A better approach, where available, is to use track level-automation to iron out the most obvious excesses, then use compression to smooth the end result. Where mix automation isn't available (for example, with a traditional stand-alone recorder and analogue mixer setup), you'll need to ride the faders while mixing and apply the compression at the same time — just like the old days!
The type of compressor you use also makes a difference, because compression straddles the line between effect and processor, in that it can alter the perceived character of the sound, as well as control its dynamic range. All compressors increase the average level of the sound as a direct result of bringing the level of the quieter notes closer to the level of the louder ones. The subjective effect, though, is also influenced by the attack and release times of the compressor, and by the amount and type of distortion it adds. As a very general rule, the most transparent compressors (those that reduce the dynamic range without affecting the subjective sound too much) use VCA gain-control elements, while those using FETs, tubes and opto devices tend to add a little more character. Generally, a vocal compressor needs to be set up with a fairly fast attack time (just a few milliseconds) and with a release time in the order of a quarter to half a second, but you can sometimes fake the more obvious aspects of an opto compressor's character by using a long attack time and a fast release on a compressor that is normally fairly transparent-sounding.
Of course, you need to be aware that the more compression you use (in other words the higher the gain-reduction meter reading), the more gain will be applied to low-level sounds relative to high-level sounds, a consequence of which is that unwanted low-level sounds, such as noise or the spill from headphones, will become more obvious. Where the recording is being made in an imperfect room, the room ambience will also become more pronounced when you add compression, which is why it is essential to record vocals in an acoustically treated space, even if the treatment is only a duvet behind the singer.
More aggressive compression can be used for certain rap and 'death metal' vocal styles, and if you haven't already tried it, it's worth downloading the free Talkback Compressor from the SSL web site (www.solid-state-logic.com). This is available from the Resources / Downloads and Manuals section after you've filled in a few registration details, and works as a VST or AU plug-in on Mac or PC. It is modelled on the talkback mic compressor used in SSL consoles, which leapt to fame when it was used to help create the big Phil Collins compressed and gated drum sound. There are plenty of other compressors, but this one is seriously unsubtle, and works a treat on vocals that need more attitude!
Reverb is the other vocal essential, if only to replace the natural ambience lost through recording in a dry environment. Few small studios have the sort of natural acoustics that suit pop vocals, so the usual approach is to use an acoustically dead vocal booth, then add back the desired ambience during mixing using some form of artificial reverberation. The main thing to consider here is that what sounds like a lot of reverb when you solo the vocal may be barely noticeable when the whole mix is playing, so always adjust the final reverb level in context. The choice of reverb is an artistic decision: what sounds most natural and 'room-like' doesn't necessarily sound the most musically pleasing. That's why algorithmic reverbs and plates are still often used in preference to convolution reverbs, which are based on real spaces such as concert halls. Where convolution reverbs are used, it is often with impulse responses taken from hardware reverb units. This makes sense when you consider that we've become so used to the sound of electronic reverb over the past few years that we perceive it as being artistically right. While it is difficult to comment generally about reverb, the current trend seems to be to use less obvious reverb treatments than in the '80s or '90s, so short, bright plates and lively ambiences are often used in addition to, or in place of, more conventional reverb treatments.
One vocal effect that has been used quite a lot in recent years is the 'telephone' filter. This can be as simple as rolling off both the low and high end, using steep filters to squeeze the audio into a band roughly between 250Hz and 2kHz. While it would sound odd to treat a whole vocal part this way, it can still be effective for short passages. However, you can also achieve some interesting variations on this effect in a very different way.
The convolution process is best known for its ability to capture reverbs and ambiences, but it is equally applicable to short delays or devices that produce predominantly tonal changes. Plug-ins such as Altiverb from Audio Ease can be used to create this type of effect, in this case if you download additional impulse responses from the Audio Ease web site (free to registered users). One of these sets includes impulse responses (IRs) taken from small transistor radios, telephones and so on. The transistor radio IR sounds extremely convincing when used to squeeze a voice into a narrow part of the spectrum, and because the IR is able to capture the more complex tonal attributes of the system being measured, the result is somehow more believable than if you used simple EQ filtering. If you don't have Altiverb, there are alternatives, such as Logic Pro's Space Designer or the freeware, PC-only SIR, as well as online resources such as Noisevault (http://noisevault.com/nv), where you can download IRs.
If you have the tools to capture your own IRs (either built into the convolution plug-in, such as with Space Designer, or a stand-alone impulse-capture application like Fuzz Measure), you can easily create your own effects by taking IRs from small speakers, transistor radios and guitar practice amps. While you're at it, you can also take IRs from toy microphones with springs inside, cassette recorders (to give you that real squashed tape effect) and even tape echo units, if you can get your hands on one for a couple of hours.
Talking of tape echo units, echo has become almost as important a vocal effect as reverb. This can vary from the very obvious surf echo of the '60s and Pink Floyd excesses of the '70s and '80s, to short, slapback echo as immortalised by artists from Elvis to John Lennon. While conventional digital delays offer low-noise, wide-bandwidth, pristine echo, they don't sound nearly as musical as tape echo. This is because tape echo has a softer sound that gets warmer and less distinct every time the sound is fed back and repeated, which makes the delay sit further back in the mix, supporting the dry sound, rather than fighting for a place at the front of the mix.
Tape echo units are now fairly rare and quite expensive, but fortunately there are numerous hardware and software solutions that use programming to mimic the distortion, filtering and pitch instabilities of tape echo. Not only is their cost a fraction of that of their hardware counterparts, but the tape won't break during some vital solo! You can also get impulse responses from tape echo machines for some convolution reverbs, but because of the long delay times involved, this can be excessively hungry on CPU resources unless you're after a simple slapback. I'll often use a subtle repeat echo mixed in with a reverb to fatten a vocal, but I also like that slapback effect for certain productions where you need just a single short delay (typically 80 to 150ms) high in the mix.
Of course if you have an open-reel tape machine lying around that uses separate record and play heads, you can also use this as an echo unit, simply by feeding it from the send on a mixer, setting it to monitor the replay head, hitting record and then bringing its output back on another spare mixer channel. If you turn up the same-numbered send control on that channel, you'll feed some of the tape's output back to its own input, producing repeat echoes. The fader controls the echo level, while the send control governs the time the repeats take to die away. The delay time and subsequent repeat time depends on the tape speed. A speed of 15ips (inches per second) usually gives a nice slapback effect, but if you need a longer delay, you can patch a conventional digital delay unit before the tape machine and set it to 100 percent wet. The tape machine will colour the sound in exactly the same way as when used on its own, but now you have a delay time equal to the tape delay plus the digital delay.
Another fun trick is to solo your vocal track and then record it to a standard cassette machine. Next, record the output of the cassette back into the computer or workstation and put it on a new track alongside the vocal. Slide the newly recorded 'cassette track' so that it comes just after the original dry track and you have genuine tape delay. The timing instability of cassettes means that this may drift out of time over long periods, but for creating slapback echo within a typical song, it should work fine. If not, break the delay track into sections and adjust the timing for each section. You can also add digital delay to the tape delay track if you want to create a repeating echo.
Used in context, more outlandish vocal effects can be a lot of fun. Perhaps the oddest is genuine reverse reverb. We've covered this before, so I won't dwell on it in too much detail here, but in the good old days, you'd record a vocal, turn the tape over so that it played backwards, add reverb and record the reverb to a spare track. When you flip the tape back the right way around, the original vocal is back where it should be but the reverb is now backwards and builds up before each word in a very surreal way.
In this DAW age, you can do much the same thing by processing a copy of an audio track to reverse it, adding 100 percent wet reverb, printing or bouncing the reverb to make it permanent, then reversing the resulting reverb track before running it back alongside the original track. You may need to adjust the timing of the reverb track relative to the dry track for the best results. You can also do the same thing very simply using any convolution reverb that allows you to reverse the impulse response, by copying the audio part to a new track, adding the reverse convolution reverb (again, 100 percent wet), then sliding the treated track forward so that the reverb builds up just before the start of the dry audio track.
Another favourite of mine is to pitch-shift the audio up an octave before feeding it to a delay or reverb, a technique that adds a surreal shimmer to the sound. I've discussed this before in the context of guitars, but it can also be very effective on vocals. Alternatively, if you want to be more subtle, you could try using pitch correction on the reverb feed and adjusting the severity of the pitch correction, so that the reverb sound is just slightly different in pitch and character to the original. What you end up with is not quite normal reverb and not quite artificial double-tracking but combines a bit of both.
To force vocals to play more of a textural role in rhythmic music, the familiar tempo-driven chopping (square-wave tremolo) effect can produce great results cycling at eight or 16 chops to the bar, and of course this can be used in conjunction with other effects, to create a more complex sound. For example, adding a long, heavy reverb and then chopping up the result can produce a sound that combines the qualities of vocals and keyboards. Alternatively, you could chop up the audio feeding the reverb.
I was recently experimenting with a vocal track using Logic's Platinumverb plug-in and discovered that it can generate extremely convincing double-tracked, slapback vocals in the style of John Lennon. You simply set the early reflections delay time to between 70 and 110ms, reducing the reverb time to less than half a second and then winding up the reverb level to around 60 percent of the dry level. The settings I used can be seen in the screenshot at the bottom of the page. Using a cluster of early reflections to create the repeat, rather than a single delay, makes the effect much more convincing, and if you also roll off some low end by using a sharp filter at around 200Hz, you can get very close to that trippy, 'Lucy In The Sky With Diamonds' sound.
Another discovery I made at the same time came about when I was trying to dream up a way to add interest and depth to a classic rock song, performed using just an electric guitar, an acoustic guitar and one vocal line. I wanted the effect of a keyboard pad but without using any extra parts, so I copied the vocal line to a new track, inserted a long reverb of about six seconds and 100 per cent wet, then inserted a compressor to keep the reverb level fairly high. The final step was to drop in a stereo rotary-speaker plug-in set to its slow speed, then mix the resulting treated reverb back under the track so that it was only just audible. Surprisingly, the rotary speaker effect hid the origins of the sound pretty well and the perceived result was much more like a low-level keyboard pad than a treated vocal. This is one I'll definitely be using again!
Of course you don't always need to come up with original weird ideas, and there are now processors that can manipulate the voice for you in a variety of ways. Roland's hardware VP70 Voice Processor, for example, is great for creating quasi-robotic effects, while companies including Antares, TC Electronic and Celemony offer sophisticated voice-shaping processes that remodel the formant structure of the singer's vocal tract to change the character of the voice without making it sound too artificial. There's still plenty of scope for going over the top if you want to create something unnaturally growly or squeaky, but if you use such devices carefully, you can record multiple vocal parts with one singer and then make them sound like an ensemble of different people. In my experience, going as far as to turn a male voice into a female voice or vice-versa rarely sounds entirely authentic, but more subtle shifts in timbre are handled pretty well. The same is true of automatic harmony devices as championed by TC Helicon: as long as the harmony parts are not too complex or too forward in the mix, they can sound very plausible, especially if you use the randomising features that introduce human-like pitch and timing offsets into the harmony parts. They sound even more convincing if you layer one or two genuinely sung harmony parts over the top, just as a sampled string patch sounds more realistic when you overdub a couple of violins.
There's such a lot you can do to process vocals in an interesting way that it would take much more space than I have here to explore them all. A good approach is to take processors designed for other purposes and just try them on vocals to see what happens. Guitar-amp simulators provide a practical way to add controlled distortion, and rotary speakers deliver a very trippy sound, as the Beatles discovered way back in the '60s. Distortions and overdrive effects that you might use without thinking on guitar can create interesting effects on a delayed vocal, particularly when combined with other effects.
Whatever sort of music you usually make, it's always a good idea to push outside of your comfort zone, experiment and try something new, or try combining existing techniques in unfamiliar ways. The resulting effect is often more than the sum of its parts and might just give you the unique sound you've been striving for!