No matter what kind of songwriting you do, the vocal tracks are typically the most important element for connecting with the listener and conveying emotional content. It's therefore crucial that you produce the best possible vocal tracks, even when the singer doesn't deliver that elusive "perfect" take.
The easiest way to get a great vocal recording is to create a composite track. In other words, you record multiple takes-each on its own track-and then combine the best parts of each take into a single "perfect" performance.
Using multitrack analog tape for this process has always been a hassle. Bouncing the takes by muting bad sections and soloing good sections can be time-consuming and inaccurate. Mix automation helps, but you still have the problem of generation loss, the deterioration in audio quality that happens each time you rerecord an analog track. Digital tape recording has made the process quite a bit less painful. With the ADAT's seamless, automated punch-in and punch-out recording, edits are more precise and digital recordings don't suffer from generation loss.
Hard-disk recording, however, offers the greatest flexibility and makes comping even easier because you can visually edit audio waveforms. In some cases, you can edit right down to the individual sample level. Even so, comping a vocal track with digital audio editing software is not just a matter of cutting and pasting. You have a number of options during recording and editing that can make the difference between a professional track and a patchwork one.
RECORDING TIPSDuring the recording session, your object, of course, is to make each take sound as good as possible. But you also want them to sound as much alike as possible. If all the takes sound similar, later edits will blend seamlessly.
To begin with, don't use time-based effects (such as reverb, chorus, or echo) during the recording. When you start pasting in sections from different tracks, things like reverb tails and LFO modulation for chorusing won't match up correctly. Dry tracks provide a more uniform sound. On the other hand, if you add effects after editing, they can help mask the edits between audio segments.
Depending on your singer's mic technique, you may want to use a bit of compression while recording each take. That will give each take a similar amplitude range. A modest setting with a fairly high threshold, fast attack, and 2:1 or 3:1 ratio should keep levels under control without squashing the life out of the recording.
In case your singer gets a little too rambunctious with those high notes, it's a good idea to patch a limiter into the recording chain to prevent signal overloading. Set up the limiter with a fast attack and release and a threshold that's a couple of decibels below the loudest peak your sound card can handle. With these settings, the limiter will simply block out distortion without otherwise altering the signal.
Noise can sometimes be a problem during recording. Even though there are methods for eliminating some noises after they occur (which I'll discuss later), prevention is always the best way to handle the problem. For example, always use a pop filter on your mic to eliminate plosives (sudden bursts of air) that your singer may produce.
Furthermore, if your mic has a low-cut switch, you might want to activate it. The human voice has very little low-frequency content, and you don't want your mic picking up things like ventilation noise or room rumble. You might also consider using a noise gate. They can produce abrupt level changes, but if you set a very low threshold and a decay time of about 250 to 400 ms, the changes won't be so obvious. Moreover, the reverb you'll add to the vocal track later should smooth out the transitions.
EDITING BASICSComposite tracks are typically created using one of three methods: cutting; copying and pasting; or using amplitude envelopes. Cutting involves destructively removing all of the bad sections from each take. You are then left with a number of tracks that contain scattered audio segments, which make up the final composite take. Cutting is easy, and it's available in any editing program, but it's not very flexible because you can't go back and make changes.
Instead of seeking out the bad sections, the copy-and-paste method involves finding and marking all of the good sections in each take. You then copy the best sections and paste them into a composite track. This method is easy, and it's flexible because it doesn't alter your source tracks.
But using amplitude envelopes is the best method by far. With this technique, all of your edits are fully adjustable at all times, and there's no need for cutting or copying and pasting. You simply assign amplitude envelopes to each take, draw in the volume changes needed to mute the bad sections, and then mix down to a new composite track. If you decide that you need to make changes, you simply adjust the envelope points and remix. Unfortunately, amplitude envelopes aren't available in all audio-editing programs, so you might have to use one of the other methods.
Whichever method you choose, always be sure to use the "snap to zero crossing" function in your editing program. By ensuring that all of your edit points land on the nearest zero-point crossing in the audio waveform, this function minimizes glitches that can occur between segments. To map out the good (or bad) sections of each take, use start and end markers. Then you can easily select the audio between each set of markers for cutting or copying and pasting. When using amplitude envelopes, however, you don't need markers because drawing the envelopes visually marks each section.
WHERE TO CUTWhere you make your edits is just as important as how you make them. Finding the right points on a waveform display can be difficult unless they all occur in the silence between lyrical phrases. Within a phrase, however, many of the words flow into one another. For example, in the phrase "How can I find a place to call my own?" the words "find a" and "my own" blend together, so there's really no way to do a clean edit between them (see Fig. 1).
The best place to edit is immediately before a hard consonant. The quick burst of air used to pronounce hard consonants separates them from the word before, so the edit will work even if there isn't a pause between the words. In our example, the best edit points are right before the words "can," "to," and "call." Hard consonants are relatively easy to find on a waveform display because of their sharp peaks.
Another useful, but more difficult, technique is to place your edit points on fricatives, consonants pronounced by forcing breath through a constricted mouth formation. The letters f and s are fricatives. Because of the noise they produce, it's possible to hide an edit point within them. You'll be editing within the word itself, so you should use this technique only when absolutely necessary. Smoothly matching up audio segments this way is often difficult, especially when the words are being sung.
KILL THE NOISEEven if you take preventative measures, noise can make its way into your tracks. But a few clicks and pops shouldn't deter you from using an otherwise great section of audio. These transient disruptions can easily be removed, and you don't need any special tools to do it. Most of the pops your singer produces while performing plosive lyrics can be filtered out by a pop filter. Clicks, however, arise from a variety of sources, including saliva snapping in the singer's mouth.
In a waveform fully zoomed out, pops and clicks have a similar "spike" look. It's best to locate them by first listening to the audio segment and then zooming in on the suspected area. When zoomed in, a click still looks like a spike, but a pop resembles a distorted waveform (see Fig. 2).
You don't want to simply delete these transient noises. That could upset the rhythm of the lyrics. Instead, reduce the volume of the pop or click by about 10 to 15 dB. You may also want to try replacing the click or pop with a tiny bit of the waveform that comes immediately before or after it. You'll have to experiment a little to see what sounds best. You also have to be cautious when making a selection. Clicks are easy to select, but pops usually come right before the start of a word. Take care not to cut any part of the word itself.
FINAL CUTIf vocals are a big part of your music, making them sound as good as possible is a high priority. Creating a composite vocal track is one of the best ways to do that. It may take a little time for you to get the knack of making clean hard-consonant edits or recognizing and removing pops and clicks, but it's definitely worth the effort.
There may not be such a thing as a perfect vocal recording. There always seems to be something that could be done just a little bit better. But if you take the time to apply the techniques outlined here, your vocal tracks can step a bit closer to perfection.
Scott R. Garrigus is an author, musician, multimedia expert, singer, and voice-over performer. You can contact him at www.garrigus.com or via e-mail at firstname.lastname@example.org.