The Performance Curve

How do you know which take is the one?

by Craig Anderton

The essence of recording comes down to this: capturing great performances. All related tools, techniques, and technologies become irrelevant without that crucial element.

Granted, some tools help promote better performances, such as the way loop recording allows the performer to get into a "groove." But ultimately, a good performance also has much to do with production skills, and knowing how to get the most out of a performer. However, is that process based solely on gut feelings, or is there some quantifiable element that can help ensure getting the best possible take?

THE PERFORMANCE CURVE

I've worked with a lot of artists over the years, and found that certain artists tended to reach peaks at a particular time in the process of recording their takes. The following charts describe the different types of performances I've seen in the studio.

The double-peak: After observing what happened with my own performances when recording composite vocal, guitar, and keyboard parts, then picking the best sections, I noticed that the quality of my takes follows a definite pattern. The first couple takes are pretty good, then they start to go downhill before taking an upward path again. Eventually they hit a peak that sometimes exceeds the initial one, then past a certain point, deteriorate at a pretty rapid rate. I call this a "double-peak" curve because it has a peak at the beginning, and a peak toward the end.

It's uncanny how often this happens. It doesn't even matter which instrument I'm playing. But that's just me; I've also produced quite a few artists over the years, and noticed other distinct patterns.

The quick starter: This performer starts strong, has several good takes in a row, then doesn't really improve on the performance over time. Many times, these are musicians who play live a lot. They're conditioned to get things right and "give it their all" because live, you're get only one take.

The long ramp-up: In this case, the performer takes a while to "warm up" and get into the groove. This often happens with musicians who compose in the studio. As they feel their way around the part, they become more comfortable with it. After they hit their stride, sometimes you'll get a killer take; sometimes you'll get a series of takes that are all pretty good, and when composited together, produce a definitive performance.

The anything goes: This is the kind of performer who goes strictly from the gut. Rather than follow a particular curve over the course of several takes, they hit high and low points within individual takes, as the mood hits them. These are the most time-consuming performances to comp, because you might end up taking different phrases from early, middle, and late takes. Yet the final results can be really good, because there are a series of spontaneous moments that produce multiple high points during a take. However, you want to make sure it "breathes," and has some less intense sections to provide contrast with the moments of high emotion. In any event, these are people where you record everything because you never know...

The rock steady: I first encountered this type with classical guitarist Linda Cohen. All her takes are consistently good, so the only real question is whether she can do one that's better.

There are other patterns as well, but these are main ones I've recognized.

WELL, SO WHAT?

This may sound a bit abstract, but there are practical ramifications. For example with Linda, she knows when she's done a good take, at which point she tends to want to move on (classical guitarists don't get huge budgets for studio time, so time is of the essence). I usually agree with her, but once I had heard her do a tune better in rehearsal than how she did it in the studio, even though her take was technically flawless. I wanted to ask her to do another one, but knew she'd think it was superfluous.

So, I said into the talkback mic that unfortunately, the record button hadn't been enabled on her previous take. She was kind of bummed but she's a pro, so she did another take. However, my "mistake" sort of shook things up; while her part was again perfect, it had a little more feel (I presume it was because she had gone through these conflicting emotions of thinking she'd done a real good part, only to find it hadn't been recorded). Of course, the previous take had been recorded, so if the newer one wasn't as good, there wouldn't have been any problem.

However, you wouldn't try that approach with the "long ramp-up" performer, because it takes them so long to get where they're going that they'd likely strangle you if they thought you'd made such a major mistake. With their type of curve, you're best off saying "the last take was really good, but they keep getting better, so let's try just one more."

The "quick starter" is something else. If the takes aren't happening, I prefer to move on to a different song entirely, then return to the one where I want a better take. Coming back to it seems to sort of "reboot" this type, which takes advantage of their "quick starter" mentality.

For the "anything goes" style, I usually don't ask for new takes, but tend to go more for punches in specific sections ("That was good, but I think there needs to be a bit more energy when the second verse comes in"). These performers seem to break performances down into smaller pieces rather than thinking in terms of takes, so they fit well with a punch-in oriented approach.

With the "double peak" type, as long as the takes keep getting better in the second series of peaks, keep recording. Once you hit two or three takes in a row that don't improve on previous takes, move on. It's unlikely you're going to get anything more that's worth recording.

THE VOLUME JUMP TRICK

Here's another sneaky trick I was taught by an extremely talented engineer: Just when you think the performer is about to peak, turn up the headphone volume ever so slightly – no more than a dB or so. This raises the person's energy level an extra little bit, and often inspires what ends up being the best take.

But you only get one, or possibly two, chances to do this. It's the novelty of the change that makes the difference. So, you have to gauge precisely when that Cartier-Bresson-like "perfect moment" is about to occur. Knowing the performance curve helps you decide. With "anything goes," I wait until I have enough takes in the can that I know it's possible to put together a good part. Then I'll goose the volume a tad and do a few more takes. Sometimes these are ideal for adding that slight extra "edge" on the final verse or chorus, or elsewhere for that matter.

The "long ramp-up" type is the most difficult to anticipate. You have to choose the moment that's just before their best take. Some performers have such consistent performance curves you can almost do it by the numbers-- for example, you know that the ninth or tenth take is almost always the best one. In other cases, you just have to trust your feelings about when to do the boost.

With a "double peak" performer, it's usually pretty obvious when the second peak is happening. That's when to do the level boost.

For the "quick starter," I record a take or two, then bump up the volume a tad to see if I can get that "magic take" just before things start to fade. For the "rock steady," I'll say, "Okay, we have what we need, but let's do one more for luck," and turn up the volume a bit. Hey, if it gets a good performance, anything goes.

BREAK TIME

The performance curve also influences when the performer should take a short break, which most musicians feel the need to do occasionally during the course of a session. With the "quick starter" type, have the glass of water already set up next to the mic; once you start, you don't want to stop. The "long ramp-up" performer can sometimes benefit from working breaks into the process. This seems to impart a somewhat fresh perspective when the performer returns; they'll proceed in the direction they were going, but with a slightly different "vibe." This may give more options in the final composite performance (e.g., you can drop in the second verse from one of the post-break takes to add a bit more variety).

With the "double-peak" type, the best place to work in a break is if the second peak is slow in coming. Sometimes a break will "break the ice" and cause the second peak to shake loose. If it doesn't, then it's probably best to move to a different tune. Sometimes the planets just are not in alignment to do the perfect performance, and part of producing is recognizing when that happens.

THAT'S A TAKE

Granted, there are a lot of variables, so the above are more guidelines to get you thinking than ironclad rules. Having said that, once you become aware of this phenomenon you might be surprised at how often it is an ironclad rule. Just like some people are night people and some are day people, it seems some people settle naturally into a performance curve that doesn't vary much, if at all.

So next time you're recording, see if a performance curve manifests itself. You might really be able to use that knowledge to your advantage.

Craig Anderton is Editor Emeritus of Harmony Central. He has played on, mixed, or produced over 20 major label releases (as well as mastered over a hundred tracks for various musicians), and written over a thousand articles for magazines like Guitar Player, Keyboard, Sound on Sound (UK), and Sound + Recording (Germany). He has also lectured on technology and the arts in 38 states, 10 countries, and three languages.