Sampling Giveaways

Before you hyperventilate looking for the freebies, maybe I should have titled this “Sampling Dead Giveaways.” Circumstance occasionally finds me evaluating sequenced orchestral emulations (both mockups and for-real product) to help spot those elements which betray the artificiality. Unfortunately, there are usually a few. This isn’t a knock on those who face the difficult task of faking the beauty and complexity of a live ensemble with little more than high-resolution “sonographs” of its colors and effects. If you have never really sat in the player’s chair and learned the tricks of the trade behind the sound and phrasing already in the psyche of your listener, here are a few tips (not exhaustive and in no particular order) on how to improve your sampling and sequencing snookery:

Non-tapered note endings: To achieve graceful and authentic (not the gulping-for-air variety) note endings, live players almost always taper off the phrase-ending notes, and in a manner relative to the tempo. While most synthesis technology allows tempo-agnostic control over release times, most samples do not. The note-off command in the midi spec makes no allowance for rolling off certain note endings with a very short diminuendo, so you will want to find a way to achieve one.

“Ghosted” short notes: All things being equal, the short notes in a mixed phrase tend to get swallowed up in the reverberant sound. Players have learned to slightly accent them to give them the same perceived presence as their longer siblings. When sequencing, listen for “ghosted” notes and goose their velocities as needed to bring them out.

Uncool Swing: Good swing is highly tempo-dependent. The slower the time, the more uneven the long/short-long/short ratios of the note durations. The faster the time, the more even, to the point where be-bop became essentially straight eighths. Compensating for weak keyboard or improv chops by sequencing slow and playing back faster can, if not done with great care, result in a weird groove.

Wind Instrument Organs: No matter how unbelievably homogeneous professional orchestral wind and brass players strive to become, no two sound exactly the same. Why should your samples? When you stack the same waveforms vertically in an ensemble, you’ve got an organ, not a consort. If you want a three-voice trombone section, use three different samples on three tracks. With unisons, unless you have a section sample that matches your individuals in timbre and character, build your unison one track at a time.

Unheeded Sample Anomalies: In their attempts to make their samples more life-like, many companies introduced effects into their products which, with constant repetition, can become downright annoying. A much-vaunted string library from years ago recorded a reverse-taper (swell) into their note beginnings. It sounded cool the first couple of times you heard it, but hundreds of times later it felt like Johnny-one-note-on-the-kazoo. Beyond the obvious considerations of intonation, quality of recording, and practicality, really give your library a going-over for consistency and, believe it or not, neutrality.

Inadequate Panning: If you can get your hands on both a midi mockup that got approved and the finished recording by live players, do it, and study them carefully. As most mockups are done in some state of haste, shortcuts and compromises are inevitable and they will teach volumes. With the live players, close your eyes and notice how you can almost point at individuals and sections in the stereo field. Then notice how difficult that becomes with poorly- or hastily-panned and processed sequences.

Diversity, Diversity, Diversity: If the three most important considerations in business are “location, location, location,” for convincing orchestral simulation it’s the big “D.” Real fundamental evidence of mediocre sequencing is an overall, dull sameness to the sound of the performance and the production. By the sheer nature of the beast, live players abhor playing the same notes, at the same dynamic, with the same phrasing and the same color, in the same combination, at the same tessitura, with the same… Well, you get the picture. Part of the solution is to avoid the temptation to cut-and-paste anything, unless you really are on a deadline (with an emphasis on dead.) Another part involves playing your notes rather than stepping them in, even if you have to “record slow/play back faster.” Innovations such as Kontakt’s “scripting,” which allows control and automation of advantageous anomalies (human-like, for instance) into the behavior of the sample library, will also go a long way toward realistic tracks.

Well, enough for now, with perhaps more to come in the future. As I said, it’s a tall order to use a machine and not spill the beans on “Is it live or is it Memorex?” However, a little common sense, some reverse engineering of the performance process, and a lot of hard work and attention to detail can go far in helping you “get real” with your tracks.