Wednesday, May 4, 2011

It's All in the . . . Timing

There’s an old joke about a comedian convention where the celebrity comics are scheduled to tell jokes at the event’s big banquet. Everyone in the audience knows all the jokes already because the jokes have been catalogued for the convention. To save time, the speakers decide to just say the joke's number instead of telling the whole joke. The first speaker is a star, a legend in the business. “Forty-seven!” he calls out to laughter all around. “One-hundred and sixty-two!” The audience explodes with deep belly laughs and leaps to its feet in an ovation. The second speaker, a newcomer to all this, thinks, “Wow, this is easy.” When it’s his turn, he strides onto the stage and in his best stage voice proclaims, “Twenty-six!” Silence. Undaunted, he tries again: “Eighty-four!” Silence again, and the audience is now shifting uncomfortably in its seats. With a mighty roar the newbie lets loose with his finale, “One hundred and thirty-five!!” Crickets. After his ordeal, the shaken comic approaches a veteran and asks, “What happened? I did everything the guy ahead of me did. But I bombed. Why?” The old comedian puts his hand on the younger man’s shoulder and says gently, “Kid, you just didn't have the timing.”

Music, like comedy, is all about the timing. And it’s just as hard to master when playing a minuet as it is to entertain an audience of seasoned showbiz types. A student can memorize the pitches and rhythms of Beethoven’s Piano Concerto No. 4 in G major, but it takes a concert-level artist to bring out the special musical expressiveness of the piece that makes it a transcendent listening experience. To do that requires a command of timing. Everyone has seen amazing videos on YouTube of young children executing incredibly difficult pieces with technical mastery. But the common complaint that fills up the comment spaces is that the performance lacks feeling and emotion. What they’re talking about is “expressiveness,” and that, it turns out, is closely related to timing.

At least that’s what new scientific research asserts. In a project conducted at McGill University, a scientist has determined that timing, and to a lesser extent volume changes, are the key ingredients listeners use to determine whether a piece of music is expressive. In other words, after working hard your whole life to a metronome to develop your timing, you may have to abandon it to become more expressive.

Dr. Daniel Levitin, Director of the Laboratory for Music Perception, Cognition, and Expertise at McGill University, devised a series of experiments to test how humans measure expressiveness in music. In one of them, he uses a set of Chopin Nocturnes, first played by a pianist with timing variations according to his taste, and then programmed as a MIDI file with no variation in timing or duration (all notes were quantized to their written, sheet-music value). Then a third track was created, which was quantized exactly halfway between the human and fully quantized versions. Not surprisingly, musicians and non-musicians ranked the original, musician-performed version as the most expressive of the three.

That by itself may not seem that revelatory, but that was just the first question in the experiment. In subsequent tests, percentages were changed, and factors other than timing were brought into the mix. In one listening test, expressive factors in the human performance were amplified by 150%, with volume and tempo variations exaggerated from the original human rendering. Interestingly, listeners found that too much expression had a counterproductive result—losing effectiveness if you take it too far. Human players, it seems, know how to appeal to human listeners in a way that machines can’t replicate. At least not yet.

Dr. Levitin gave his listening subjects music examples where only the volume was changed while the timing was maintained as it was in the original, and then the converse—where the timing was altered while the volume was kept constant. These comparisons are more difficult to hear, but listeners still came down on the side of timing as being the expressive factor in the majority of the results. In other words, of the two principal factors involved in conveying expression—volume changes and timing alterations—it is timing that humans find has the more expressive power.

In these experiments, all the listener had to do was pick the more expressive example of three. But since we’re all musicians here, let’s raise the bar. Take the test yourself, but instead of just choosing, say, answer A as the most expressive, go ahead and identify the other two choices, according to the info in the experiment's introduction. Try to identify not just the most expressive of the three, but which is the straight MIDI rendition and the other the “150% humanized” (or exaggerated) version, or which track has only volume changes and which one has timing changes. The differences are sometimes subtle, so to make it challenging, see if you can do it on one pass. It gets more difficult when the choices involve not just timing or volume, but timing and volume together, in various combinations. (The tests also included randomness in addition to uniform percentages.)

Go to the New York Times website and take the test. The article thoughtfully includes pdfs of the sheet music for further study. After you run through the experiments, try a different approach to your music. If you've been running scales to a metronome, working out with drum machines, or invoking your synchronized arpeggiator to quantized grooves, try injecting some rubato into your music. It will give you a new take on expressiveness, and it might even improve your ability to tell a joke.

No comments: