He invents statistical methods for helping missiles hit their targets more precisely, and she plays the oboe in symphony orchestras. Together, they hope to achieve a synthesis of techniques.
Aside from the fact that missiles and oboes are more or less cylindrical, these objects have little in common. But Dr. James C. Spall and his wife, Katherine Ceasar-Spall, may have found a deeper kinship between missiles and oboes, one that could lead to improving them both.
Spall is a statistical analyst with the strategic systems department of the Johns Hopkins University's Applied Physics Laboratory in North Laurel. On behalf of the Defense Department, Spall applies mathematical techniques called optimization theory and regression analysis to weapons systems to find ways to improve the accuracy of missiles. Now he is using mathematical techniques to analyze the process of making oboe reeds.
Ceasar-Spall, who has played the oboe with the Richmond Symphony Orchestra, the Baltimore Symphony Orchestra and other music organizations, teaches the difficult instrument to students. They live in Ellicott City.
The oboe requires a player to vibrate both sides of a double reed. Its distinctively subdued and somewhat melancholy sound is a vital element in works by romantic masters such as Brahms and Beethoven.
But Ceasar-Spall, like most other oboists, is compelled to spend a large part of her time making reeds for her instrument - more time than she actually spends playing it.
The worst of it, Spall and Ceasar-Spall say, is that most of the reeds an oboist makes turn out to be unusable. If a way could be found to predict whether a partly finished reed was doomed to failure, roughly three-quarters of the time that is now wasted trying to improve intrinsically bad reeds could be saved by discarding them before completing the laborious "refining" process.
Oboe reeds are notoriously unpredictable. A finished reed that starts its concert career by emitting a few perfect notes or bars of music may abruptly fail, whining piteously as it dies. Another reed may yield pure and beautiful notes for a symphony or even a season.
Commercially made oboe reeds are expensive and often of poor quality, Ceasar-Spall said, so most professional oboists make their own from a special type of cane.
"Oboe cane is a species called Arundo donax, a plant similar to bamboo," Ceasar-Spall said. "The best plants traditionally come from the French Riviera, although the cane plantations there were heavily damaged during World War II, and most oboists agree they have never been as good as they were before the war. In 1979, there was a great batch, but no cane of the same quality has come along since."
To make a reed, the oboist slits a length of cane lengthwise into three pieces; each slice, with luck, may eventually find its way into an oboe. A razor-sharp gouge is then used to trim cane away from the inner side of a slice until it is just the right thickness. At that point, most oboists score the slice lengthwise and bend the halves together. Any excess cane along the line where the edges join is trimmed away. The reed is tied tightly around a metal tube that is placed in the mouth end of the oboe, then scraped until it is playable, a time-consuming step.
Many a night Spall watched in sympathy as his wife toiled over her never-ending task of making reeds, until it occurred to him that a mathematical model might be capable of predicting successes and failures before they became apparent to the reed maker. The result was a technical paper published last summer by Spall and his wife in the Journal of the American Society for Testing and Materials; the gist of the paper was reported in the much more widely read Chemical and Engineering News.
In the report, Spall described four mathematical models he had devised for evaluating oboe reeds. Each method was tested by Ceasar-Spall, and the couple cited one of these models as proving to be reasonably predictive.
The object of regression analysis is to create a mathematical framework within which several independent variables are simultaneously gauged for their interactive effects on an outcome.
The variables chosen by Spall for his models are all based on subjective assessments by an individual reed maker, not on objective quantitative measurements. That, he concedes, is a weakness of his system.
The reed maker is asked to assign a numerical value to the amount of opening at the end of a reed, the appearance of the wood grain, the pressure needed to gouge the cane, the shininess of the cane, and most of all, the sound the reed makes when it is first blown, before it is fine-tuned and fitted to an instrument.
Spall acknowledged that a better way of measuring reed variables was needed to refine the model or develop a better one.
Pub Date: 12/14/97