top of page


My research examines non-articulatory tongue motion during clarinet performance, with a focus on multiphonic production. I am compiling a catalogue of dyad multiphonics (see resource page) and providing strategies on learning to produce each multiphonic. The result from this research is a new resource for performers, teachers, and composers who are interested in incorporating clarinet multiphonics in their repertoire.

Research abstract:

Clarinet multiphonics have become increasingly popular among composers since they were first introduced in the 1950s. However, it is a topic poorly understood by both performers and composers, which sometimes leads to the use of acoustically impossible multiphonics in compositions. Producing multiphonics requires precise manipulations of embouchure force, air pressure, and tongue position. These three factors are invisible to the naked eye during clarinet performance, leading to many conflicting theories about multiphonic production strategies, often based on subjective perception of the performer. This study attempts to observe the latter factor—tongue motion—during multiphonic production in situ using ultrasound. Additionally, a multiphonic catalog containing 604 dyad multiphonics was compiled as part of this study. The author hypothesized that nearly all, if not all, of the multiphonics can be produced using one of four primary production strategies. The four production strategies are: (A) lowering the back of the tongue while sustaining the upper note; (B) raising the back of the tongue while sustaining the upper note; (C) changing the tongue position to that of the lower note while sustaining the upper note; and (D) raising the root of the tongue (a sensation similar to constricting the throat) while sustaining the upper note. To distill production strategies into four primary categories, the author documented his perceived tongue motion over twenty repetitions of playing every multiphonic in the catalog. These perceptions were then confirmed or corrected through ultrasound investigation sessions after every five repetitions. The production strategies detailed in this study are only for finding the correct voicing to produce the multiphonics. The catalog compiled during this study is organized using two different organizational systems: the first uses the traditional method of organizing by pitch; the second uses a fingering-based system to facilitate the ease of finding multiphonics in question, since notated pitches of multiphonics often differ between sources.

The complete document is available here.

Below is a short video showing a simplified version of how the imaging process is carried out.

The transducer is stabilized under the chin using the headset. The image of the tongue is displayed on the monitor. The right side of the image corresponds to the front of the mouth.

bottom of page