VITA 2010

	Vienna Talk 2010 on Music Acoustics "Bridging the Gaps" September 19–21

List of Vienna Talk 2010 presenters

Amir, Noam (P,O)

Presentation type: Poster

Improvements on the use of maximum length sequences in acoustic pulse reflectometry

Acoustic Pulse Reflectometry (APR) is a common method used to measure the impulse response of a system, such as a room, tube, or musical wind instrument. Classically, it is based on the generation of one or many acoustic pulses and measuring the reflections generated in the object. Today, APR encompasses the use of more complex excitation signals, the most common of them being either a swept-sine or a Maximum Length Sequence (MLS). Both methods have been found to excite nonlinearities in the instrumentation system, most often in the loudspeaker. When using MLS, these nonlinearities crop up as spurious, artifactual peaks in the impulse response. In this paper we describe an improved MLS signal, termed here WMLS. This signal is obtained through a specific form of lengthening the original MLS signal using a tunable parameter. The net result is an impulse response with greatly reduced non-linear artifacts. Correct selection of the WMLS signal can bring the artifacts down to the level of the background noise, effectively eliminating them altogether. We present a comparison of measurements using both standard MLS and WMLS excitations, demonstrating the feasibility of WMLS and the improvement in the resulting signal to noise ratios.

Presentation type: Oral / Invited

Characterizing and improving vibrato in the singing voice

Vibrato is widely used by singers and instrument players to embellish their performance, most often in sustained notes. In this paper we describe our ongoing research on various aspects of this subject. Our initial attempts to characterize vibrato stemmed from a search after quantitative measures for vibrato quality, in order to compare vibrato before and after vocal warmup. A set of measures in time and frequency domain was defined, and used to evaluate the vibrato quality of a group of singing students. These objective measures were later correlated with subjective evaluations by singing teachers, and combined through an optimization process by linear weighting to obtain a single figure representing vibrato quality. This was implemented in a real-time “vibratometer” giving the singer visual feedback on their vibrato. Our further research examines vibrato characteristics of several singing styles – classical, jazz and pop. Sustained notes sung by well known singers in commercial recordings are analyzed. At this stage we present the methodology used to distinguish the singer’s vibrato from the multitude of accompanying instruments, and various parametric methods for quantifying the vibrato contour.