Highlights Of The Head Acoustics User Group Meeting Of November 9, 2000

Contents

• Sound Quality of Theaters and Music Halls
• On the Perception of Transients: Discrimination of Spectral Shape and Sensitivity to Temporal Sequences
• Utilizing the MS Office Tools within ArtemiS
• Integrating ArtemiS and SQlab in a Multi-Location Laboratory Setting

Sound Quality of Theaters and Music Halls

Dr. Chris Jaffe, Founder and Principal of Jaffe Holden Acoustics and Distinguished Visiting Professor at Rensselaer Polytechnic Institute.

Dr. Jaffe provided an overview on the acoustical design considerations and acoustical measurements for performing arts spaces. He also presented many examples of successful projects that his firm has worked on throughout the world.

Studies have shown that the great concert halls of the world share similar acoustical properties that can be defined, quantified and then reproduced when designing a new hall or auditorium. The three key metrics used to measure the acoustical performance of a concert hall are: liveliness, intimacy and warmth. Liveliness is a function of the reverberation time, intimacy is related to the initial time delay gap, and warmth is defined as a ratio of the bass reverberation time to that of the mid-frequencies. Studies of the great music halls have consistently shown that their acoustical response with respect to these three metrics falls within a specific range.

Historically, the optimal design for a concert hall has been a "shoe box" design; long and narrow, with the stage located at one end. There are obvious drawbacks to this design, especially when applying it to a large scale. The challenge facing architects and acoustical engineers is to duplicate the acoustical response of the shoebox design in a contemporary, flexible and audience friendly design.

Acoustical engineers must be able to translate desired acoustical qualities into physical design parameters that an architect can adhere to. The architect can then vary the spatial features such as the location of reflective and absorptive surfaces relative to the audience to achieve the desired acoustical response. If physical limitations preclude incorporating the desired design features, the correct acoustical response and ambience can be achieved electronically using a system of microphones and loudspeakers.

Recently Dr. Jaffe and Wade Bray (HEAD acoustics) worked with a group of graduate students at Rensselaer Polytechnic Institute to measure the acoustical response of Bass Performance Hall in Fort Worth, Texas. The group used an SQlab, two HEAD Measurement Systems, and several accelerometers, along with a laptop loaded with ArtemiS. The purpose of the study was to determine the impulse response and transfer function between the hall, the stage and the reverberation chamber behind the shell. The students also used a NoiseBook to evaluate speech intelligibility at various locations in the seating area.

On the Perception of Transients: Discrimination of Spectral Shape and Sensitivity to Temporal Sequences

Dr. Gregory Wakefield - Associate Professor of Electrical Engineering and Computer Science and Co-Director of the MusEn Project at the University of Michigan.

The effect of phase on the perception of transients.

Relative to steady state analysis, very little is known about the perception of transients. In typical power spectral density analysis much of the information regarding phase is disregarded. Recent studies with 4-ms clicks have tried to determine the importance of phase information when analyzing transients. In other words is the phase important, or is it just spectral shape?

The results of these studies show that the phase of a transient can influence the ability to distinguish spectral changes. The magnitude of these effects varies between test subjects and is more pronounced in the lower frequencies (below 4000 Hz). The effect of the phase changes is also wiped out in signals with durations less than two milliseconds.

Source perception and object identification.

Research at The University of Michigan using singers in the School of Music as subjects, sought to identify the characteristics of a singers voice that allow you to identify the singer throughout their entire range.

A three-note scale of various students was recorded and a modal distribution, akin to the spectrogram, was used for analysis. (The Modal distribution is a nonlinear distribution that does not distort frequency or time like a conventional spectrogram and is therefore a very useful tool for analyzing signals in music.) A source filter model was then developed using the extracted frequency modulation from the singer's voice as the driving function and collapsing the remaining spectral variations of the glottal source and vocal tract into a composite transfer function. An average of the spectral components across all of the singers was calculated and removed from that of each singer. The resulting residual spectrum could be used to identify the singer with approximately 95% accuracy.

The characteristics found in the residual spectrum have many practical applications. A singer's signature spectrum can be morphed with a passage from a song to let the singer hear how their voice should sound in different registers. It may also be possible to develop an objective metric to diagnose and provide insurance coverage for spasmodic dysphonia using characteristic residual diagrams of the disease.

Utilizing the MS Office Tools within ArtemiS

John Harrington, Lead Acoustical Engineer at Dell Computer

The PC industry has historically used sound power as a measure of the acoustical performance of their products. Dell Computer recently instituted a binaural sound quality program that utilizes ArtemiS and a HEAD Measurement System (HMS III). Dell also used the export functions of ArtemiS, Visual Basic Programming, and HEADbase to develop a complete data collection, analysis, reporting and storage system.

Dell's new data analysis and reporting system utilizes a four-monitor computer system to simultaneously open and run ArtemiS, HEADbase, Microsoft Office, and NT Explorer. An MS Excel template was created utilizing the Excel export function in and Visual Basic. The template automates the various graphing, report writing and custom analysis functions for Dell's sound quality program.

Rather than developing a new database for their file tracking and organizing, Dell utilized HEADbase. John cited the following advantages of HEADbase:

• It is easy to create different databases.
• HEADbase is easily networked.
• A powerful and easy to use text search is included.
• HEADbase supports drag and drop functions with MS Office products.
• It is inexpensive off-the-shelf technology (saving months, if not years in development time).

Since it's inception, Dell's sound quality program resulted in a 40% reduction in loudness as well as a significant reduction in tonal components for new desktop computers.

Integrating ArtemiS and SQlab in a Multi-Location Laboratory Setting

Tony Nava, NVH Development Engineer at Visteon Corporation

The Climate Control Systems Group at Visteon recently standardized the way they collected, analyzed, stored and reported multi-channel data with a NVH System including ArtemiS, SQlab and NoiseBook. Prior to implementing this new system their multi- channel recordings were stored on DAT and analyzed using a variety of spectrum analyzers, real-time analyzers and a Binaural Analysis System (BAS).

Visteon has several NVH data collection locations including 3 anechoic chambers in 2 buildings, and field test units. Analysis is performed at each site and the various formats utilized by these different systems made data storage and transfer between the sites difficult. The non-standardized data and output formats also made report generation cumbersome and time consuming. The problems were compounded by the fact that BAS, being a stand-alone system, had to be used on location and could only be used by one engineer at a time.

Installing ArtemiS on a network with a multiple user license proved to be an ideal solution for Visteon. The network installation allows multiple users to perform analysis wherever it is required and facilitates the transfer of data and reports between the various locations. ArtemiS also supports a variety of data formats making it easy to share data and analyses with customers (both internal and external) and is compatible with existing BAS data. ArtemiS's export utilities automate the task of providing customers detailed reports that can include, among other things, graphical files, audio files, charts and raw data files.

Visteon also standardized their data acquisition equipment by using SQlab and Noisebook. An SQlab is now installed at each anechoic chamber and an additional unit is used for field-testing. This set up gives the engineers and technicians a uniform interface and common data and storage formats at each location. SQlab has proven to be a flexible, user-friendly system for recording multiple sound, vibration and tach inputs. When sound only (or sound / tach only) recordings are required Visteon uses NoiseBook. The data files created by NoiseBook are completely compatible with the data files obtained with ArtemiS and the SQlab and can be integrated seamlessly into the rest of the system.

Back to the HEAD acoustics news page | Back to the user group meetings page