Human Perception of Combined Sound and Vibration

Engineering Integrity Society

Tuesday, 19 April 2005

Millbrook Proving Ground


ABSTRACTS

"Subjective Equivalence of Sound and Vibration in Vehicles"
J. Giacomin and M. Ajovalasit, Perception Enhancement Systems Research Group
Department of Mechanical Engineering, The University of Sheffield

Human subjective response to sound and vibration is an important consideration in the design of automobiles. For sound, methods such as dB(A), Stevens Loudness and Zwicker loudness are regularly used to evaluate subjectively perceived intensity. For vibration, the frequency weightings defined by standards ISO 2631 and ISO 5349 are widely applied for the same purpose. There are currently, however, no methods for evaluating the two stimuli in combination. When considering the effect of the combined stimuli, a first question is the equivalence relationship between the two. Knowledge of this relationship has several engineering uses including the identification of which stimuli, the sound or the vibration, requires the greater investment of resources during the vehicle development programme.

Research was performed to determine the equivalence relationship between steering wheel vibration and sound in automobiles. The primary objective was to define the relationship as a function of the statistical properties of the stimuli. Secondary objectives included the identification of optimal intensity metrics, and an evaluation of the stimuli range occurring in automobiles. The experiments were performed using a simulator and headphones, and involved 20 subjects for each test. Stimuli were used which reproduced a constant speed test over a tarmac road surface, a constant speed test over a metal bar and stoplight idle conditions. Regression expressions of subjective equivalence were developed for all three stimuli types using both unweighted and frequency-weighted data. Subjective response to the vibration was found to increase in relative importance with respect to the sound for short duration, transient, stimuli.

"Human response to combined steering vibration and sound, and fundamentals of cross-modal and contextual interactions"
Dr Neil J Mansfield (Loughborough University)

Human occupants of vehicles are bombarded with signals through all sensory systems simultaneously. Some of these (e.g. auditory and tactile signals) are inextricably linked such that they are almost always experienced in combination whilst others are not necessarily connected (e.g. visual and auditory systems) but simultaneously demand the attention of the driver. It is often desirable to study perceptual channels in isolation with juries, in the laboratory or during field trials. The benefit of simplification is to allow for focused product optimisation, but the cost is a reduction in realism. The presentation will use case studies of laboratory and test track research to demonstrate the importance of considering cross-modal interactions (i.e. interactions between different senses), the importance of study context and the importance of the attention of the study participants in their opinions of their physical environment.

"Developing best practice for use of an interactive NVH simulator"
P Jennings (Warwick University)

An NVH Simulator has been developed to offer an alternative approach to conventional automotive sound quality techniques, such as jury evaluation in a listening room or on-road studies. This facility immediately offers opportunities for NVH evaluation through its functionality. It will improve decision making within vehicle development, and will allow a greater understanding of customer perception.

However to maximise the benefits, a robust methodology for its use needs to be developed. A number of methodological and psychological factors can affect the outputs. This presentation will describe results from, and plans for, a programme of research work that is being carried out at the University of Warwick to develop the necessary methodology.

"A Functional Approach to NVH Driving Simulation"
R Williams & M Allman-Ward (Sound & Vibration Technology)

In today’s highly competitive automotive market, a strong brand identity is essential. The character and refinement of the sound of a vehicle is becoming increasingly important to automotive OEMs as a product differentiator. It is recognised that whilst objective metrics can describe some aspects of human perception of sound (such as loudness), currently there are no robust metrics that can describe whether the character of a sound is appropriate for the brand Engineers therefore have to assess their data subjectively, i.e. listen to it or experience it, in order to assess possible customer reaction to the product and to make the correct engineering decisions.

Traditionally, two methods are available to evaluate sound quality subjectively: driving the actual car on the road, or listening to sound recordings of important driving conditions. It has been well documented that the other stimuli present when driving a car influence the perception of the sound and that evaluating sounds whilst sitting passively in a listening room does not give the same impression.

In order to address these problems, SVT developed NoViSim, a vehicle noise and vibration simulator. In NoViSim, sound, vibration and visual stimuli are provided whilst the assessor freely "drives" the vehicle through a virtual environment. It therefore adds two new dimensions to the to the vehicle NVH process, interactivity and context. The aim of NoViSim is that sufficient additional stimuli are provided to ensure that the subjective assessment of the sound is the same as it would be when driving the real vehicle on the road.

NoViSim has already been used for engineering purposes, including high level powertrain sound quality target setting and the selection of alternative suspension designs at the concept stage in a new vehicle programme. NoViSim also has excellent potential as a research tool since it offers a totally configurable multi-modal environment where any combination of sound, vibration and visual stimuli can be experienced whilst driving.

Current research work including understanding how customers make decisions about sounds and the development of robust methods for jury evaluation in an interactive environment is described in a separate paper. This presentation focuses on the functionality and the engineering application of the simulator.

"Application of source path contribution methods to the NVH perception in vehicles"
K.B.Ginn, Mehdi Batel, Alun Crewe (Brüel & Kjær A/S, Denmark)

The first stage in the evaluation of sound and vibration in vehicles is the acquisition of good sound and vibration data. This paper illustrates by means of practical examples, some of the newest tools at the disposal of the NVH engineer. Innovative transducers and instrumentation were used in various methods of Source Path Contribution analysis. Data acquisition was performed on a test track as well as on a laboratory test rig. Acquired data was used both for Sound Quality assessment of the car and Source-Path-Receiver modelling. The results presented include noise and vibration contributions from various sources in the cabin and how these results could be used to assess the driver’s perception of vehicle comfort.

“Methodologies for the analysis of NVH perceived quality”
V. Falasca & F. Ferrian (Centro Ricerche Fiat)

The attention to the perceived quality, e.g. the quality as it is evaluated by the final customer, is becoming more and more important for the car makers.

Acoustic and vibrational aspects play traditionally an important role in determining the “image” of the car as it is perceived by the final customer.

As a consequence, it is necessary to define a criteria for the evaluation and thus design of vehicle vibroacoustic performances.

In the first part, the presentation will describe methodologies implemented in CRF in order to:

The second part will explain how the development of these methodologies has permitted to identify two comfort quality indexes: IQV (Vibrational Quality index) and IQA (Acoustic Quality Index). It will be described how these indices have been finalised and applied.

Finally, the last part will describe how bench testing can be used toghether with Quality indexes to help and shorten design phase of vibroacoustic vehicle bahaviour.

In particular will be presented a bench testing methodology able to make an objective and quick assessment of the vehicle body flexibility. The focus will be placed on the definition of a synthetic numeric index, calculated from accelerometric transmissibility functions, measured on trimmed vehicle on 4-poster test rig.

Experimental procedure, data analysis, index extraction and some applications of the methodology will be presented.


"Driveability“ - Efficient Design of Brand Specific Attributes
Dr. Peter Schöggl, Dep. DV, Head of Vehicle
Erik Bogner, Manager Driveability and Simulation
(AVL List GmbH, Austria)

Car owners nowadays place very high demands on the quality of the driving experience. Manufacturers are faced with the huge challenge of constantly increasing product quality coupled with producing special brand distinctiveness in ever shorter development times. New processes and new automobiles are needed to do that.

Shifting development processes from the road to the rig, expanding the application of realtime simulation and “virtual engineering” together offer tremendous potential. They will not be possible, however, until objective assessments are available for all the target groups because subjective vehicle evaluations cannot be obtained in the laboratory.

The report describes driveability criteria that are responsible for driving experience and illustrates examples of optimum configurations. It also describes processes with which the parameters of driveability quality and driveability character can be made objectively measurable. The report shows how development/improvement can be achieved also in the laboratory and the measures that can be used to obtain a special driving experience quality.

Measurement parameters for driveability on the test bed make it possible to significantly tighten development processes if realtime models are available at the same time. Engine and drivetrain modules can be evaluated and optimised in terms of their influence on comfort. The engine, drivetrain and vehicle can then be evaluated much earlier. Engine management parameters can be tuned manually or automatically and at the same time emissions and fuel consumption can be taken into account. Parametrisable realtime models make virtual product optimisation possible.

Shifting development work from road to rig offers effective advantages in terms of development costs and development time. It means there is then more time for vehicle development and brand-specific design to achieve the greatest possible customer satisfaction.