Sound quality evaluations using interactive simulation : innovation report
Giudice, Sebastiano D. (2009) Sound quality evaluations using interactive simulation : innovation report. EngD thesis, University of Warwick.
WRAP_THESIS_Giudice_2009.pdf - Submitted Version - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Official URL: http://webcat.warwick.ac.uk/record=b2285615~S15
Sound Quality engineering (SQE) is a discipline that should be embedded within Noise Vibration
and Harshness (NVH) engineering. Its purpose is to tailor and enhance a vehicle’s sound in order for it
to meet and / or exceed the customers’ expectations of the car and of the brand. NVH engineers need to
use the opinions of customers, key decision makers in the organisation and those of their colleagues to
set objective NVH targets for new vehicles. Their opinions are captured through jury evaluations.
Interactive simulation is a new approach to presenting sounds to a jury. It enables assessors to
evaluate existing and concept cars generated in real time and before the manufacture of physical
prototypes. This document summarises the research carried out for the EngD research programme. Its
aim was to develop methodological approaches for setting up evaluations using interactive simulation,
optimise its data capture and analysis capabilities and provide insight, to NVH engineers, into how
assessors evaluate sound quality.
The first stage of the research involved benchmarking how the interactive simulation tool was being
used by its developers, and how NVH targets are set within an OEM. This provided the foundation
upon which to build the new methodology. The benchmark was compared against methodological
approaches used in experimental psychology and in other sensory perception practices. This identified
that principles for the design of evaluations had not been considered and appropriate statistical analysis
techniques were not being implemented. Therefore it was not possible to ensure if the differences
observed in the results were significant or not. It also became apparent that as each assessor was free to
drive vehicles however they wished, the NVH engineers would not be able to link the assessor’s
subjective impressions with the acoustical stimuli used to form an opinion. This was due to the lack of
observational methods that could be applied. In addition, the use of this form of interactivity was novel
compared to the approaches available before the introduction of the simulation. Therefore it was not yet
understood how it could influence the outcome of the evaluations.
An iterative approach was adopted to develop both tools and methods. Following the
benchmarking stages, experimental design principles were implemented and a structured briefing
method was formulated for the first time. These contributed to the overall methodologies and were
also used to ensure the studies conducted as part of this research programme were free from
experimental biases. This stage was followed by identifying a statistical analysis framework which can be
used to study the assessors’ subjective impressions. These contributions enable the NVH engineer to
understand if the differences observed between sounds are significant or not.
The next phase was to understand how to best capture subjective opinions. Recommendations for
this were based on the purpose of the evaluation. For the benchmarking stages of the NVH target
setting process, many cars are often evaluated and therefore the duration of the evaluation can be a
concern. However, the desired level of accuracy of the results is not as high as it is for the validations
stages, upon which key engineering decisions are based, and also fewer cars are evaluated. Taken
together it was possible to recommend the use of scaled interfaces for the benchmarking stages and the
paired comparison method for the validation stages. The former takes less time to complete than the
latter, which is however more accurate. The data capture capabilities were further enhanced through the
development of a driver observation module; this enabled the assessor’s assessment strategies to be
recorded. The data collected provided insight into how assessors drove with the aid of new visualisation
and analysis techniques developed.
Given the availability of these new methods and tools it was then possible to use the simulator to
observe the influence of interactivity on the outcome of the evaluation. This demonstrated that
assessment strategies can vary depending on the question asked to the assessor. The study indicated
that assessors associated the refinement character of a car with driving conditions other than those
presented to them in traditional evaluations.
Through the research and developments of this EngD programme, NVH engineers can now
observe with confidence if differences between sounds are significant and they can see and hear how
the assessor made decisions. Therefore they can now link subjective impressions formed with the
stimuli evaluated by the assessor, hence an effective way of using the simulator has been proposed.
|Item Type:||Thesis or Dissertation (EngD)|
|Subjects:||H Social Sciences > HF Commerce
T Technology > TL Motor vehicles. Aeronautics. Astronautics
|Library of Congress Subject Headings (LCSH):||Automobile sounds -- Evaluation, Consumer panels|
|Official Date:||January 2009|
|Institution:||University of Warwick|
|Theses Department:||School of Engineering|
|Supervisor(s)/Advisor:||Jennings, Paul Anthony, 1964-|
|Sponsors:||Engineering and Physical Sciences Research Council (EPSRC)|
Actions (login required)