Perceptual studies in concert hall acoustics

Aki Haapaniemi

Research output: ThesisDoctoral ThesisCollection of Articles

Abstract

The studies that established the known perceptual dimensions of concert hall acoustics relied on questionnaire ratings at live concerts, limited by bias effects and impossibility of perceptual comparisons, and listening tests using basic auralization methods that had limited fidelity. With modern methods, the concert hall sound field can be captured and delivered to the laboratory with a much higher degree of realism. The work of this thesis aims to further understanding of perceptual aspects of concert hall sound fields using these methods. It was found that, given a similar listening position in various concert halls, the differences between halls are more salient in the early rather than late reflections. In a discrimination task with early or late response as variable, performance was worse when discrimination was based on a variable late response, and took more time. Timbre was often quoted as a differentiating factor, suggesting that timbre differences have an important role in characterizing concert halls. Results of participant-set optimal levels between the direct sound, early reflections, and late reflections show that the shoebox-shaped hall was already closer to optimal balance, while the vineyard-style hall clearly lacked early reflections. Results were essentially independent of musical stimulus, although two contrasting excerpts were used. The differences were largest between halls,and smaller between the different listening distances. The corresponding parameter changes suggest that the distant positions in both halls lacked clarity. Analysis points to timbre as a factor alongside clarity and source width for setting the balance between direct sound and early reflections. The seat-dip effect (SDE) has been assumed to degrade the bass response of halls. However, the results of discrimination and preference tests show that the initial SDE spectrum is inaudible when sufficient reflections are provided. This leaves more freedom in choosing a hall design that provides adequate low frequencies via reflections, rather than having to design a seating area that minimizes the SDE. A wide dip was found more easily detectable than a deeper but narrower dip, and preference for uncolored spectrum emerged only for the halls with a higher clarity index. Timbre was found as a discriminating factor between direct sound spectra, alongside the perceived amount of bass. The direct sound localization threshold was found to be affected by the hall and stimulus characteristics, and an interaction between them. The estimate for a localization threshold in halls was -2.7 to 1.7 dB direct-to-reverberant ratio in the 700-4000 Hz range. The threshold was notably lower for a hall with prominent early lateral reflections; a weighting that marginalized over lateral directions improved the fit between parameter values at threshold across halls, suggesting that lateral directions may be of less interference to localization than directions closer to the source.The difference between precise and imprecise thresholds was about 7 dB, suggesting that previous reports of a sudden change between imprecise and precise localization may be overstated.
Original languageEnglish
QualificationDoctor's degree
Awarding Institution
  • Aalto University
Supervisors/Advisors
  • Lokki, Tapio, Supervising Professor
  • Lokki, Tapio, Thesis Advisor
Publisher
Print ISBNs978-952-60-8279-0
Electronic ISBNs978-952-60-8280-6
Publication statusPublished - 2018
MoE publication typeG5 Doctoral dissertation (article)

Keywords

  • room acoustics
  • concert halls
  • room impulse response
  • spatial audio
  • seat-dip effect
  • auditory localization

Fingerprint Dive into the research topics of 'Perceptual studies in concert hall acoustics'. Together they form a unique fingerprint.

Cite this