The Binaural Project

This is the startpage for the Binaural Project site.

The Binaural Project was launched in 2011, with the aim to gain more insight in the relationship between perceptual aspects in concert listeners and features in the binaural signal. A basic activity has been to acquire binaural data for the analysis, in particular to collect binaural recordings form concerts with symphony orchestras in big halls.  So far, the Binaural Project has concentrated on spatial listener aspects. However, all hearing aspects relevant to music performance or speech, listener aspect as well as musicians’ hearing aspects, are in the general scope of the project. 


Figure 1, left to right:: Scheme of the measurement set up; In-ear microphones connected to wav-recorder; Microphones placed in ear canal.

INTRODUCTION

There is a common understanding among acousticians1 that a symphony orchestra can produce a broad sound image in one concert hall and a narrow or frontal sound image in another concert hall, even if playing the same piece. This perceptive effect is referred to as Apparent Source Width (ASW), or Source Broadening. A similar common understanding goes for perceived Listener Envelopment (LEV): One concert hall offers the sensation that musician and listener is in the same space, enveloped by the same physical environment, while another hall fails to do so. A concert hall’s ability to add source broadening and envelopment to the listening experience during a music performance, can be measured by analysing room impulse responses, according to ISO-33822. Early lateral fraction (LF) and early inter-aural cross-correlation (IACCE) are indicators for source broadening, while late lateral strength (LLG) is an indicator for listener envelopment. Beranek (2008)3 suggested that listener envelopment can be predicted from a combination of late strength (Glate) and late inter-aural cross-correlation (IACCL). More recently4, Skalevik has investigated the time-varying inter-aural cross-correlation, IACC(t), in binaural recordings during symphony orchestra performances in concert halls, including several well-known halls. The objective is to see whether spatial listener aspects correspond to any features in IACC(t). One would like to know how big the hall-to-hall differences are compared to the variation due to music content. Another interesting question arisen during the Binaural Project is whether or not source broadening and source localization are mutually exclusive.

HALLS

As of April 2017,  a total of 10 hours of binaural signal have been aquired, from a total of 10 big concert halls (in addition to data from smaller halls):

City

Hall

Bergen

Grieghallen

Berlin

Philharmonie

Boston

Symphony Hall

Chicago

Orchestra Hall

Helsinki

Music Centre

Kristiansand

Kilden

New York

Avery Fisher Hall

Oslo

Konserthuset

Paris

Philharmonie

Stavanger

Fartein Valen

TERMS AND DEFINITIONS

Some relevant terms and definitions are presented in Table 1.

Table 1 Terms and definitions

Term

Description

 

BRIR

Binaural Room Impulse Response, measured from an impulse with a pair of microphones at the ears of a dummy head or head of a real person

IACC

Inter-aural cross-correlation, a value between -1 and +1, the correlation between two normalized signals in a binaural pair; it can be interpreted a measure of the balance between Median Energy and Lateral Energy, see below, where +1 indicates all Median Energy, no Lateral Energy, while –1 indicates all Lateral, no Median Energy, and ‘0’ means that Median Energy equals Lateral Energy.

 

IACCE

IACC ‘early’, i.e. IACC from the first 80ms of a BRIR

 

IACCL

IACC ‘late’, i.e. IACC from the BRIR part after 80ms

 

IACCA

IACC ‘all’, i.e. IACC from the whole BRIR

 

IACC(t)

IACC as a function of time, i.e., if not otherwise noted, a sequence of IACC calculated for every 100ms of a binaural signal

 

Median plane

The vertical plane perpendicular to the lateral axis

Lateral

Median

Energy

Arrows indicate concentric cones defining directions from where incedent sound contributes equally to Lateral Energy and Median Energy, depending on frequency

 

The concentric cones are known as the Cones of Confusion, since incident sounds from two different directions along a cone is hard to distingush from one another 

 

LM(t)

Median signal level, the sum signal of the binaural signal, and the level of energy arriving from the Median plane and which contributes to increased IACC, ultimately IACC=1 in absence of lateral reflections

 

dL(t)

Level step at time t, i.e. the difference in signal level (dB) between the RMS levels from the periods [t-0.1, t] and [t-0.2, t-0.1]

 

dLM(t)

Level step in LM(t)

 

IACCE(t)

Statistic of all IACC(t)-values occurring when dLM(t) >6dB, interpreted as direct sound and early reflections

 

IACCL(t)

Statistic of all IACC(t)-values occurring when dLM(t) < -3dB, interpreted as reverberant sound, i.e. late reflections

 

IACCE3, etc.

Extension ‘3’ indicates average from octave bands 500, 1000 and 2000Hz

 

MF

Median energy fraction, 0.5∙(1+IACC)

 

BLF

Binaural Lateral Fraction 0.5∙(1-IACC)

 

LF

Early lateral fraction, the ratio of early lateral energy to all early energy, measured with a figure-8 microphone and an omni-directional microphone, from the first 80ms of an impulse response

 

LLG

Late lateral Strength, i.e. the lateral energy in the part after the first 80ms from an impulse response, normalized to the energy of direct sound from the same source at 10m distance

 

Glate

Late Strength, i.e. the total energy in the part after the first 80ms from an impulse response, normalized to the energy of direct sound from the same source at 10m distance

 

ASW

Apparent Source Width, or Source Broadening, the perceptual effect that a sound source appears wider than it is, and the it tends to broaden with increased loudness, depending on early lateral reflections, thus measurable with LF or 1-IACCE

 

LEV

Listener envelopment, the sonic impression that musician and listener is enveloped in the same space, depending on the strength of late lateral reflections, thus measurable with LLG or a combination of 1-IACCL and Glate

 

m, s, CI

Statistical parameters: average, standard deviation and confidence interval around m at 5% significance level, thus for population N, CI is m±1.96∙s/N0.5

 

 

...to be continued….

 

References

1. Marshall, A.H., Barron, M., "Spatial responsiveness in concert halls and the origins of spatial impression", App. Acoustics, 2000;62(2):91-108.

2. International Standard, “ISO-3382 Acoustics – Measurement of Room Acoustic Parameters – Part 1, Performance Spaces”, 1st edition (2009)

3. Beranek, L., Concert Hall Acoustics 2008, J. Audio Eng. Soc., Vol. 56, No. 7/8, 2008, Concert Hall Acoustics 2008  (source 1) ( source2  download in new window)

4. Skålevik, M., Can source broadening and listener envelopment be measured directly from a music performance in a concert hall? Proc. Inst. of Ac. (IOA), Vol. 37. Pt.3 2015, paper 34 http://www.akutek.info/Papers/MS_Spaciousness-meter.pdf

5. http://www.akutek.info/demo_files/binaural_hearing.htm

6. http://www.akutek.info/demo_files/binaural_localization.htm

7. http://www.akutek.info/demo_files/spatial_impression.htm

8. http://www.akutek.info/demo_files/binaural_localization_cues.htm

9. About binaural hearing models, the Jeffress model, see below

 

 

 

Text Box:

HOME

 

START

Page created 05.04.2017

Latest change 06.04.2017

Binaural project link

BRIR buildup