Conference PaperPDF Available

Calibrated auralization simulation of the abbey of Saint-Germain-des-Prés for historical study

Authors:
  • dieBauingenieure
Conference Paper

Calibrated auralization simulation of the abbey of Saint-Germain-des-Prés for historical study

Abstract and Figures

Over recent decades, auralizations have become more prevalent in historic research and archaeological acoustics. With these techniques it is possible to explore the acoustic conditions of buildings which have been significantly modified over time, providing that the original geometry and the acoustic characteristics of their surfaces are known. In this manner, historians are provided with the opportunity to explore lost acoustic environments of important buildings. Calibration of auralizations is necessary if one wishes to build a scientific tool rather than a simple audio novelty. In this context, a study was carried out on the Parisian Saint-Germain-des-Prés. The abbey church was begun in the 11th century, with major modifications undertaken in the 12th and again 17th centuries which resulted in changes in the acoustic conditions. A geometrical acoustic (GA) model of the church was created and calibrated, as discussed in the following section. Sec. 3, describes the validation of the calibration by means of an auralization listening test. The acoustic environment of the church as it stood before the 17th century modifications was compared to that of the current Saint-Germain-des-Prés. The calibrated GA model was modified to reflect the church’s configuration during this period in Sec. 4. When simulation results of the current and pre-modern configurations were compared, it was observed that the abbey church of Saint-Germain-des-Prés used to have perceptually shorter reverberation (T20 and EDT) and higher clarity (C50 and C80), especially in the principally occupied areas.
Content may be subject to copyright.
Proceedings of the Institute of Acoustics
Vol. 37. Pt.3 2015
CALIBRATED AURALIZATION SIMULATION OF THE
ABBEY OF SAINT-GERMAIN-DES-PRÉS FOR
HISTORICAL STUDY
BNJ Postma Audio Acoustics Group, LIMSI-CNRS, Orsay, France
A Tallon Art Department, Vassar College, Poughkeepsie, USA
BFG Katz Audio Acoustics Group, LIMSI-CNRS, Orsay, France
1 INTRODUCTION
Over recent decades, auralizations have become more prevalent in historic research and
archaeological acoustics. With these techniques it is possible to explore the acoustic conditions of
buildings which have been significantly modified over time, providing that the original geometry and
the acoustic characteristics of their surfaces are known.1 In this manner, historians are provided
with the opportunity to explore lost acoustic environments of important buildings.
Calibration of auralizations is necessary if one wishes to build a scientific tool rather than a simple
audio novelty. In this context, a study was carried out on the Parisian Saint-Germain-des-Prés. The
abbey church was begun in the 11th century, with major modifications undertaken in the 12th and
again 17th centuries which resulted in changes in the acoustic conditions.2
A geometrical acoustic (GA) model of the church was created and calibrated, as discussed in the
following section. Sec. 3, describes the validation of the calibration by means of an auralization
listening test. The acoustic environment of the church as it stood before the 17th century
modifications was compared to that of the current Saint-Germain-des-Prés. The calibrated GA
model was modified to reflect the church’s configuration during this period in Sec. 4. When
simulation results of the current and pre-modern configurations were compared, it was observed
that the abbey church of Saint-Germain-des-Prés used to have perceptually shorter reverberation
(T20 and EDT) and higher clarity (C50 and C80), especially in the principally occupied areas.
2 CALIBRATION CONCEPT
As with any scientific simulation, it is necessary to calibrate GA models. An overview of the
calibration procedure is presented in what follows. Subsequently, acoustic measurements which
served as references for the calibration are discussed. Finally, the creation and calibration of the
GA model are considered. The calibration was performed according to the previously reported
7 step procedure.3
1. RIR measurements are carried out in the studied venue. The results of these measurements
are used as a reference for the calibration.
2. The geometrical model is created and remains unchanged during calibration.
3. Preliminary acoustical properties are assigned to all surfaces, resulting in a GA model.
4. Since stochastic implementations of Lambert scattering in GA software leads to run-to-run
variations, the GA model's repeatability is quantified. These variations are then taken into
account when simulation and measurement are compared.
5. The sensitivity of the GA model to adjustments of scattering coefficients is quantified.
6. Acoustical surface properties are modified, taking into account the determined sensitivities, in
order to arrive at global mean differences between measurement and simulated results for
reverberation and clarity parameters of less than 1 just noticeable difference (JND).
7. Acoustic properties of local key surfaces are adjusted to minimize the standard deviation
(SD) of the differences for reverberance and clarity parameters.
190
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
This procedure relies on several assumptions:
Selected acoustic parameters for reverberation and clarity are sufficient metrics.
Calibration according to objective parameters within 1 JND results in a valid simulation.
Calibration of a GA model for a sufficient ensemble of discrete points (source and receiver
positions) provides sufficient confidence in the quality of the simulated RIR at other positions.
Additionally, since previous studies reported variations in RIR analysis algorithm implementations4,
it is important to use a single analysis tool for estimating the acoustic parameters when comparing
results from measured and simulated RIRs. In order to avoid issues regarding automatic noise
detection algorithms when applied to the ideal background noise-free simulated RIRs, a low level
white Gaussian noise (65 dB) was added to the simulated RIRs prior to parameter analysis.
This calibration procedure shows similarities to a recent study on another church.5 However that
study did not address calibration for auralization, but was concerned with calibration of the GA
model for reported parameters by the simulation software only, in order to predict the effect of
changes in acoustic conditions. The selected metrics were global mean EDT and C80 with 1 JND of
measured values. While these parameters agreed well, it is worth noting that T30 values differed by
over 20% for some positions. No local calibration was carried out to reduce variance.
2.1 Associated measurement
Acoustic measurements were carried out in order to serve as a reference for the calibration with the
following details. Additional details can be found in Postma and Katz3.
Signal- The Exponential Swept Sine method was employed. The sweep frequency went from 20 Hz
to 20 kHz, duration of 10 s.
Sound source- The audio output was sent to an amplifier (Servo 120a, SAMSOM) and sequentially
to two miniature dodecahedral sound sources (model 3D-032, Dr-Three).
Microphones- Four omnidirectional microphones (model 4006, DPA) and an artificial head (Neuman
KU 80 equipped with model 4060, DPA) were used.
Measurement positions- Fig. 1 shows the measurement plan. Two source positions were chosen
representing typical usage (pulpit:S1 and altar:S2). 32 receiver positions covered normal
attendance (1-2, 4-7, 9-12, 14-17, 19-20), the high altar (21-22, 24-25), and the sanctuary (26-
27, 29-30). For the dummy head, center-line positions (3, 8, 13, 18, 23, 28) were measured.
Measured parameters- RIR's were analyzed using LIMSI's in-house MatLab impulse response
analysis (IRA) toolkit. For the purpose of this study, six ISO Standard6 parameters were
calculated: T20, EDT, C50, C80, IACC early (e), and IACC late (l).
2.2 GA model
CATT-Acoustic (v.9.0.c, TUCT v1.1a) was employed to create the GA model and perform
simulations.7 The geometry of the Saint-Germain-des-Prés was determined from a 3D laser scan
point cloud as well as architectural plans and sections (see Fig. 2). The surface materials in the
abbey church were determined by means of visual inspection. Absorption coefficients were adopted
Fig.1: Source and receiver measurement plans in the Saint-Germain-des-Prés.
191
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
from publicly available databases. Scattering coefficients of surfaces were generally modeled using
the option estimate which provides a simple estimation of these coefficient based on a given
characteristic depth representative of the surface's roughness. The binaural GA simulation
incorporated the previously measured HRTF of the dummy head used during the measurement.
Ten repetitions were run of the initial configuration. Analysis of the SD calculated per position for
each acoustic parameter quantified the run-to-run variation. The sensitivity of the GA model to the
scattering coefficient was studied by running simulations of the initial GA model followed by
simulations with all scattering coefficients set to 0%, then to 99%, with absorption coefficients
unchanged. Then, absorption coefficients of the materials with the largest surface areas were
adjusted, since small variations lead to a considerable effect. After the mean reverberation
parameters (T20 and EDT) were adjusted to within 1 JND of the measured values, scattering
coefficients were adjusted to achieve mean parameters (C50 and C80) within 1 JND. Fig. 2
compares the mean measured T20, EDT, C50, and C80 to those of the calibrated GA model.
Finally, acoustical properties of local key surfaces were adjusted to minimize the SD of the
differences between measured and simulated results for the reverberance and clarity parameters.
3 SUBJECTIVE LISTENING TESTS
To evaluate the assumptions on the validity of the calibration procedure, a listening test was carried
out comparing measured and simulated RIR auralizations. It should be noted that prior to
commencing listening tests, some additional processing is required concerning the measured RIR.
3.1 Preparation of the measured RIR
The frequency response characteristics of the measurement system were compensated for by
creating an equalization filter. The measurement chain (one microphone only) was installed in an
anechoic room (IRCAM, Paris) and the RIR of the omnidirectional speaker was measured at
increments in the horizontal plane. The resulting RIRs were time-windowed to 512 pt, to remove
any reflection artifacts, from which the FFT was calculated and the mean over all directions of the
magnitude determined. A filter was generated to match the inverse of this response using the
recursive filter design yulewalk method. Non-linear frequency weighting followed a bark scale
approximation, constraining the filter's level of detail to follow human hearing sensitivity. The
resulting filter was applied to all measured RIRs prior to any spectral analysis.
Subsequently, differences in SNR between frequency bands were compensated for. The RIR
was decomposed into 1/3rd-octave band components (spanning 100 16000 Hz). The noise floor
was detected by determining the SNR for each 1/3rd-octave band. The signals were then windowed
at the point 5 dB above the noise floor, eliminating the trailing noise. The decay rate (reverberation
time) was then calculated over the entire window, and this decay rate was used to synthesize the
continuation of a noise-free reverberant tail. Since a reliable decay estimate reasonably requires at
Fig.2: (Left:) GA model of the Saint-Germain-des-Prés. (Volume ~22200 m3, 4518 surfaces).
(Right:) Comparison between measured and simulated mean T20, EDT, C50, and C80 (±1 JND).
192
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
least 15 dB SNR, 1/3rd-octave bands with SNR < 20 dB were discarded (muted in the final RIR).
This typically resulted in omitting the lowest two 1/3rd-octave bands: 100 and 125 Hz (these were
therefore also omitted from the simulated RIRs). An equal power cross-fade between the measured
and synthesized responses was applied over the last 10 dB decay of the measured and the first
10 dB decay of the synthesized response to provide a smooth transition, limiting audible artifacts.
Fig. 3 depicts the spectral magnitudes of the simulated and measured RIRs on measurement
position of source 2 combined with receiver 02 (S2R02) after this procedure.
3.2 Stimuli
The resulting measured and simulated RIRs were convolved with two anechoic audio extracts
appropriate to the acoustic function of the room: female soprano singing Abendempfindung, by
W.A. Mozart; male tenor performing A Chloris, by R. Hahn. Fig. 3 depicts the spectral composition
of the chosen extracts. As the two lowest 1/3rd-octave bands were omitted from the RIRs, the
extracts were chosen to have minimal energy in these bands. RMS of the measured and simulated
convolutions was used for normalization.
3.3 Test protocol
The test was set up as an AB comparison. Stimuli were compared for both omnidirectional and
binaural receiver configurations for two source and receiver positions (Omni: S1R02, S1R12,
S2R02, S2R12; Binaural: S1R03, S1R13, S2R03, S2R13) resulting in 16 tested pairs. Four
configurations were repeated to monitor the repeatability of responses, resulting in 20 pairs.
Additionally, participants were given three training pairs to ensure they understood the task. Results
for these three training pairs were not analyzed.
Participants were asked to rate the similarity of samples according to Reverberance, Clarity,
Distance to the source, Coloration, and Plausibility. For binaural receiver pairs, participants were
asked to additionally rate the similarity of Apparent Source Width (ASW) and Listener Envelopment
(LEV). It should be noted that the binaural head orientation in all configurations was towards S2.
Participants responded using a continuous graphic 100 pt scale, ranging from ‘A is much more ...’ to
‘B is much more ...’ corresponding respectively to values of 50 and +50, with a center 0 response
indicating no perceived difference. Presentation order and AB correspondence to simulation and
measurement were randomized. Participants were able to listen to the compared pairs as many
times as desired. Auralizations were presented via headphones (Sennheiser model HD 600) at an
RMS level of 75 dBA. The experiment took place in an isolation booth, ambient noise level
<30 dBA. The 12 participants (mean age: 39.6 SD: 16.7) all reported normal hearing.
Fig 3: 1/3-octave RMS power spectrum for (Left) measured and simulated RIR for the early
(0-200 ms) and late (200-3000 ms) parts of the RIR: S2R02. (Right) anechoic stimuli.
193
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
3.4 Test results
Initial attention is given to the repeatability of responses, determined from the absolute difference
between the 4 repeated configurations (see Fig. 4a). The mean difference between repetitions over
attributes was 9.7 pt. Individual attribute repeatability mean values were used as tolerance ranges
to estimate whether a subjective acoustic attribute differed between measurement and simulation.
Overall results were then compared (see Fig. 4b). Considering repetition tolerance, measured
binaural auralizations were judged slightly brighter and the mean slightly ASW wider than simulated
RIR auralizations. Results for the remaining attributes were near 0, within repetition tolerances.
Comparing omnidirectional and binaural configurations, a one-way ANOVA indicated significant
differences for Clarity, Coloration, and Plausibility (Reverberance: F=0.02, p= 0.88; Clarity: F=8.80,
p<10-2; Distance: F: 0.03, p=0.87; Coloration: F=52.00, p<10-2; Plausibility: F=7.42, p=0.01).
In an attempt to understand the foundations of the subjective responses, Table 1 presents objective
parameter averages for measured and simulated RIRs at the auralized positions for EDT, C80,
IACC(e), and IACC(l), which relate to Reverberance, Clarity, ASW, and LEV respectively.6
Perceptual results agree with these objective parameters.
Subsequently, the results were analyzed per position. Figs. 4c and 4d show that the degree of
variance is smaller for the omnidirectional receiver condition. The omnidirectional auralizations are
centered on 0. The exception to this observation is S1R12 for attributes Coloration, Clarity, and
Distance which could be explained by the difference in C80 (see Table 1). The binaural condition
responses exhibit more variation while the majority of the attributes were within the repetition
tolerances (~10 pts). Specifically, the measured RIR at S1R13 was perceived slightly more unclear,
further away, brighter, and more enveloping. The measured RIR at S2R03 was found to be brighter
and to have a wider ASW, and the measured S2R13 was judged to have a wider ASW than the
simulated RIR counterparts. It should be noted that the exact positions between omnidirectional and
binaural receiver conditions differed slightly (approx. 2 m). Taking this into consideration, significant
differences in perceptual similarity judgements between receiver types were found for Clarity at
S1R12/13 (F=5.91, p=0.02) and S2R02/03 (F=9.74, p<10-2) as well as for Coloration at S1R02/03
(F=13.53, p<10-2), S1R12/13 (F=27.18, p<10-2), and S2R02/03 (F=22.55, p<10-2).
In general, binaural auralizations for Clarity agreed slightly better than omnidirectional auralizations.
This is due to the outlier position S1R12. Source and receiver for this position are near the highly
ornate pulpit and several columns. It is possible that the scattering properties for these elements
were erroneous, leading to excessive early reflections and subsequently a perceivable higher
Clarity. Coloration similarity results agreed better for omnidirectional auralizations than binaural
auralizations. A possible explanation is slight misalignments of the dummy head during the
measurement relative to S2. Concerning Plausibility, both omnidirectional and binaural auralizations
were judged equally plausible for measured and simulated responses, within the repetition
Table 1: Measured and simulated single number frequency average objective parameters (EDT, C80,
IACC (e), IACC (l)) per position (Differences in bold are higher than 1 JND).
Position
C80 (dB) JND: 1.0
(500-1000Hz)
IACC(e) JND: 0.075
(500-4000Hz)
IACC(l) JND: 0.075
(500-4000Hz)
Meas
Sim.
Diff.
Meas
Sim.
Diff.
Meas
Sim.
Diff.
Meas
Sim.
Diff.
S01R02
5.76
5.56
+0.20
-6.9
-5.9
-1.0
-
-
S01R12
5.27
5.12
+0.15
-1.2
1.7
-2.9
-
-
S02R02
6.96
7.02
-0.06
-8.9
-8.6
-0.3
-
-
S02R12
6.76
6.73
+0.03
-8.1
-7.8
-0.3
-
-
S01R03
5.65
5.53
+0.11
-5.4
-6.5
+1.1
0.351
0.296
+0.055
0.089
0.304
-0.215
S01R13
5.04
4.55
+0.49
-0.0
1.4
-1.4
0.387
0.514
-0.127
0.093
0.308
-0.216
S02R03
7.11
7.20
-0.09
-9.5
-8.1
-1.4
0.463
0.730
-0.267
0.104
0.359
-0.255
S02R13
6.98
6.77
+0.21
-7.8
-6.1
-1.7
0.766
0.863
-0.096
0.093
0.381
-0.288
194
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
tolerance. Finally, ASW was judged slightly wider for the measured binaural auralizations than for
their simulated counterparts.
4 PRE-MODERN AND CURRENT SAINT-GERMAIN-DES-PRES
With the GA model validated, modifications were made to represent its pre-modern configuration. A
17th century plan, the only direct evidence of the state of the building during the Enlightenment, was
used as a reference for geometrical modifications. Beside these geometrical changes, the use and
position of liturgical adornments are different relative to today’s minimal usage.
Fig. 5 depicts the pre-modern plan of the abbey church of Saint-Germain-des-Prés. The
architecture changed in the 17th century, with the easternmost bay of the south nave aisle
converted into a chapel dedicated to Saint Maur. Furthermore, the sanctuary, the high altar, and the
center of the fifth bay of the nave were fully enclosed by screens, defining the principal liturgical
zones of the pre-modern building. A final screen, positioned to separate the sanctuary from the
choir, was probably acoustically and visually transparent.
Due to lack of information about adornments in Saint-Germain-des-Prés during the 16th and 17th
century, the Cathedral of Notre Dame de Paris, for which the documentary evidence is relatively
Fig.5: 17th century plan of the Saint-Germain-des-Prés8.
Fig 4: Perceptual results for all subjects on subjective similarity of measured and simulated RIR
auralizations. (a) Absolute differences for repeated pairs. (b) Omnidirectional vs. binaural receivers.
Measured on the left, Simulated on the right. Results by position for (c) omnidirectional and
(d) binaural receivers. Box limits represent 25% and 75% quartiles, (+) outliers,
(Ο) median, and ( | ) mean values.
195
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
rich, was selected as a reference in order to make an ‘educated guess' as to which materials were
used and where they were positioned.9 In Notre Dame, screens also enclosed the sanctuary and
high altar, presumably extending upward halfway the shaft of the sanctuary's columns and covered
with draperies. Notre Dame was furnished for festivities especially in the area of the sanctuary by
tapestries and the floor immediately in front of the altar was covered with rugs. Finally, it was
assumed that the painted plaster was already present since this was typical for churches of that
age. Details of the GA model material definitions can be found in Postma and Katz3.
The principal performers and auditors at Saint-Germain-des-Prés were monks, situated in the
principal areas of the high altar and sanctuary. Therefore, current and historical listening conditions
are compared when a source is positioned in these areas and the receivers positioned inside and
outside these principal areas. Fig. 6 shows that the pre-modern configuration had a perceptually
shorter EDT and higher C50, an aspect which is more prominent in the principal areas.
Shorter reverberation times and higher clarity are associated with higher speech intelligibility. It is
probable, therefore, that the performance of the liturgy in the choir of Saint-Germain-des-Prés
changed in synchrony with the acoustics. For example, it would have been possible, with this
increased clarity, to have performed the standard chant repertory at a higher tempo. The increased
clarity of the performance space may have also encouraged the composition of new musical forms
inspired by, and adapted to, the changed acoustical environment. Further research on the
implications of these observations is the focus of ongoing studies.
Fig. 6: Summary of EDT and C50 results for the 17th century and the current configuration GA model
considering receivers inside the principal areas (S2-R17,19,22,24,26,27) and those at the other
positions (S2-R1-2,4-7,9-12,14-16,20-21,25) (number of rays: auto (average = 127400, SD=0),
length IR: auto (average=7850 ms, SD=74 ms)). (See legend of Fig. 4 for notations.)
Fig.5: 17th century plan of the Saint-Germain-des-Prés.8
196
Proceedings of the Institute of Acoustics
Vol. 37. Pt.2 2015
5 CONCLUSION
A GA model of the abbey church Saint-Germain-des-Prés was calibrated according to a
standardized method. The calibration was performed with the assumption that when reverberation
and clarity parameters are within 1 JND of the measured value a valid model is created. To
evaluate the validity of this assumption a listening test was carried out. Finally, the model was
adjusted to represent its architecture during festivities in the Enlightenment era.
The listening test showed that using the methodical calibration procedure led to a perceptually valid
GA model of the Saint-Germain-des-Prés, in addition to the objective measure validity based on
reverberation and clarity metrics. Some trends on perceptual attribute differences between
measured and simulated auralizations were found which slightly exceeded participant repeatability
tolerances, specifically Coloration and ASW for binaural auralizations. Additional studies are
necessary to understand the reason for these differences. As Coloration was judged sufficiently
similar for omnidirectional auralizations, HRTF interpolation and processing employed to convert the
measured data to CATT format should be investigation. Studies are currently underway to examine
the suitability of the calibration method for other venue types, such as theatres.
Exploration of the estimated acoustical conditions of the pre-modern state of the abbey church of
Saint-Germain-des-Prés as compared to the calibrated model, through analysis of the
omnidirectional simulated RIRs indicate that it is reasonable to assume that the church had a
shorter reverberation time and higher clarity in the principal liturgical areas, especially during the
most important celebrations when greater quantities of sound absorptive materials were deployed.
The space today is thus acoustically unlike that of the Middle Ages, and performance practices of
monks who used the space daily were adapted to an acoustical space that favoured greater clarity.
Acknowledgments - This work was funded in part by the ECHO project (ANR-13-CULT-0004,
echo-projet.limsi.fr).
6 REFERENCES
1. B.F.G. Katz and E. Wetherill, Fogg art museum lecture room, a calibrated recreation of the
birthplace of room acoustics. Proc. Forum Acusticum, 21912196, Budapest (2005).
2. A. Tallon, “L'espace acoustique de l'abbatiale de Saint-Germain-des-Prés[The acoustic
space of the abbey of Saint-Germain-des-Prés], in Saint-Germain des Prés. Mille ans d'une
abbaye à Paris, ed. Matthieu Guyot, Paris: Académie des Inscriptions et Belles-Lettres,
(forthcoming).
3. B.N.J. Postma and B.F.G. Katz, “Creation and calibration method of virtual acoustic models
for historic auralizations,” Virtual Reality, no. SI: Spatial Sound, pp. 120, 2015,
doi:10.1007/s10055-015-0275-3.
4. B.F.G. Katz, International round robin on room acoustical response analysis software 2004.
J. Acoust. Soc. Am. 4:158-164, doi:10.1121/1.1758239 (2004).
5. G. Iannace, U. Berardi, and C. Ianniello, Study of a Historical Church Based on Acoustic
Measurements and Computer Simulation. Proc. 22nd ICSV, 1-8, Florence (2015).
6. ISO, Geneva, Switzerland: International Organization for Standardization: ISO 3382-
1:2009(E). Measurement of the Reverberation Time of Rooms with reference to other
Acoustical Parameters (2009).
7. B. Dalenbäck: CATT-A v9: User’s Manual CATT-Acoustic v9. CATT, Gothenburg (Sweden):
(2011).
8. Bibliothèque national de France: Plans manuscrits de l'église de Saint-Germain-des-Prés en
1644, “l'année de son rétablissement" [Plan documents of the abbey of Saint-Germain-des-
Prés in 1644 “year of reinstatement”]. In: Annales de l'abbaye de Saint-Germain-des-Prés
(555-1743), et pièces annexes jusqu'en 1753. MS. FR. 18816, FOL. 72, vol 265, Fondation
de L'Abbey (1898).
9. C. Wright: Music and Ceremony at the Notre Dame of Paris. Cambridge University Press
(2008).
197
... Finally, an overview of the test protocol is presented, building upon a previous preliminary study. 22 A. Preparation of the measured BRIR Prior to commencing the subjective listening test, some additional processing was required for the measured BRIRs. First, the frequency response characteristics of the measurement system have been compensated for by creating an equalization filter. ...
... The results of a preliminary test 22 indicated that the measured binaural auralizations were judged as "brighter" than their simulated counterpart. No such difference was observed for the omni-directional RIRs. ...
... In contrast to the preliminary test, 22 where participants could only start listening at the beginning of the auralization, the test interface now allowed participants to be able to select the starting play point during listening. Additionally, the definitions for coloration and tonal balance had been revised. ...
Article
Recently, auralizations have become more prevalent in architectural acoustics and virtual reality. However, there have been few studies examining the perceptual quality achievable by room acoustic simulations and auralizations. Such studies have highlighted potential problems in creating perceptually equivalent simulations when compared to measured auralizations in terms of parameter estimation. In order to accomplish realistic auralizations, calibration of the geometrical acoustics model can be considered a necessary step. In situations where the studied space exists, well calibrated auralizations can be employed for multiple purposes, such as multi-modal virtual reality explorations, studies of the acoustical influence of renovations, and historic research. Using this case type as a base, a perceptual study evaluating state-of-the-art binaural auralizations has been carried out. Three test sites of different complexity and acoustics were selected: the abbey church Saint-Germain-des-Prés, the cathedral Notre-Dame de Paris, and the Théatre de l’Athénée. Models were calibrated according to omni-directional source-receiver measurements for reverberation and clarity parameters. In the subjective listening test, measured and simulated binaural auralizations were compared according to eight acoustic perceptual attributes. Results showed that the methodical calibration procedure employed in combination with attention to control factors led to ecologically/ perceptually valid auralizations.
... Room acoustical measurements were carried out to serve as a reference for the calibration. Details of the measurement system are described in [8]. Fig. 2bshows the measurement plan with S1-4 representing the source positions and R's depicting the omnidirectional and binaural microphone receiver positions. ...
... To validate the GA model calibration, a listening test was carried out comparing measured to simulated binaural auralizations. The test protocol was based on a previous study [8]. For this test, the omnidirectional source directivity was used, as it corresponded to the measured configuration. ...
Conference Paper
Full-text available
As part of the 850-year anniversary of the Notre-Dame cathedral in Paris, there was a special performance of 'La Vierge', by Jules Massenet. A close mic recording of the concert was made by the Conservatoire de Paris. In an attempt to provide a new type of experience for those unable to attend, a virtual recreation of the performance using these roughly 45 channels of audio source material was made via auralization. A computational acoustic model was created and calibrated based on in-situ measurements for reverberation and clarity parameters. A perceptual study with omnidirectional source and binaural receiver validated the calibrated simulation for the tested subjective attributes of reverberation, clarity, source distance, tonal balance, coloration, plausibility, apparent source width, and listener envelopment when compared to measured responses. Instrument directivity was included in the final simulation to account for each track's representative orchestral section based on published data. Higher-Order Ambisonic (3rd order) room impulse responses were generated for all source and receiver combinations using the CATT-Acoustic TUCT software. Virtual navigation throughout a visual 3D rendering of the cathedral during the concert was made possible using an immersive rendering architecture with BlenderVR, MaxMSP, and an Oculus Rift Head-Mounted Display. This paper presents the major elements of this project, including the calibration procedure, perceptual study, system architecture, lessons learned, and the technological limits encountered with regards to such an ambitious undertaking.
... An overview of a significant portion of this research has been reviewed in Girón et al. [1], discussing different experimental procedures, results, and their theoretical interpretations. A number of notable studies have been carried out in spaces of significant historical importance: St. Peter's Basilica [2], Haghia Sofia and Süleymaniye Mosque [3], St. John's Baptistery [4], Saint-Germain-des-Prés Abbey [5], and St. Paul's Cathedral [6]. ...
Article
Full-text available
The Cathédrale Notre-Dame de Paris is amongst the most well-known worship spaces in the world. Its large volume, in combination with a relatively bare stone construction and marble floor, leads to rather long reverberation times. The cathedral suffered from a significant fire in 2019, resulting in damage primarily to the roof and vaulted ceiling. Despite the notoriety of this space, there are few examples of published data on the acoustical parameters of this space, and these data do not agree. Archived measurement recordings from 1987 were recovered and found to include several balloon bursts. In 2015, a measurement session was carried out for a virtual reality project. Comparisons between results from these two sessions show a slight but significant decrease in reverberation time (8%) in the pre-fire state. Measurements were recently carried out on the construction site, 1 year since the fire. Compared to 2015 data, the reverberation time significantly decreased (20%). This paper presents the preliminary results of these measurements, providing a documentation of the acoustics of this historic worship space both prior to and since the 2019 fire.
... On the contrary, just like some architectural elementsrood screen, wooden panels of the stalls, etc. -isolated the choir from the rest of the building, the acoustic sensation provided by the remoteness of the chant created a symbolic distance between the sacred and the people. [5,25] . ...
Article
Full-text available
This interdisciplinary study brings together acousticians and anthropologists to examine the memory of the acoustics of Notre-Dame de Paris before the fire of April 2019, using a qualitative approach to collect the testimonies of 18 people involved in the sound usages of the cathedral. Testimonies were analyzed in light of research conducted in the anthropology of the senses, sensory perception, memory, and cultural heritage. Analysis highlights an apparent contradiction between the remarkable acoustics of the monument before the fire and the impression of musicians. These musicians reported a struggle to tame the cathedral's sound space, to hear each other well enough to craft their performances and to reach an acceptable level of clarity in their musical practice. These phenomena are examined with acoustic measurements and numerical simulations using a calibrated geometrical acoustics model of the cathedral before the fire, which allows for an objective exploration of the acoustic characteristics of Notre-Dame. This analysis concludes that the well-known reverberation of Notre-Dame, as well as the acoustic barrier of the transept and the poor acoustic return on the podium (the usual place for concert performers) negatively impact singers' comfort. This highlights the tension between the original architectural design of the cathedral and its modern religious and cultural usages. However, the regular occupants have developed a deep familiarity with these constraints during their ritual and musical practices, adjusting to the acoustics in a unique way. Such a tradition of adaptation must be considered as a part of cultural practice, not to be overlooked during the reconstruction.
Article
This interdisciplinary study brings together acousticians and anthropologists to examine the memory of the acoustics of Notre-Dame de Paris before the fire of April 2019, using a qualitative approach to collect the testimonies of 18 people involved in the sound usages of the cathedral. Testimonies were analyzed in light of research conducted in the anthropology of the senses, sensory perception, memory, and cultural heritage. Analysis highlights an apparent contradiction between the remarkable acoustics of the monument before the fire and the impression of musicians. These musicians reported a struggle to tame the cathedral’s sound space, to hear each other well enough to craft their performances and to reach an acceptable level of clarity in their musical practice. These phenomena are examined with acoustic measurements and numerical simulations using a calibrated geometrical acoustics model of the cathedral before the fire, which allows for an objective exploration of the acoustic characteristics of Notre-Dame. This analysis concludes that the well-known reverberation of Notre-Dame, as well as the acoustic barrier of the transept and the poor acoustic return on the podium (the usual place for concert performers) negatively impact singers’ comfort. This highlights the tension between the original architectural design of the cathedral and its modern religious and cultural usages. However, the regular occupants have developed a deep familiarity with these constraints during their ritual and musical practices, adjusting to the acoustics in a unique way. Such a tradition of adaptation must be considered as a part of cultural practice, not to be overlooked during the reconstruction.
Article
Full-text available
At the end of the 19th century, Wallace Clement Sabine undertook the task of correcting the acoustics of the Fogg Art Museum lecture hall. His work on this acoustically difficult (semi‐circular, domed, arched) auditorium was the basis for his monumental work on reverberation time and absorption, beginning the science of room acoustics as it is now known. The room underwent various renovations over 70 years, concluding with its demolition in 1973. Just prior to this event, one author was fortunate to have measured several room impulse responses. Through the use of architectural plans and the few existing photos, a computational room acoustical model was created. This model was calibrated using the historically important measurements as a reference. The geometrical model incorporates the room’s various architectural phases. Using computer auralizations, the room is now accessible in acoustical terms, with the ability to ‘‘perience’’ an educated rendering of the acoustics of the room as Sabine found it in 1895. The results of this study, particularly the geometrical model, are intended for public use and are made generally available for room acoustics students and researchers who chose to follow in Sabine’s footsteps (www.catt.se/FoggArtMusum.htm). Portions of this presentation were presented at Forum Acousticum 2005.
Article
Virtual reality provides the possibility for interactive visits to historic buildings and sites. The majority of current virtual reconstructions have focused on creating realistic virtual environments, by concentrating on the visual component. However, by incorporating more authentic acoustical properties into visual models, a more realistic rendering of the studied venue is achieved. In historic auralizations, calibration of the studied building’s room acoustic simulation model is often necessary to come to a realistic representation of its acoustical environment. This paper presents a methodical calibration procedure for geometrical acoustics models using room acoustics prediction programs based on geometrical acoustics to create realistic virtual audio realities, or auralizations. To develop this procedure, a small unfinished amphitheater was first chosen due to its general simplicity and considerable level of reverberation. A geometrical acoustics model was calibrated according to the results of acoustical measurements. Measures employed during the calibration of this model were analyzed to come to a methodical calibration procedure. The developed procedure was then applied to a more complex building, the abbey church Saint-Germain-des-Prés. A possible application of the presented procedure is to enable interactive acoustical visits of former configurations of buildings. A test case study was carried out for a typical seventeenth-century configuration of the Saint-Germain-des-Prés.
Article
The intent of this study is to examine the variations between current implementations of standard room acousticmeasures for impulse response measurements. An international round robin has been conducted using a single real measured impulse response, rather than a synthesized response. This offers a more rigorous test of analysis procedures. While there is good agreement at higher frequencies, large variations are found at lower frequencies in which the noise level within the measurement is greater. Some errors are attributed to the existence or robustness of noise-floor detection.
L'espace acoustique de l'abbatiale de Saint-Germain-des-Prés " [The acoustic space of the abbey of Saint-Germain-des-Prés], in Saint-Germain des Prés. Mille ans d'une abbaye à Paris
  • A Tallon
A. Tallon, " L'espace acoustique de l'abbatiale de Saint-Germain-des-Prés " [The acoustic space of the abbey of Saint-Germain-des-Prés], in Saint-Germain des Prés. Mille ans d'une abbaye à Paris, ed. Matthieu Guyot, Paris: Académie des Inscriptions et Belles-Lettres, (forthcoming).
Study of a Historical Church Based on Acoustic Measurements and Computer Simulation
  • G Iannace
  • U Berardi
  • C Ianniello
G. Iannace, U. Berardi, and C. Ianniello, 'Study of a Historical Church Based on Acoustic Measurements and Computer Simulation.' Proc. 22nd ICSV, 1-8, Florence (2015).
CATT-A v9: User's Manual CATT-Acoustic v9
  • B Dalenbäck
B. Dalenbäck: CATT-A v9: User's Manual CATT-Acoustic v9. CATT, Gothenburg (Sweden): (2011).
Plans manuscrits de l'église de Saint-Germain-des-Prés en 1644, " l'année de son rétablissement" [Plan documents of the abbey of Saint-Germain-des- Prés in 1644 " year of reinstatement
  • Bibliothèque National De France
Bibliothèque national de France: Plans manuscrits de l'église de Saint-Germain-des-Prés en 1644, " l'année de son rétablissement" [Plan documents of the abbey of Saint-Germain-des- Prés in 1644 " year of reinstatement " ]. In: Annales de l'abbaye de Saint-Germain-des-Prés (555-1743), et pièces annexes jusqu'en 1753. MS. FR. 18816, FOL. 72, vol 265, Fondation de L'Abbey (1898).
International Organization for Standardization: ISO 3382- 1:2009(E) Measurement of the Reverberation Time of Rooms with reference to other Acoustical Parameters
  • Geneva Iso
ISO, Geneva, Switzerland: International Organization for Standardization: ISO 3382- 1:2009(E). Measurement of the Reverberation Time of Rooms with reference to other Acoustical Parameters (2009).
Fogg art museum lecture room, a calibrated recreation of the birthplace of room acoustics
  • B F G Katz
  • E Wetherill
B.F.G. Katz and E. Wetherill, 'Fogg art museum lecture room, a calibrated recreation of the birthplace of room acoustics.' Proc. Forum Acusticum, 2191-2196, Budapest (2005).