Content uploaded by Andor Fürjes
Author content
All content in this area was uploaded by Andor Fürjes on May 25, 2020
Content may be subject to copyright.
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
1
Tales of more than One Thousand and One Measurements
(STI vs. room acoustic parameters – a study on extensive measurement data)
Authors
Andor T. Fürjes,
aQrate Ltd. (Hungary, 8083 Csákvár Haraszt u. 15.) furjes.andor@aqrate.hu
Attila B. Nagy
Reziduum Ltd. (Hungary, 1012 Budapest, Mátray u. 8/b), nagyab@reziduum.hu
May 2020
Abstract
Looking at the history of standards on room acoustic qualities, there has been and still there is an
ongoing interest to define the minimum set of requirements for room acoustics comfort. Speech
intelligibility is a major issue and STI (speech transmission index) gained popularity to express speech
intelligibility in room acoustic specifications even if it is not a room acoustic measure.
This paper is a short summary of results of an extensive measurement data set. Measurement data is
collected from several types of rooms located in Hungary, carried out by different acousticians using
slightly different equipment and software. Measurement data is then evaluated by using a MatLab
script using theories of relevant standards.
Results are finally evaluated to compare STI versus room acoustic parameters.
1 Introduction
Room acoustics is about the phenomenon when waves of a sound source propagate and are repeatedly
reflected from surfaces of an enclosure. Due to energy losses during propagation, the energy of sound
waves diminishes and results in an infinitely long decay in most cases.
Room acoustic parameters basically characterize this temporal response, assuming the room is a Multi-
Input and Multi Output linear time-invariant causal system. Commonly understood room acoustic
parameters are defined in ISO 3382-1 and ISO 3382-2 standards.
Overall speech intelligibility is usually characterized by using STI (speech transmission index) as defined
by EN 60268-16 standard. STI was originally meant to describe overall speech intelligibility when the
(electroacoustic) system is noisy and nonlinear, and also models nonlinear subjective phenomena such as
spectral and level masking or empirical corrections for aged, or non-native listeners. Due to the above,
calculation and measurement of STI is based on only spectral (and not temporal) characteristics of the
response.
Despite fundamental dissimilarities STI is often compared to room acoustic parameters to prove that STI
is needed to complete room acoustics specifications. The most common argument is that reverberation
times do not correlate well to STI and to ensure good intelligibility STI or C50 requirements must be
added to the specification.
The authors find, however, that a direct comparison of reverberation times to STI is basically not right
(see fundamental differences above). Experience also suggests that even if specifying only reverberation
times seems somewhat ‘old-school’ nowadays, a single maximum reverberation time requirement is
adequate if the aim is simply to avoid a surely bad outcome.
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
2
To support this finding and to see how other room acoustic parameters relate to STI in real situations, the
authors did analyze an extensive measurement data set of room acoustic measurements. Data has been
collected from several types of rooms located in Hungary, measurements were carried out by different
acousticians using different equipment and software. Evaluation and analysis were carried out by means
of MatLab scripts based on theories of relevant standards.
2 Method
The method of the study is fairly straightforward: we took impulse response measurements from practices
of our own and our colleagues, evaluated impulse responses calculating basic room acoustic parameters
and STI.
Please note, that since STI calculations are based on the indirect method (evaluation of impulse
responses), effects of nonlinearities and noise are not considered. Also, because we focused on comparing
room acoustic parameters to STI, no threshold masking effect was considered either. Since effects of
spectral auditory masking were found to be negligible for natural speech sound levels, and room acoustics
design shall not be gender-specific, gender-specific settings and equalization were also not considered.
The following room acoustic parameters were calculated according to ISO 3382-1:
- early decay time (EDT10r)
- reverberation time (T10r, T20r)
- clarity (C50, C80).
In addition to parameters of ISO 3382-1, the following parameters were also evaluated just for curiosity:
- early decay time (EDT10), derived from the timing of the direct sound and -10 dB points on EDC;
- reverberation time (T10, T20) derived from the timing of the -5 dB and -15 dB or -25 dB points on EDC
respectively;
- reverberation time (T15, T15r) derived from the -5 dB and -20 dB or -25 dB points on EDC, direct
reading and linear regression fitted respectively;
- reverberation time (Tsnr) derived from the -5 dB and estimated signal-to-noise ratio points on EDC,
fitted using linear regression;
- C8 early-to-late energy ratio with 8 ms time limit, approximating direct-to reverberant ratio.
To match evaluation, same IIR band-pass filters were applied matching the ANSI S1.11 specification for
both STI and room acoustic parameters.
Signal-to-noise ratio was estimated by a pessimistic non-iterative empirical algorithm based on shape-
matching of energy decay (Schroeder) curve.
To compare the single-value parameter STI to frequency dependent parameters, different mean or
weighted values were used:
- m2: mean value from 500 Hz and 1 kHz octave bands
- m3: mean value from 500 Hz, 1 kHz and 2 kHz octave bands
- m4: mean value from 250 Hz, 500 Hz, 1 kHz and 2 kHz octave bands
- A: value derived from A-weighted impulse response evaluation.
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
3
Results are compared both for each measurement and for each room. For each room, room acoustic
parameters are averaged, and STI mean minus standard deviation is calculated (denoted by addition of σ,
e.g. STIstd,σ).
3 Measurement conditions
Measurement data was intentionally collected from diverse conditions to avoid any conclusion that would
be otherwise specific to certain situations including room type, size, geometry, condition, etc. The only
condition was that the sound path from source to receivers is unobscured. Figure 1. shows photos of some
of the measurement sites. Measurement data of 1202 measurements from 46 different locations/facilities
and 121 different spaces (volumes from 20 m3 to >100.000 m3) was collected. Some of the rooms were
acoustically treated or dampened, other were completely bold and lively. Some of the rooms were
furnished, others were basically empty (e.g. sports halls). Functions of spaces include offices, classrooms,
music school rehearsal rooms, conference halls, theatres, stadiums, a swimming pool, and studios for
sound recording, filming, screening.
The equipment to collect impulse responses was diverse, too:
- sound sources: small 2-way active loudspeakers (<4” diameter, mimicking size of a human talker),
larger coaxial two-way loudspeakers (with nominal 90×50° H×V directivity) on-axis and off-axis,
compact omnidirectional sources (monopole) or dodecahedron loudspeakers (as per standard);
- microphones: single omnidirectional microphones and Ambisonic tetrahedral 4-channel microphones;
- measurements system (only to collect impulse responses): WinMLS or EASERA software, or custom
post-processing of recordings of responses to logarithmic swept sine wave excitation.
Special cases are stadiums, where PA systems were used as the sound source and since PA systems are
distributed systems there, the excitation includes already a series of echoes due to large distances of the
loudspeakers without any sound reflection. In other words, responses of stadiums are a combination of
decaying responses of the semi-open architectural environment and the large-scale distributed
loudspeaker system.
4 Evaluation of the results
Figures 2…15 show scatter plots of various room acoustic parameters against STI for each measurement.
Scatter plots also include linear or logarithmic regression lines.
The following trends are clearly shown:
a) room acoustic parameters related to simple energy ratios (including EDT10, T10, T15, T20, D50, C50,
C80) have better correlation to STI than parameters calculated using linear regression (including
EDT10r, T10r, T15r, T20r, Tsnr);
b) there seems to be only minor benefit of means m3 or m4 over m2 (500 Hz and 1 kHz mean time)
in terms of correlation to STI;
c) best correlation to STI is seen at C80,m4;
d) even if direct correlation to STI is lower for reverberation times, there seems to be a clear line of
minimum STI for each reverberation time.
Figure 1 (next page). Photos of some of the measurement sites.
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
4
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
5
Figures 16…30 show scatter plots of various room acoustic parameters against STI for each room. Scatter
plots also include linear or logarithmic regression lines.
The following trends are clearly shown if results of averages of means of room acoustic parameters are
compared to average minus deviation of STI taken for each room:
e) average of mean decay and reverberation times have much better correlation to STIavg-σ for each
room than mean decay and reverberation times to STI for each measurement;
f) best correlation to STIavg-σ is seen at EDT10,A, EDT10,m2 EDT10,m3 and EDT10,m4;
g) earl-late energy ratios and center time have significantly lower correlation.
According to the results, worst expected STI can be estimated from reverberation times for each
measurement in the form of STI≥a·ln(x)+b, where x is the measured decay or reverberation time and a
and b are adjusted to measured data of x.
Summary of adjusted terms is shown in Table 1. Using these estimates typically less than 1% of
measurements proved to provide lower STI than expected lowest value. Seemingly there is hardly any
difference in the terms of T10, T15 and T20.
An example of the application of this lower limit estimation is shown in Figure 31.
Table 1. Terms of expression to estimate lower limit of STI
based on different decay or reverberation time metrics.
A m2 m3 m4
a -0,15 -0,16 -0,17 -0,18
b +0,53 +0,56 +0,56 +0,56
a -0,17 -0,16 -0,18 -0,19
b +0,54 +0,56 +0,56 +0,56
a -0,18 -0,18 -0,19 -0,20
b +0,54 +0,55 +0,56 +0,55
a -0,19 -0,18 -0,20 -0,20
b +0,54 +0,55 +0,56 +0,55
a -0,19 -0,18 -0,20 -0,20
b +0,54 +0,55 +0,56 +0,55
a -0,19 -0,18 -0,19 -0,20
b +0,54 +0,55 +0,56 +0,56
a -0,19 -0,18 -0,19 -0,20
b +0,55 +0,55 +0,56 +0,55
a -0,19 -0,19 -0,19 -0,20
b +0,55 +0,55 +0,56 +0,55
parameter derivation
regression
direct
regression
EDT
10
T
10
T
15
T
20
direct
mean/weighting
regression
direct
regression
direct
term
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
6
5 Conclusions
Based on the evaluation results of more than one thousand and one measurements, we found the following
conclusions:
- STI has a strong correlation with basic room acoustical parameters, such as C50, C80 and EDT10;
- Reverberation time has proved to be completely suitable to express room acoustical limits of a given
room. Similarly to room length and width: values may fall within, but not outside of the limits. This
is the reason why reverberation time and its variants do not correlate well with STI directly.
Measurement results show, that there is a minimum STI value guaranteed for a given reverberation
time – meaning this depends purely on room acoustical properties –, whilst larger STI values may
result from smaller source-receiver-distance and from the directivity of source and receiver.
Reverberation time on its own is appropriate for expressing minimum requirements, such as ‘shall not
be worse than’. On the other hand, reverberation time is not suitable for requirements such as ‘shall
be exactly X’.
- As EDT correlates well with STI, further and more thorough investigation to practically utilize EDT
is needed, as it is more easily measurable than T20 and C50, and is much more easily measurable than
STI.
- There is very little difference between the correlation of STI and room acoustic parameters of different
mean metrics (A, m2, m3, m4). There is practically zero difference between these metrics if they are
used to express minimum requirements (‘shall not be worse than’). The most practical parameter
seems to be m2 mean value as it is the easiest to measure and to calculate among all. If looking at
measurability only, A-weighting would be the most practical choice.
From the above the authors found STI to be superfluous to be used in room acoustics as it is not strictly
a room acoustical parameter and it is not required for expressing room acoustical effects on speech
intelligibility. Please note, that requirements of noise levels is usually found elsewhere in acoustic
specifications (background noise, spill from adjacent spaces, traffic noise, HVAC noise, equipment noise,
etc.).
However, room acoustical parameters do not describe noise (noisiness) and non-linearity, therefore STI
is still the best choice if probability of degradation in speech intelligibility is to be assessed and effects of
nonlinearity or noise shall be expressed implicitly.
6 Acknowledgement
The author would like to thank colleagues Éva Arató Borsi and Gergely Borsi for providing additional
room impulse response measurement data for this study.
The author would like to thank the Section for Acoustics of the Hungarian Engineering Chamber to
support the preparation of a room acoustic design guideline, during which numerous observations were
revealed.
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
7
Figure 2. Comparison of mean EDT10 values vs. STI of all measurements.
Figure 3. Comparison of mean EDT10,r values vs. STI of all measurements.
y = -0,13ln(x) + 0,62
R² = 0,87
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3, 20 6,40
STI std. (-)
edt10,A (sec)
edt10,A vs.sti
y = -0,12ln(x) + 0,65
R² = 0,84
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6 ,40
STI std. (-)
edt10,m2 (sec)
edt10,m2 vs.sti
y = -0,15ln(x) + 0,63
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
edt10,m3 (sec)
edt10,m3 vs.sti
y = -0,16ln(x) + 0,62
R² = 0,92
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
edt10,m4 (sec)
edt10,m4 vs.sti
y = -0,12ln(x) + 0,64
R² = 0,79
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3, 20 6,40
STI std. (-)
edt10r,A (sec)
edt10r,A vs.sti
y = -0,12ln(x) + 0,66
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
edt10r,m2 (sec)
edt10r,m2 vs.sti
y = -0,14ln(x) + 0,65
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
edt10r,m3 (sec)
edt10r,m3 vs.sti
y = -0,15ln(x) + 0,64
R² = 0,80
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
edt10r,m4 (sec)
edt10r,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
8
Figure 4. Comparison of mean T10 values vs. STI of all measurements.
Figure 5. Comparison of mean T10,r values vs. STI of all measurements.
y = -0,13ln(x) + 0,65
R² = 0,78
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std. (-)
t10,A (sec)
t10,A vs.sti
y = -0,12ln(x) + 0,67
R² = 0,74
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
t10,m2 (sec)
t10,m2 vs.sti
y = -0,14ln(x) + 0,66
R² = 0,71
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t10,m3 (sec)
t10,m3 vs.sti
y = -0,15ln(x) + 0,65
R² = 0,73
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
t10,m4 (sec)
t10,m4 vs.sti
y = -0,13ln(x) + 0,66
R² = 0,74
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std. (-)
t10r,A (sec)
t10r,A vs.sti
y = -0,12ln(x) + 0,68
R² = 0,69
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
t10r,m2 (sec)
t10r,m2 vs.sti
y = -0,13ln(x) + 0,66
R² = 0,60
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t10r,m3 (sec)
t10r,m3 vs.sti
y = -0,13ln(x) + 0,65
R² = 0,61
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
t10r,m4 (sec)
t10r,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
9
Figure 6. Comparison of mean T15 values vs. STI of all measurements.
Figure 7. Comparison of mean T15,r values vs. STI of all measurements.
y = -0,13ln(x) + 0,66
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std. (-)
t15,A (sec)
t15,A vs.sti
y = -0,12ln(x) + 0,67
R² = 0,74
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
t15,m2 (sec)
t15,m2 vs.sti
y = -0,14ln(x) + 0,66
R² = 0,70
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t15,m3 (sec)
t15,m3 vs.sti
y = -0,15ln(x) + 0,65
R² = 0,72
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
t15,m4 (sec)
t15,m4 vs.sti
y = -0,12ln(x) + 0,68
R² = 0,68
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
t15r,m2 (sec)
t15r,m2 vs.sti
y = -0,12ln(x) + 0,66
R² = 0,56
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t15r,m3 (sec)
t15r,m3 vs.sti
y = -0,13ln(x) + 0,66
R² = 0,58
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
t15r,m4 (sec)
t15r,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
10
Figure 8. Comparison of mean T20 values vs. STI of all measurements.
Figure 9. Comparison of mean T20,r values vs. STI of all measurements.
y = -0,13ln(x) + 0,66
R² = 0,76
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std. (-)
t20,A (sec)
t20,A vs.sti
y = -0,13ln(x) + 0,67
R² = 0,75
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
t20,m2 (sec)
t20,m2 vs.sti
y = -0,14ln(x) + 0,66
R² = 0,67
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t20,m3 (sec)
t20,m3 vs.sti
y = -0,15ln(x) + 0,65
R² = 0,69
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
t20,m4 (sec)
t20,m4 vs.sti
y = -0,13ln(x) + 0,67
R² = 0,72
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std. (-)
t20r,A (sec)
t20r,A vs.sti
y = -0,12ln(x) + 0,68
R² = 0,68
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
t20r,m2 (sec)
t20r,m2 vs.sti
y = -0,12ln(x) + 0,66
R² = 0,54
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t20r,m3 (sec)
t20r,m3 vs.sti
y = -0,13ln(x) + 0,66
R² = 0,54
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
t20r,m4 (sec)
t20r,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
11
Figure 10. Comparison of mean Tsnr values vs. STI of all measurements.
Figure 11. Comparison of mean center time tc values vs. STI of all measurements.
y = -0,12ln(x) + 0,69
R² = 0,65
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std. (-)
tsnr,A (sec)
tsnr,A vs.sti
y = -0,12ln(x) + 0,70
R² = 0,63
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6, 40
STI std. (-)
tsnr,m2 (sec)
tsnr,m2 vs.sti
y = -0,12ln(x) + 0,68
R² = 0,47
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
tsnr,m3 (sec)
tsnr,m3 vs.sti
y = -0,12ln(x) + 0,68
R² = 0,45
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0 ,80 1,60 3,20 6,40
STI std. (-)
tsnr,m4 (sec)
tsnr,m4 vs.sti
y = -0,17ln(x) + 1,37
R² = 0,85
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,A (msec)
tc,A vs.sti
y = -0,16ln(x) + 1,34
R² = 0,84
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,m2 (msec)
tc,m2 vs.sti
y = -0,16ln(x) + 1,37
R² = 0,86
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,m3 (msec)
tc,m3 vs.sti
y = -0,17ln(x) + 1,39
R² = 0,86
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,m4 (msec)
tc,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
12
Figure 12. Comparison of mean D50 values vs. STI of all measurements.
Figure 13. Comparison of mean C8 values vs. STI of all measurements.
y = 0,01x + 0,29
R² = 0,79
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 60 7 0 80 90 100
STI std. (-)
d50,A (%)
d50,A vs.sti
y = 0,01x + 0,33
R² = 0,90
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 60 70 80 90 100
STI std. (-)
d50,m2 (%)
d50,m2 vs.sti
y = 0,01x + 0,31
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 6 0 70 80 90 100
STI std. (-)
d50,m3 (%)
d50,m3 vs.sti
y = 0,01x + 0,28
R² = 0,90
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 6 0 70 80 90 100
STI std. (-)
d50,m4 (%)
d50,m4 vs.sti
y = 0,01x + 0,68
R² = 0,19
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5 ,0 15,0 25,0
STI std. (-)
c8,A (dB)
c8,A vs.sti
y = 0,01x + 0,73
R² = 0,41
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c8,m2 (dB)
c8,m2 vs.sti
y = 0,01x + 0,72
R² = 0,36
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c8,m3 (dB)
c8,m3 vs.sti
y = 0,01x + 0,71
R² = 0,32
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c8,m4 (dB)
c8,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
13
Figure 14. Comparison of mean C50 values vs. STI of all measurements.
Figure 15. Comparison of mean C80 values vs. STI of all measurements.
y = 0,02x + 0,56
R² = 0,79
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5, 0 1 5,0 25,0
STI std. (-)
c50,A (dB)
c50,A vs.sti
y = 0,02x + 0,60
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c50,m2 (dB)
c50,m2 vs.sti
y = 0,03x + 0,59
R² = 0,92
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c50,m3 (dB)
c50,m3 vs.sti
y = 0,03x + 0,58
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c50,m4 (dB)
c50,m4 vs.sti
y = 0,02x + 0,49
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5, 0 1 5,0 25,0
STI std. (-)
c80,A (dB)
c80,A vs.sti
y = 0,02x + 0,55
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c80,m2 (dB)
c80,m2 vs.sti
y = 0,02x + 0,53
R² = 0,94
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c80,m3 (dB)
c80,m3 vs.sti
y = 0,02x + 0,51
R² = 0,95
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. (-)
c80,m4 (dB)
c80,m4 vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
14
Figure 16. Comparison of SNR (dBA) values vs. STI of all measurements.
y = 0,00x + 0,56
R² = 0,02
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 60 70 8 0 90 100
STI std. (-)
snr,A (dB)
snr,A vs.sti
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
15
Figure 17. Comparison averages of mean EDT10 vs. average minus deviation STI of all rooms.
Figure 18. Comparison averages of mean EDT10,r vs. average minus deviation STI of all rooms.
y = -0,17ln(x) + 0,57
R² = 0,93
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10,A (sec)
Average edt10,A vs.sti,σ
y = -0,15ln(x) + 0,60
R² = 0,93
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10,m2 (sec)
Average edt10,m2 vs.sti,σ
y = -0,16ln(x) + 0,59
R² = 0,94
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10,m3 (sec)
Average edt10,m3 vs.sti,σ
y = -0,17ln(x) + 0,59
R² = 0,94
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10,m3 (sec)
Average edt10,m4 vs.sti,σ
y = -0,17ln(x) + 0,59
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10r,A (sec)
Average edt10r,A vs.sti,σ
y = -0,15ln(x) + 0,61
R² = 0,88
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10r,m2 (sec)
Average edt10r,m2 vs.sti,σ
y = -0,15ln(x) + 0,61
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10r,m3 (sec)
Average edt10r,m3 vs.sti,σ
y = -0,16ln(x) + 0,60
R² = 0,90
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
edt10r,m3 (sec)
Average edt10r,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
16
Figure 19. Comparison averages of mean T10 vs. average minus deviation STI of all rooms.
Figure 20. Comparison averages of mean T10,r vs. average minus deviation STI of all rooms.
y = -0,17ln(x) + 0,60
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10,A (sec)
Average t10,A vs.sti,σ
y = -0,15ln(x) + 0,62
R² = 0,88
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10,m2 (sec)
Average t10,m2 vs.sti,σ
y = -0,16ln(x) + 0,61
R² = 0,90
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10,m3 (sec)
Average t10,m3 vs.sti,σ
y = -0,17ln(x) + 0,61
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10,m3 (sec)
Average t10,m4 vs.sti,σ
y = -0,17ln(x) + 0,60
R² = 0,90
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10r,A (sec)
Average t10r,A vs.sti,σ
y = -0,13ln(x) + 0,63
R² = 0,73
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10r,m2 (sec)
Average t10r,m2 vs.sti,σ
y = -0,14ln(x) + 0,63
R² = 0,78
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10r,m3 (sec)
Average t10r,m3 vs.sti,σ
y = -0,15ln(x) + 0,62
R² = 0,80
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t10r,m3 (sec)
Average t10r,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
17
Figure 21. Comparison averages of mean T15 vs. average minus deviation STI of all rooms.
Figure 22. Comparison averages of mean T15,r vs. average minus deviation STI of all rooms.
y = -0,17ln(x) + 0,61
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std.σ (-)
t15,A (sec)
Average t15,A vs.sti,σ
y = -0,15ln(x) + 0,62
R² = 0,88
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15,m2 (sec)
Average t15,m2 vs.sti,σ
y = -0,16ln(x) + 0,62
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15,m3 (sec)
Average t15,m3 vs.sti,σ
y = -0,17ln(x) + 0,61
R² = 0,90
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15,m3 (sec)
Average t15,m4 vs.sti,σ
y = -0,15ln(x) + 0,62
R² = 0,82
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15r,A (sec)
Average t15r,A vs.sti,σ
y = -0,13ln(x) + 0,64
R² = 0,71
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15r,m2 (sec)
Average t15r,m2 vs.sti,σ
y = -0,14ln(x) + 0,63
R² = 0,73
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15r,m3 (sec)
Average t15r,m3 vs.sti,σ
y = -0,15ln(x) + 0,63
R² = 0,75
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t15r,m3 (sec)
Average t15r,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
18
Figure 23. Comparison averages of mean T20 vs. average minus deviation STI of all rooms.
Figure 24. Comparison averages of mean T20,r vs. average minus deviation STI of all rooms.
y = -0,17ln(x) + 0,61
R² = 0,88
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std.σ (-)
t20,A (sec)
Average t20,A vs.sti,σ
y = -0,16ln(x) + 0,63
R² = 0,88
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t20,m2 (sec)
Average t20,m2 vs.sti,σ
y = -0,16ln(x) + 0,62
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t20,m3 (sec)
Average t20,m3 vs.sti,σ
y = -0,17ln(x) + 0,62
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t20,m3 (sec)
Average t20,m4 vs.sti,σ
y = -0,15ln(x) + 0,63
R² = 0,79
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std.σ (-)
t20r,A (sec)
Average t20r,A vs.sti,σ
y = -0,13ln(x) + 0,64
R² = 0,74
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t20r,m2 (sec)
Average t20r,m2 vs.sti,σ
y = -0,14ln(x) + 0,64
R² = 0,76
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t20r,m3 (sec)
Average t20r,m3 vs.sti,σ
y = -0,15ln(x) + 0,63
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
t20r,m3 (sec)
Average t20r,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
19
Figure 25. Comparison averages of mean Tsnr vs. average minus deviation STI of all rooms.
Figure 26. Comparison averages of mean center time tc vs. average minus deviation STI of all rooms.
y = -0,15ln(x) + 0,65
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
tsnr,A (sec)
Average tsnr,A vs.sti,σ
y = -0,14ln(x) + 0,66
R² = 0,76
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
tsnr,m2 (sec)
Average tsnr,m2 vs.sti,σ
y = -0,15ln(x) + 0,66
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
tsnr,m3 (sec)
Average tsnr,m3 vs.sti,σ
y = -0,16ln(x) + 0,66
R² = 0,77
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1, 60 3,20 6,40
STI std.σ (-)
tsnr,m3 (sec)
Average tsnr,m4 vs.sti,σ
y = -0,19ln(x) + 1,42
R² = 0,82
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 3 2 64 128 256 512 1024
STI std. (-)
tc,A (msec)
Average tc,A vs.sti,σ
y = -0,17ln(x) + 1,38
R² = 0,85
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,m2 (msec)
Average tc,m2 vs.sti,σ
y = -0,18ln(x) + 1,40
R² = 0,86
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,m3 (msec)
Average tc,m3 vs.sti,σ
y = -0,19ln(x) + 1,43
R² = 0,85
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
8 16 32 64 128 256 512 1024
STI std. (-)
tc,m4 (msec)
Average tc,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
20
Figure 27. Comparison averages of mean D50 vs. average minus deviation STI of all rooms.
Figure 28. Comparison averages of mean C8 vs. average minus deviation STI of all rooms.
y = 0,01x + 0,21
R² = 0,82
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 6 0 70 80 90 100
STI std. σ (-)
d50,A (%)
Average d50,A vs.sti,σ
y = 0,01x + 0,27
R² = 0,93
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 6 0 70 80 90 100
STI std. σ (-)
d50,m2 (%)
Average d50,m2 vs.sti,σ
y = 0,01x + 0,25
R² = 0,92
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 6 0 70 80 90 100
STI std. σ (-)
d50,m3 (%)
Average d50,m3 vs.sti,σ
y = 0,01x + 0,22
R² = 0,91
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0 10 20 30 40 50 6 0 70 80 90 100
STI std. σ (-)
d50,m4 (%)
Average d50,m4 vs.sti,σ
y = 0,01x + 0,69
R² = 0,12
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std., σ (-)
c8,A (dB)
Average c8,A vs.sti,σ
y = 0,03x + 0,80
R² = 0,56
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c8,m2 (dB)
Average c8,m2 vs.sti,σ
y = 0,02x + 0,78
R² = 0,50
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c8,m3 (dB)
Average c8,m3 vs.sti,σ
y = 0,02x + 0,77
R² = 0,44
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c8,m4 (dB)
Average c8,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
21
Figure 29. Comparison averages of mean C50 vs. average minus deviation STI of all rooms.
Figure 30. Comparison averages of mean C80 vs. average minus deviation STI of all rooms.
y = 0,02x + 0,54
R² = 0,79
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std., σ (-)
c50,A (dB)
Average c50,A vs.sti,σ
y = 0,02x + 0,57
R² = 0,89
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c50,m2 (dB)
Average c50,m2 vs.sti,σ
y = 0,02x + 0,56
R² = 0,88
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c50,m3 (dB)
Average c50,m3 vs.sti,σ
y = 0,02x + 0,55
R² = 0,86
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c50,m4 (dB)
Average c50,m4 vs.sti,σ
y = 0,02x + 0,50
R² = 0,82
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std., σ (-)
c80,A (dB)
Average c80,A vs.sti,σ
y = 0,02x + 0,53
R² = 0,82
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c80,m2 (dB)
Average c80,m2 vs.sti,σ
y = 0,02x + 0,53
R² = 0,74
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c80,m3 (dB)
Average c80,m3 vs.sti,σ
y = 0,02x + 0,52
R² = 0,72
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
-25,0 -15,0 -5,0 5,0 15,0 25,0
STI std. σ (-)
c80,m4 (dB)
Average c80,m4 vs.sti,σ
A. T. Fürjes, A. B. Nagy,
Tales of One Thousand and One Measurements (STI vs. room acoustic parameters – a study on extensive measurement data)
22
Figure 31. Comparison of mean T20,r values vs. STI of all measurements and
the adjusted lower limit estimation of STI based on mean T20,r values.
y = -0,13ln(x) + 0,67
R² = 0,71
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,20 6,40
STI std. (-)
t20r,A (sec)
t20r,A vs.sti
y = -0,12ln(x) + 0,68
R² = 0,68
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,2 0 6,40
STI std. (-)
t20r,m2 (sec)
t20r,m2 vs.sti
y = -0,12ln(x) + 0,66
R² = 0,54
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,2 0 6,40
STI std. (-)
t20r,m3 (sec)
t20r,m3 vs.sti
y = -0,13ln(x) + 0,66
R² = 0,54
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
0,10 0,20 0,40 0,80 1,60 3,2 0 6,40
STI std. (-)
t20r,m4 (sec)
t20r,m4 vs.sti