JPH03221819A - Method for calculating sound field characteristic - Google Patents

Abstract

PURPOSE:To highly accurately evaluate various kinds (types) of sound fields by multiplying the square h<2>(t) of an impulse response from a sound source position up to a listener's position by a function W(t) for time weight and calculating the intensity modulated transmission function mW(F) of the sound field based upon a specific equation. CONSTITUTION:In order to make great account of the transient characteristics of a time area including a direct sound and an initial reflected sound obtained immediately after the arrival of the direct sound, the square h<2>(t) of an impulse response from the sound source position up to the listener's position is multiplied by the function for time weight, i.e. a function (time window)W(t) to be attenuated with the lapse of time, and the intensity modulated transmission function mW(F) indicated by the equation is found out by using the multiplied value.

Description

DETAILED DESCRIPTION OF THE INVENTION "Field of Industrial Application" The present invention relates to a method for calculating sound field characteristics used, for example, to evaluate the clarity of a sound field.

"Prior Art" Conventionally, various methods have been proposed and used to evaluate the clarity of a sound field. These techniques are applied (
(measurement, prediction, etc.), each has its own strengths and weaknesses. These methods have a common drawback of poor evaluation accuracy. Three typical methods conventionally used will be described below.

(a) Klein's method represents the amount of speech intelligibility loss in a 0゜reverberant sound field (Articulation Loss)
s:AL) (%), reverberation time (T), room volume (V),
This is a prediction method using the distance between the sound source (speaker) and the listener (D) and the directivity coefficient (Q) of the speaker used as the sound source, and is expressed by the following equation.

V The feature of this method is that the speech intelligibility of the reverberant sound field can be easily estimated using only the overall and average physical quantities of the reverberant sound field. However, although this method allows the approximate speech intelligibility of the reverberant sound field to be predicted, its accuracy is poor.

(b) Lochner & Burger method 3
The S/N method represented by 2) is one of the methods that uses an acoustic transfer function from the sound source (speaker) position to the listener position in the reverberant sound field. In this method, the energy of the direct sound from the sound source (speaker) to the listener and the early reflected sound component within a certain time (usually a value of 50 to looms) after the arrival of the direct sound is taken as the signal component (S). This method considers the reverberant energy component of the noise (N@et) as noise (N@et), and measures speech intelligibility from this S/N and rf value. An example of this S/N and ff value when the above-mentioned time is 95 m5 is shown below.

Here, (1) is the impulse response from the sound source (speaker) to the listener, and a (t) is the weighting function corresponding to the human hearing characteristics.
tion). Although this method requires more time and effort in measurement and calculation than the method (a), it has good accuracy. However, the accuracy is still not sufficient.

(C) Intensity modulation transfer function (Modulation Tr
answerFunction: MTF)
method using +31 +41 fs) Optical transfer function (Op
tical Transfer Function, Q
'l'F)+61 is applied to the evaluation of the sound field regarding speech intelligibility, and the MTF, that is, m(F), is calculated using the following formula (1), where i - indicates a complex number, F: frequency, and this Calculate S/N and ff from the value of m (F), and further calculate the final value 5peech Transmission
Index (S TI) or RASTI (Rap
id-3TI)'' and evaluates the speech intelligibility of the sound field using this value.

Although this method is a little more complicated to calculate than methods (a) and (b), it is relatively accurate, and measuring instruments using this principle (
RASTI measuring instrument) (7) is often used.

“Problem to be solved by the invention” The m (F) value represents the Fourier component of the frequency F of h” (t) (echogram: square value of impulse response) during the time interval O-■, and the echo Equal weighting is given to the entire time domain of the gram, which means that m
(F) indicates that the value represents the time-average amount of the sound field. On the other hand, when considering speech intelligibility in a reverberant sound field from an auditory perspective, the transient characteristics of the sound field in the direct sound and early reflection sound regions are important, and the reverberation components in the part temporally distant from the direct sound are It is thought that clarity will not be affected much. For this reason, when this method is applied to a reverberant sound field where the shape of the echogram is substantially different, it is expected that discrepancies will occur in the evaluation of speech intelligibility. This matter will be explained by experimental results in the "Examples" section of this invention.

This invention has the drawback that the method (C) above does not emphasize the transient characteristics of the sound field in the time domain, including the direct sound and the early reflected sound immediately after the arrival of the direct sound. This is intended to emphasize the transient characteristics of the time NMi including the direct sound and the early reflected sound immediately after the arrival of the direct sound by temporally weighting the sound, and will be described in detail below.

"Means for Solving the Problem" In this invention, in order to emphasize transient characteristics in the time domain including direct sound and early reflection sound immediately after the direct sound arrives, the square h2 of the impulse from the sound source position to the listener position ( t) is multiplied by a function that performs temporal weighting, that is, a function that decays over time (time window) W(t), and using this, the intensity modulation transfer function m, (F ).

Note that ST finally calculated from the values of m and (F)
The process of deriving I or RASTI is the same as the conventional method.

The effectiveness of the method according to the present invention will be explained below using examples.

We present the results of intelligibility experiments using a real sound field using a reverberation chamber and a synthetic sound field created by a computer.

"Example 1" A reverberation room with a reverberation time of about 2.5 seconds (500 Hz) and an echogram h2(1,) of an exponential type and a rectangular shape as shown in FIG.
FIG. 2 shows the results of a speech intelligibility test for each parameter in a synthetic sound field with impulse responses h(t) of ngular type, triangular type, and inverse trigonal type. The impulse response h(t) is below 7”TET
It was determined by X (LO M series). Regarding the actual sound field, the impulse response was determined by measurement. In the same figure (a), the horizontal axis is RASTI obtained by the conventional method, and the vertical axis is Articulation Loss, which represents the amount of speech intelligibility loss.
s (A L) (%). In the same figure (b), the horizontal axis is the time window W (t), exp (-1/τ), τ-200ξ
ml, I(
RASTI (Windoiye
d RASTI), and (the vertical axis is the same as in Figure (a). From Figure 2 (a), it can be seen that for different types of echograms, even with the same RASTI value, the Articulation
It can be seen that the values of Loss are different. In other words, it can be seen that the sound intelligibility of the sound field is different. It can also be seen that the data is different from the data of the reverberation room, which is the actual sound field. On the other hand, as can be seen from FIG. 2(b), the Wi
When the data is organized using ndowedRASTI, it can be seen that the results of various types of synthetic sound fields and the data of the reverberation room, which is the actual sound field, are almost on the same line. In other words, it is shown that sound fields having the same RASTI have the same speech intelligibility regardless of the type of echogram.

“Example 2” As in Example 1, the reverberation time was 2.5 seconds (500
Hz) and the speech intelligibility of a synthetic sound field with an exponential type echogram as a reverberation component and a direct sound added to it (refer to the construction drawing: direct sound ten reverberation index type). The results are shown in Figure 3(a).

Shown in (b). For reference, the results of an exponential type sound field that does not include direct sound as shown in Example 1 are also shown.

However, rather than using the exponential data of Example 1, the experiment was repeated with different experimental parameters (reverberation time). The results were the same as in Example 1, and it is thought that the reproducibility of the experimental data was compromised.

Figure 3(a) is RASTI by conventional method, Figure 3(b)
is the Windowed RAS T I according to this invention.
This has been organized by Similar to the results in Example 1, when arranged using the Windowed RAS T I according to the present invention, it can be seen that the results of the actual sound field and the results of the synthetic sound field are on the same line. In addition, in Fig. 3(a), an exponential synthetic sound field (Direct + Expo
Comparing the results of a synthetic sound field (exponential type) and an exponential synthetic sound field that does not have a direct sound (exponential type), it is found that the result of an exponential synthetic sound field that includes a direct sound is closer to the result of the real sound field. Recognize. This is a matter of the prior art, and in the conventional evaluation using MTF (STI or RASTI), when evaluating a sound field with greatly different echogram shapes, discrepancies occur in the evaluation of speech intelligibility. As stated above, this can be interpreted as a lie.

In other words, an exponential synthetic sound field (Direct + Exponential
type) is closer to the actual sound field result.

As described above, according to Example 1.2, the method of this invention, m, (F)
(i.e., Windowed RASTI)
By, the conventional MTF, m (F) (i.e. RAST
It was found that the speech intelligibility in the sound field can be evaluated more accurately than I).

In these Example 1.2, the time window W(t) is exp
(-t/r), r = 200ms, but the value of τ was changed depending on the type of sound source (in this experiment, the intelligibility of single syllables was measured, but for example, the intelligibility of words and conversation). Or,
Or it can be replaced with other functions. Also,
In this example, RA S T I (Rap
id-3TI), but it is also possible to use the m w (F) of this invention for the conventional Sr1 method, and the Windowed S
It is clear that T I is an accurate evaluation value of speech intelligibility in a sound field. Furthermore, m' and (F) of the present invention can be used not only for evaluating clarity, but also for evaluating the spread of a sound field, for example. Furthermore, the weighting function W(t) may be changed depending on the frequency.

"Effects of the Invention" As explained above, according to the method of calculating sound field characteristics according to the present invention, it can be used, for example, to evaluate intelligibility, and can be used accurately for various types (types) of sound fields. It can be evaluated. By using this method, the speech intelligibility of various existing halls, theaters, gymnasiums, lecture halls, assembly halls, movie theaters, etc. (including electroacoustic equipment) can be evaluated accurately1. By using this method in conjunction with computer-based simulation (calculating the echogram or impulse response between the sound source position and the listener position using algorithms such as the sound ray method) for these buildings at the design stage, the sound Speech intelligibility in a field can be calculated and can be used in acoustic design of architecture and electroacoustic equipment.

It can also be used for acoustic design of emergency broadcast equipment.

[References]

(1) W, K1ein+ “Articula
tion Loss of Con5゜nants a
s a Ba5is for tM Design a
nd Judgment of 5ound Reinf
arrangement systems + J, ＾ud
io Eng, Soc,, vol, 19. pp
, 920-922 (1971 Dec) (2) J.
P, A, Lochner a, nd J, F,
Burger, “The Influence of
Reflections on Auditor
To m8coustics J, 5ound
Vibration, vol, Lpp.

426-454 (1964) (3) T, Houtgast, J, M, 5t
eenken and R, Pr.

mp, ”Pedicting 5peech int
elligibility in r.

oms from the modulation
transfer function, I.

2 Genera'l room acoustics
+ Acust+ca vol, 46 (1980) (4) T, Houtgast and
J,' M, 5teenken+ “8 Re
vieIIIof the MTF Concept
in Room Acoustics and Its
Use for Estimating 5peec
h Intelligib-ility in Aud
itoria, J., Acoust, S.
oc, ^n,.

vol, 77 (1985 Mar,) (5) M
, R,5 chroeder, “Modulat
ion Trai+5ferFunctions:De
finition and measurement
t”Abstract in IEEE, A
SSP 26. P2S5 (1978) (6) K.
Rosenh'auer and K.J.
Ro'5enbruch.

“The measurement of the 0
ptical transfer functions
of 1enses+ Rep, Prog, P
hys, +vol.

30, ppl-25 (1967) (7) J, M, 5teenken and
T, Houtgast,' ”Ra5ti:a t
ool for evaluating audit
oria, Bruel & Kjaer Tec
hnical Revlvol, 3 (1985
)

[Brief explanation of drawings]

Figure 1 is a diagram showing the experimental conditions for the speech intelligibility test in Example 1.2 (an experiment in a reverberation room as a real sound field, and an experiment in a computer-generated sound field), and Figure 2 shows the results of the real sound field;
and the results of speech intelligibility tests for sound fields with impulse responses h(t) such that the echogram h2(t) is exponential, rectangular, triangular, or inverted triangular among synthetic sound fields. (a) is a diagram showing the results of RASTI according to the conventional method, (b) is a diagram showing the results of RASTI according to the present invention.
Figure 3 shows the result of the real sound field and the echogram h of the synthetic sound field.
2(t) shows the results of a speech intelligibility test when it has an exponential type echogram as a reverberation component and a synthetic sound field in which direct sound is added to it, and only has an exponential type reverberation component, and (a
) shows the results of the conventional method, and (b) shows the results using the present invention.

Claims (1)

[Claims] (1) A method of calculating the characteristics of a reverberant sound field using an acoustic transfer function from the sound source position to the listener position, the square h of the impulse response from the sound source position to the listener position. ^2(t) is multiplied by the temporal weighting function W(t), and from this the intensity modulation transfer function m_w(F) of the sound field is calculated as follows: ▲ There are mathematical formulas, chemical formulas, tables, etc. ▼ F is A sound field characteristic calculation method characterized in that frequency and i represent a complex number.