WO2014206212A1

WO2014206212A1 - Digital image generation method through 3d spatial distribution of sound quality objective parameters

Info

Publication number: WO2014206212A1
Application number: PCT/CN2014/079954
Authority: WO
Inventors: 卢奂采; 金江明; 胡伟杰
Original assignee: 浙江工业大学
Priority date: 2013-06-26
Filing date: 2014-06-16
Publication date: 2014-12-31
Also published as: CN103389155A; CN103389155B

Abstract

A digital image generation method through 3D spatial distribution of sound quality objective parameters is performed by the following steps: Step 1. recording holograpahic acoustic pressure data on a hologrpahic measurement surface by using a microphone array; Step 2. reconstructing acoustic quantities in a 3D acoustic field; Step 3. calculating the distribution of a sound quality objective parameter (such as loudness, acuteness, and roughness) at each point in the 3D space according to acoustic quantity information inside the 3D space obtained in Step 2 and based on an acoustic pressure-sound quality objective parameter mapping model, and giving the distribution in the form of a 3D digital image, thereby implementing 3D spatial visualization of sound quality objective parameters.

Description

Dimensional spatial distribution digital image generation method

The invention relates to near-field acoustic holography technology, sound source recognition and positioning technology, sound field visualization technology, measurement and evaluation technology of objective parameters of sound quality and noise control technology.

Background technique:

The traditional noise control is mainly aimed at reducing the sound pressure level of the sound field response. The research results show that the reduction of sound pressure level can not fully improve the subjective hearing experience of human voice. The effect of noise on people is not only related to the sound pressure level, but also to the frequency composition of the sound, the physical characteristics of the human auditory system, and the psychological characteristics. The two sounds of the same sound pressure level will cause a huge difference in the psychoacoustic sensitivity of people due to the different frequency composition. Therefore, the evaluation of noise needs to introduce a quality index that reflects the subjective and objective feelings of the person. A reference to noise evaluation.

The invention is applied to the visualization of the objective parameter distribution of the sound quality in the three-dimensional sound field and the spatial localization of the sound source which is most closely related to human hearing. When the existing sound quality analysis and measurement method evaluates the sound quality of the sound in the specified three-dimensional space, only the objective parameter information of the sound quality of a certain measurement position specified in the three-dimensional space can be obtained, and the distribution of the sound quality information in the entire three-dimensional space cannot be obtained. As well as visualizing the image, it is not only impossible to evaluate the sound quality of the sound field in the entire three-dimensional space, but also to give the spatial position of the sound source related to the sound quality.

Summary of the invention: The present invention overcomes the following shortcomings of the prior art: 1. Overcoming the existing optical quality objective parameter measurement, calculation analysis and evaluation methods and techniques alone, can not give the sound quality distribution and the visualization image in the entire three-dimensional space by one measurement, Defects in sound source position information related to objective parameters of sound quality cannot be given. 2. Overcoming the use of near-field acoustic holography to analyze the three-dimensional spatial sound field distribution without considering the subjective auditory sensation of human beings, can not give the defect of the sound source position in the sound field which is closely related to human subjective hearing sensation. The invention provides a method for generating a three-dimensional digital image of an acoustic quality objective parameter combined with a near-field acoustic holography method and a sound quality objective parameter analysis method. The sound quality objective parameter three-dimensional digital image generating method proposed by the invention has the sound field distribution, and can also give the sound source spatial position information which affects the subjective auditory feeling of the human being, and provides the sound in the whole three-dimensional space. The quality of the objective parameter distribution of the image, thus providing a visual, direct guidance for noise reduction and sound quality improvement.

The invention proceeds as follows:

1. Record the holographic sound pressure data on the holographic measurement surface using a microphone array (the microphone array can be a rigid surface spherical microphone array, a hollow sphere array, a microphone array conforming to the sound source structure or a planar array). A spherical microphone array is arranged in the closed sound field, and sound field sound pressure information is measured and recorded. Holographic sound pressure data is obtained from an array of planar microphones and other arrays of conformal microphones of any shape in open and semi-open sound fields.

2. Three-dimensional sound field acoustic volume reconstruction According to the measured holographic sound pressure data, the distribution information (sound pressure, normal particle velocity and normal sound intensity, etc.) of the acoustic field of the three-dimensional space is obtained by the near-field acoustic holography method. The form of the three-dimensional image is given.

The holographic sound pressure measured by a spherical microphone array is used to reconstruct the acoustics in three dimensions: distribution, the sound field transformation formula is as follows: n = 0

The 4TM in the formula is determined by the following formula:

In the above formula: (r, ^) is the spherical coordinate of any point in the three-dimensional space in the sound field; α is the radius of the spherical microphone array, k is earned, k = , « is the angular frequency, c is the speed of sound, ω 2π;

/ for the frequency. Α (" is the holographic sound pressure data collected on the microphone array; _A (r, A^iy) is the reconstructed sound pressure at the specified position (r, A) in the three-dimensional space; Y _n ^m {0, (^ % spherical harmonic The function, ( ) is the ball Bessel function, is the ball Hank function; represents the conjugate, "," represents the derivative, and N is the number of spherical harmonic expansion items. When specifying the position of the sound field reconstruction point in the entire three-dimensional space, then The sound pressure distribution of the entire sound field is obtained.

3. According to the sound field information in the three-dimensional space obtained in 2, calculate the distribution of the objective parameters (such as loudness, sharpness, roughness, etc.) of the sound quality of each point in the three-dimensional space, and give them in the form of three-dimensional images to realize the objective quality of sound quality. 3D visualization of parameters.

The single point sound quality loudness calculation model in the sound field is as follows: C GE^ + 2E _THRQ ) _ ( 2E _THRQ ) E _THR Q < E _T < 10

N'=

In the above formula, N/ is the characteristic loudness of the zth filter, which is the audible valve energy level, which is the energy level of the signal, and the unit is dB, C is the fixed value of 0.046871, when _500, the fixed value is 2.3067, cochlear The low-frequency gain G is 1, a is 0.2, and when <500Hz, E _THR Q, « can be obtained according to the discrete data interpolation calculation provided by ANSI. G is the cochlear low frequency gain. Therefore the total loudness formula is:

According to the three-dimensional spatial distribution results of the sound pressure in the sound field calculated in 2, combined with the calculation model of the single point sound quality objective parameters in the space, the coupled three-dimensional matrix mapping model of the sound pressure field and the loudness field in space is established.

^W l " ^372

^W l "^372

= \Ν[{τ,θ,φΧ(τ,θ, ) into 1x372

Or abbreviated as

Where: (r,^ ) is the spherical coordinate of the third point in the three-dimensional space, which is the angular frequency.

For frequency, mO.M, M are the number of frequencies corresponding to different sound sources. Is the sound pressure at any frequency at the third point in the sound field, Ρ, the sound pressure at each frequency at the third point in the sound field A vector consisting of reconstructed values, W is an auditory filter matrix consisting of 372 filters W, representing the response of the human ear to all frequencies in the audible band, and N' is the characteristic loudness vector. The sum of the points in the characteristic loudness vector N' can be obtained by the formula (4) to obtain the loudness of the specified point of the sound field, and the three-dimensional distribution result of the sound field loudness can be obtained by repeating the calculation process for the three-dimensional spatial node of the sound field. Other sound quality objective quantity three-dimensional distribution results can also be obtained using an analysis procedure similar to the loudness calculation, see Figure 1.

Through the above method, the objective parameter values of the sound quality at various positions in the measured space can be given, and the spatial distribution can be given in the form of a three-dimensional image, thereby identifying the position of the sound source having the greatest influence on the subjective hearing of the human. BRIEF DESCRIPTION OF THE DRAWINGS:

Figure 1. Flow chart for calculating the three-dimensional distribution of objective parameters of sound quality according to the present invention

Figure 2. Schematic diagram of a spherical microphone array

Figure 3. Schematic diagram of spherical microphone array and sound source distribution of dual source

Figure 4(a). Comparison of the loudness and sound pressure of the two-point source sound field of 3.5kHz (69dB) and lkHz (75dB)

Figure 4(b). Comparison of the sharpness and sound pressure of the sound field of the two-point sound source of 3.5 kHz (69 dB) and l kHz (75 dB)

Figure 4(c). Comparison of loudness and sound pressure calculations for sound fields of 3.5 kHz (70 dB) and 7 kHz (76 dB) sound sources

Figure 4(d). Comparison of the sharpness and sound pressure calculation results of the sound field of the two-point sound source of 3.5 kHz (70 dB) and 7 kHz (76 dB). The present invention will be further described below by way of specific examples. Referring to the drawings: The method for generating a three-dimensional digital image of an objective parameter of sound quality proposed by the present invention, the calculation result of which gives the distribution of the sound field, and also gives the spatial position information of the sound source which affects the subjective auditory feeling of the person, and provides the whole The image of the objective parameters of the sound quality in the three-dimensional space provides visual and direct guidance for noise reduction and sound quality improvement.

The invention proceeds as follows:

The holographic sound pressure measured by the spherical microphone array is used to reconstruct the acoustic quantity distribution in the three-dimensional space. The sound field transformation formula is as follows:

P _t (r ^, ∑

n =

The 4TM in the formula is determined by the following formula: p _t (α, θ, φ) Y _n ^m (θ, φ)* sin θάθάφ

0 JO ^

In the above formula: (r, ^) is the spherical coordinate of any point in the three-dimensional space in the sound field; a is the radius of the spherical microphone array, k is earned, k c, « is the angular frequency, c is the speed of sound, ω 2 2π;

/ for the frequency. Α (" is the holographic sound pressure data collected on the microphone array; _A (r, A^iy) is the reconstructed sound pressure at the specified position (r, A) in the three-dimensional space; „ ^m ( ) is the spherical harmonic function, ( ) For the ball Bessel function, ( ) is the ball Hank function; represents the conjugate, "," represents the derivative, and N is the number of spherical harmonic expansion items. When the position of the sound field reconstruction point in the entire three-dimensional space is specified, the whole is obtained. Sound pressure distribution of the sound field.

The single point sound quality loudness calculation model in the sound field is as follows:

C GF -\2F E <E _I < 10

N'=

In the above formula, N/ is the characteristic loudness of the Zth filter, which is the audible valve energy level, which is the energy level of the signal, and the unit is dB, C is the fixed value of 0.046871, when _500, the fixed value is 2.3067, the cochlea The low-frequency gain G is 1, a is 0.2, and when <500Hz, E _THR Q, « can be obtained according to the discrete data interpolation calculation provided by ANSI. G is the cochlear low frequency gain. Therefore the total loudness formula is:

According to the three-dimensional spatial distribution results of sound pressure in the sound field calculated in 2, combined with the calculation model of the objective parameters of single point sound quality in space, the coupled three-dimensional matrix mapping model of sound pressure field and loudness field in space is established.

Or abbreviated as:

Where: (r,^) is the spherical coordinate of the third point in the three-dimensional space, which is the angular frequency, (o _m = 2 f _m , is the frequency, m OM , M is the number of frequencies corresponding to different sound sources. The sound pressure at any frequency at the first point is the vector composed of the sound pressure reconstruction values at the frequencies at the third point in the sound field, and W is the auditory filter matrix composed of 372 filters W, representing the person The response of the ear to all frequencies in the audible frequency band, N' is the characteristic loudness vector. The sum of the parts in the characteristic loudness vector N' can be obtained by the formula (4) to obtain the loudness of the specified point of the sound field, and the three-dimensional space node of the sound field Repeating this calculation process can obtain the three-dimensional distribution of sound field loudness. Other sound quality objective quantity three-dimensional distribution results can also be obtained by an analysis process similar to loudness calculation, see Figure 1. In this example, spherical arrays are used for measurement. Array, as shown in Figure 2, on the sphere 1. As shown in Figure 3, arrange two pulsating ball sources in space: The parameters of sound source 1 are set to lkHz, 75dB, placed 0.3m on the positive X-axis of the space rectangular coordinate system (0.3m, 0, 0); The parameter of the sound source 2 is set to 3.5 kHz, 69 dB, and is placed at 0.3 m (-0.3 m, 0, 0) on the X negative half axis of the space rectangular coordinate system, that is, the angle between the two sound sources is 180. °, the radius of the spherical microphone array (Figure 2) is "0.1m. The three-dimensional spatial distribution of sound pressure, loudness and sharpness at a radius of 0.2 m is reconstructed using the calculation method given by the present invention. Fig. 4(a) is a comparison diagram of the calculation results of the three-dimensional spatial distribution of sound pressure and loudness when the sound source 1 of 1 kHz, 75 dB, and the sound source 2 of 3.5 kHz and 69 dB are simultaneously present. Fig. 4(b) is a comparison diagram of the calculation results of the three-dimensional spatial distribution of the sound pressure and the sharpness when the sound source 1 of the 1 kHz, 75 dB, and the sound source 2 of the 3.5 kHz, 69 dB sound source are simultaneously present.

2. The same two-sound source sound field model is used, but the parameter of the sound source 1 is set to 7 kHz, 70 dB, and placed at 0.3 m (0.3 m, 0, 0) on the x positive half-axis of the space rectangular coordinate system; The parameter of source 2 is set to 3.5 kHz, 76 dB, and is placed at 0.3 m (-0.3 m, 0, 0) on the x negative half-axis of the space rectangular coordinate system, and the angle between the two sources is still 180°. Figure 4 (c) and Figure 4 (d) are 7 kHz, 70 B sound source 1 and 3.5 kHz, 76 dB sound source 2 two sound sources using this method to reconstruct the spherical sound pressure, loudness and A comparison of the results of the three-dimensional spatial distribution of sharpness.

Comparing and analyzing the results given in the above figure, different from the traditional acoustic holography method for identifying the localized sound source according to the sound pressure, the method provided by the present invention can obtain the three-dimensional spatial distribution information of the objective parameters of the sound quality of the sound field, and recognize the subjective subject matter. The sound source position information closely related to hearing, Figure 4 shows the three-dimensional spatial distribution map of the objective parameters of sound quality such as loudness and sharpness, realizing the sound source localization based on human subjective auditory feeling, comparing Fig. 4(a) and Fig. 4(b), figure 4(c) and Fig. 4(d), it can be found that the spatial position with the highest sound pressure is not the same as the spatial position with the loudest loudness, and the spatial position with the loudest loudness is not the same as the spatial position with the sharpest sharpness. Therefore, it is necessary to locate the key sound source according to the subjective auditory feeling and take corresponding measures to achieve the purpose of sound field noise reduction and sound quality improvement.

Claims

An inter-distributed digital image generation method, as follows;

Step 1. Using a microphone array, the microphone array can be a rigid surface spherical microphone array, a hollow sphere array, a microphone array conforming to the sound source structure or a planar array, recording holographic sound pressure data on the holographic measurement surface.

A spherical microphone array is arranged in the closed inner sound field, and acoustic information such as sound field holographic sound pressure is measured and recorded. Holographic sound pressure data is obtained from an array of planar microphones and other arrays of conformal microphones in open and semi-open sound fields.

Step 2. Three-dimensional sound field acoustic volume reconstruction

According to the measured holographic sound pressure data, the distribution information (sound pressure, normal particle velocity and normal sound intensity) of the acoustic field of the three-dimensional space is obtained by the near-field acoustic holography method, and is given in the form of a three-dimensional image;

The acoustic volume of a three-dimensional space is reconstructed using a holographic sound pressure measured by a spherical microphone array.

P _t

The 4TM in the formula is determined by the following formula:

(2) In the above formula: (r, ^ ) is the spherical coordinate of any point in the three-dimensional space in the sound field; α is the radius of the spherical microphone array, which is the wave number, k = 0) lc , « is the angular frequency, c is the speed of sound, ω Two 2π;[ ,

/ for the frequency. Α (" » is the holographic sound pressure data collected on the microphone array; _A (r, A^ iy) is the reconstructed sound pressure at the specified position (r, ^) in the three-dimensional space; „ ^m ( ) is the spherical harmonic function, ( ) is the ball Bessel function, ( ) is the ball Hank function; represents the conjugate, "," represents the derivative, and N is the number of spherical harmonic extensions. When the position of the sound field reconstruction point in the entire three-dimensional space is specified, Sound pressure distribution of the entire sound field;

Step

3. Calculate the three-dimensional space according to the sound pressure distribution in the three-dimensional space obtained in step 2. The distribution of the objective parameters of the sound quality between the points is given in the form of a three-dimensional image to realize the three-dimensional visualization of the objective parameters of the sound quality;

Sound field

N:

In the above formula, N, 'is the characteristic loudness of the z-th filter, which is the audible valve energy level, which is the energy level of the signal, the unit is dB, C is the fixed value 0.046871, when _ 500Hz, E _THRQ is the fixed value 2.3067, the cochlear low-frequency gain G is 1, "0.2", and when <500H _Z , E-Q, « can be obtained according to the discrete data interpolation calculation provided by ANSI. G is the cochlear low frequency gain. Therefore the total loudness formula is:

According to the three-dimensional spatial distribution results of the sound pressure in the sound field calculated in the second step, combined with the calculation model of the objective parameters of the single point sound quality in the space, the coupling between the sound pressure field and the loudness field in the space is established. The matrix mapping model is:

^W l - ^W 372

[ p _t {τ,θ,φ,ω _χ ) -- p _t (r, θ, φ. ■ [Ν[(Γ,Θ, )- ₂ (Γ,Θ, )] 1x372

Or abbreviated as:

P,W = N' (6) where: (r,^) is the spherical coordinate of the third point in space, is the angular frequency, o _m =2 f _m , is the frequency, m two l..M, M is different The number of frequencies corresponding to the sound source is composed. The sound pressure at the specified frequency at the z-th point in the sound field is the vector composed of the sound pressure reconstruction values at the z-th points in the sound field, and W is the auditory filter matrix composed of 372 filters W. Representing the response of the human ear to all frequencies in the audible band, which is the characteristic loudness vector;

(4) The sum of the elements in the characteristic loudness vector N' can obtain the loudness of the specified point of the sound field, and the calculation process of the three-dimensional space node of the sound field can obtain the three-dimensional distribution result of the sound field loudness.