JP2000023299

Patent Translate
Powered by EPO and Google
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2000023299
[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a
sound image localization control device and a sound image localization control system that make
it feel as if a sound image is localized at a desired arbitrary position different from the position of
an actual speaker.
[0002]
2. Description of the Related Art A sound image forming method for controlling a sound image by
controlling signal levels (amplitudes) and phases of both ears has conventionally been disclosed
in, for example, Japanese Patent Application Laid-Open No. 2-298200. That is, in this sound
image forming method, the signal from the sound source is subjected to frequency analysis, and
the level difference and the phase difference (time difference) depending on the frequency are
given to both left and right channel signals to digitally control the localization of the sound
image.
[0003]
In order to localize a sound image, generally, a method of controlling a head-related transfer
function (HRTF) with a digital filter is used. The basic principle is shown in FIGS. 1 (a) and 1 (b).
As shown in FIG. 1 (a), this method is a method of collecting a sound image generated in the
10-05-2019
1
original sound field using a simulated head simulating a listener and supplying it to the listener.
As shown in FIG. 1 (b), the head related transfer functions (HRTF-R, HRTF-L) measured at each
incident angle of the sound wave are replaced with the electric filters 101 and 102, and the
direct sound collected in the original sound field is obtained. By processing the sound and the
reflected sound with the filters 101 and 102, it is possible to provide the listener with a sound
image equivalent to the system using the artificial head of FIG. 1 (a). In this case, control is
performed using sound image localization filters as the filters 101 and 102 in order to
approximate head transfer functions (HRTF-R, HRTF-L). At this time, in order to move the sound
image localization position, the transfer characteristics of all the necessary sound image
localization positions are measured in advance, this is held as a coefficient set of the sound image
localization filter, and the coefficients of the sound image localization filter (digital filter) are an
angle Switch by unit and use.
[0004]
First, the method of setting the characteristic value of the sound image localization filter will be
described. In general, the transfer function of the sound image localization filter is obtained by
generating white noise in an environment given to the system and measuring its impulse
response by arithmetic processing. : Shingaku Technical Report EA 93-1 1993-04 "). FIG. 2 shows
an example of the head related transfer function obtained in this manner. The coefficient value of
the digital filter that realizes the characteristics of FIG. 2 can be held as a coefficient set to realize
the sound image localization filter.
[0005]
FIG. 3 is a view showing an example of the arrangement of a conventional sound image
localization control apparatus. Referring to FIG. 3, the sound image localization control device
includes a CPU 201 that controls the whole, a coefficient ROM 202 that stores a coefficient set of
the sound image localization filter, an interface unit 203 that receives a position control signal,
and Rch (R And a sound image localization filter unit 205 for Lch (L-channel).
[0006]
Here, the Rch sound image localization filter unit 204 converts the Rch sound source input signal
(analog signal) into a digital signal, an A / D converter 211, a sound image localization filter
(digital filter) 212, and a sound image The coefficient buffer 213 is stored by loading coefficients
of the localization filter 212 from the coefficient ROM 202 and stored, and a D / A converter 214
10-05-2019
2
that converts an output signal (digital signal) from the sound image localization filter 212 into an
analog signal. .
[0007]
The Lch sound image localization filter unit 205 converts an Lch sound source input signal
(analog signal) into a digital signal, an A / D converter 221, a sound image localization filter
(digital filter) 222, and a sound image localization. A coefficient buffer 223 stores coefficients of
the filter 222 loaded from the coefficient ROM 202 and a D / A converter 224 which converts an
output signal (digital signal) from the sound image localization filter 222 into an analog signal.
[0008]
In such a configuration, the coefficient set of the sound image localization filters 212 and 222 is
stored in the coefficient ROM 202, loaded from the coefficient ROM 202 to the coefficient buffer
213 and 223 in accordance with the position control instruction from the CPU 201, I have
control.
Therefore, in this conventional method, the characteristic values of the sound image localization
filters 212 and 222 are determined by the coefficient set of the digital filter that realizes the
frequency characteristic obtained from the impulse response in a certain environment stored in
the coefficient ROM in advance. The coefficient set of the sound image localization filter for
realizing the characteristic value could not be changed.
[0009]
Also, after converting previously measured spatial impulse responses of a large number of people
and real ear headphone responses into feature parameter vectors corresponding to human
auditory characteristics, clustering is performed to use a reduced number of reduced data. There
is also a method (Japanese Patent Laid-Open No. 5-252598), but in this method, the
characteristic value of the sound image localization filter actually depends on the environment in
which the coefficient is set and the ear canality of the individual. Sometimes good characteristics
were not obtained.
[0010]
10-05-2019
3
Also, as a sound image localization filter, it is necessary to realize a complex frequency
characteristic as shown in FIG. 2, and for that purpose, a high-order FIR filter is required, and a
large scale hardware is required for its operation. In addition, storage capacity for a large amount
of filter coefficients was required.
On the other hand, if an IIR filter is used for this sound image localization filter, there is an
advantage that the change of characteristics is easy and versatile, and the coefficient memory can
be reduced.
However, IIR filters are difficult to design for realizing complicated frequency characteristics, and
tend to be unstable such as limit cycle oscillation.
[0011]
For example, when a transfer function having complex frequency characteristics as shown in FIG.
2 is designed with an IIR filter, it has been reported that a simple arbitrary characteristic can be
designed with a bi-quad digital filter (by Kasuga Masao) Literature "Real-time processing method
of digital equalizer The Institute of Electronics and Communication Engineers Information System
National Convention, pp 228, 1979"). This method is to design a target transfer function by
cascading bi-quad transfer functions having two zeros and poles described later. On the other
hand, there is also a method of designing an IIR filter by sampling frequency characteristics (the
document by Ochi Hiroshi "Design of an IIR filter Interface 206-213, (1996.11)"), but a large
number of filter orders are required. There are also problems such as not being of a certain
order.
[0012]
The above-mentioned system is basically a method of designing an analog transfer function and
Z-converting it (US Pat. No. 4,188.504). expressed.
[0013]
[Equation 1]
[0014]
10-05-2019
4
In the case of designing in this method, three parameters of fc (center (cutoff) frequency), Q
(sharpness) and L (gain) which are design parameters are appropriately changed in order to
approximate the target frequency characteristics. Then, a method of approximating to the design
characteristic is taken.
However, this method is a trial and error method that approximates while minimizing the
difference between the target characteristics and the characteristics of the frequency
characteristics of the design filter at appropriate frequency points, without clear guidelines, and
many frequency points. They have to be calculated over time, and have the disadvantage of being
inefficient.
Furthermore, in the case where the poles are close to each other, design problems such as
difficulty in changing parameters by Q are practically significant.
[0015]
Also, when changing the localization position of the sound image as described above,
conventionally, a position control signal is issued from the CPU 201, and the coefficient set of the
sound image localization filter is stored in memory in advance to switch the position. This is done
as needed by calling and switching, or by downloading coefficients to a designated memory area
(Japanese Patent Laid-Open No. 5-252598).
[0016]
However, when changing the coefficients of the sound image localization filter, it takes a certain
time for the filter characteristic to reach the desired characteristic value, so that switching noise
occurs when switching the sound image localization position. It had happened.
[0017]
SUMMARY OF THE INVENTION In order to avoid these problems, two sound image localization
filters are prepared, and when switching the localization angle, a fixed time until the operation of
the above filter becomes stable There is a method to switch after the elapse of.
Further, there is a system in which a variable attenuator is prepared and a cross fade is
10-05-2019
5
performed in order to move the localization position smoothly at the time of the switching
(Japanese Patent Application Laid-Open No. 6-245300).
An example of this circuit configuration is shown in FIG. Referring to FIG. 4, the sound image
localization control device includes a CPU 301 that controls the whole, a coefficient ROM 302
that stores a coefficient set of a sound image localization filter, an interface unit 303 that
receives a position control signal, and Rch (R And a sound image localization filter unit 305 for
Lch (L-channel).
[0018]
Here, the Rch sound image localization filter unit 304 converts the Rch sound source input signal
(analog signal) into a digital signal, an A / D converter 311, and two sound image localization
filters (FIR filters) 312a. , 312b and the coefficients of each of the sound image localization filters
312a and 312b are loaded from the coefficient ROM 302 and stored respectively from the
coefficient buffers 313a and 313b and the output signals (digital signals) from the sound image
localization filters 312a and 312b as analog signals It has D / A converters 314 a and 314 b to
convert and a fader 315.
[0019]
Also, the Lch sound image localization filter unit 305 converts the Lch sound source input signal
(analog signal) into an A / D converter 321, and two sound image localization filters (FIR filters)
322a, Coefficient buffers 322a and 323b in which the coefficients of the sound image
localization filters 322a and 322b are loaded from the coefficient ROM 302 and stored,
respectively, and output signals (digital signals) from the sound image localization filters 322a
and 322b are converted to analog signals And D / A converters 324 a and 324 b and a fader
325.
[0020]
Each of the faders 315 and 325 is configured as a cross fade means by two variable attenuators
and an addition means.
[0021]
In such a configuration, the CPU 301 reads out the required coefficient set from the coefficient
ROM 302 according to the position control signal of the sound image, and the two coefficient
10-05-2019
6
buffers possessed by the two FIR filters 312a and 312b of the sound image localization filter unit
304 for Rch. In one of the coefficient buffers 313a and 313b, and in one of the two coefficient
buffers 323a and 323b possessed by the two FIR filters 322a and 322b of the sound image
localization filter unit 305 for Lch.
Therefore, the two FIR filters 312a and 312b of the sound image localization filter unit 304 for
Rch operate with the coefficient for the old position and the coefficient for the new position, and
the two FIR filters 322a of the sound image localization filter unit 305 for Lch. , 322 b operate
with the coefficients relating to the old position and the coefficients relating to the new position,
by performing fade-in (new position) and fade-out (old position) processing for a certain time
using the faders 315 and 325 Even when the position of the sound image changes, smooth
replacement from the old position to the new position is possible with respect to the output
signal from the sound image localization filter.
[0022]
FIG. 5 is a diagram for explaining the timing of the cross fading process.
For example, consider the case where the sound image localization process is currently
performed at a position of 60 degrees, and the sound image localization process is performed at
a position of 90 degrees.
Now, it is assumed that one of the FIR filters 312a and 322a is in operation with a coefficient for
60 degrees supplied and the other FIR filters 312b and 322b are in non-operation. In this state,
as shown in FIG. 5 (e), if there is an instruction to switch the sound image localization position
from the position of 60 degrees to the position of 90 degrees in the CPU 301, the CPU 301
executes the timing shown in FIG. The coefficients for 90 degrees are supplied to the other FIR
filters 312b and 322b. Furthermore, the cross fade control signal is output from the CPU 301 at
the timing shown in FIG.
[0023]
Then, according to the cross fade control signal, the outputs of one of the FIR filters 312a and
322a are faded out as shown in FIG. 5 (d) by the faders 315 and 325, and the other FIR filters
10-05-2019
7
312b and 322b. The output of is faded in as shown in FIG. 5E and switched while crossfading
from one of the FIR filters 312a and 322a to the other of the FIR filters 312b and 322b. If
switching is performed while crossfading is performed over several tens of milliseconds,
switching coefficients can be switched to smoothly change the sound image localization position
without generating switching noise.
[0024]
However, although the FIR filters 312a, 312b, 322a and 322b in this case require high-order
filters depending on their impulse response times, their coefficient memory capacity is also large
and the hardware scale is large. . Therefore, preparing a plurality of sound image localization
filters for each of Rch and Lch (two for Rch and two for Lch) as described above has a
disadvantage that the circuit scale becomes very large.
[0025]
According to the present invention, a small-sized sound image localization filter is used, and for
each of Rch and Lch, the sound image localization position is smoothly controlled with a small
hardware scale by preparing only one small-sized sound image localization filter. It is an object of
the present invention to provide a possible sound image localization control device and a sound
image localization control system.
[0026]
SUMMARY OF THE INVENTION In order to achieve the above object, the invention according to
claim 1 comprises A / D conversion means for converting an input sound source signal into a
digital signal, each sound image localization position Of the sound image localization filter
simulating the head related transfer function measured for the above, the coefficient storage
means for storing the coefficients of the sound image localization filter, and the characteristics of
the sound image localization filter by changing the coefficients stored in the coefficient storage
means Control means for controlling the auditory position of the sound source signal, and D / A
conversion means for converting the digital output signal of the sound image localization filter
into an analog signal, wherein the sound image localization filter comprises For the sound image
localization filter simulating the head transfer function, the control means uses characteristic
parameter values for approximating the frequency characteristics of the given target head
transfer function. It is characterized in that it comprises a coefficient calculating means for
calculating a coefficient value of the sound image localization filter I.
10-05-2019
8
[0027]
The invention according to claim 2 is characterized in that, in the sound image localization
control apparatus according to claim 1, the coefficient calculating means changes the
characteristic value of the sound image localization filter by changing the characteristic
parameter value. And
[0028]
The invention according to claim 3 is the sound image localization control apparatus according
to claim 1 or 2, wherein the control means further causes the head to change the auditory
position of the sound source signal. Parameter interpolation value calculation means for
calculating an interpolation value between the characteristic parameter value at the position
before change and the characteristic parameter value at the position after change for the
characteristic parameter value approximating the frequency characteristic of the partial transfer
function; It is characterized by being equipped.
[0029]
According to the fourth aspect of the present invention, a single sound image localization filter,
coefficient switching means for switching the coefficients of the sound image localization filter to
change the characteristics of the single sound image localization filter, and coefficient switching
means Coefficient holding means for storing switched coefficients, input buffer means for storing
a sound source signal as an input signal of a sound source localization filter, output buffer means
for storing an output signal of a sound image localization filter for a fixed time, fade in / fade out
of output buffer means A fade-in / fade-out control means for controlling a function, and when
switching coefficients of a single sound image localization filter, an output signal before
switching coefficients in a single sound image localization filter and an output after switching the
coefficients It is characterized in that control is performed to cross fade the signal.
[0030]
DETAILED DESCRIPTION OF THE INVENTION Embodiments of the present invention will be
described below with reference to the drawings.
The present invention provides a design method for realizing complex frequency characteristics
using an IIR filter as a sound image localization filter in order to achieve the above object, and a
sound image localization control apparatus is configured using this method. .
10-05-2019
9
Furthermore, when moving (changing) the position of the characteristic value of the sound image
localization filter, interpolation operation between the reference position is performed based on
the characteristic parameter (characteristic parameter value approximating the frequency
characteristic of the head transfer function). By doing this, it is made to move smoothly between
positions.
[0031]
FIG. 6 is a view showing an example of the arrangement of a sound image localization control
apparatus according to the present invention.
Referring to FIG. 6, the sound image localization control device includes a control unit 1 for
controlling the whole, a sound image localization filter unit 4 for Rch (R-channel), and a sound
image localization filter unit 5 for Lch (L-channel). And have.
[0032]
Here, the Rch sound image localization filter unit 4 converts the Rch sound source input signal
(analog signal) into a digital signal, the sound image localization filter (digital filter) 12 and the
sound image A coefficient buffer 13 in which the coefficients of the localization filter 12 are
stored, and a D / A converter 14 for converting an output signal (digital signal) from the sound
image localization filter 12 into an analog signal.
[0033]
The Lch sound image localization filter unit 5 converts an Lch sound source input signal (analog
signal) into a digital signal, an A / D converter 21, a sound image localization filter (digital filter)
22, and a sound image localization. A coefficient buffer 23 in which the coefficients of the filter
22 are stored, and a D / A converter 24 for converting an output signal (digital signal) from the
sound image localization filter 22 into an analog signal.
[0034]
FIG. 7 is a view showing a specific example of the control means 1 of FIG. 6. In the example of
FIG. 7, the control means 1 of FIG. 6 is an interface unit 33 to which a position control signal is
input, a CPU 31, and parameters. Sound image localization based on the parameters calculated
by the parameter generation unit 34 that generates the parameter, the initial setting parameter
10-05-2019
10
storage unit 35 in which the parameter generated by the parameter generation unit 34 is
initialized, the parameter operation unit 36, and the parameter operation unit 36 A coefficient
generation calculation unit 37 that generates and calculates coefficients of the filters 11 and 22
and a parameter interpolation value calculation unit 38 that performs interpolation calculation of
parameter values are included.
[0035]
Here, the parameter generating unit 34 is an initial parameter based on the impulse response
measuring unit 41 for measuring the frequency characteristic of the head related transfer
function, and the frequency characteristic of the head related transfer function measured by the
impulse response measuring unit 41. And a parameter extraction unit 42 for extracting the
[0036]
The initial parameters extracted by the parameter extraction unit 42 are characteristic values of
a sound image localization filter forming a head transfer function measured in advance as shown
in FIG. Have parameters (fc, Q, L).
Here, fc is the center (cutoff) frequency of the sound image localization filter, Q is the sharpness,
and L is the gain.
[0037]
The initial parameters are sent to the parameter calculator 36 through the CPU 31, and the
parameter calculator 36 performs parameter calculations to obtain frequency characteristics, and
the frequency characteristics calculated by the parameter calculator 36 are coefficients. It is sent
to the generation operation unit 37.
Then, the coefficient generation calculation unit 37 calculates the coefficient values of the sound
image localization filters 12 and 22 to generate necessary coefficient sets, sends the generated
coefficient sets to the coefficient buffers 13 and 23, and outputs the sound image localization
filters 12 and 22. It is supposed to work.
10-05-2019
11
[0038]
Further, when changing the position, the parameter interpolation value calculation unit 38 uses
the interpolation value obtained by dividing the two positions at a certain interval for the abovementioned parameters (fc, Q, L), and changes its frequency characteristics. It is obtained and
given to the coefficient generation operation unit 37.
[0039]
Next, the processing operation of the sound image localization control apparatus having such a
configuration will be described.
First, an outline of the processing operation of the sound image localization control device of
FIGS. 6 and 7 will be described.
For example, when a position control signal according to a sound image localization instruction
(for example, an instruction to localize a sound source at a position 120 degrees right backward)
from a main CPU such as a game machine is input via the interface unit 33, this position control
signal In accordance with the above, the CPU 31 reads out from the initial setting parameter
storage unit 35 the initial parameters necessary to calculate the head-related transfer
characteristics with respect to the position information, and supplies the initial parameters to the
parameter calculation unit 36.
The parameter calculator 36 calculates the frequency characteristic based on the given initial
parameter, and supplies it to the coefficient generator 37.
The coefficient generation calculation unit 37 calculates the coefficient values of the sound
image localization filters 12 and 22 based on the frequency characteristic from the parameter
calculation unit 36 and transfers the calculated values to the coefficient buffers 13 and 23.
Specifically, the analog frequency characteristics obtained as described later are Z-transformed
and designed. This method can be easily obtained by converting the analog transfer function H
(s) into the transfer function Hz (Z) in the discrete domain using the S-Z conversion method (refer
to the document AV · Masao Kasuga Digital Signal Processing for OA, pp. 109-113. The
coefficient sequence is determined from the Z conversion equation obtained by this. The sound
10-05-2019
12
image localization filters 12 and 22 perform convolutional operation processing on the time axis
according to the coefficients set in the coefficient buffers 13 and 23 for the input sound source
input signals (sound source input signals of Rch and Lch). The signals calculated by the sound
image localization filters 12 and 22 are sent to D / A converters 14 and 24 and reproduced from
a pair of speakers. In the case of a game machine, for example, the coefficients of the sound
image localization filter are switched at any time by a sound image localization command from
the CPU 31 so as to correspond to the movement of the game object according to the operation
of the operator.
[0040]
First, the method of obtaining the characteristic values of the sound image localization filters (IIR
filters) 12 and 22 will be described in more detail. Also in the present invention, the design
scheme of the IIR filter is basically designed in the same manner as the above-described prior art
scheme, and is designed by Z conversion. That is, assuming that the sound source input signal is
X (s) and the transfer functions to the left ear EL (s) and right ear ER (s) of the listener are HL (s)
and HR (s), respectively, the following equation is obtained. Be
[0041]
[Equation 2]
[0042]
Note that HR (s) / HL (s) is the ratio of the characteristics between both ears.
As seen in Equation 2, the s-function system on the right side is a space transfer function having
these complicated frequency characteristics, and this is realized by a digital filter. As this digital
filter, an IIR digital filter is adopted by providing versatility in one type, taking into consideration
the number of multiplications, the filter coefficient memory, the number of registers, etc. and
achieving stable filter multiplexing. The filter function is a biquad type of zeros and poles, and
this is designed as a continuous function system, and this is Z-transformed to obtain a digital
filter. These properties are given by the center (cutoff) frequencies Fc, Q (sharpness), L (gain),
which can be changed by appropriate modifications. The transfer function H (z-1) is given by
Equation 1. However, in the case of designing a complex function system as shown in FIG. 2, this
design method does not establish a design method, and at present, it has to be designed by a
10-05-2019
13
heuristic method.
[0043]
In order to approximate an arbitrary characteristic, three design parameters Fc, Q and L will be
changed. As a method of changing this, there is a method of setting a desired characteristic as a
target value and approximating it while minimizing the difference between this and a design
value at an appropriate frequency point. However, this method is not efficient because it needs to
calculate them over many frequency points.
[0044]
Therefore, in the present invention, the document by Satoshi Miyauchi, Keito Mekada, Koji
Hasegawa, Masao Kasuga, Harutake Yasuda "Design of an IIR filter with complex amplitude
frequency characteristics and Proceedings of the Acoustical Society of Japan Conference 3-3-2-2
, As described in 571.572 (1997.3) , selecting a point near the point where the property of
the function changes, that is, the point where the positive and negative of the second derivative
of the function are inverted, Noting that it is appropriate as an interpolation point connected to,
the inflection point of the function is adopted as a frequency point used for design calculation. In
other words, by changing the three design parameters fc, Q, and L so as to minimize the
difference between the design value and the desired value at a total of three points at the center
frequency and the inflection points located before and after that, arbitrary characteristics are
obtained. Approximate
[0045]
Also, the design method of the filter is as follows. That is, for example, when designing a complex
function system having single-peak characteristics shown in FIG. 8, first, the filter function to be
designed is second-order differentiated, and these two inflection points p and q are selected.
Furthermore, the center frequency fc is selected to select these three points as sample frequency
points for calculating the frequency characteristics of the approximate design of the function.
Then, the difference between the design characteristic and the target characteristic at these
points is obtained by calculation, and the parameter Q is controlled to approximate the frequency
characteristic of the target so as to minimize the difference.
10-05-2019
14
[0046]
Referring to FIG. 9, a flow chart of the filter design scheme according to the present invention is
shown. Referring to FIG. 9, first, initialization parameters are set (step S1). Next, the parameter
value Q is changed (step S2), and an error value with the target filter (target characteristic) is
calculated (step S3). Next, it is determined whether the error value is smaller than a
predetermined threshold value (step S4). If not smaller, the error value is set to the threshold
value (step S5), and the process returns to step S2 again to change the parameter value Q. The
above process is repeated, and when the error value becomes smaller than the threshold value,
the optimum parameter value Q can be set (since the parameter value Q approximating the target
frequency characteristic is obtained), the process ends. Do.
[0047]
A more specific design flow chart is performed according to the following procedure. (1) Input
the target frequency characteristics. (2) Next, the number of filter stages for achieving the target
characteristic and rough design parameters (fc, Q, L) are input. (3) Next, the filters are ranked in
descending order of L value of the input filter, and the following processing is performed in that
order. -Align the center frequency fc of the input design characteristics with the center frequency
of the target characteristics. -Match the L value of the input design characteristics to the target
characteristics. Find two inflection points, one for the higher frequency and one for the lower
frequency. Change the Q value so as to reduce the difference between the frequency
characteristics at the two inflection points. (4) If it is determined that the amount of error has
converged (for example, 0.1 dB or less), the process ends.
[0048]
By the above procedure, characteristic parameter values required for the target sound image
localization filter are determined.
[0049]
Next, more specific processing operation of the sound image localization control apparatus of the
present invention will be described.
10-05-2019
15
The initial parameters are stored in the initial parameter storage unit 35. This initial parameter is
a characteristic value of the sound image localization filter which forms a head transfer function
measured in advance as shown in FIG. 2 and has parameters (fc, Q, L) in a required position unit
(for example, 30 degrees unit). doing. This initial parameter is obtained by the parameter
extraction unit 42 based on the frequency characteristic of the head related transfer function
measured by the impulse response measurement unit 41.
[0050]
Thus, after the initial parameters are stored in the parameter storage unit 35, the initial
parameters are sent to the parameter operation unit 36 through the CPU 31, and the parameter
operation unit 36 performs parameter operation to obtain frequency characteristics. It is
obtained and given to the coefficient generation operation unit 37. The coefficient generation
operation unit 37 converts the analog frequency characteristic value obtained as described above
into a transfer function in the discrete domain by Z conversion, and obtains a necessary
coefficient set. The obtained coefficient set is sent to the coefficient buffers 13 and 23 to operate
the sound image localization filters 12 and 22.
[0051]
Next, the interpolation calculation method of the head related transfer function between
positions when changing positions will be described. In the present invention, when changing the
position, for example, the parameter interpolation value calculation unit 38 performs
interpolation calculation of parameter values held in units of 30 degrees. Specifically, for the
above-mentioned parameters (fc, Q, L), the frequency characteristic is determined using an
interpolation value obtained by dividing between two positions at certain intervals. That is, this
frequency characteristic is determined by obtaining each interpolation value of the parameters
(fc, Q, L) by the parameter interpolation value calculation unit 38 shown in FIG. 7, and the sound
image localization filter for each frequency characteristic at a certain time interval The coefficient
generation calculation unit 37 calculates the coefficients of Then, the generated coefficients are
sent to the coefficient buffers 13 and 23 each time.
[0052]
10-05-2019
16
As described above, using the interpolation value obtained by dividing the two positions by a
certain interval in the parameter interpolation value calculation unit 38, the frequency
characteristic is determined and the coefficient is calculated to change the position of the sound
image. Can be moved smoothly. As for the filter problem with respect to the change of the
coefficient set between the divided positions at this time, the positions can be switched more
smoothly by using the fade-in / fade-out function described later.
[0053]
FIG. 10 is a view showing an example of the arrangement of a sound image localization control
apparatus having a fade in / fade out function. The sound image localization control apparatus of
FIG. 10 uses a single sound image filter for each of Rch and Lch. It has a function to realize cross
fade when switching coefficients. Referring to FIG. 10, this sound image localization control
device includes a CPU 51 that controls the whole, a coefficient ROM 52 storing coefficient sets of
sound image localization filters, an interface unit 53 to which a position control signal is input,
and Rch (R- And a sound image localization filter unit 55 for Lch (L-channel).
[0054]
Here, the sound image localization filter unit 54 for Rch comprises an A / D converter 61 for
converting a sound source input signal (analog signal) of Rch into a digital signal, and one sound
image localization filter (FIR filter) 12 A coefficient buffer 64 in which coefficients of the sound
image localization filter 12 are loaded from the coefficient ROM 52 and stored respectively, a
selector SEL, an input buffer 62 storing sound source signals to be input signals of the sound
source localization filter 12, sound image localization Output buffers 65 and 66 for storing the
output signal of the filter 12 for a fixed time, a buffer control unit 63 for controlling the input
buffer 62, the output buffers 65 and 66, and the selector SEL, an output signal from the fader 67
and the sound image localization filter 12 And D / A converter 68 for converting (digital signal)
into an analog signal.
[0055]
Further, the sound image localization filter unit 55 for Lch comprises an A / D converter 71 for
converting a sound source input signal (analog signal) of Rch into a digital signal, a sound image
localization filter (FIR filter) 22 and A coefficient buffer 74 in which each coefficient of the sound
image localization filter 22 is loaded from the coefficient ROM 52 and stored, a selector SEL, an
input buffer 72 for storing a sound source signal to be an input signal of the sound source
localization filter 22; 22. Output signals from the sound image localization filter 22 (the output
buffers 75 and 76 for storing the output signals of 22 for a predetermined time, the buffer
10-05-2019
17
control unit 73 for controlling the input buffer 72, the output buffers 75 and 76, and the selector
SEL) And D / A converter 78 for converting a digital signal into an analog signal.
[0056]
Each of the faders 67 and 77 is configured as a cross fade means by two variable attenuators and
an addition means.
[0057]
In such a configuration, in the normal case, the selector SEL is set to the sound image localization
filters 12 and 22 side and the output buffers 65 and 75 side, and the A / D converters 61 and 71
according to the sample period. The digital signal sequence output from the signal processing
unit is processed by the sound image localization filters 12 and 22 described above, and is sent
to the output buffers 65 and 75.
The output buffers 65 and 75 are FIFO type registers, and the signals from the sound image
localization filters 12 and 22 are output with a delay according to a sample period for a certain
fixed time (fade out time).
Also at this time, the digital buffers from the A / D converters 61 and 71 are stored in the input
buffers 62 and 72.
[0058]
Next, when the sound image localization position change signal (position control signal) is input
to the interface unit 53, the coefficients of the coefficient buffers 64 and 74 are first replaced.
After that, the selector SEL is switched, and the output of the sound image localization is
calculated by the sound image localization filters 12 and 22 at high speed using the digital signal
string of the input buffers 62 and 72 stored up to then. It is stored in the buffers 66 and 76. The
calculation of the sound image localization filter does not depend on the sample period, and since
the known data of the input buffers 62 and 72 can be used as it is, high-speed calculation is
possible.
10-05-2019
18
[0059]
The contents of the output buffers 65 and 75 and the contents of the output buffers 66 and 76
thus obtained become the output data for the old coefficient and the output data for the new
coefficient, and in the faders 67 and 77, If fade-in / fade-out processing is performed using these
two data as shown in FIG. 11, it is possible to smoothly move the sound image without causing a
problem at the time of coefficient switching.
[0060]
As described above, according to the first aspect of the present invention, A / D conversion
means for converting input sound source signals into digital signals, and each sound image
localization position The characteristics of the sound image localization filter are manipulated by
changing the coefficients stored in the sound image localization filter that simulates the
measured head-related transfer function, the coefficient storage unit that stores the coefficients
of the sound image localization filter, and the coefficient storage unit Control means for
controlling the position of the sound source signal in the auditory sense, and D / A conversion
means for converting the digital output signal of the sound image localization filter into an
analog signal, wherein the sound image localization filter comprises a single IIR filter Is used for
the sound image localization filter simulating the head related transfer function, and the control
means uses sound parameter determination to approximate the frequency characteristics of the
given target head related transfer function. It has coefficient calculation means for calculating the
coefficient value of the filter, and the sound image localization filter is simplified by using a
single IIR filter instead of the FIR filter which conventionally required large-scale hardware.
Sound image localization control can be performed by hardware.
In particular, according to the present invention, it is possible to provide a sound image
localization apparatus which is small in circuit scale and can be mounted on game machines,
personal computers and the like, and is excellent in the sound image localization feeling.
[0061]
According to the second aspect of the present invention, in the sound image localization control
device according to the first aspect, the coefficient calculating means changes the characteristic
value of the sound image localization filter by changing the characteristic parameter value.
Conventionally, the characteristic value of the sound image localization filter, which could not be
easily changed since the coefficient set for the impulse response measured in advance was held,
10-05-2019
19
is the characteristic parameter (center frequency, cut-off frequency, sharpness, gain)
representing the frequency characteristic. It can be easily changed by changing it.
[0062]
According to the invention of claim 3, in the sound image localization control apparatus
according to claim 1 or 2, when the control means further changes the auditory position of the
sound source signal, Parameter interpolation value calculation for calculating an interpolation
value between the characteristic parameter value at the position before change and the
characteristic parameter value at the position after change for the characteristic parameter value
approximating the frequency characteristic of the head related transfer function By including the
means, it is possible to easily calculate the interpolation value of the parameter between the two
positions of the sound image localization filter, and to move the sound image smoothly.
[0063]
Further, according to the invention of claim 4, a single sound image localization filter, coefficient
switching means for switching the coefficients of the sound image localization filter to change
the characteristics of the single sound image localization filter, and the coefficient switching
Coefficient holding means for storing the coefficients switched by means, input buffer means for
storing the sound source signal as an input signal of the sound source localization filter, output
buffer means for storing the output signal of the sound image localization filter for a fixed time,
fade in of the output buffer means · A fade-in / fade-out control means for controlling a fade-out
function, and when switching the coefficients of a single sound image localization filter, after
switching the output signal and coefficients before switching the coefficients in a single sound
image localization filter Control to cross-fade the output signal of the Using a plurality of largescale FIR filter when switching of the functions of the fade-out, it is possible to perform a single
sound image localization filter.
Also for the IIR filter, although the time until the filter characteristics stabilize is short, similar
effects can be obtained.
[0064]
Brief description of the drawings
[0065]
10-05-2019
20
1 is a diagram for explaining the basic principle of the sound image localization control system.
[0066]
2 is a diagram showing an example of a head related transfer function.
[0067]
3 is a diagram showing an example of the configuration of a conventional sound image
localization control device.
[0068]
4 is a diagram showing an example of the configuration of a conventional sound image
localization control device having a function of performing cross fade.
[0069]
5 is a diagram for explaining the timing of the cross fading process.
[0070]
<Figure 6> It is the figure which shows the constitution example of the sound image localization
control device which relates to this invention.
[0071]
7 is a diagram showing a specific example of the control means of FIG.
[0072]
<Figure 8> It is the figure which shows the complex function system which consists of the
unimodal characteristic.
[0073]
<Figure 9> It is the flowchart in order to explain the filter design system of this invention.
[0074]
<Figure 10> It is the figure which shows the constitution example of the sound image localization
10-05-2019
21
control device which relates to this invention.
[0075]
11 is a diagram for explaining the fade in · fade out process in the sound image localization
control device of FIG.
[0076]
Explanation of sign
[0077]
1 control means 4 sound image localization filter unit 5 for Rch (R-channel) sound image
localization filter unit 11 for Lch (L-channel) 11, 21 A / D converter 12, 22 sound image
localization filter 13, 23 coefficient buffer 14, 24 D / A converter 31 CPU 33 interface unit 34
parameter generation unit 35 initial setting parameter storage unit 36 parameter operation unit
37 coefficient generation operation unit 38 parameter interpolation value operation unit 41
impulse response measurement unit 42 parameter extraction unit 51 CPU 52 coefficient ROM 53
interface unit 61, 71 A / D converter 62, 72 Input buffer 63, 73 Buffer control unit 64, 74
Coefficient buffer 65, 75 Output buffer 66, 76 Output buffer 67, 77 Fader 68 D / A converter
10-05-2019
22

Download Report