JP2008104229

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 JP2008104229
The present invention provides a loudspeaker system that includes drivers of various physical
sizes and can achieve predetermined and constant directivity over a large area of both vertical
and horizontal planes. The present invention is a multi-channel loudspeaker system providing a
compact loudspeaker arrangement and a filter design methodology operating in the digital signal
processing domain. Furthermore, loudspeaker systems can be designed to include drivers of
various physical sizes, and can achieve the above-mentioned constant directivity for large areas
in both vertical and horizontal planes. [Selected figure] Figure 1
Loudspeaker row system
[0001]
The present invention relates generally to multi-way speaker systems. In particular, the invention
relates to a multi-way speaker system comprising an array of multiple drivers capable of
achieving high quality sound.
[0002]
Related Art High quality loudspeakers in the audio frequency range typically use multiple
specialized drivers using dedicated components for the audio frequency band, eg, tweeters
(generally 2 kHz to 20 kHz) and midrange drivers (generally 200 Hz) KHz 5 kHz) and woofer
(generally 20 Hz to 1 kHz). The driver's acoustic output is only on one line perpendicular to the
10-05-2019
1
loudspeakers (due to the need for spacing as required due to the physical dimensions of the
specialized driver comparable to the wavelength of the acoustics being dissipated Usually, socalled acoustic centers) are integrated into the frequency-independent response of the intended
plane. Off its axis, the frequency response is somewhat distorted by the interference caused by
the different path lengths that the sound waves reach from the driver to an assumed point in
space. Historically, many attempts have been made to create loudspeakers with smooth off-axis
response and controlled sound fields for large spaces. For example, D'Appolito uses a geometric
method (ie, a central tweeter and two woofers placed symmetrically along the vertical axis) to
eliminate roving errors in multi-way loudspeakers Configuration) was proposed. Several
loudspeaker manufacturers adopted the method and even extended it with an array of midrange
drivers and woofers arranged symmetrically around one or two central twitters. Manufacturers
designed by D'Appolito and adopting D'Appolito's approach use digital filters that substitute
analog filters in passive or analog crossover circuits or in the digital domain . Analog or passive
crossover circuits necessarily introduce phase distortion. Moreover, with this design, spacing is
not optimal and generally results in too large to completely avoid off-axis deviation from the
ideal smooth response.
[0003]
In an alternative solution, the basic design concept is to apply a very steep, "brick wall" finite
impulse response (FIR) filter to avoid large band transitions, thereby making the error unheard .
However, apart from audible discontinuities, the response of the polarity of the associated driver
individual may still differ at the transition point. Thus, with this design solution, it may be
difficult to achieve a given, smooth polarity of operation throughout the audible range. In yet
another alternative, Van der Wal can achieve logarithmically spaced transducer arrays with very
well-controlled directivity (which is nearly constant across a wide frequency range in one
dimension) Suggests that. Some embodiments of this technology are described in US Pat. No.
6,128,395. As with the previous techniques, this design technique is constrained. Because (i) the
log interval is determined according to only a certain formula, (ii) the filter design is only valid in
certain cases, and (iii) significant if the actual interval deviates from the logarithmic interval
Errors can occur because they can be unavoidable due to the physical size of the driver or design
constraints. Furthermore, this design is limited to one type of driver (i.e. a full range driver),
limiting its application to public address systems. U.S. Patent No. 6,128,395
[0004]
Thus, the prior art provides a loudspeaker system that includes drivers of various physical sizes
and can achieve a predetermined constant directivity over a large area of both vertical and
horizontal planes. There is still a need for loudspeaker configurations and filter designs that
10-05-2019
2
overcome the limitations of.
[0005]
SUMMARY OF THE INVENTION The present invention is a multi-way loudspeaker-speaker
system capable of producing high quality sound from a single, compact, line-arranged
loudspeaker, which is typical of left, right, front and back. It can be utilized in a conventional
surround sound entertainment system having surround sound channels and a center channel.
[0006]
In one embodiment, the line array is disposed in a single housing with a sealed compartment
separating the specific drivers from one another to prevent driver coupling or assembled as a
single unit And a plurality of tweeters, a midrange driver, and a woofer.
The line array may be a single channel array with various signal paths from the input to
individual loudspeaker drivers or multiple drivers.
Each signal path comprises a digital input and includes a digital FIR filter and a power D / A
converter connected to either a single driver or multiple drivers.
[0007]
The performance, positioning and placement of the line array loudspeaker driver may be
determined by a filter design algorithm that determines the coefficients of each FIR filter of each
of the loudspeaker signal paths. A cost minimization function is applied to the predetermined
frequency points using the initial driver position and the initial directivity goal function, which
determines frequency points on a logarithmic scale within the relevant frequency range. If the
results obtained from the application of the cost minimization function do not meet the system
performance requirements, then the driver's location can be changed to the next, and the cost
minimization function is applied again until the obtained results meet the system requirements It
can be done. Once the results obtained meet the system requirements, the linear phase filter
coefficients of each FIR filter in a single signal path are computed using Fourier approximation or
other frequency sampling method.
10-05-2019
3
[0008]
The multi-way loudspeaker of the present invention may include built-in DSP processing, a D / A
converter, and an amplifier, and may be coupled to a digital network (eg, the IEEE 1394
standard). Furthermore, the multi-way speaker system of the invention can be designed as a wall
mountable surround system due to its compact size.
[0009]
Multi-way speaker systems can use different sized drivers to reduce distortion and allow for high
power operation. This is because, for an equivalent broadband driver arrangement, specialized
drivers can operate optimally in their dedicated frequency band. The multi-way speaker design of
the present invention can also provide better control of room response with a smooth off-axis
response. This system, as well as the control of total sound power, can additionally control the
frequency response of the reflected sound, thereby suppressing floor and ceiling reflections.
[0010]
The loudspeaker according to the present invention is placed at a position approximately
crossing the x and y axes of the loudspeaker and is symmetrical along the loudspeaker with one
central driver and both the x and y axes with respect to the central driver. The loudspeaker
having a central driver and at least two drivers different in size, and symmetrically disposed with
respect to the central driver receiving digital output signals from the at least one power D / A
converter, respectively A driver and the two drivers, and a digital output signal filtered through
at least one digital FIR filter.
[0011]
In one embodiment, the central driver is a tweeter.
[0012]
In one embodiment, the at least two drivers are woofers.
[0013]
In one embodiment, the at least two drivers are midrange drivers.
10-05-2019
4
[0014]
One embodiment further comprises at least two additional drivers located at points further from
the central driver than the at least two drivers.
[0015]
In one embodiment, the at least two additional drivers are woofers.
[0016]
In one embodiment, the at least two drivers are tweeters, the loudspeaker further comprises at
least two additional transducers, and the central driver and the at least two drivers are
symmetrical with respect to the central driver. Located between additional transducers.
[0017]
In one embodiment, the at least two additional transducers are mid-range speakers.
[0018]
In one embodiment, the at least two additional transducers are woofers.
[0019]
Another loudspeaker according to the invention is a central tweeter located at a point on the
loudspeaker designed as a base point, and at least two mid-range drivers located symmetrically
with respect to the base point, the size being greater than that of the central tweeter At least two
woofers that are larger than the at least two midrange drivers and larger in size than the at least
two midrange drivers, and located farther from the central tweeter than the at least two
midrange drivers; And at least two woofers arranged symmetrically with respect to a base point,
wherein the central tweeter, the at least two mid-range drivers, and the at least two woofers
output digital output signals from at least one power D / A converter. Receiving, the digital
output signal comprising at least one digital FI It is multiplied by the filter through the filter.
In one embodiment, the system further includes at least two additional tweeters arranged
symmetrically with respect to the central twitter and located between the central tweeter and the
10-05-2019
5
at least two midrange drivers.
[0020]
In one embodiment, near the opposite edge of the loudspeaker such that the central tweeter, at
least two mid-range drivers and the at least two woofers are located between the at least two
additional woofers. And at least two additional woofers located at.
[0021]
Yet another loudspeaker according to the invention is at least one central tweeter and at least
two additional tweeters, one of the at least two additional tweeters being at least two on each
side of the central tweeter. An additional tweeter, at least two midrange drivers, one of the at
least two midrange drivers being at least two midrange drivers located on each side of the at
least two additional twitters; A woofer, wherein one of the at least two woofers further comprises
at least two woofers located on each side of the at least two midrange drivers, the at least one
central tweeter and at least two additional ones. A tweeter and at least two midrange drivers,
Both two woofers receive digital output signals from at least one power D / A converter,
respectively, and the digital output signals are filtered through at least one digital FIR filter.
[0022]
A method of designing a line array loudspeaker system according to the invention comprises:
determining the initial driver position; setting the initial directivity target function of the system;
a cost minimization function based on the initial directivity target function And applying linear
phase filter coefficients for each filter of the system.
[0023]
In one embodiment, the initial driver position is coordinates with respect to the origin of the
loudspeaker.
[0024]
In one embodiment, the frequency points are established on a logarithmic scale with a
predetermined frequency range based on the predetermined initial directivity target function.
[0025]
In one embodiment, the cost minimization is a function applied at the frequency points that starts
incrementally from the lowest frequency.
10-05-2019
6
[0026]
In one embodiment, the method further comprises the step of confirming the result obtained
from the cost minimization function against the desired performance reference value.
[0027]
In one embodiment, adjusting the initial position of the driver if the result obtained from the cost
minimization function is not optimal, establishing a new initial position of the driver based on the
adjusted driver initial position. And reapplying the cost minimization function based on the new
initial position of the driver.
[0028]
In one embodiment, the Fourier approximation is used to determine the linear phase filter
coefficients.
[0029]
Other systems, methods, features and advantages of the present invention will be or become
apparent to one with skill in the art upon consideration of the following figures and detailed
description.
It is intended that all such additional systems, methods, features and advantages be included
within this description, be within the scope of the present invention, and be protected by the
accompanying claims.
[0030]
DETAILED DESCRIPTION The present invention may be better understood with reference to the
following drawings.
Elements in the figures are not necessarily to scale, emphasis instead being placed upon clearly
illustrating the principles of the invention.
10-05-2019
7
Furthermore, in the figures, like reference numerals are assigned to corresponding parts
throughout the different figures.
[0031]
FIG. 1 shows a block diagram of the signal flow for each one-dimensional (1D) multi-way
loudspeaker 100 embodiment of the present invention and the loudspeaker driver of system
100.
As shown in FIG. 1, the multi-way loudspeaker 100 may design the loudspeaker as six paths
comprising: (i) a central tweeter 102, connected to a first power D / A converter 103, ( ii) two
additional tweeters 104 and 106 connected to the second power D / A converter 105, and (iii)
two midrange drivers 108 connected to the third power D / A converter 107. And 110, (iv) two
middle band drivers 112 and 114 connected to the fourth power D / A converter 109, and (v)
two power lines connected to the fifth power D / A converter 111. Four woofers 120, 122, 124
and 126 connected to woofers 116 and 118 and (vi) sixth power D / A converter 113.
The connections between the loudspeakers for each amplifier indicate different directions of the
multiway loudspeakers.
Thus, the loudspeaker may be designed as a single channel multiway loudspeaker.
[0032]
In figure (1), the drivers (also called transducers) are separate enclosed compartments 128, 130,
132, 134, 140, as indicated by the separators 136, 138, 144, 146, 150 and 152. It can be
installed in a housing 154 constituted by 142 and 148.
By placing the drivers in separate enclosed compartments, the connection of adjacent drivers is
minimized.
10-05-2019
8
Although various compartments can be seen in FIG. 1, the loudspeaker system can be designed
such that the compartments are not visible to the consumer when implemented on the finished
product.
The compartment 128 containing the woofers 120 and 122 can be separated by a separator 136
from the compartment 132 containing the woofer 116.
Similarly, the compartment 130 containing the woofer 126 and 124 can be separated by the
separator 138 from the compartment 134 containing the woofer 118.
The midrange drivers 112 and 114 included in the compartments 140 and 142 may be
separated from the compartments 132 and 134 by separators 144 and 146, respectively.
All tweeters 102, 104, 106 and midrange drivers 110 and 108 may also be included in
compartment 148 and separated from compartments 140 and 142 by separators 150 and 152.
[0033]
According to FIG. 1, central tweeter 102, tweeters 104 and 106, midrange drivers 110, 108, 112,
114, 116 and 118, low frequency woofers 120, 122, 124 and 126 are linearly aligned along the
y-axis. It is installed symmetrically with respect to 102.
In a typical arrangement, tweeters 102, 104 and 106 having an outer diameter of about 40 mm,
midrange drivers 110, 108, 112, 114, 116 and 118 having an outer diameter of about 80 mm,
and woofers 120, 122, 124 and 126 having an outer diameter of about 120 mm. May be
included.
Typically, transducer cone sizes may vary based on the desired application and the desired array
size.
Furthermore, the transducer may utilize neodymium magnets, but it is not necessary for the
10-05-2019
9
described application to utilize that particular type of magnet.
[0034]
The central tweeter 102 may be located on the y-axis at a center point 0, which is the
intersection of the x-axis and the y-axis.
The tweeters 104 and 106 may be placed at their centers about +/− 40 mm from the center
point.
The midrange drivers 110 and 108 may be placed with their centers at about +/- 110 mm from
the center point 0. The midrange drivers 112 and 114 may be placed at their centers
approximately +/- 220 mm from the center point. The low frequency woofers 116 and 118 may
then be placed about +/− 350 mm in their centers from the center point. The low frequency
woofers 120 and 124 may be centered about +/- +/- 520 mm from the center point. The low
frequency woofers 122 and 126 may then be placed about +/- +/- 860 mm in their centers from
the center point.
[0035]
FIG. 1 also shows a block diagram 160 of the signal flow of the multi-way loudspeaker system.
While FIG. 1 shows the six directions 162, 164, 166, 168, 170 and 172 of the signal flow, the
channels can be split in more than one direction. The signal stream may comprise a digital input
174 which may be implemented using a standard interface format (e.g. SPDIF or IEEE 1394 and
derivatives thereof) to drive the driver through various paths or directions as shown in FIG. It is
possible to connect. Each path or direction 162, 164, 166, 168, 170 and 172 connects to either
a single or multiple loudspeaker driver and a digital FIR filter 176 and power D / A converters
103, 105, 107, 109. , 111 and 113 may be included. Power D / A converters 103, 105, 107,
109, 111 and 113 can be realized as a cascade of conventional audio D / A converters (not
shown) and power amplifiers (not shown) or directly It can be implemented as a class D (class-D)
power amplifier (not shown) with digital inputs. FIR filter 176 may be implemented by a digital
signal processor (DSP) (not shown). The loudspeaker driver may be a tweeter as shown, a
midrange driver or a woofer (e.g. those shown).
10-05-2019
10
[0036]
In operation, the outputs of each of the multiple FIR filters 176 are connected to multiple power
D / A converters 103, 105, 107, 109, 111 and 113, which are then connected to the baffle of the
housing 154. It is supplied to the installed multistage loudspeaker driver 102, 104, 106, 108,
110, 112, 114, 116, 118, 120, 122, 124 and 126. One or more drivers (eg, 120, 122, 124 and
126) may be coupled in parallel with the path or signal path 162 including the power D / A
converter 113.
[0037]
Figure 2 is another one-dimensional multi-way loudspeaker, similar to the loudspeaker in Figure
1, but with two midrange drivers instead of four and four woofers instead of six Include. In
particular, FIG. 2 shows a single-channel one-dimensional four-pass loudspeaker 200 having a
central tweeter 202 surrounded by two additional tweeters 204 and 206. In addition,
loudspeaker 200 includes two mid-range drivers 208 and 210 and four woofers 214, 216, 218
and 220. The tweeters 202, 204 and 206, the midrange drivers 208 and 210, and the four
woofers 214, 216, 218 and 220 are all arranged linearly along the y-axis symmetrically with
respect to the central tweeter 202.
[0038]
Three signal paths (not shown) may be provided to the compartment 226. The first path may be
provided to central tweeter 202, the second path may be provided to tweeters 204 and 206, and
the third path may be provided to midrange drivers 208 and 210. Just above and below
compartments 226 are compartments 222 and 224 containing woofers 214 and 218 (divided by
the separators indicated by lines 228 and 230 respectively) and woofers 216 and 220
respectively. The woofers 214, 218, 216 and 220 may all be supplied by the fourth path.
[0039]
Typical arrangements of the multi-way loudspeakers shown in FIG. 2 are tweeters 202, 204 and
206 of about 40 mm outside diameter, midrange drivers 208 and 210 of about 80 mm outside
diameter, and woofer 214 of about 160 mm outside diameter, 216, 218 and 220 may be
10-05-2019
11
included. As mentioned above, transducer cone sizes may vary based on the desired application
and the size of the desired array. The number of signal paths and the number of drivers of a
particular type may also vary.
[0040]
The central tweeter 202 may be located on the y-axis at a center point 0 (the intersection of the x
and y axes in FIG. 2). The tweeters 204 and 206 may then be placed with their centers +/- 40
mm from the center point.
[0041]
Next, the midrange drivers 208 and 210 may be placed with their centers at center point 0 to
about 110 mm. For low frequencies, woofers 214 and 216 may be placed their centers +/- about
240 mm from the center point. Next, the low frequencies woofers 218 and 220 may be placed
their centers +/- about 380 mm from the center point.
[0042]
FIG. 3 is a flow chart of a filter design algorithm 300 used to design a loudspeaker system of the
present invention. The purpose of the filter design algorithm 300 is to determine the coefficients
of each FIR filter of each signal path of the loudspeakers. As shown in detail below, an initial
driver position and an initial directivity goal function are initially determined (step 310). The
initial position or design layout of the speakers and drivers will be determined according to the
application's many variables (eg, desired speaker size, intended application or use, manufacturing
constraints, aesthetic or other product design) Be done. Next, driver coordinates are defined for
each driver along the major axis. Next, an initial guess of the directional goal function is
established, which involves establishing frequency points on a logarithmic scale within the
relevant interval. The cost function is then minimized at predetermined frequency points (step
312). If the result does not meet the performance requirements of the system (step 314), then
the location of the driver is then changed and the cost minimization function is reapplied (step
316). This cycle can be repeated until the result meets the requirements. Once the results meet
the requirements, linear phase filter coefficients are computed (step 318). Additional operations
(step 320) may also be performed to equalize the drivers, compensate for phase shift, and
improve beam steering.
10-05-2019
12
[0043]
In a first step 310, an initial driver position and an initial directivity objective function are set. As
mentioned above, the number, position, dimensions and orientation of the driver are mainly
determined by the product design form. Once the orientation has been determined, then the
initial coordinate values are the initial driver coordinates p (n) of the N driver on the main axis, n
= 1. . . It may be defined for N. For example, in the one-dimensional (1D) array shown in FIG. 1, N
= 13: p (n) = [−. 86,-. 52,-. 35,-. 22,-. 11,-. 04, 0,. 04.11,. 22,. 35,. 52,. 86] m (meter).
[0044]
To determine the initial directivity target function, an initial guess of the directivity target
function T (f, q) must be defined, and the directivity target function is determined based on the
desired performance of the driver at a particular angle q Be done. FIG. 4 is a graph illustrating an
example set of target functions for angle dependent attenuation at five specific angles q. The
directivity objective function identifies the desired noise level attenuation in decibel (y-axis)
notation, and the measurement is at a sufficiently large distance (larger than the size of the
loudspeaker) in the anechoic environment to the origin (central tweeter) It can be measured at
various frequencies at an angle q degrees away from the perpendicular line. The frequency
vector f identifies a set of frequency points (e.g. 100) on a logarithmic scale within the relevant
interval (e.g. 100 Hz ... 20 kHz).
[0045]
The angle vector q (i), i = l,. . , N q specify a set of angles at which optimization is achieved. Figure
4, on the other hand, shows the first guess for directivity at five set angles: (Nq = 5): q = [0, 10,
20, 30, 40] °, in most cases it is two angles Only, i.e., (Nq = 2) may be sufficient to specify
directivity. In this case, the targeted directivity can be determined by the outer angle, for example
40 degrees and 0 degrees (zero directivity specification on axis) ie q = [0, 40] °.
[0046]
10-05-2019
13
With the exception of the on-axis objective function, the objective function for each angle is
double logarithmic up to a value T <0 decibel of the frequency fc (eg fc = 350 Hz) specified from
T = 0 decibel (dB) at f = 0 It descends linearly on the scale and maintains a constant value
thereafter. The on-axis objective function 402 remains constant at 0 dB over the entire frequency
range. All of the target pointing functions at 404 at 10 degrees, 410 at 410 degrees, 412 at 30
degrees, and 414 at 40 degrees start at T = 0 dB and descend on the double logarithmic scale
until the function reaches fc (Figure 4 at 350 Hz), and then maintain a constant value throughout
the remaining analysis frequency range.
[0047]
After the initial driver position and the initial directivity target function have been determined,
the next step 312 starts with the lowest step-wise frequency increment (e.g. 100 Hz) and each
step following the solution obtained as the first solution To minimize the cost function F (f) at a
given frequency vector point f by using the following equation:
[0048]
Use the following
[0049]
Where Hm (n, f, q) is the set of measured amplitude characteristics of the normalized targeted
driver n, frequency f and angle q for the response obtained on the axis (angle 0) An example is
shown in FIG.
FIG. 5 shows frequency response measurements 500 of a single-fit titter at various vertical
displacement angles normalized to the axis.
In FIG. 5, line 502 shows the on-axis response, line 504 is the frequency response measured at
10 degrees, line 506 is the response at 20 degrees, line 508 is the response at 30 degrees, Line
510 is the frequency response measured at 40 degrees, all measured in the frequency range
between 1 kHz and 20 kHz.
[0050]
10-05-2019
14
Furthermore, the minimization is realized by changing the real-valued frequency point C opt (n, f)
of the channel filter, where n is the driver index in the interval [0, 1] and f is the interval [0, 1 ]
Frequency. In addition, the constraints C opt (n, f) = 0, f> fo, f <fu must be achieved and depend
on the characteristics of the particular driver n. For example, in the case of the woofer, the upper
operating limit is fo = 1 kHz, the lower operating limit in the twiter is fu = 2 kHz, and for the
midrange driver fu = 300 Hz, fo = 3 kHz Obtained.
[0051]
The above procedure for minimizing the cost function can be achieved by the function
"fminsearch", which is part of the Matlab (R) software package owned and distributed by The
Math Works. The "fminsearch" function of the Matlab software package uses the Nelder-Mead
simplex algorithm or a derivative of them. Alternatively, an exhaustive search over a defined grid
over constrained parameter ranges may be applied. Other methodologies can also be used to
minimize the cost function.
[0052]
If the deviation between the obtained result and the target is small enough or acceptable as
determined by the person skilled in the art to apply a particular design, then the FIR filter
coefficients of each signal path of the line array will be Obtained. FIG. 6 is a graph 600 of
acceptable obtained results of a line array similar to that shown in FIG. 1 determined along the yaxis. The graph shows the resulting filter frequency response V (f, q) after passing through step
314 of FIG. Passing means that the result meets the requirements. In FIG. 6, line 602 represents
the on-axis response V (f, q (1)), line 604 is the frequency response V (f, q (2)) at 10 degrees, and
line 606 is 20 degrees. Is the frequency response V (f, q (3)), line 608 is the frequency response
V (f, q (4)) at 30 degrees, and line 610 is V (f, q (5) at 40 degrees. ))), All shown in the frequency
range between 50 Hz and 20 kHz.
[0053]
FIG. 7 shows the six signals of the line array loudspeaker system shown in FIG. 1 once the cost
minimization function has been applied and the results obtained are sufficiently small or found to
be within the acceptance range of the desired application 8 is a graph 700 illustrating the
10-05-2019
15
resulting frequency response C opt (n, f) of each of the paths. The line represented by L1 or 702
is the frequency response of the first signal path feeding the central channel tweeter 102 (FIG. 1)
and the second signal path feeding L2 or 704 to the tweeters 104 and 106 (FIG. 1) L3 or 706 is
the frequency response of the third signal path feeding the midrange drivers 110 and 108 (FIG.
1), and L4 or 708 feeds the midrange drivers 114 and 116 (FIG. 1). L5 or 710 is the frequency
response of the fourth signal path, L5 or 710 is the frequency response of the fifth signal path
feeding woofer 116 and 118, L6 or 812 is the sixth signal path feeding woofer 120, 122, 124
and 126 Frequency response of
[0054]
If the deviation between the obtained result and the target is not acceptable in the particular
design application (ie, if it is too large), then the driver position or geometry, and / or the
parameters q (i) and the target The fc of the function T (f, g) (see FIG. 3) must then be changed.
Once modified, the cost minimization function must be applied again, and the process must be
repeated until the obtained results and goals are small enough or results that are acceptable for
the application.
[0055]
As shown in FIG. 3, once the driver position and driver geometry are deployed, so that the
algorithm obtains results within the acceptance range of the target function, each signal path n =
1. . . N FIR filter coefficients must then be determined (depicted as step 318 in FIG. 3). One way
to determine the FIR coefficients is to use Fourier approximation (frequency sampling) to obtain
a linear phase filter at a given angle. When applying Fourier approximation or other frequency
sampling methods, the angle selection must be made so that the approximation is accurate
enough.
[0056]
The Fourier approximation can be achieved by the function "firls", which is part of the software
package Matlab (R) owned and distributed by The Math Works. Similar methodologies can be
used by implementing other software systems to minimize the cost function.
10-05-2019
16
[0057]
FIG. 8 is a graph 800 showing the frequency response of linear phase FIR filter 804 after FIR
filter coefficients have been obtained according to the method described above, with the
frequency response of one signal path 802 being equivalent to L4 or 708 of FIG. It is.
[0058]
In addition, modifications can be made to the FIR filter to equalize the measured frequency
response of one or more drivers (especially with a tweeter, midrange driver).
The impulse response of this type of filter can be obtained by known methods, which, as
mentioned above, must be convolved with the impulse response of the linear phase channel filter
when determining the FIR filter coefficients. Furthermore, the voice coil (acoustic center of the
driver) can not be aligned. To compensate for this, appropriate delays can be incorporated into
the filter by adding leading zeros to the FIR impulse response.
[0059]
Furthermore, delays can be added to each channel according to the following equation: Δt = p /
c · sin α, (p = driver coordinates, c = 345 m / s) where the main acoustic beam (alternately to the
main axis, Can be oriented at a desired direction at an angle α.
[0060]
Furthermore, the geometrical arrangement of the one-dimensional layout is such that the design
process can be performed in two dimensions (ie along the x and y axes) by symmetrical
geometrical arrangement as described above. It can be changed.
Due to symmetry, the same directivity characteristics occur along the y-axis (vertical) except at
higher corner frequencies.
[0061]
10-05-2019
17
While various embodiments of the invention have been described, it will be apparent to those
skilled in the art that many more embodiments and implementations are possible within the
scope of the invention. Accordingly, the invention is not to be restricted except in light of the
attached claims and their equivalents.
[0062]
FIG. 6 illustrates an example of a one-dimensional six-pass loudspeaker system installed
symmetrically along the y-axis at a base point, and is a block diagram illustrating signal flow to
each loudspeaker driver of the system. FIG. 5 illustrates an example of a one-dimensional (1D)
four-pass loudspeaker system using nine loudspeaker drivers symmetrically placed along the yaxis at a root point. Fig. 5 is a flow chart of a filter design algorithm used to design a loudspeaker
system. FIG. 7 is a graph illustrating a directional target function of angle dependent attenuation.
FIG. 7 is a graph showing measured values of amplitude characteristics of a single-impelled
twister at various out-of-axis displacement angles. FIG. 2 is a graph showing acceptable obtained
results of a line sequence similar to that shown in FIG. 1 determined along the y-axis. FIG. 7 is a
graph showing the frequency response of the digital filter assigned to the signal path of the line
array design shown in FIG. 1 after application of the cost minimization function. FIG. 8 is a graph
showing the smoothed frequency response of the third signal path shown in FIG. 7 with the
frequency response of the linear FIR filter after the FIR filter coefficients have been established
and applied.
Explanation of sign
[0063]
DESCRIPTION OF SYMBOLS 100 multi way loudspeaker 102 center tweeter 103 1st power D / A
converter 104, 106 additional tweeter 105 2nd power D / A converter 107 3rd power D / A
converter 108, 110 middle-range driver 109 1st Four-power D / A converter 111 Fifth power D /
A converter 112, 114 Mid-range driver 113 Sixth power D / A converter 116, 118 Woofers 120,
122, 124, 126 Woofers 128, 130, 132, 134 , 140, 142, 148 compartments 136, 138, 144, 146,
150, 152 separator 154 housing
10-05-2019
18

Download Report