JP6083322B2 - Channel divider and audio playback system including the same - Google Patents

Description

  The present invention relates to a channel divider used for reproducing an audio signal by a multi-way speaker system and an audio reproducing system including the channel divider, and in particular, corresponding to a speaker system of 3 ways or more, a low-pass filter, a band-pass filter, and In addition, the present invention relates to a channel divider in which each high-pass filter is configured by a linear phase filter that does not cause a Gibbs phenomenon, and the reproduction sound quality of a multi-way speaker system can be appropriately changed.

  In a multi-way speaker system including a woofer for low-frequency range playback and a tweeter for high-frequency range playback, it is common to include a network circuit that divides the audio signal amplified by the amplifier in accordance with the playback frequency band of each speaker unit. It is. In recent years, a network system corresponding to a woofer and a network circuit corresponding to a tweeter can be independent of each other, and a speaker system corresponding to a bi-wiring connection provided with an input terminal connected to each of them has become widespread. In bi-wiring connection, bi-amp driving using a power amplifier connected to the network circuit of the woofer and another power amplifier connected to the network circuit of the tweeter is adopted.

  On the other hand, as a method of realizing multi-amplifier (bi-amplifier) driving without using the network circuit of the speaker system, there is a case where a channel divider is provided in front of the power amplifier and the audio signal is divided into bands by this channel divider. In the audio playback system of the multi-amplifier-multi-way speaker system as described above, the channel divider divides the input audio signal into at least a low-frequency side output signal and a high-frequency side output signal and outputs it to the multi-amplifier. To do. By using a channel divider, LPF (low-pass filter) or HPF (high-pass filter) having a steeper transition band than the network of the speaker system can be set, the crossover frequency can be set relatively freely, etc. (For example, patent document 1). In addition, when a channel divider is used, the low-frequency amplifier circuit and the high-frequency amplifier circuit are independent, reducing the cross-modulation distortion that occurs when the low-frequency component and the high-frequency component are superimposed, reducing the playback sound quality. It is said that there is an advantage that it is excellent.

  In recent years, there has been an attempt to give a DSP a channel divider function so that a speaker system without a network circuit can be connected using an AV receiver including a DSP and a multi-amplifier (Patent Document 2, Patent Document). 3). When the channel divider is realized by a digital filter, an FIR filter or an IIR filter may be used. In order to use the channel divider, it is necessary to appropriately set the crossover frequency by measuring the reproduction band of each speaker unit of the speaker system using a microphone or the like (Patent Documents 4 and 5).

  When the multi-way speaker system includes a woofer for low-frequency reproduction and a tweeter for high-frequency reproduction, and a squawker for mid-range reproduction (sometimes called a mid-range speaker or sometimes a full-range speaker). The channel divider needs to be able to set a BPF (band pass filter). In order to set the BPF, a first crossover frequency that defines the band division between the woofer and the squawker and a second crossover frequency that defines the band division between the squawker and the tweeter are determined. The BPF may be configured by deriving from the first crossover frequency and the second crossover frequency. However, the BPF is an HPF that sets the low frequency side crossover frequency and the low frequency side crossover frequency. An LPF for setting the over frequency may be connected in series.

  In conventional channel dividers, the LPF on the low-frequency side, the BPF on the middle-frequency range side, and the HPF on the high-frequency range side can be selected from several standard filter characteristics, and have a phase shift characteristic of a linear phase. A non-recursive filter (finite length impulse response filter, FIR filter) may be used. In particular, an FIR filter obtained by approximating a cyclic filter (infinite length impulse response filter, IIR filter) by cutting it to a finite length causes a Gibbs phenomenon in which the pass level of the stopband increases due to errors caused by the cutting. There is a problem. On the other hand, as an FIR filter in which the Gibbs phenomenon does not occur, there is a filter having a linear phase phase shift characteristic called a Willkinson filter, a Herrmann filter, or an Apodizing filter. In the case of the phase shift characteristic of the linear phase, the delay is constant with respect to the frequency, and the phase delay characteristic and the group delay characteristic are flat.

  However, in the case of a multi-way speaker system of 3 Way or more, the setting of the respective filter characteristics becomes complicated, so that a smooth amplitude frequency characteristic can be obtained at either the first crossover frequency or the second crossover frequency. However, there is a problem that a dip may occur at the other frequency and the amplitude frequency characteristic may change. In order to use a filter having a phase shift characteristic of these linear phases for a channel divider corresponding to a multi-way speaker system of 3 Way or more, if the FIR filter is not appropriately configured, the distortion of the waveform of the output signal becomes large, and the reproduced sound is reproduced. There is a problem that it appears as a timbre change. Further, when the multi-way speaker system has a network circuit, it is necessary to cope with it. Therefore, even when a desired gain characteristic is set in the filter of the channel divider, it is preferable to configure an FIR filter that does not cause the Gibbs phenomenon appropriately.

JP 2005-109969 A JP 2002-111399 A JP 2005-184149 A Japanese Patent No. 4321315 Japanese Utility Model Publication No. 5-39097

  The present invention has been made to solve the above-described problems of the prior art, and an object of the present invention is to provide a channel divider and an acoustic reproduction system for use in reproducing an audio signal by a multi-way speaker system. Corresponding to the above speaker systems, the low-pass filter, band-pass filter, and high-pass filter are each composed of linear phase filters that do not cause Gibbs phenomenon, so that the reproduction sound quality of the multi-way speaker system is appropriately It is an object of the present invention to provide a channel divider and a sound reproduction system that can be changed.

The channel divider of the present invention includes a low-pass filter unit that divides an input audio signal into a first output signal on the low frequency side and outputs the first output signal to a first output terminal, and an audio signal higher than the first output signal. A high-pass filter section that divides a band into second output signals on the sound range side and outputs the second output signals to the second output terminal; and a sound signal that is on the high sound side of the first output signal and on the low sound side of the second output signal A band pass filter unit that divides a band into a third output signal and outputs the band pass filter unit to a third output terminal, wherein the low pass filter unit, the high pass filter unit, and the band pass filter unit are: Both are composed of a finite-length impulse response filter, and the Bernstein has a sum of 1 using a continuous function g (x) of a real number x in the range [0, 1] as a coefficient of the finite-length impulse response filter. In the following mathematical expression 1, which is a term expression, the real number x is defined as the following mathematical expression 2, the continuous function g (x) is defined as the desired reference composite characteristic, and the order m is defined as the following mathematical expression 3, A coefficient obtained by giving a delay equal to or more than a predetermined number of samples to the coefficient of the finite-length impulse response filter according to the term is set in the low-pass filter unit, and the coefficient of the finite-length impulse response filter according to the third term of Equation 1 below is set A coefficient obtained by giving a delay equal to or greater than a predetermined number of samples is set in the high-pass filter unit, and a delay equal to or greater than the predetermined number of samples is given to the coefficient of the finite-length impulse response filter according to the second term of Equation 1 below. The coefficient obtained in this way is set in the band pass filter section.

(However, in the formula, 0 <x <1, n is an integer of 1 or more. 0 <m1 <m2 <... <Mb <N: an integer of 1 or more. )

(However, in the formula, 0 <x <1, fs: sampling frequency.)

(However, in the formula, N is an integer of 1 or more. Fc: crossover frequency.)

(However, in the formula, 0 <m1 <m2 <... <Mb <N: an integer of 1 or more.)

  Preferably, in the channel divider of the present invention, the continuous function g (x) indicating a desired reference synthesis characteristic is equal to 1, the waveform of the input audio signal, the first output signal, the second output signal, and the third output. The waveform of the sum signal with the signal matches.

  Preferably, in the channel divider of the present invention, the band-pass filter unit is composed of a band filter by a plurality of M finite-length impulse response filters that output a plurality of M (M: an integer of 2 or more) third output signals. The

  Preferably, the channel divider of the present invention includes a level adjustment circuit for adjusting the level of the first output signal, the second output signal, or the third output signal, and the first output signal, the second output signal, or the third output signal. And a delay circuit that adjusts the delay time of the first output signal, the second output signal, or the third output signal.

  In addition, the sound reproduction system of the present invention corresponds to the above-described channel divider, an amplifier including an amplifier circuit corresponding to the first output terminal, the second output terminal, and the third output terminal of the channel divider, and at least the first output signal. And a speaker system including a tweeter corresponding to the second output signal, a tweeter corresponding to the third output signal, a squawker or a woofer, and capable of bi-wiring connection with the amplifier.

  The operation of the present invention will be described below.

  In the channel divider of the present invention, the input audio signal is divided into at least a first output signal on the low frequency side, a second output signal on the high frequency side of the first output signal, and a high frequency side of the first output signal. And band-dividing into a third output signal that is lower than the second output signal and outputting the result to the first output terminal, the second output terminal, and the third output terminal, respectively. The channel divider includes an amplifier including an amplifier circuit corresponding to each output terminal, a woofer or subwoofer corresponding to at least the first output signal, a tweeter corresponding to the second output signal, and a tweeter corresponding to the third output signal. Alternatively, a sound reproduction system including at least a squawker or a woofer and a speaker system capable of bi-wiring connection with an amplifier is configured.

  Note that the 3-way multi-way speaker system includes a case where a sub-woofer that reproduces the low-frequency side is added to a 2-way multi-way speaker system including a woofer and a tweeter. Therefore, the user can adjust the channel divider to change the frequency band and the reproduction level of each speaker unit that are reproduced in an overlapping manner, so that there is an advantage that the reproduction sound quality of the speaker system can be easily adjusted. The network circuit of a general speaker system is ± 6 to 18 dB / Oct. The channel divider LPF, BPF, and HPF are ± 24 to 96 dB / Oct. The above transition region characteristics are also possible, and by introducing a channel divider, the attenuation rate of the transition region and the stop region can be increased.

  The channel divider of the present invention includes a low-pass filter unit that divides an input audio signal into a first output signal on the low frequency side and outputs the first output signal to a first output terminal, and an audio signal higher than the first output signal. A high-pass filter section that divides a band into second output signals on the sound range side and outputs the second output signal to the second output terminal; and a sound signal that is on the high sound range side of the first output signal and on the low sound range side of the second output signal A band-pass filter unit that divides the band into third output signals and outputs the third output signals to the third output terminal. In the channel divider of the present invention, each of the low-pass filter unit, the high-pass filter unit, and the band-pass filter unit is configured by a finite-length impulse response filter that is a linear phase filter that does not cause a Gibbs phenomenon.

Here, the coefficients of the filters of the finite-length impulse response filters in the low-pass filter unit, the high-pass filter unit, and the band-pass filter unit are obtained by the above Equations 1 to 3. In other words, the coefficient of each finite-length impulse response filter is a real number in the above-described Equation 1 which is a Bernstein polynomial in which the sum using a continuous function g (x) of a real number x in the range [0, 1] is 1. x is defined as Equation 2 above, continuous function g (x) is defined as a desired reference composite characteristic, and order m is defined as Equation 3 above, and the coefficient of the finite-length impulse response filter according to the first term of Equation 1 is used. A coefficient obtained by giving a delay equal to or greater than the predetermined number of samples is set in the low-pass filter unit, and a delay equal to or greater than the predetermined number of samples is given to the coefficient of the finite impulse response filter according to the third term of Equation 1 above. The coefficient obtained in this way is set in the high-pass filter unit, and a delay equal to or greater than the predetermined number of samples is added to the coefficient of the finite-length impulse response filter according to the second term of Equation 1 above. Setting the Ete obtained coefficients to the band pass filter section. Note that the delay equal to or greater than the predetermined number of samples may be a delay equal to or greater than one sample, and the delay required for the FIR filter coefficient obtained as a result of using Equation 2 above to satisfy causality. Or more than the number of samples equal to.

  As a result, the channel divider of the present invention is composed of a low-pass filter, a band-pass filter, and a high-pass filter with linear phase filters that do not cause a Gibbs phenomenon so as to be suitable for a 3-way multi-way speaker system. The playback sound quality of the multi-way speaker system can be changed appropriately. In particular, if the continuous function g (x) indicating the desired reference synthesis characteristic is made equal to 1, the waveform of the input audio signal, the first output signal, the second output signal, and the third output signal The waveform of the sum signal can be matched. Therefore, when the multi-way speaker system of 3 Way or more is supported, the amplitude frequency characteristic does not change and a peak or dip does not occur near the crossover frequency.

  Further, the third output signal that is the output of the bandpass filter unit may be further divided into a plurality of M (M: an integer of 2 or more). That is, in the channel divider, the band pass filter unit includes a band filter by a plurality of M finite-length impulse response filters that output a plurality of M third output signals. Therefore, it is possible to cope with a multi-way speaker system of 4 Way (in the case of M = 2) or more, and similarly, a peak or dip does not occur near the crossover frequency without changing the amplitude frequency characteristic.

  The channel divider of the present invention includes a level adjustment circuit for adjusting the level of the first output signal, the second output signal, or the third output signal, and the phase inversion of the first output signal, the second output signal, or the third output signal. And a delay circuit that adjusts a delay time of the first output signal, the second output signal, or the third output signal. The range to adjust the playback sound quality will be expanded, and it will be compatible with various multi-way speaker systems.

  The channel divider and the sound reproduction system of the present invention are configured with a low-pass filter, a band-pass filter, and a high-pass filter, each of which is a linear phase filter that does not cause a Gibbs phenomenon, corresponding to a speaker system of 3 ways or more. Thus, the playback sound quality of the multi-way speaker system can be changed appropriately.

It is a figure explaining the sound reproduction system containing the amplifier apparatus 1 by preferable embodiment of this invention. Example 1 4 is a graph for explaining frequency characteristics (gain characteristics) of a low-pass filter LPF, a band-pass filter BPF, and a high-pass filter HPF of the channel divider 12 of the amplifier device 1. Example 1 5 is a graph for explaining a frequency characteristic (gain characteristic) of a sum of channel dividers 12 of the amplifier device 1. Example 1

  Hereinafter, a channel divider and an audio reproduction system including the same according to preferred embodiments of the present invention will be described, but the present invention is not limited to these embodiments.

  FIG. 1 is a diagram illustrating a sound reproduction system according to a preferred embodiment of the present invention. Specifically, the sound reproduction system includes an amplifier device 1 including a channel divider and speaker systems 7L and 7R connected to the amplifier device 1, and FIG. 1 is a block diagram showing the internal configuration of each. FIG. 2 is a graph for explaining the frequency characteristic (gain characteristic) of the low-pass filter unit 13 of the channel divider 12 of the amplifier device 1, and FIG. 3 is a frequency characteristic (gain characteristic) of the sum of the channel divider 12. It is a graph explaining about. In addition, illustration and description are abbreviate | omitted about the one part structure unnecessary for description, an internal structure, etc.

  The sound reproduction system includes an amplifier device 1 and speaker systems 7L and 7R connected to the amplifier device 1, converts digital signal data data input to the amplifier device 1 into stereo audio signals L and R, and the amplifier device. This is a sound reproduction system for reproducing stereo sound by a speaker system 7 composed of two speakers 7L and 7R after being amplified at 1. The amplifier device 1 includes a DSP and a multi-amplifier, and can be connected to a multi-amplifier that operates a channel divider. The speaker system 7 is a 3-way bi-wiring SS including a woofer WO, a squawker SQ, and a tweeter TW, and is bi-wiring connected to the amplifier device 1 by a speaker cord. Therefore, the user who uses the amplifier device 1 can adjust the reproduction quality of the speaker system 7 by adjusting the channel divider to change the frequency band and the reproduction level in which the woofer and the tweeter overlap and reproduce. Become.

  The amplifier device 1 includes a DSP (digital signal processor) 10 that performs signal processing of digital signal data data, a D / A converter 2 that receives outputs from the DSP 10 for four channels, and converts them into analog signals, and these analog signals. And an amplifier circuit 3 for amplifying each of the signals and outputting them to the speaker system 7. The stereo audio signals L and R may be supplied to the DSP 10 via an A / D converter (not shown) as stereo signals (left signal L and right signal R) supplied in analog. The amplifier device 1 includes a CPU 4 that is a control circuit that controls the whole, an operation unit 5 that is connected to the CPU 4 and receives an instruction input from a user, and a display circuit 6 that includes a display. Specifically, the amplifier device 1 can be configured by a DSP that supports multi-channel sound, an AV receiver that incorporates a multi-channel amplifier circuit, and the like. The operation unit 5 includes an input device such as a switch, a jog dial, or a remote control device. The display circuit 6 may be a built-in FL display, a liquid crystal display, or the like, or a display device connected to another. Of course, the amplifier device 1 may be constituted by another sound reproducing device including the DSP 10, the D / A converter 2, the multi-channel amplifier circuit 3, and the CPU 4 such as a microcomputer. Further, the DSP 10 included in the amplifier device 1 may be independent as a single channel divider device.

  The speaker system 7 is a 3-way bi-wiring SS including a woofer WO, a squawker SQ, and a tweeter TW, and includes a network circuit and an input terminal corresponding to each speaker unit. The low frequency range reproduction woofers WO of the left speaker 7L and the right speaker 7R are connected to the input terminal tL, respectively. Further, the squawker SQ for mid-range reproduction is connected to the input terminal tM. In addition, the tweeter TW for high sound range reproduction is connected to the input terminal tH.

  Since the woofer WO and the squawker SQ and the tweeter TW of the speaker system 7 mainly have predetermined frequency bands capable of reproducing sound, the crossover frequency fc1 between the woofer WO and the squawker SQ is between the squawker SQ and the tweeter TW. Is divided by a channel divider with the crossover frequency fc2 as a boundary. The woofer WO for low frequency reproduction mainly reproduces the frequency band below the crossover frequency fc1, the squawker SQ mainly reproduces the frequency band below the crossover frequency fc1 to fc2, and the tweeter TW represents the crossover frequency. A frequency band of fc2 or higher is mainly reproduced. The speaker system 7 is adjusted using a channel divider so as to realize a relatively flat synthesized sound pressure frequency characteristic.

  The DSP 10 includes therein a decoder 11 that converts input data data into stereo signals L and R, and a channel divider 12 that includes a digital filter. The DSP 10 outputs the output terminals (DL 1, DL 2, DL 3, DR 1, DR 2, DR 3) for 6 channels of the channel divider 12 to the D / A converter 2. The channel divider 12 includes a low-pass filter unit 13, a band-pass filter unit 15, a high-pass filter unit 14, a first output adjustment circuit 13a, and a third output corresponding to the left signal L and the right signal R, respectively. An adjustment circuit 15a and a second output adjustment circuit 14a are included. The CPU 4 controls the filter settings of the low pass filter unit 13, the band pass filter unit 15 and the high pass filter unit 14 of the channel divider 12 of the DSP 10. Stereo signals L and R which are output signals of the decoder 11 of the DSP 10 are branched and input to a low-pass filter unit 13, a high-pass filter unit 14 or a band-pass filter unit 15 which are FIR filters, respectively.

  The low-pass filter unit 13, the band-pass filter unit 15 and the high-pass filter unit 14 of the channel divider 12 of the present embodiment are respectively a low-pass filter LPF or a high-pass filter HPF or a band-pass filter constituted by FIR filters. It is a digital filter configured to include a filter BPF. The low-pass filter unit 13 band-divides the input audio signal into a first output signal on the low-frequency range side, and outputs the first output signal to the first output terminal D1. The high-pass filter unit 14 divides the input audio signal into a second output signal on the high-frequency range side, and outputs the second output signal to the second output terminal D2. In addition, the band pass filter unit 15 divides the input audio signal into a third output signal that is higher in the frequency range than the first output signal and lower in the frequency range than the second output signal, and the third output terminal D3. Output to.

  The output of the low-pass filter unit 13 is input to the first output adjustment circuit 13a and output to the output terminals (DL1, DR1) of the DSP 10 as the first audio output signal D1. The output of the high-pass filter unit 14 is input to the second output adjustment circuit 14a and output to the output terminals (DL2, DR2) of the DSP 10 as the second audio output signal D2. The output of the band pass filter unit 15 is input to the third output adjustment circuit 15a, and is output to the output terminals (DL3, DR3) of the DSP 10 as the third audio output signal D3. The first output adjustment circuit 13a, the second output adjustment circuit 14a, and the third output adjustment circuit 15a expand the range for adjusting the reproduction sound quality so that the first sound output signal D1 can be applied to a wide range of speaker systems 7. Each of the second audio output signal D2 and the third audio output signal D3 includes a level adjustment circuit that performs level adjustment, a phase inversion circuit that performs phase inversion, and a delay circuit that adjusts the delay time.

  In the channel divider 12 according to the present embodiment, for example, as shown in the graph of FIG. 2 (standardized amplitude frequency characteristic), the low pass filter unit 13 (LPF) and the band pass filter unit 15 are illustrated. The audio signal input by setting (BPF) and the high-pass filter unit 14 (HPF) can be divided into bands. The filter characteristics of the low-pass filter unit 13, the band-pass filter unit 15 and the high-pass filter unit 14 shown in FIG. 2 are the respective output signals (first audio output) when an impulse is input as the input audio signal. Signal D1, second audio output signal D2, and third audio output signal D3). Here, the crossover frequency fc1 between the woofer WO and the squawker SQ is set to 2.5 kHz, and the crossover frequency fc2 between the squawker SQ and the tweeter TW is set to 7 kHz. As described above, the low-pass filter unit 13, the band-pass filter unit 15, and the high-pass filter unit 14 each include a low-pass filter LPF, a high-pass filter HPF, or a band-pass filter BPF configured by FIR filters. Therefore, the group delay characteristic is a linear phase characteristic that becomes a constant value with respect to a change in frequency. That is, the low-pass filter unit 13, the band-pass filter unit 15, and the high-pass filter unit 14 are linear phase filters, respectively.

  In the channel divider 12 of the present embodiment, when the low-pass filter unit 13, the band-pass filter unit 15 and the high-pass filter unit 14 are each configured by a finite-length FIR filter, the pass level of the stop band increases. In order to suppress the occurrence of the Gibbs phenomenon, the filter coefficient of the FIR filter is set as follows.

Specifically, a Herrmann Filter known as LPF or HPF in which no Gibbs phenomenon occurs is used. The LPF of Herrmann's Formula can be expressed by Equation 5. Here, each symbol is as follows. Equation 6 shows a binomial coefficient. Since the square of the sine function and the square of the cosine function can be expressed as Equation 7 and Equation 8, respectively, Equation 5 can be designed as a finite-length FIR filter. Using Equation 7 as a real number x, a channel divider is designed using Bernstein's polynomial (x = sin2 (ωΔt)).
f s: sampling frequency
L: Arbitrary integer, filter bandwidth
K: Arbitrary integer, polynomial order, steepness of filter transient

Equation 9 shows a Bernstein polynomial. In several Formula _{9, b v, N (x} ) can be expressed as Equation 10.

Here, each symbol is as follows.
n: Arbitrary integer βν: Arbitrary real number

  In addition, the Bernstein polynomial of Equation 11 using the continuous function g (x) in the range where the real number x is [0, 1] is uniform as in Equation 12 where x is in the range [0, 1]. Converge. When the Bernstein coefficient is 1, there is a relationship of Equation 13.

  Therefore, in order to use the Bernstein polynomial as a finite length FIR filter in the channel divider, the real number x is determined as in the above Equations 2 to 14. The polynomial formed by these two equations becomes a finite-length linear phase FIR filter. Using the relationship of the Bernstein polynomial in Equation (11), it is decomposed as in Equation (15). However, in the formula, 0 <m1 <m2 <... <mb <N: an integer of 1 or more.

Equation 15 can be rewritten as Equation 16. That is, the first term of equation 16 corresponds to the low-pass filter corresponds to the third term is the high-pass filter of Equation 16, corresponding to the second term bandwidth communication over the filter unit of the formula 16 . The first term of Equation 16 is equivalent to the Herrmann's Formula LPF expressed by Equation 5, and has a relationship of L + K = N. By determining this order mk, a channel divider can be designed. The order mk can be obtained from Equation 3 when the crossover frequency is fc and the sampling frequency is fs.

  Equation 15 corresponds to the case where the waveform of the input audio signal matches the waveform of the sum signal of the first output signal, the second output signal, and the third output signal in the channel divider. In other words, if the continuous function g (x) indicating the desired reference composite characteristic is made equal to 1, no peak or dip will occur near the crossover frequency, and the woofer WO and the squawker of the speaker system 7 will not be generated. The amplitude frequency characteristic of the sum characteristic of the reproduced sound from the SQ and the tweeter TW can be kept unchanged. The speaker system 7 can be adjusted to achieve a relatively flat synthesized sound pressure frequency characteristic using a channel divider. In the output stage of the channel divider 12, the amplitude frequency characteristic does not change, and no peak or dip is generated near the crossover frequencies fc1 and fc2, and the reproduction sound quality of the 3-way speaker system 7 is appropriately changed. Can do.

In the channel divider of the present invention, the number of polynomial terms v for a predetermined frequency is obtained by the following Equation 17, and the desired gain is input to g (v / N) to give the sum amplitude frequency characteristics, The gain characteristic near the over frequency can be adjusted. In that case, Equation 15 can be expanded by using Equation 11 to obtain Equation 1 above.

However, if the function g is a continuous function, its frequency characteristics will be the expected value, but if the function g is a discontinuous function, it will not be the desired value, so the value must be corrected as follows: . When the target gain is G, the band center is Mc, and the band width is Mw, the frequency response of the filter designed with only this band term is expressed as in Equation 18.

  FIG. 3 shows the sum of the frequency characteristics (gain characteristics) of the channel divider 12 when the gain characteristics of the low-pass filter section 13, the band-pass filter section 15 and the high-pass filter section 14 are as shown in FIG. It is a graph explaining about. That is, a gain of +6 dB is set at the crossover frequency fc1 of the low-pass filter unit 13 and the band-pass filter unit 15, and a gain of +3 dB is set at the cross-over frequency fc2 of the band-pass filter unit 15 and the high-pass filter unit 14. And the coefficients of the FIR filters of the low-pass filter unit 13, the band-pass filter unit 15 and the high-pass filter unit 14 of the channel divider 12 are set. Here, the integer N = 1000. Thus, there is an advantage that the filter of the channel divider 12 can be designed without being affected by the Gibbs phenomenon. Since the coefficients of the FIR filter have a finite tap length, it is not necessary to apply a window function, and the filter can be designed only in the frequency domain. Each FIR filter coefficient may be set so as to provide a delay equal to or greater than the number of samples equal to the delay necessary to satisfy the causality.

  The channel divider 12 is not limited to the one applied to the 3-way bi-wiring SS including the woofer WO, squawker SQ, and tweeter TW as in the above embodiment. The 3-way multi-way speaker system includes a case where a sub-woofer for reproducing the low frequency range is further added to a 2-way multi-way speaker system including a woofer and a tweeter. In that case, the first output signal D1 divided into the low sound band side is output to the subwoofer, and the second output signal D2 divided into the high sound band side is output to the tweeter of the 2-way multi-way speaker system. The third output signal D3 divided into bands between them may be output to the woofer of the 2-way multi-way speaker system.

  The channel divider 12 is a 4-way or higher multi-way speaker (not shown) including a subwoofer and / or a mid-range (squaker) and / or a super tweeter. A configuration including a divided output may be used. Of course, the channel divider 12 of this embodiment is also effective in the case of a multi-way speaker including a network circuit. Similarly, the channel divider 12 is effective when the speaker system 7 includes a full range speaker. When a substantial crossover frequency fc with another speaker is set in response to a level drop caused by the limit of the reproduction sound pressure frequency characteristics of a full range speaker or a speaker such as a woofer, squawker, and tweeter. In addition, a low-pass filter unit, a band-pass filter unit, or a high-pass filter unit of the channel divider 12 can be set in accordance with this.

When the multi-way speaker system of 4 Way or more is supported, the channel divider 12 causes the band pass filter unit 15 to further divide the third output signal D3 into a plurality of M (M: integer of 2 or more) (not shown). ) A plurality of M band division filter units may be configured. In that case, the second term of Equation 1 may be further expanded so as to correspond to, for example, a 4-way multi-way speaker system when M = 2. In the channel divider of the present invention, the range for adjusting the reproduction sound quality is widened, and it becomes possible to cope with various multi-way speaker systems.

  The channel divider of the present invention can be applied not only to a stereo apparatus that reproduces a stereo audio signal, but also to an audio reproduction system including a multi-channel surround sound reproduction apparatus.

DESCRIPTION OF SYMBOLS 1 Amplifier apparatus 2 D / A converter 3 Amplifier circuit 4 CPU
5 Operation unit 6 Display circuit 7 Speaker system 10 DSP
DESCRIPTION OF SYMBOLS 11 Decoder 12 Channel divider 13 Low pass filter part 13a 1st output adjustment circuit 14 High pass filter part 14a 2nd output adjustment circuit 15 Band pass filter part 15a 3rd output adjustment circuit

Claims (5)

A low-pass filter unit that divides the input audio signal into a first output signal on the low-frequency range and outputs the first output signal to the first output terminal; and a second high-frequency-side filter for the audio signal. A high-pass filter section that divides the output signal into a band and outputs the output signal to the second output terminal; and a third high-frequency filter side of the audio signal that is higher than the first output signal and lower than the second output signal. A band divider including a band-pass filter unit that divides the output signal into a band and outputs it to the third output terminal,
The low-pass filter unit, the high-pass filter unit, and the band-pass filter unit are all configured by a finite-length impulse response filter, and the coefficients of the finite-length impulse response filter are set to [0, 1]. In the following Equation 1, which is a Bernstein polynomial in which the sum using a continuous function g (x) of a real number x as a range is 1, the real number x is defined as the following Equation 2, and the continuous function g (x) is determined as desired. The coefficient obtained by defining the reference synthesis characteristic and the order m as the following Equation 3 and giving a delay equal to or greater than the predetermined number of samples to the coefficient of the finite impulse response filter according to the first term of the following Equation 1. A coefficient obtained by setting a low-pass filter unit and giving a delay equal to or greater than a predetermined number of samples to the coefficient of the finite-length impulse response filter according to the third term of Equation 1 below. Was set to the high-pass filter, it sets the coefficients obtained by applying a predetermined number of samples equal to or greater than the number of delay in the coefficient of the finite impulse response filter according to the second term of Equation 1 below to the band pass filter section To
Channel divider.
(However, in the formula, 0 <x <1, n is an integer of 1 or more. 0 <m1 <m2 <... <Mb <N: an integer of 1 or more.)
(However, in the formula, 0 <x <1, fs: sampling frequency.)
(However, in the formula, N is an integer of 1 or more. Fc: crossover frequency.) The continuous function g (x) indicating the desired reference synthesis characteristic is equal to 1, and the sum of the waveform of the input audio signal, the first output signal, the second output signal, and the third output signal The signal waveform matches,
The channel divider according to claim 1. The band-pass filter unit is configured by a plurality of M finite-length impulse response filters that output a plurality of M (M: an integer of 2 or more) third output signals.
The channel divider according to claim 1 or 2. A level adjusting circuit for adjusting a level of the first output signal or the second output signal or the third output signal; and a phase for performing phase inversion of the first output signal, the second output signal or the third output signal. An inverting circuit; and a delay circuit that adjusts a delay time of the first output signal, the second output signal, or the third output signal.
The channel divider according to any one of claims 1 to 3.   5. The channel divider according to claim 1, an amplifier including an amplifier circuit corresponding to the first output terminal, the second output terminal, and the third output terminal of the channel divider, and at least the first A speaker system including a woofer or subwoofer corresponding to an output signal, a tweeter corresponding to the second output signal, and a tweeter, squawker or woofer corresponding to the third output signal, and capable of bi-wiring connection with the amplifier. And an audio playback system including: