JPS6350195A - Audio signal transmission system - Google Patents

Abstract

PURPOSE:minimize distortion due to crossover network by making the spectrum analysis of inputted audio signals and dividing the signals according to the result of analysis. CONSTITUTION:The system is provided with a digital filter 2, a spectrum analyzing section 3, a band dividing circuit 4, power amplifiers 5a, 5b, a woofer 6a and a mid and high range speaker 6b. When signal dividng audio signals inputted corresponding to plural outputting means according to frequency band, the spectrum analysis of inputted audio signals is made and the signals are divided according to the result of analysis. Accordingly, the division of frequency band can be made very accurately by the spectrum analysis, and the frequency division of any frequency is made possible. Though it becomes not a real time processing due to spectrum analysis, scarcely any problem is caused aurally.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an audio signal transmission system, and more particularly to an audio signal transmission system in which an output section is provided with a plurality of output means having different physical characteristics.

[Conventional technology]

Generally, the audible frequency band of audio signals is 20 to 20,
It is approximately 000Hz, and it is extremely difficult to reproduce this entire range of audible frequencies with high fidelity using one speaker.

Therefore, multiple speakers are usually used, each of which is responsible for reproducing a different frequency band, to reproduce the entire range of audible frequencies. The following two methods have been considered as a method of dividing and supplying an input signal according to frequency bands to the plurality of speakers, that is, a method of dividing a so-called crossover network.

That is, one is a passive network that performs post-division via a power amplifier in the output stage, and the other is a multiway system that performs division before being input to the power amplifier. In general, passive networks can be constructed relatively inexpensively, but multi-way systems are easier to achieve in terms of high fidelity reproduction, and multi-way systems are generally adopted in many cases.

Furthermore, there are two types of circuit configuration methods for multi-way crossover networks: an analog type that uses a combination of R, L, and C, and a digital type that converts to a -degree digital signal and processes it, and many methods have been proposed for both. There is.

[Problem to be solved by the invention]

In the analog multiway system described above, it is difficult to achieve both out-of-band attenuation characteristics and phase characteristics in the crossover frequency characteristics, and the frequency characteristics may also vary due to variations in component characteristics. On the other hand, digital multiway systems can achieve both the above-mentioned out-of-band attenuation characteristics and phase characteristics, but there is a limit to the crossover frequency, making it impossible to realize a crossover network, especially in the low range. .

An object of the present invention is to provide an audio signal transmission system that can solve the above-mentioned problems. That is, to provide an audio signal transmission system that can realize desired frequency characteristics in a crossover network without deteriorating phase characteristics, can freely set the crossover frequency, and can minimize distortion caused by the crossover network. The purpose is to

Means for Solving Problem C] With such an objective, the present invention provides an audio signal transmission system including an input section, an output section, and a transmission section interposed between them, in which the output section is physically Including a plurality of output means with different characteristics, when dividing an input audio signal corresponding to the plurality of output means by frequency band, a spectrum analysis of the input audio signal is performed, and according to the analysis result. The configuration is such that the signal division is performed by

[Effect]

By configuring as described above, frequency band division can be performed extremely accurately by spectrum analysis, and frequency division can be performed at any frequency. Also, spectrum analysis eliminates real-time processing, but this does not cause much of a problem when it comes to hearing.

〔Example〕

The present invention will be explained below using examples.

FIG. 1 is a diagram showing a schematic configuration of a system as an embodiment of the present invention. In the figure, reference numeral 1 denotes an input section, from which input audio and signals are output as digital signals. 2 is a digital filter, 3 is a spectrum analyzer, 4 is a band division circuit, 5a and 5b are power amplifiers, 6a is a bass speaker, and 6b is a medium/high frequency speaker.

A specific example of the configuration of the spectrum analysis section 3 and the band division circuit 4 will be described below. FIG. 2 is a diagram showing a specific configuration example of a spectrum analysis section and a band division circuit for band-dividing an input audio signal into a low frequency band and a middle and high frequency band with 200 Hz as a boundary. 12.14 in Figure 2
is a digital low-pass finite-length impulse response filter (
13 is a sub-sampler, 15 is a zero sample adder, 16 is an interpolation type digital low-pass filter, 17 is a delay corrector, and 18 is a signal calculator for medium and high frequencies.

Next, the operation of each section in FIG. 3 will be explained.

First, a digital audio input signal is supplied to the FIR low pass filter 12 . To simplify the explanation that follows, 4 is output from a well-known compact disc (CD) as a representative digital input signal.
, 1 KHz sample frequency. This input signal is filtered by a filter 12 where signal components of 2KT (signal components of z or more are cut).
13 subsampled by the subsampler
/10 sampling frequency (4.41 KHz). This digital audio signal, which includes signal components up to 2KHz and is sampled at 4.41KHz, is sent to the next stage of FIR.
The signal is supplied to a low-pass filter 14. Output is 200Hz
A sampling frequency signal of 4.41 KHz is obtained by cutting off the above signal components.

Next, nine zero sample data are added between each data to restore the sampling frequency to the output of the filter 14 at step 15. Now the sampling frequency is the original 44.1 KH
Return to Z. Furthermore, the 9 zero data are changed into data obtained by interpolating between data at both ends using a sine wave or the like using a ←*compensating digital low-pass filter. Now 2
This means that a low-frequency digital output with frequencies above 0.0 Hz cut is obtained.

On the other hand, the output for medium and high tones is calculated from the input and the output for low tones. First, the input is a 17 delay compensator and is synchronized with the bass output. That is, the time until the aforementioned bass output is obtained is delayed. Next, 18 mid-high range signal calculators calculate a mid-high range digital output from the synchronized input and low range output.

In this way, by using various digital low-pass filters and manipulating the sampling frequency, it is possible to digitally process even low-frequency sounds of, for example, 200 Hz, that is, one period of 5 ms, without imposing a large load on computing. In the above-described configuration, the system consisting of various digital low-pass filters, samplers, etc. corresponds to a spectrum analysis section that analyzes and sieves low and middle and high frequencies on the order of milliseconds. Further, with reference to FIG. 3, a 113 Hz digital crossover network, which is another specific example of the spectrum analysis section and band division section in FIG. 1, will be explained. 19 is input 4 4
, is an average value calculator that calculates the average value of 200 pulses of a digital audio signal with a sampling frequency of 1 KHz every 1/220.5 seconds. On the other hand, the same input data is supplied to the buffer memory 2o.

In order to return the data generated every 1/220.5 seconds output from the average value calculator 19 to the original sampling frequency of 44.1 KHz, the zero sample data adder 15 adds 199 points between each data. Add zero data. Next, the 199 zero sample data are processed by the interpolation type digital low-pass filter 16 by interpolating between the sample data at both ends using a sine curve. This means that low-frequency digital output data with frequencies above 113 Hz cut is obtained.

On the other hand, the middle and high frequency signals are calculated as the difference between the input data and the bass output data as in the previous example.

It is also essential to properly synchronize them in the buffer memory 20. 44.1, the transmission time of 200 data in a pulse of 7 seconds corresponds to about half a wavelength of 113 Hz, which is about 4.5 ms. Even if computing takes 2.5ms, the total time is 7ms, 34
For a sound wave at 0 m/sec, the propagation distance is only about 2.4 m. Even if it were used in a live public address system, it would not be perceived by human senses and would not cause any discomfort.

Furthermore, although the explanation so far has been based on the Compact Disc standard, the above system can also be used with other digital audio signals, such as the 31.5KHz of the 8rnm video tape recorder (VTR) standard, and the digital audio recorder (DAT). The above-described system can be similarly applied to 48 KHz digital audio signals, and the above-described system can also be applied to analog inputs by artificially digitizing them.

Adaptive digital filters (hereinafter referred to as ADFs) are known that can adaptively change the characteristics of a digital filter, although the characteristics of the digital filter do not change during -H. A.D.
F is capable of changing tap switching, coefficients of a coefficient unit, etc. in accordance with an input digital signal or preset control data, thereby changing its characteristics.

FIG. 4 shows the basic configuration of the ADF.

In FIG. 4, 21 is a digital filter whose characteristics can be switched according to control data from the control circuit 22, and 23 is an adder/subtractor. xj is input signal data, yj
is output signal data, and dj is target data indicating target characteristics. The control circuit 22 switches the coefficients and taps of the coefficient unit in the filter 21 according to input signal data xj, output signal data yj, and target data dj. By doing this, it is possible to obtain filters having various frequency characteristics, delay characteristics, etc.

FIG. 5 shows a system as another embodiment of the present invention,
This uses the ADF described above.

The digital audio signal inputted from the input section 31 is supplied to the ADFIA processing section 32 . The ADF processing unit 32 ultimately converts the bass speaker 11a1 to the midrange speaker 11.
b. High-pitched speaker 1. A power amplifier 10a is connected to each speaker so that an ideal output can be obtained from each IC.
, 10b and 10c are supplied with three systems of output signals. This ADF processing unit 32 has multiple ADs in parallel and multiple ADs in series.
It is configured by connecting F. This reason is FIR
Whether it is a filter or an IIR (infinite impulse response) filter, the number of delay stages that can perform accurate processing is about 3 to 4 stages, whereas a digital filter with a number of delay stages of about 3 to 4 can handle steep frequency characteristics. This is due to the fact that it is impossible to achieve. That is, by connecting a plurality of ADFs in series, steep frequency characteristics can be realized. Also, when changing the frequency characteristics in various ways, it is virtually impossible to achieve complex frequency characteristics, such as characteristics with multiple peak frequencies, with a single digital filter group. It is from.

FIG. 6 shows a configuration example of the ADF processing section 32 shown in FIG. 5. In FIG. 6, 35 is a terminal to which the digital audio signal from the input section 31 is supplied, 36 is a terminal to which target data from the target data setting circuit 33 is supplied, and 41a to 46
a, 41b to 46b, and 41c to 46c are ADFs connected in parallel and series in a matrix. The configuration of each ADF is, for example, as shown in FIG.

Adders 47, 48, and 49 output audio signals output to the low-range, middle-range, and high-range speakers lla, llb, and llc to terminals 37, 38, and 39, respectively. The target data inputted from the terminal 36 includes data for individually adjusting the characteristics of each ADF.

How to use the system as described above will be explained below.

One of the basic drawbacks of multiway systems is that one instrument or one person's voice is played from different speakers depending on the frequency. This causes the sound image to move or become unclear. Various attempts have been made to avoid this phenomenon, but none of them are perfect. However, one solution can be obtained by introducing ADF.

FIG. 7 shows the frequency and volume ranges included in music and speech. In the case of a professional soloist, it is thought that the sound can be 3 to 6 dB louder than the speaking voice shown in the figure, but the frequency range and volume are still clearly smaller than that of a full orchestra. Therefore, Figure 5 shows the midrange reproduction system, especially the speaker l.
Choose lb with as wide a band as possible. In general, the distortion rate is low unless you raise the volume extremely, for example 10~16c.
Use a single cone of m.

Now, suppose that an audio signal containing a mixture of a full orchestral part and a mainly vocal solo part is inputted from the input section 31 in FIG. This input signal is determined by the frequency band (for example, the region where 95% of the energy exists) and the volume.
It is possible to classify into In other words, the full orchestra part has a wide band and a large volume signal, and the part centered on the vocal solo has a relatively narrow band and a limited volume range (7th
(see figure).

Therefore, the characteristics of this input audio signal are determined by the ADF processing unit 3.
The control circuit (see FIG. 4) of the first-stage ADFs 41a to 46a in step 2 makes the determination and switches the frequency characteristics of each ADF. For example, when the center part of a solo vocal is input, the AD determines the output signal to the midrange speaker.
The characteristics of F43a to 43c and 44a to 44c are set so that the passband is wider than when a full orchestral part is input, and the characteristics of the ADF are set to determine the output signals to the low frequency speaker and the high frequency speaker. Set the type that narrows the passband. With this configuration, a full orchestra, where high frequency band, large volume, and low distortion are important, can be output in multi-way, even if the sound image is particularly important.

The above usage method was performed by controlling the characteristics of the ADF according to the input signal, but the target data will be explained below. The setting parameters for this target data are the nature (type) of the source;
, speaker characteristics, reproduction sound field characteristics, user preferences, etc.

First, there are many genres of sources, including classical, jazz, pop, rock, and vocals. Although the recording/mixing process is somewhat tailored to suit each genre, the playback system itself also has its own unsuitability. For example, the so-called tonshari type, which emphasizes the bass and treble, is said to be suitable for pop and rock music. Therefore, appropriate target data is set for each genre and stored in a ROM, etc., so that when the user selects the genre selector during playback, the target data determined by each genre is supplied to each ADF. I'll do it.

Next, as speaker characteristics, target data for each ADF is set based on the frequency characteristics, directivity, damping factor, impedance, etc. of the connected speaker system. Furthermore, target data for the reproduction sound field is set, including speaker settings, acoustic characteristics of the reproduction sound field, multi-processing in a surround system, etc. User preferences are also related to all of the above. Generally, there are three ways of thinking about ideal regeneration.

PHF school-----Person who values reproduction with high physical fidelity. Generally, this idea is central to the field of electrical signal processing.

SHF School - Reproduces sound that feels exactly like the original sound. Trance information users such as speakers are still unfinished and physically incomplete. Therefore, compromises and seasonings are made somewhere. SHF is often seen in classical music lovers.

GR school - 111 This group tries to create comfortable, good music without paying too much attention to the original sound. It is mainly seen among light music lovers.

An even more interesting fact is that true intentions and tatemae are very different. When I ask many audio enthusiasts what the best sound is, they say it's PIF. However, the required speaker system has a sound quality of 90% or more. In other words, the real intention is S.
It's either HF or GR, but the official name is PHF.

In order to respond to such user psychology, PHF is basically
It is important to have a transparent configuration. Then, set the target signal value according to your preference.

Of course, there are many different preferences among SHF, GR, etc. There are people who like harmonious sounds that emphasize pleasant sound vibrations and harmony, people who like crisp sounds, people who like loud sounds, people who like resonant sounds, and so on. These preferences are replaced with controllable physical quantities and supplied to each ADF as target data.

Furthermore, some classical music enthusiasts wish to reproduce the differences between concert halls. Naturally, target data based on the acoustic characteristics of each hall will be given to each ADF. Further, the application of these various ADFs and the use of the ADF by the input itself described above may be used alone or in combination. In addition, these adjustments, that is, the target data, are frequency characteristics,
Contains data for controlling delay characteristics, sound source position, directivity, etc.

Further, as the systemization progresses, the amount of target data increases, and each target data becomes more complex and requires a large capacity memory. It is therefore also advantageous to store the control inputs and the respective target signal values in a portable memory such as a ROM, card or chip. In this case, it is also useful to similarly store the scenery of the performance venue, explanations of the target data, etc. in the aforementioned portable memory.

Furthermore, when transmitting to users via broadcasting or various recording media, it is also possible to transmit the sound field characteristics of the performance venue as target data.

In a system using the ADF as described above, each AD
The control circuit of F performs spectrum analysis of the input audio signal, and when dividing these signals into frequency bands, it is possible to make this division characteristic variable. Furthermore, by controlling the above-mentioned ADF using other parameters, it becomes possible to reproduce audio signals of various characteristics.

(Effects of the Invention) As explained above, according to the audio signal transmission system of the present invention, a crossover network with desired characteristics can be realized without causing distortion.

[Brief explanation of drawings]

FIG. 1 is a diagram showing a schematic configuration of a system as an embodiment of the present invention, FIG. 2 is a diagram showing a specific configuration of a spectrum analysis section and a band division circuit in FIG. 1, and FIG. FIG. 4 is a diagram showing the basic configuration of the ADF. FIG. 5 is a schematic configuration of a system as another embodiment of the present invention. FIG. 6 is a diagram showing an example of the configuration of the ADF processing section in FIG. 5. FIG. 7 is a diagram showing the frequency and volume ranges included in music and speech. 1.31 --- Input section 2 --- Digital filter 3 ---
−−−−−−−− Spectrum analysis section 4 −−−−−−
---One band dividing section 5a, 5b, 10a, 10b, 10c-
1111---Power amplifier 6a, 6b, lla, llb, 1lc-
11---------Speakers 12, 14------Digital low-pass FIR filter 15---------Zero sample data adder 1
6-----, ----1 interpolation type digital low-pass filter

Claims (1)

[Claims] An audio signal transmission system including an input section, an output section, and a transmission section interposed between them, wherein the output section includes a plurality of output means having different physical characteristics, and an input section corresponding to the plurality of output means. 1. An audio signal transmission system characterized in that, when dividing an input audio signal into frequency bands, a spectrum analysis of the input audio signal is performed, and the signal division is performed according to the analysis result.