RU2145446C1 - Method for optimal transmission of arbitrary messages, for example, method for optimal acoustic playback and device which implements said method; method for optimal three- dimensional active attenuation of level of arbitrary signals - Google Patents

Abstract

FIELD: radio engineering. SUBSTANCE: method involves converting messages into electric signals of message source, matched filtration of electric signals from message source, their amplification, converting into signals of same type, transmission of signals through channel with arbitrary parameters into reception point, reception and converting of signals into received electric signal, transmission of signal to processing point, processing of electric signals of message source and received electric signal, generation of signals to control matched filtration and additional active noise reduction electric signals which are amplified, converted into signals of same physical type and emitted to channel with arbitrary parameters up to message reception point. Said method in particular discloses acoustic playback method and method for optimal three-dimensional active attenuation of level of arbitrary signals. Device has signal source and playback channel which is designed as low-pass amplifier and loudspeaker, which are connected in series, sounding device, signal processing unit and communication line. Signal processing unit provides possibility of generation additional signals for active noise reduction. EFFECT: increased precision of information transmission. 19 cl, 7 dwg

Description

 The invention relates to cybernetics and can be used, for example, in radio engineering.

State of the art
A known method of transmitting messages in a channel with random parameters, for example in a radio channel, which consists in processing the received signals using correlation methods. The disadvantage of this method is that it requires knowledge of multidimensional distribution laws or correlation functions describing the message and interference. The method is applicable for messages and interference, amenable to mathematical modeling in the form of stationary random processes.

 The known method of sound reproduction and a system for its implementation, which consists in coordinated filtering of the electrical signal of the source, amplification, conversion of the amplified electrical signal of the source into an audio signal, its emission, reception and transformation at the listening point of the total sound signal - direct and reflected sound waves of the emitted source signal, as well as sound waves of interference and noise into an electrical listening signal, transmitting the received electrical signal to the place of its processing, about the operation of the electrical signals of the source and listening, the formation of control matched filtering signals (RU, A 2038704).

 This method of sound reproduction improves the accuracy of signal reproduction at the listening point due to automatic predistortion of the amplitude-frequency parameters of the source signal in accordance with distortions of the amplitude-frequency characteristics (AFC) of the audio information transmission channel.

Disadvantages:
1) the method and system allow noise reduction only at frequencies and at times of sound reproduction of signals;
2) the system for implementing the method should be equipped with a signal source with normalized output characteristics;
3) the system does not allow for the correction of the phase-frequency characteristic (PFC) of the signal;
4) the system does not allow localization of sounds in the space surrounding the listener with a plausible effect of "surround"sound;
5) the system does not allow for individual correction of the amplitude-frequency characteristics, filtering of noise and noise of the signal source in the automatic optimization of reproducible signals.

Disclosure of Invention
The present invention is based on the task of creating such methods and systems for their implementation that can improve the accuracy of information transmission in a channel with random parameters, for example, to improve the quality of sound reproduction taking into account interference and noise, changes in the shape, volume, acoustic properties of a room, location of speakers, location listening points, orientations at the listening point of the listener's head, distortions in the signal source, features of auditory perception of sound pressure frequencies, n improve the frequency response and phase response of low-frequency amplifiers and loudspeakers, and thus reduce the standard deviation (RMS) of the signal in the system, for example, increase the noise reduction efficiency at the listening point, accurately adjust the volume, balance, amplitude-frequency and phase-frequency characteristics of the sound reproduced at the listening point signal, expand the range of the effect of "surround" sound and increase the accuracy of localization of sounds in relation to the listener.

 The problem is solved in that in the known method of sound reproduction, which consists in coordinated filtering of the electric signal of the source, amplification, conversion of the amplified electric signal of the source into an audio signal, its emission, reception and transformation at the listening point of the total sound signal - direct and reflected sound waves of the emitted signal source, as well as sound waves of interference and noise into an electrical listening signal, transmitting the received electrical signal to its place Work, processing the electrical signals of the source and listening, forming control signals matching filtering, according to the invention generate additional electrical signals for active noise reduction, which amplify, convert into sound signals and emit to the point of reception of the total sound signal.

 The problem is solved in that in the known sound reproduction system containing a signal source and a sound reproduction channel, made in the form of a low-frequency amplifier and loudspeaker connected in series, a probing device, a signal processing unit, configured to coordinate filtering of the source signal, communication line, while the output of the probing device via a communication line is connected to the first input of the signal processing unit, the output is connected to the second input of the signal processing unit d source signal and the output signal processing unit connected to the input of low-frequency amplifier according to the invention a signal processing unit configured to generate at its output additional signals for active noise reduction in the set-point probe device.

System implementation options are possible to:
the signal source was made with at least one additional output for multi-channel audio reproduction, additionally introduced according to the number of additional outputs of the signal source, at least one additional audio reproduction channel, made in the form of an additional low-frequency amplifier and an additional loudspeaker connected in series, as well as an additional sounding device , an additional signal processing unit, configured to co-filter si the source and the formation of additional signals for active noise reduction, an additional communication line, the output of the additional sounding device through an additional communication line is connected to the first input of the additional signal processing unit, the additional output of the signal source is connected to the second input of the additional signal processing unit, the output of the additional signal processing unit is connected with the input of an additional low-frequency amplifier;
the source of the signal was made with the possibility of switching the output signals to change the order of their connection to the sound channels;
in addition, a signal generator connected to the third input of the signal processing unit was introduced;
in addition, a signal generator was introduced, connected to the third input of the signal processing unit and the third input of the additional signal processing unit;
a signal processing unit, a communication line, a sensing device, an additional signal processing unit, an additional communication line, an additional probing device were functionally combined into an optimal signal processing unit for a block system configuration;
the signal processing unit and the additional signal processing unit were made in the form of a multi-channel analog-to-digital converter, a computer with software and a multi-channel digital-to-analog converter connected in series, while the number of inputs of the analog-to-digital converter is two times larger and the number of channels of the digital-to-analog converter is equal to the number sound channels;
the signal source was configured to control the amplitude-frequency characteristics of the signals at its outputs;
the signal source was made with the possibility of noise reduction;
the signal source was configured to automatically control the levels of its output signals;
the signal source was made with the possibility of automatic level control and non-automatic regulation of the amplitude-frequency characteristics of the signals at its outputs;
the signal source was configured to control signal levels at its outputs to control the volume level at the listening point;
a sounding device and an additional sounding device or a communication line and an additional communication line were configured to control the transmission coefficient to control the volume level at the listening point;
gear ratios have been finely compensated;
a signal processing unit and an additional signal processing unit were configured to automatically control the amplitude-frequency or phase-frequency characteristics from the second inputs to the outputs of the blocks to optimize the energy or time parameters of the signals;
a signal processing unit and an additional signal processing unit were configured to automatically control the amplitude-frequency and phase-frequency characteristics from the second inputs to the outputs of the blocks for full parameter optimization of the signals.

 The problem is solved using the method of actively lowering the level of signals of any physical nature, which consists in receiving and converting these signals into an electric signal, transmitting the received electric signal to the place of its processing, processing the electric signal to actively lower the signal level, amplifying it, converting it into a signal the same physical nature and radiation to a point in the space of signal reception.

 The problem is solved in that in the known method of transmitting messages in a channel with random parameters, which consists in converting messages into electrical signals of a message source, coordinated filtering of electrical signals of a source, their amplification, converting electrical signals into signals of the same physical nature, transmitting these signals through a channel with random parameters to the point of receiving messages, receiving and converting the signal into a received electrical signal, transmitting a received electric signal According to the invention, additional electrical signals are generated for active noise reduction, which amplify, convert into signals of the same physical nature and emit into a channel with random parameters up to message receiving points.

 The invention is based on the principles of coordinated filtering of signals when transmitting messages in a channel with random parameters and the principle of the synthesis of additional signals that are compatible with the parameters of noise and noise, which allows for the active reduction of the level of signals (i.e., these noise and interference) at the point of message reception.

 The modes of coordinated filtering and synthesis of additional signals are provided through the use of feedback, for example, acoustic and special signal processing.

 The signal processing is based on the principle of the sequential implementation of processes for measuring the current values of signal parameters and transmitting messages in a form adjusted based on the results of the analysis of signal parameters.

 The indicated actions are automated.

 In contrast to the well-known sound reproducing systems, the proposed optimal sound reproduction system carries out optimization of pre-emphasis - coordinated filtering of the source signals and optimization of the parameters of the additionally synthesized signals. As a result of the sound radiation of the pre-emphasized source signals and additionally synthesized signals at the listening point, the sound vibrations repeat the sound vibrations of the original sound sources with the greatest possible accuracy, that is, optimally.

 These advantages, as well as features of the present invention will become apparent during a subsequent review of the following possible embodiments of the invention with reference to the accompanying drawings.

Brief Description of the Drawings
FIG. 1 illustrates a method for optimal sound reproduction.

 FIG. 2 depicts a functional diagram of a single channel optimal sound reproduction system.

 FIG. 3 depicts a functional diagram of a multi-channel (stereo) optimal sound reproduction system.

 FIG. 4 depicts the same as in FIG. 2, with an additional node - a signal generator.

 FIG. 5 depicts the same as in FIG. 3, with an additional node - a signal generator.

 FIG. 6 is an example of a full-parameter signal processing unit.

 FIG. 7 illustrates the principle of active noise reduction.

The best embodiment of the invention
When playing in a specially unsuitable room, for example, in a living room, a passenger compartment, a cabin of a ship, etc., signal distortion occurs due to a number of reasons.

 The sound reproduction room may have a different shape, volume and acoustic properties. The acoustic properties of the room may vary, for example, the position of the curtains, doors or windows may be different.

 In the room, the listener can randomly place the speakers, as well as arbitrarily choose a listening position and head orientation at the listening position.

 Due to these factors, linear distortions of the signal appear at the listening point: frequency, phase, and transient.

 Frequency distortions are a consequence of the multipath coming to the listening point of sound waves: direct and re-reflected. For different frequencies at the listening point, one can observe an increase in the volume of signals with respect to the volume at other frequencies during the in-phase arrival of waves, and a decrease in volume during the out-of-phase addition of oscillations of direct and reflected sound waves. As a result, the real relationships between the amplitudes of the components of the complex oscillation are violated and the spectrum of the information signal changes.

 Subjectively, these distortions are manifested in the unnatural sound of musical signals, for example, low-frequency sounds bubbling with a disproportionately high intensity, there is an imbalance in the volume of the sound of instruments or vocals on individual notes, speech becomes illegible.

The unevenness of the frequency response of the signal at the listening point can be on the order of $\genfrac{}{}{0ex}{}{\text{+}}{\text{-}}$ $\genfrac{}{}{0ex}{}{\text{ten}}{\text{thirty}}$ dB or more, which is comparable to the dynamic range of most music signals.

 Phase distortion occurs as a result of changes in the time delays in the arrival of sound vibrations between the various frequency components of a complex sound signal. For example, being at a different angle to a multiband speaker system (speaker) with spatially spaced, for example, low-frequency, mid-frequency and high-frequency speakers, the listener will perceive sound vibrations with different phase shifts in the low, medium and high frequencies. As a result, the shape of the audio signal is distorted at the listening point. These distortions are most noticeable with the spatial separation of the speakers in the passenger compartment. Subjectively, distortions are perceived as delays, for example, of low-frequency components, the sound becomes spatially indefinite. The most noticeable effect on the fidelity of perception of musical signals is phase-frequency distortion in multichannel sound reproduction systems. Distortions have a strong negative psychophysiological, emotional effect on listeners with a high musical culture or education, good hearing and musical memory. Distorted sound reproduction annoys the listener by the discrepancy between the perceived “sound picture” and the images of real “sound pictures” stored in his memory, which he received while studying at a music school or attending concert performances by musicians. Systematic listening to sound, musical signals with strong distortion, as well as disappointment in the unrealistic sound reproduction on expensive equipment can lead to diseases of the nervous system of the listener and cause a deterioration in human auditory, musical and aesthetic capabilities.

 Transient distortions for the listening room are usually determined by the reverberation time, as the interval during which the total signal energy decreases by a million times against the original value in the absence of other sound energy in this space.

 During sound reproduction in the listening room, secondary noise and noise may occur. For example, at high volume, secondary noise and noise may appear in the form of rattling of glass, dishes, decorative panels and other unreliably fixed objects. Excited by vibrations of useful signals, these objects can create additional sounds that distort the shape of the sound signal at the listening point.

 In addition, there may be interference and noise external in nature to the sound reproduction room, for example, noise from the street, from adjacent walls, noise from heating or lighting systems, noise from a running engine or other car systems, etc.

 The level of external noise is about 30 - 70 dB, and in some cases, for example, when driving in a convertible, it can reach values hazardous to human health on the order of 90 - 100 dB.

 Thus, the premise of sound reproduction as an integral part of any path of a spatial-sound reproducing system introduces significant distortions and degrades the overall quality of sound reproduction of the entire system.

 The distortions introduced by the premises are of a random nature and are caused by many random circumstances. Therefore, in mathematical modeling, sound signals at the listening point can be represented as a non-stationary, non-ergodic, multidimensional random process.

 When solving the problem of optimizing sound reproduction, it is necessary to take into account the distortions introduced by all the links of the sound reproducing path: signal source - low-frequency amplifier - loudspeaker - room - listener.

 Currently, recording methods, recording devices, storage media and signal sources, the distortion level of which is comparable to or even less than the resolution of the human hearing to record these distortions, have been created. For example, the best samples of compact disc (CD) players, created using digital signal processing methods, allow you to get almost the maximum signal quality at their output. Further improvement of the electrical parameters of signal sources, for example, reducing non-linear distortions below 0.01% or increasing the dynamic range above 120 dB, does not make practical sense for a person, since these improvements become invisible to the listener.

 The main areas of improvement of signal sources and storage media are miniaturization of structures, increasing efficiency, reducing weight, increasing playback time, improving ergonomics of structures.

 The distortions introduced by the best examples of low-frequency amplifiers are also small. Further improvement of the electrical parameters of the best samples of low-frequency amplifiers becomes almost invisible to the hearing even for experts from among the leading conductors of orchestras, singers and musicians with absolute hearing.

 The efforts of developers of low-frequency amplifiers are mainly aimed at increasing efficiency, reducing overall dimensions, improving the reliability of low-frequency amplifiers.

 The distortions introduced by the best speaker models significantly exceed the level of distortion introduced by the signal source and low-frequency amplifier.

 For a comprehensive solution to the problem of optimizing sound reproduction, it is necessary to take into account and compensate for signal distortions that occur in all elements of the sound reproduction path.

 Having not comprehensively solved the problem of optimizing sound reproduction, it is pointless to classify the quality of household sound reproducing systems, since the listener does not care at which link of the sound reproduction path there are distortions, and it is important for him to hear an audio signal that differs in form from the original source by a value guaranteed in accordance with the quality class of the sound reproducing system.

 The functional method of studying natural events in the field of sound reproduction and classification of sound reproducing systems, according to which a conditional division into groups (classes) of sound reproducing equipment is carried out according to the technical (electrical) indicators of its main elements, is not suitable for household sound reproducing systems. The classification of the quality of sound-reproducing equipment only taking into account its own electrical indicators of the equipment blocks, indicators characterizing the properties and structure of the system blocks, not affecting the relationships and relations between the system blocks and the external environment that affect the processes in the system, is not entirely legitimate and misleads the consumer relatively real the feasible quality of spatial sound reproduction in the domestic operation of the equipment.

 Obviously, from the listener's point of view, the determining parameter, for example, of the amplitude-frequency distortions should be the standard for the frequency response of the signal at the listening point as an indicator that directly characterizes the quality of sound reproduction, and not the standards for the frequency response of the individual blocks and elements in the sound reproducing system, such as amplifier, loudspeaker, etc., which are parameters that indirectly determine the quality of sound reproduction.

 The traditional, functional approach to assessing the quality of sound reproduction is legitimate in appropriate conditions, in special studio conditions (in anechoic chambers). In other words, constant quality parameters can be applied in relation to stationary or, in extreme cases, to quasi-stationary processes, when destabilizing factors slightly change (distort) sound signals and these distortions are not detected (are not felt) by a person, but in systems with random, strongly varying parameters it is necessary to use other principles for the selection of parameters and criteria in the classification of systems.

Thus, the random nature of the origin of strong distortions of sound signals objectively requires that the quality of modern household sound-reproducing systems, including equipment, the so-called HI-FI and HIgh-End classes, should be defined as a class of random “quality” of sound reproduction, when , for example, the frequency response of the signal at the listening point of a domestic building or car interior can reach $\genfrac{}{}{0ex}{}{\text{+}}{\text{-}}$ $\genfrac{}{}{0ex}{}{\text{ten}}{\text{thirty}}$ dB, and the most probable signal-to-noise ratio (noise) does not exceed, as a rule, 5 - 30 dB.

 A general deterioration in the quality of sound reproduction is also facilitated by some sound processing algorithms implemented by processors, for example, in discretely switched filters, widespread mini-systems or, for example, in Dolby Surround matrix coding systems for multichannel sound transmission of television programs or video films in a "home theater" .

 Indeed, the application of linear signal transformation algorithms with fixed parameters of these transformations with respect to signals whose parameters can change randomly over time, in principle, cannot improve the quality (reliability) of audio signals and leads to an increase in distortion (increase in entropy). Such signal transformations lack logical justifications for their use as processing algorithms that increase the quality of sounds for the listener.

 We show that the use of step-switched filters in household sound reproducing systems is not legitimate.

 According to the PHILIPS catalog (p. 14) for 1997, the company justifies the use of such a unit in its equipment with the following arguments: "Using the magic SOUND button (the button for turning on the digital sound processor) you can create the mood you want. And you don't just hear it: you feel it. It covers you ... ". The following are graphs of the frequency response of the filter, corresponding to its state of POP, CLASSIC, JAZZ, ROCK, VOCAL, as well as explanations of the type: “for pop music, the sound processor is configured to play sound basses and energetic high notes. As you can see from the graph, they are amplified mainly , frequencies from the range 100 - 200 Hz. " At the same time, the listener is invited to switch the filter to the operating mode corresponding to the sound signal during the operation of the sound reproducing system.

 With regard to such devices, the following can be noted.

 Firstly, the conditional division of all kinds of sound signals into a finite set of genres will lead to a mismatch between the spectra of real signals and the rigidly fixed amplitude-frequency response of the corresponding filter, for example, the spectral components of the bass of a man or singer performing a vocal part on high notes can differ by several octaves. and with the instrumental performance of various classical works, the difference in the spectra of signals can be even greater.

 Secondly, the criterion by which a listener could determine before belonging to a signal whether it belongs to a particular genre, for example, when listening to radio broadcasts, is incomprehensible. The listener will need some time, as well as a certain musical culture, to determine by ear the musical genre to which the sound signal belongs. The listener needs to be distracted from listening to switch the filter, and therefore the emotional and aesthetic perception of sound information will deteriorate. Such filter switching will irritate the listener because such actions are tedious work.

 Thirdly, it is unlikely that in the process of sound reproduction, the listener will become like an automaton and quickly engage in filter switching, and, as a result, he will listen to signals in a very distorted form.

 Fourthly, it is unlikely that sound engineers working in recording studios are not able to introduce the required pre-emphasis of the frequency response of the phonogram using the studio equipment.

 And finally, fifthly, the scientific basis of such an approach is not clear.

 It can be assumed that the idea of creating step-switched filters as “quasi-consistent” with filter signal parameters was based on the generalized Kolmogorov-Wiener linear filtering theory to solve the problem of increasing the signal-to-noise ratio. However, the application of the provisions of this theory, mainly oriented to solving problems in radio communications, for the task of improving the quality of sound reproduction in domestic conditions, strictly speaking, is not right, since this theory is valid for stationary random processes, and, as already noted, in everyday conditions of sound reproduction have to deal with non-stationary random processes. In addition, for the synthesis of an optimal filter, consistent with the parameters of the signals and noise, it is necessary to know the spectrum of the useful signal, the interference spectrum, and the attenuation coefficient of the information transmission channel. Step-change filters do not perform the functions of such devices.

 For these reasons, it also seems ineffective to use smoothly tunable filters — graphic equalizers — to optimize signals. Obviously, the listener, even as a professional sound engineer or musician, is not able to optimally filter out the spectral components of interference at various levels with the required accuracy, for example, to compensate for the amplitude-frequency distortion of a signal arising in such elements of the sound reproduction path as a low-frequency amplifier, loudspeaker, room , since there is no criterion for such manual adjustments (the installation criteria do not indicate installation criteria EQ knobs), and various arbitrary variations in adjusting the frequency response of the signals are also possible.

Matrix coding systems have historically emerged to optimize sound reproduction in large cinema halls. The need for multichannel sound reproduction in films is explained by the fact that, with respect to the screen, part of the audience is usually not located in the center of the hall, therefore, with two-channel sound reproduction, a corresponding shift of the apparent sound source occurs for them. Recently, various modifications of matrix coding systems have become widespread, which allow simulating various acoustic properties of rooms and the location of sounds in relation to the listener. The listener chooses the indicated effects and the corresponding signal processing algorithms based on their own random criteria, for example, out of curiosity. How will the symphony orchestra sound in an acoustic design resembling sounds in a basement or in a stadium?
Thus, using matrix coding systems, the listener receives in the "home theater" sound information in a very distorted form.

 The distortion effect is enhanced by digital signal compression as a way to increase the signal-to-noise ratio in some car radio models.

 A comparison of the basic principles of Norbert Wiener cybernetics with modern methods used in sound reproduction shows the inconsistency of these methods when recording (inputting), processing and outputting audio information.

 Modern sound reproducing systems, including the so-called HI-FI and HIgh-End classes, are the simplest non-purposefully working machines. With their help, the optimal sound reproduction is fundamentally unattainable, i.e. high quality sound reproduction.

 In order to optimize sound reproduction, it is necessary to improve all stages: input; processing and output of audio information in accordance with the well-known principles of cybernetics.

 In accordance with the principles of cybernetics, high-quality spatial sound reproduction should be understood as such sound reproduction, in which the sound vibrations affecting the hearing organs exactly repeat the sound vibrations of the original sources of sounds that previously acted on this person or his virtual image - his imaginary movement backward along the time axis.

 This definition does not contradict the principles of adjacency, similarity, cause and effect, logic, and requires for its implementation the presence of the basic elements of a highly organized system (machine): information sensors; its processing unit and corresponding algorithms for this processing; information output devices.

 From the principle of causation, it follows that for high-quality sound reproduction from a signal source, the output of which is presented in the form of electrical signals (in analog form - in the form of time functions of voltage or current amplitude, in digital form - in the form of a binary-coded digital sequence signals) it is necessary, firstly, to accurately enter sound information into the system. At the output of the signal source, information loss should be minimized.

 The greatest likelihood of the input information can be provided through a microphone head. Ideally, the listener's head should be used. With small, almost unregistered average listener distortions, for example, in the apparatus of the initial level of likelihood, a model of the head of some average listener, for example, a model of the head of Bach or a modern singer or conductor, can be used.

 In the equipment of a high level of credibility, it is also possible to introduce a parametric series, for example, the size of the head in accordance with generally accepted standards for hats or other criteria for personifying input information on CD, MD, etc.

 For artificially synthesized sound information, that is, sounds created by a recording director through multichannel electrical signals with arbitrary use of linear and non-linear methods of processing them and which never existed in nature as a “sound picture”, in accordance with the principle of cause and effect The tasks of high-quality sound reproduction are modified, since it is impossible to find a logically justified criterion and accurately enter, process and reproduce (output) sounds (info mation), have never existed in nature.

 For artificially synthesized sound signals, the statement of the problem of virtual high-quality sound reproduction is legitimate, which implements, as accurately as possible, in accordance with the electrical signals of the source (carrier) of signals (information), the director's idea of the form of the “sound picture” he artificially created during processing of sound signals.

 Secondly, the processing of audio information in a system with random parameters requires an on-time analysis and comparison, according to certain criteria, of the parameters of the signals from the source, which can be considered “ideal” or reference signals, and distorted signals — signals at the listening points.

 In accordance with the principles of similarity and adjacency, parameters and characteristics of signals having the same shape and physical meaning can be compared and analyzed. Electrical signals may be analog, digital, or digital. Comparison of signals should be carried out in the same form. For example, if two harmonic oscillations are compared, more precisely, the oscillations are close in shape to harmonic oscillations, then the following parameters can be compared: frequency, duration, amplitude, time delay of the beginning of these oscillations relative to a certain point in time (for oscillations of the same frequency, this parameter can make sense of the initial phase of these oscillations or the full phase - a temporary shift by a time longer than the duration of the period). Oscillations can also be compared in terms of derivatives derived from the specified parameters: in terms of power or energy. It is customary to compare complex vibrations consisting of a large number of elementary components by comparing their energies by integrating the spectral energy densities of the signals. Such a comparison of signals of complex shape allows them to be compared in an averaged form.

 Using the well-known principles of representing complex signals in the form of series or Fourier integrals, one can compare the spectral components (harmonics) corresponding in frequency according to the above parameters and various criteria. Such a comparison of signals of complex shape can be considered full-parameter, fully describing the information contained in the "ideal" and distorted signals. Digital signals can be analyzed by comparing the values (numbers) discretized in time and quantized by the level of the signals.

 Obviously, having, as a result of the analysis of the corresponding parameters of the signals, their current values, which are denoted in the most general form by the symbols A and B, it is possible by appropriate processing of these data to find the magnitude of their inconsistency with each other, for example, the magnitude of their difference. Knowing the magnitude of the error C and using, for example, a control method with negative feedback, one can introduce the corresponding correction of the “ideal” signal according to the corresponding parameter, converting it into an “ideal - predistorted” signal.

 Subsequently, the “ideal - predistorted” signal, passing through the elements of the sound reproducing system, is converted into a distorted signal, which becomes identical to the “ideal” in the corresponding parameter.

 Considering that a person has two hearing organs, it is advisable to use a two-channel (two-signal) system for input, processing and output of audio signals as a basic model.

 Thus, the two-channel signal processing unit must process and identify (calculate) differences in the parameters of the reference ("ideal") and distorted due to the totality of random signal factors.

 Distorted signals recorded by the microphones of a special device - a sounding device, at the listening points (at points in space near the hearing organs) with respect to the optimal sound reproducing system are feedback signals, without which, as you know, it is impossible to create the optimal technical solution - a highly organized machine.

 The need to introduce a probing device into the system can be justified by the following interpretation of the principle of similarity (similarity).

 If the electrical signals of the original sound sources are “ideal” signals in shape that exactly repeat the electrical signals at the output of the sounding device, then (taking into account a number of additional conditions and reservations, which are described in more detail below), this means that the acoustic vibrations during sound reproduction accurately repeat the acoustic vibrations of the primary sources sounds.

 If the indicated electrical signals are different, then the acoustic signals are also different, and to improve the quality (accuracy) of sound reproduction, it is necessary to adjust (adjust) the "ideal" signal so that the difference between the "ideal" and distorted signals is reduced to the minimum possible value, ideally to zero , i.e. carry out parametric optimization of signals.

 Thirdly, in order to carry out this parametric optimization and ultimately generate plausible sound signals, it is necessary to introduce some device connected to the signal processing unit into the sound path between the signal source and the amplifier — a controlled filter matched to the signal parameters — optimal filter.

 The characteristics of this filter, for example, the complex transmission coefficient, should be functionally dependent on distorting factors in the low-frequency amplifier, loudspeaker, communication lines, premises and be automatically tuned automatically in accordance with the detected signal distortions.

 With automatic correction of all possible parameters of the signals or their components that determine the parameters of the complex transmission coefficient of this filter, the optimal filter can be called full-parameter or canonical; with automatic correction of one of the parameters, the optimal filter can be called a parametric optimal filter.

 When optimizing the energy parameters, a situation is possible where the difference between the corresponding parameters may turn out to be close in magnitude to one of them.

 For example, at certain frequencies or in the entire band of sound frequencies, the energy of the useful, information signal during the analysis can be equal to zero, and the energy of the sound signal at the listening point and therefore the energy of the electric distorted signal can be different from zero. This situation is typical for sound reproduction in domestic, specially unsuitable rooms in the presence of external acoustic noise and noise. Obviously, in this case, except by active noise reduction methods, it is impossible to achieve the equality of the energy parameters of the "ideal" and distorted parameters. In other words, it is necessary by any of the possible methods to carry out the emission of additional signals in the corresponding frequency band, which at the listening points should be out of phase with respect to acoustic noise and interference.

 If in automatic mode, at the right time and at certain frequencies, additional signals are emitted and parametrically optimized based on the criterion for equal amplitudes and out-of-phase components of these signals, i.e. If the energy of the total signal, which is a superposition of acoustic noise and additionally emitted signals, is equal to zero, then the signal-to-noise ratio may increase when reproducing in a room with a high level of acoustic noise and noise.

 It follows from the previous one that with a prolonged absence of a source signal, a mode of active, spatially localized reduction of the level of acoustic noise and interference can be provided, for example, for a driver of a motor vehicle when he does not want to listen to informative sound signals from a source, as well as noise from an engine and other car systems, but wants to be in silence.

 It is clear that the above particular case of a broader problem may be the subject of independent invention.

 The moments of time when active noise reduction should be turned on and off, respectively, should be functionally dependent on the current signal-to-noise ratio in the corresponding frequency sub-bands of the analysis of the energy spectra of "ideal" and distorted signals. The specified signal-to-noise ratio (interference) can be determined by hardware, software or software and hardware depending on the selected form of signal analysis and processing: in analog, digital (discrete), combined, using a universal computer, for example, a widely used PC - IBM PC or using specialized microprocessor control systems, analog automated control systems, etc.

 In order to effectively reduce the level of acoustic noise and interference, the formation of additional signals should be carried out from distorted signals received directly at the listening point, and not at other points in space, for example, near the radiation sources of these noises.

 At high signal-to-noise ratios, the signal can be adjusted at the listening point to an “ideal” signal without active noise reduction only due to the coordinated filtering of the “ideal” signal. In this mode, the optimum filter can change the frequency response and phase response.

 At low signal-to-noise ratios, for example, when the noise energy is several times higher than the useful signal energy, by actively reducing the energy, the energy parameters of the signal at the listening point can be optimized, i.e. Frequency response of the signal. Optimization of the phase response of the signal at the listening point at low signal-to-noise ratios is impossible, therefore the accuracy of the "conversion" of noise to an "ideal" signal cannot be arbitrarily high and is determined by the specific characteristics of the useful signals, noise, control time constant, and other parameters.

 Thus, for a specific acoustic situation, dynamic changes in the characteristics of signals, and the speed of the system’s self-regulation, there is a certain threshold signal-to-noise ratio below which the accuracy of correspondence of “ideal” signals and signals at listening points will be lower than a predetermined value.

 An essential circumstance in the optimization of sound reproduction is the random nature of the delay that the distorted signal undergoes in relation to the “ideal” one. The magnitude of this delay for typical conditions of sound reproduction is determined mainly by the time it takes for sound vibrations from the speaker to the listener and is, for example, of the order of 0.01 - 0.005 s.

 As already noted, the time delay of the signal components can be different and is a random variable. Obviously, having found out in some way the indicated delays of the signals and their components and finding the relative time offset of the components relative to each other, we can then introduce the corresponding correction in the time domain of the “ideal” signal and its components, for example, using discretely switched delay lines as elements of phase adjustment in an optimal filter. When using the digital signal processing method, knowledge of these delays allows digital filtering methods to carry out a similar phase correction procedure for the "ideal" signal. Indeed, knowing two signals as a function of the voltage time at the output of the signal source and at the output of the probing device, they can be discretized in time and quantized in level in accordance with the Kotelnikov theorem and using analog-to-digital converters. Further, by performing multi-band digital filtering of signals by means of digital filters, it is possible to find delays of signal components for various frequency bands of analysis using digital correlation analysis.

 Thus, for optimal sound reproduction, it is necessary to be able to correct the “ideal” signal directly during sound reproduction, using consistent filtering of the signal in analog or (dual) digital form.

 To find random delays of signals and their components, the following methods can be used: a) a synchronization method with an additional synchronization signal, b) a threshold synchronization method, c) a correlation method.

 The synchronization method with an additional synchronization signal consists in the fact that during the sound reproduction, a synchronization signal is introduced into the signal path, for example, at the output of the signal source, along with the “ideal” signal, as an information signal. This signal can have a different shape. For example, as this signal there may be packs of high-frequency sound vibrations with a frequency of 30 - 100 kHz. The main requirements for these signals should be: the absence of interference from these signals in the range of human perception of external radiation, high noise immunity of these signals. The generation of signals in time can be carried out once at the beginning of sound reproduction or throughout the entire time of sound reproduction in the form of periodically generated, for example with a period of 10 or 60 s, start-stop synchronizing signals. Radiation of such signals can be carried out by high-frequency speakers of loudspeakers or by additional piezoelectric emitters mounted on loudspeakers.

 Using this method, the absolute delay value of the “ideal” signal for each of the audio channels can be determined. Knowing the value of the time delays of signals, it is possible to carry out a temporary correction to synchronize the moments of arrival of sound waves at the listening point and increase the accuracy of the psychophysiological perception of the place of sound localization by the listener.

 The specified method does not allow to determine the delay of the signal components and therefore does not allow to implement phase-frequency optimization algorithms for the signals. The method requires limiting the level of synchronization signals to exclude adverse psychophysiological effects on the listener. This method can be used for parametric optimization of signals.

 Finding signal delays can also be done using the threshold synchronization method. This method consists in the fact that the “ideal” and distorted signal or their corresponding components are analyzed according to the level relative to a given threshold, for example, relative to the nominal (94 dB) at the level of -20 dB or -30 dB. The threshold level, depending on the complexity of the system, can be set a priori and remain unchanged, or it can be varied and set by the system itself in accordance with the current changes in signal parameters and the algorithm of the system’s functioning. If the signal level exceeds the threshold, then a decision is made on the appearance of the signal or, accordingly, on its completion in the time domain of analysis.

 This method differs from synchronization with an additional signal in that the delay is determined not by analyzing additional synchronization signals in the time domain, but directly by the audio signals themselves. In addition, this method allows you to determine the delay of the signal components and, therefore, allows you to implement various algorithms for phase-frequency optimization of signals and, in particular, full-parameter optimization. Indeed, for each frequency sub-band of the analysis, the corresponding delays of the signal components can be found and, therefore, the necessary correction of the signals in the time (phase) region is introduced.

 The disadvantage of this method is the high probability of error when finding a delay due to noise and interference when the process of sound reproduction is carried out at low volume. Special measures and algorithms are needed to reduce the likelihood of abnormal errors in calculating delays. It is clear that these algorithms can be based on averaging in the time domain of the delay values and, consequently, will lead to an increase in the inertia of the system during its self-adaptation under the conditions of interference and noise. The advantage of this method is the simplicity of the hardware implementation, the absence of special requirements for radiating and sound-collecting equipment (loudspeakers, sounding devices).

 Finding the delays of the signals and their components in the correlation method of signal analysis is the following algorithm for performing operations on signals. One of the signals - "ideal" - is delayed in parallel branches with some time step, for example, through delay lines with a common input.

 The delayed “ideal” signals thus obtained are fed to the first inputs of the multipliers. The second inputs of the multipliers supply a distorted signal. Multiplied signals are passed through low-pass filters to filter the higher harmonics of this non-linear operation. Next, the filtered signals are compared with each other in level. The signal of the highest level corresponds to the greatest correlation (coincidence) of the "ideal" and distorted signals, and the corresponding delay in the line is, in fact, the current signal delay in the sound reproduction path. For component analysis of the signals, the signal components delayed by the optimal time can then be added up and a wideband phase-corrected signal can be obtained. In this signal, the components will be displaced relative to each other as much as they underwent the opposite relative displacement when passing through all the elements of the sound reproduction path: low-frequency amplifier, loudspeaker, listening room. At the listening point, the acoustic signal components will be reproduced in the form of the acoustic signal components of the original sound sources, i.e. high quality.

 The advantage of the correlation method for determining the optimal delays of signals and their components over the threshold method will be most noticeable with a low signal-to-noise ratio and with a relatively small number of analysis bands, since in a wide analysis band, for example, an octave wide, harmonics of different frequencies can be present at the same time, and the signal itself at the output of a broadband filter can have a complex shape. If the number of bands is large and the signals at the output of these filters approach harmonic oscillations, then in order to simplify the signal processing equipment and the algorithms for this processing, instead of the correlation signal processing method, a simpler threshold synchronization method can be used. In the digital method of signal processing using, for example, a computer, it is advisable to use the correlation method of phase-frequency optimization of signals.

 It is clear that using the above signal processing methods and choosing possible parameters for signal optimization, you can get a large number of functional circuits and signal processing algorithms. However, the foregoing allows us to highlight the essential features of the method of sound reproduction and its corresponding nodes, united by certain relationships in a spatial sound reproducing system.

 The method of optimal sound reproduction can be formulated as follows.

 The method of optimal sound reproduction, which consists in coordinated filtering of the electric signal of the source, amplification, conversion of the amplified electric signal of the source into an audio signal, its emission, reception and transformation at the listening point of the total sound signal - direct and reflected sound waves of the emitted source signal, as well as sound interference waves and noise in an electrical listening signal, transmitting the received electrical signal to the place of its processing, processing of electrical signals fishing source and listening shaping matched filtering control signals, characterized in that the additional form electrical signals for active noise reduction, which are amplified, converted into sound signals, and to emit sound receiving point aggregate signal.

 This formulation allows us to use the above sequence of actions, the feasibility of the implementation and the content of which are already justified above, as essential features of the method as a process of performing actions on material objects (electrical and acoustic signals).

 A significant difference between the proposed method of sound reproduction from the traditional, currently widely used methods is the coordinated filtering of the electrical signal of the source. Unlike step-switched filters, manually controlled equalizers, and other devices with processor-based signal processing that allow you to perform mismatched filtering of signals according to random criteria, matched filtering as an action on signals is understood in the broadest sense of its interpretation, i.e., as correction of any possible signal parameters according to logically reasonable, objective criteria.

 The processing of the electrical signals of the source and the listening signal in the method is not filled with specific content, i.e. the method does not disclose the algorithms of this processing. Possible signal processing algorithms and the corresponding structural diagrams of the devices should be described in the claims of the optimal sound reproduction devices. Possible device options will be given below.

 The sign of “consistent filtering” emphasizes that, despite the possible digital form of processing of electrical signals, the optimization of sound reproduction is carried out in real time, directly in the process of perception by the listener of sound information.

 A new distinguishing feature of the proposed method as an object of the invention is the formation of additional electrical signals for active noise reduction, their amplification, conversion into sound signals and radiation. In the proposed method, these actions are combined with a consistent filtering of signals. In more detail, the process of generating additional signals will be considered on examples of the description of the operation of embodiments of devices.

 In accordance with the provisions of Norbert Wiener cybernetics, the sign “optimal” emphasizes the belonging of the method and the corresponding system for its implementation to the class of purposefully working machines, i.e. machines of the highest form of organization. In other words, to the class of optimal machines. These machines allow you to best solve the corresponding problems. The technical result is maximum in terms of relevant indicators or characteristics. For example, providing a maximum signal to noise ratio, a minimum of standard deviation or a minimum of noise level, etc. The accuracy of the operation is determined by the current characteristics of the information transmission channel and the specific technical characteristics of the hardware of the system (machine). The indicated accuracy is determined by objective criteria (for example, the minimum standard deviation) and does not depend on subjective factors associated with human participation in the process of the system. The accuracy of the optimization of the current values of the signals intended for the implementation of the corresponding functions - coordinated filtering and noise reduction, is in no way connected with the current ability of a person to make decisions on choosing the "optimal" parameters and adjustments in the system. The essence of the optimal method is revealed by the combination and functional relationships of other essential features of the claims.

 In accordance with the above method of optimal sound reproduction, one can consider such sound reproduction as an ideal optimal sound reproduction or sound reproduction with a maximum level of quality, when in the process of listening to signals a person is not able to distinguish the listening signal from the source signal, i.e. the accuracy of matching each other's electrical listening signals and source signals is higher than a person’s resolving ability to record all possible types of such inconsistencies (distortions), as well as to pick up transient processes by ear during correction by a system of sound signals.

 The proposed formulation of the sound reproduction method introduces a new meaning into the concept of the quality class of a sound reproducing system.

 In systems with parameters dynamically changing in a random way, in contrast to deterministic systems or systems with random processes (stationary) and time-invariant statistical parameters, the concept of a quality class takes on a different meaning - this is the ability of a system to more effectively and efficiently detect and eliminate discrepancies between signal parameters , one of which is a reference.

 A sound reproducing system will be of a higher quality class relative to another if, for example, it has a larger dynamic range (higher power of a low-frequency amplifier and speakers), a higher degree of resolution in the frequency domain of signal analysis, i.e. a greater number of analysis frequency bands, a more rectangular shape of their frequency response, a more advanced signal processing algorithm, allowing, all other things being equal, to achieve a more effective and faster time-reducing noise reduction effect for various signal and noise spectra, signal-to-noise ratios in the audio reproduction band.

 According to the classification proposed by Norbert Wiener, the highest form of organization of material objects, in particular sound-reproducing systems (i.e., a quality class or complexity group), can be considered purposefully working, full-parameter, extrapolating (predicting in all possible parameters), self-learning, i.e. . independently synthesizing the optimal signal processing algorithm for specific conditions of sound reproduction and, in accordance with this algorithm, optimally performing coordinated filtering and generating additional signals for active noise reduction, a sound reproducing system with feedback.

 Self-training of a sound reproducing system allows you to implement an optimal system with a high level of "intelligence", suitable for almost any conditions of sound reproduction.

 It is important to note that a high-level sound reproducing system in the course of work can itself synthesize the optimal order of actions on signals and ultimately improve at the algorithm level the method of optimal signal processing depending on the specific conditions of sound reproduction.

 Obviously, sound reproducing systems of the highest form of organization, as well as highly organized machines for solving other problems, can be implemented using a computing node with a sufficiently high speed and memory capacity. In the best way, modern personal computers may be suitable for these purposes. A significant part of consumers has computers of a similar class. Having supplemented the basic elements of traditional sound-reproducing equipment (signal source, low-frequency amplifier (ULF), loudspeakers) with additional elements: a signal input device (ADC), a signal output device (DAC), a computer with software that performs the functions of coordinated filtering and generating additional signals for active noise reduction, as well as sounding devices and communication lines, the user can block-wise combine these devices into an optimal sound reproduction system. By gradually replacing software, computers, and other elements of the system, it is also possible to gradually increase the quality class of the system as they improve. With this configuration of the system, it becomes possible "at your fingertips", more precisely at the "tips" of ideas, without changing the mass, volume of the hardware, significantly improve the quality of sound reproduction.

 This approach in sound reproduction corresponds to the global scientific and technological trend of multimedia.

 The above principles of high-quality sound reproduction can be presented in symbolic form - in the form of mathematical formulas (theory) of optimal sound reproduction (information transfer).

 We illustrate the above method of sound reproduction using a cascade connection model and a description of the operation of the analyzed, enlarged system components.

 Let U (t) be the sound vibrations of the primary sources of sounds affecting the human hearing organ, i.e. sound vibrations of the original sources of sounds at the listening point.

 Having installed a microphone at the listening point, we denote the signal voltage u (t) at its output.

The signal u (t) can be relayed immediately or recorded and placed on a storage medium (magnetic tape, compact disc (CD), minidisk (MD), etc.). After some time Δt, the signal u (t) can be converted in the signal source (amplifier, radio, player, etc.) into the voltage of the signal source u _ist (t).

для точек прослушивания правым и левым ухом в электрические сигналы

Электрическим сигналам на выходах микрофонов во время записи будут соответствовать электрические сигналы источника ("идеальные")

Если считать, что при преобразовании звуковых колебаний в электрические сигналы, записи информации на носитель, хранении, преобразовании информации в электрические сигналы в источнике отсутствуют линейные и нелинейные искажения, а собственный коэффициент шума устройств равняется нулю (другими словами, на указанных этапах обработки сигналов исключены искажения и потеря информации), то сигнал на выходе источника сигнала можно представить в виде (1), (2).For a healthy person with two functioning hearing organs, accurate recording (recording) of sound signals is possible with two-channel conversion of sound vibrations

for listening points of the right and left ear into electrical signals

The electrical signals at the microphone outputs during recording will correspond to the electrical signals of the source ("ideal")

If we assume that when converting sound vibrations into electrical signals, recording information on a medium, storing, converting information into electrical signals, the source does not have linear and nonlinear distortions, and the device’s own noise figure is zero (in other words, distortions are eliminated at the indicated stages of signal processing and loss of information), then the signal at the output of the signal source can be represented in the form (1), (2).

(2)
где K, K_п, K_л - коэффициенты, не зависящие от u(t), u_п(t), u_л(t).u _source (t) = K • u (t-Δt) (1)

(2)
where K, K _p , K _l are coefficients independent of u (t), u _p (t), and u _l (t).

 Expressions (1) and (2) are a condition for undistorted input of information for one or two channels (a requirement for the energy and time parameters of the signals).

(3)

(4)
где

спектральные функции сигналов, определяемые согласно преобразованиям (5), (6) (для любого символа функции считается, что

)

(5)

(6)
Если в помещении прослушивания осуществляется звуковоспроизведение сигналов источника u_ист(t) или

то на выходе зондирующего устройства электрический сигнал прослушивания обозначим u'(t) или

Сигналу прослушивания соответствует его спектральная функция

или

которую можно определить с помощью преобразований Фурье.Since the form of sound reproduction of audio signals can be of any type, the signals u _ist (t), u _ist.p (t), u _ist.l (t) can be represented as a Fourier integrals (3), (4)

(3)

(4)
Where

spectral functions of signals determined according to transformations (5), (6) (for any symbol of the function, it is assumed that

)

(5)

(6)
If the room is carried out listening sound reproduction source signals u _ist (t) or

then at the output of the probing device, the electrical listening signal is denoted by u '(t) or

The listening signal corresponds to its spectral function

or

which can be determined using Fourier transforms.

где

- спектр сигнала источника,

- комплексный коэффициент передачи усилителя низкой частоты,

- комплексный коэффициент передачи громкоговорителя,

- комплексный коэффициент передачи помещения звуковоспроизведения,

- комплексный коэффициент передачи зондирующего устройства,

- приведенная ко входу приемного элемента зондирующего устройства спектральная функция (напряжения) помех (связанная с сигналом источника помеховая составляющая: гармонические составляющие, появляющиеся вследствие нелинейности таких элементов тракта, как усилитель низкой частоты, громкоговоритель; вторичные помехи, появляющиеся в помещении прослушивания в виде дребезга плохо закрепленных предметов - панелей, стекол и т.д., а также реверберационные искажения - многократные повторы сигналов источника, и некоррелированные с сигналом источника помехи - внешние акустические помехи, например шум из-за стен соседней квартиры, шум улицы и т.д.).Using a cascade model for describing the sound reproduction process, in which four-terminal devices are sequentially connected: a low-frequency amplifier and a loudspeaker, to the output of which one of the inputs has a six-terminal device (adder), which is a model that allows formally taking into account the influence of the listening room and the sounding device, and on the second input of which interference comes in with random, a priori unknown parameters, the spectrum of the listening signal can be represented as (7)

Where

- spectrum of the source signal,

- the integrated gain of the low-frequency amplifier,

- the integrated gain of the speaker,

- the integrated transmission coefficient of the sound reproduction room,

- the complex transfer coefficient of the probing device,

- spectral interference function (voltage) reduced to the input of the receiving element of the probing device (interference component associated with the source signal: harmonic components appearing due to the nonlinearity of such path elements as a low-frequency amplifier, loudspeaker; secondary interference appearing in the listening room as a chatter is bad fixed objects - panels, glasses, etc., as well as reverberation distortions - repeated repetitions of the source signals, and uncorrelated with the source signal noise interference - external acoustic interference, for example, noise from the walls of an adjacent apartment, street noise, etc.).

будет равен спектру сигнала источника

при выполнении операций согласно тождеству (8)

Выражение, заключенное в квадратных скобках формулы (8), представляет собой комплексный коэффициент передачи согласованного с параметрами элементов тракта звуковоспроизведения и зондирующего устройства фильтра. Выражение для комплексного коэффициента передачи согласованного фильтра определяется выражением (9)

Выражение (9), для удобства интерпретации комплексного коэффициента передачи согласованного фильтра, можно представить в виде двух сомножетелей (10)

(10)
Физический смысл выражения (10) заключается в том, что для осуществления согласованной фильтрации между источником сигнала и усилителем низкой частоты необходимо включить корректирующий фильтр, имеющий комплексный коэффициент передачи, определяемый по формуле (11)

а для компенсации искажений приемного элемента зондирующего устройства необходимо установить на его выходе корректирующий фильтр с комплексным коэффициентом передачи, определяемым по формуле (12)

В рамках данной модели описания звуковых сигналов можно показать, что отличие сигналов, получаемых на стадии записи первоисточников звуков

от формально приведенных ко входу микрофонов сигналов

определяется передаточной характеристикой, имеющей физический смысл, аналогичный комплексному коэффициенту передачи зондирующего устройства

Понятно, что если одно и то же зондирующее устройство использовать в одном случае для записи сигналов (вводе информации), а в другом случае при звуковоспроизведении в качестве зондирующего устройства, то указанные комплексные коэффициенты могут не учитываться, поскольку они определяют систематические ошибки при вводе информации, которые взаимокомпенсируются при звуковоспроизведении.You can see that the spectrum of the listening signal

will be equal to the spectrum of the source signal

when performing operations according to identity (8)

The expression, enclosed in square brackets of formula (8), is a complex transmission coefficient consistent with the parameters of the elements of the sound reproduction path and the probe filter device. The expression for the complex gain of the matched filter is determined by the expression (9)

Expression (9), for the convenience of interpreting the complex transfer coefficient of a matched filter, can be represented in the form of two factors (10)

(10)
The physical meaning of expression (10) lies in the fact that to implement consistent filtering between the signal source and low-frequency amplifier, it is necessary to include a correction filter having a complex transfer coefficient determined by the formula (11)

and to compensate for distortion of the receiving element of the probing device, it is necessary to install a correction filter with a complex transmission coefficient determined by the formula (12) at its output

Within the framework of this model for describing sound signals, it can be shown that the difference between the signals obtained at the stage of recording the primary sources of sounds

from formally reduced to the microphone input signals

is determined by a transfer characteristic having physical meaning similar to the complex transfer coefficient of the probing device

It is clear that if the same sounding device is used in one case for recording signals (entering information), and in the other case when playing sound as a sounding device, then these complex coefficients may not be taken into account, since they determine systematic errors in entering information, which are mutually compensated during sound reproduction.

 In practice, it is quite difficult at the hardware and methodological level to implement such an individual scheme for inputting audio information and processing signals. Therefore, it is advisable to adjust the characteristics of the “microphone-head” and the probe device according to a single principle, for example, by correcting the frequency response and reducing its shape to a flat frequency response and, accordingly, reducing the shape of the phase response to linear. Thus, a possible imbalance in the electrical characteristics of the microphones of the probing device can be established at the calibration stage in special studio conditions (in an anechoic chamber).

 From a comparison of expressions (7) and (8), it also follows that for optimal sound reproduction, it is necessary to generate additional sound signals at the location of the receiving element of the probing device (at the listening point) that have antiphase components with respect to the components of acoustic noise and noise and equal modules.

 Additional sound signals can be generated at the listening point with the help of an auxiliary sound reproduction channel containing an auxiliary low-frequency amplifier and a loudspeaker, or using the main ones: a low-frequency amplifier and loudspeakers. The latter option is more preferable for economic reasons.

и помещения

Строго говоря, при пространственном звуковоспроизведении помещение влияет не только на восприятие звучания громкоговорителей слушателем, но и на работу самих источников звука, так как отраженные звуковые волны попадают и на них - изменяется частотная характеристика громкоговорителей.When describing sound reproduction within the framework of the cascade model of accounting for the influence of elements of the sound reproduction path, there is no clear physical meaning of some formally introduced complex transfer coefficients: loudspeaker

and premises

Strictly speaking, in spatial sound reproduction, the room affects not only the perception of the sound of the speakers by the listener, but also the work of the sound sources themselves, since the reflected sound waves also fall on them - the frequency response of the speakers changes.

При этом приведенная ко входу зондирующего устройства спектральная функция сигнала может быть представлена в виде

а приведенная ко входу зондирующего устройства спектральная функция помех соответственно

Тогда в соответствии с принципом суперпозиции звуковых сигналов для точки прослушивания результирующая, приведенная ко входу зондирующего устройства спектральная функция может быть представлена в виде

и соответственно спектральная функция сигнала прослушивания на выходе зондирующего устройства будет определяться выражением (13)

Выражения (7) и (13) могут быть также применены при решении других кибернетических задач и оптимизации процессов информационного восприятия человеком сигналов другой физической природы, другими органами чувств, например, в оптическом (видео) диапазоне частот, с соответствующей заменой микрофонов зондирующего устройства на видеокамеры или видеодатчики сигналов визуального восприятия. В формулах (7) и (13) в рамках выбранной математической модели приводятся параметры и характеристики канала передачи информации, число которых в общем случае может быть любым, соответственно числу выбранных элементов канала передачи информации. Однако практическая ценность выделения и анализа элементов тракта звуковоспроизведения невелика, поскольку для нестационарных и неэргодических случайных процессов некоторые коэффициенты передач элементов тракта и спектральные функции сигналов и помех являются функционально зависящими от времени. Поэтому при решении указанных задач не имеет принципиального значения точное знание всех физических процессов и их взаимного влияния в каждом элементе тракта передачи информации.In view of this circumstance, it would be more correct to formally introduce the general complex transmission coefficient of the path elements: loudspeaker - room

In this case, the spectral function of the signal reduced to the input of the probing device can be represented as

and the spectral interference function reduced to the input of the probing device, respectively

Then, in accordance with the principle of superposition of sound signals for the listening point, the resulting spectral function reduced to the input of the probing device can be represented as

and accordingly, the spectral function of the listening signal at the output of the probing device will be determined by the expression (13)

Expressions (7) and (13) can also be used to solve other cybernetic problems and optimize the processes of information perception by a person of signals of a different physical nature, by other sensory organs, for example, in the optical (video) frequency range, with the corresponding replacement of the microphones of the probing device with video cameras or video sensors of visual perception signals. In formulas (7) and (13), within the framework of the chosen mathematical model, the parameters and characteristics of the information transmission channel are given, the number of which in the general case can be any, corresponding to the number of selected elements of the information transmission channel. However, the practical value of isolating and analyzing the elements of the sound reproduction path is small, since for non-stationary and non-ergodic random processes, some transmission factors of the path elements and the spectral functions of signals and noise are functionally time-dependent. Therefore, in solving these problems, accurate knowledge of all physical processes and their mutual influence in each element of the information transmission path is not of fundamental importance.

 If you do not know the laws of physical processes or if it is impossible to obtain the necessary information a priori of this class, problems can be solved only by fitting the corresponding parameters of the current process to the parameters of some supporting process. The entire information transmission path can be considered as a “black box”, the internal structure and parameters of which are unknown, but it is only known that its influence is reduced to a random change in the levels and phases of the components of a useful, informative signal, and the appearance of interference, the characteristics of which are also generally random .

 The distorting effect of the “black box” can be compensated by using the “white box” included in the signal circuit (in the sound reproduction path). In this case, it is necessary to experimentally select the parameters of the “white box” as quickly as possible so as to neutralize the effect of the “black box”.

 The indicated preparation of the “white box” parameters is the essence of coordinated filtering and the formation of additional signals for the active noise reduction of acoustic noise and noise.

 The location of the inclusion of the "white box" in the signal circuit is determined for economic reasons. It is advisable to install such devices in low-current circuits, for example, between a signal source and a low-frequency amplifier.

 We note a number of general points that should be taken into account when developing specific algorithms for the operation of sound reproducing systems within the framework of the proposed method of sound reproduction.

 It is advisable to implement the simplest signal processing algorithms for economic reasons in hardware - in systems of the initial or average level of sound reproduction quality. For example, when using a relatively small number of analysis bands, of the order of 5-10, it is advisable to use automatic equalizers to coordinate the energy parameters, and to optimize the time delays, the corresponding devices implemented in hardware.

 The most complex signal processing algorithms with a large number of operations and the corresponding elementary modules (filters, delay lines, multipliers, adders, inverters, comparison circuits, switches, etc.) are more appropriate to use in digital equipment or use virtual elementary modules - the corresponding subroutines in the process of algorithmization using a computer.

 Indeed, to implement full-parameter processing in a two-channel system and use, for example, 50 analysis bands, with the number of discrete delays in the time domain of about 200, the total number of elementary modules of the system can be of the order of 200,000 - 400,000. The cost of such a system, implemented using analog devices, can greatly exceed the cost of standard input and output devices and PCs.

 When using digital processing methods using a computer, the hardware implementation of the signal processing unit (the “white box”) of the optimal sound reproducing system is constructed according to the well-known generalized functional diagram in the form of series-connected: input, processing, and input devices (signals).

 Functional diagram of the "white box", revealing the essence of information processing, should contain the following components and the relationship between them.

 Firstly, the “white box” in both analog and digital versions must have at least two inputs and one output. One of the inputs should receive a signal from the output of the signal source, the other - a distorted signal - a listening signal.

 Secondly, the information signal of the source in the general case should branch out inside the "white box" into two branches, the outputs of which are connected to the inputs of the controlled switch, the output of which is the output of the "white box".

 One of the branches of the "white box" should be made in the form of a non-distorting signal line.

 The signal line must have a constant frequency response and a linear phase response in the frequency band of the signal processing. Given that any real signal passing through the line has a delay at its output with respect to the input signal, the signal line can also be called a broadband delay line. It is advisable to choose the value of the signal delay in this line equal to the signal travel time along the second branch to ensure the synchronization mode of the signals at the outputs of these branches, i.e. at the inputs of a controlled switch. Otherwise, when the controlled switch is switched, a signal shift in the time domain may appear in the form of additional signal distortions - its part disappears or part of the signal repeats at the listening point.

 As another branch, an analog machine or a computing unit, for example a PC or a specialized computer, should be used.

 To coordinate with a computer the signals from the output of the signal source and the listening signal, it is necessary to convert them from an analog form into a digital signal of the corresponding standard. For these purposes, you can use standard analog-to-digital converters (ADCs). The signal at the output of the signal processing unit (the “white box”), intended to be connected to the input of the low-frequency amplifier, must have an analog form. Therefore, in a “white box” device with digital signal processing, it is also necessary to have a digital-to-analog converter (DAC), for example, a standard DAC.

 Various options for the implementation of functional circuits using these elements are possible. For example, you can offer the option with the least number of functional nodes (canonical) in the form of a common ADC channel at the input of two branches, a separate ADC channel for another input, which will convert a distorted analog signal from the output of the communication line to a digital signal. It is advisable to implement a non-distorting signal line (delay line) in the first branch and a controlled switch by software, and the DAC should be installed at the computer output, the signal from the output of which should be connected to the input of the low-frequency amplifier.

 Other options for implementing and combining these elements are possible. For example, as a delay line, you can use a digital delay line implemented in hardware or install a DAC at the output of the delay line and the output of a computer, and the controlled switch can be made in the form of a hardware-implemented electronically controlled switch (multiplexer) controlled from an additional computer output and serving for switching analog branch signals.

 Other equivalent functional schemes for implementing the signal processing unit are possible.

 Thus, in the two-channel (stereo) version of the system implementation, it is necessary to use a two-channel ADC as the input matching device (information input device), the inputs of which are connected to the outputs of the signal source, and the output to the computer input, and a two-channel ADC, to the inputs of which the signals are connected outputs of communication lines, and the output to the input of the computer. In other words, the input device must be a four-channel ADC. At the computer output, a two-channel ADC must be installed for connection to low-frequency amplifiers.

 The use of two branches of signal transmission and processing is necessary and sufficient for realizing full-parameter optimal signal processing in real time.

 When implementing a non-full-parameter optimal signal processing, for example, when implementing coordinated filtering with only one parameter - the energy parameter, the first branch and the controlled switch can be excluded. When implementing consistent filtering by optimizing only the time delays of the signals or their components, it is necessary to use two signal propagation branches.

 These branches are designed to organize two modes of operation of the sound reproduction system. In the first mode, the signal passes through the line with an a priori known delay, then through all subsequent elements of the sound reproduction path with an unknown delay. At the listening point, the signal is a distorted listening signal. This mode is auxiliary and is necessary for sounding a sound reproduction channel in the current time interval and creating conditions under which reliable information about signal distortions can be obtained.

 The second branch serves to organize the input of information about the distorted listening signal and the reference signal of the source, i.e. segments of the implementation of these signals, the analysis of these signals, the organization of a coordinated filtering mode, the formation of additional signals for active noise reduction, summing the predistorted signal of the source and additional signals. The specified total signal is an output signal and, for example, through a program-controlled switch, is fed to the output device of the signal processing unit - DAC, and then to the input (inputs) of the low-frequency amplifier.

 The feasibility of this method of organizing the operation of the signal processing unit (the "white box") is justified by the following considerations.

In the theory of discrete and digital signals, one of the main parameters is the parameter T _c - signal duration. For a person as a recipient ("receiver") of spatial sound signals, which we denote by the symbols F (t), formally is the time interval from the moment of birth to death. Therefore, when solving specific cybernetic problems, i.e. the synthesis of optimal machines, allowing to achieve maximum efficiency of human activity in space and time, for example, the task of improving the quality of sound reproduction (the synthesis of the optimal, sound time machine in the past), the formal mathematical apparatus of the theory of discrete and digital signals should be adapted accordingly, so that the general solution of the problem corresponded to the conditions of physical feasibility and acceptability for a person - an element of the optimal system "man - machine". In other words, it is necessary to find conditions for harmonizing the principles of operation and the functioning parameters of machines with the psychophysiological characteristics of a person to perceive, process and display information (draw conclusions, compare, make decisions, move, etc.).

 In general, the solution to the cybernetic problems of this class (optimization problems for the parameters of non-ergodic and non-stationary random processes using the parameters of other supporting processes) can be found using the following approach.

We will use the previously introduced notation of the formalized components of the sound reproducing system in the model of cascading their connection and the symbolic representation of the sound signals perceived by a person throughout life as a function F (t), where F (t) ≡ 0 for t _p ≥ t ≥ t _cm (Fig. . 1).

For definiteness, the problem formulation will be assumed that the time interval between t _p and t _cm will be located at least one time interval [t _1; t ₂ ] between time t ₁ - the beginning and t ₂ - the end of the sound signals F (t), which should, as far as possible, repeat the sound signals of the original sources of sounds of the same duration T _c = t ₂ - t ₁ (the Doppler effect is not taken into account how unlikely for this task), but that existed previously in the time domain. In other words, the goal of solving the problem is the synthesis of an optimal signal processing algorithm.

Note that in the general case, the signal duration T _s , the number of these signals and the nature of the location on the time axis are not a priori determined, each individual listener determines the number and duration of sessions of optimal (high-quality) listening to audio information from the signal source. The indicated parameters are random variables.

Связь между этими спектральными функциями приводится в выражениях (7), (8), (13), пределы интегрирования равны T_с.Sound vibrations F (t) and U (t) over a time interval of duration T _s correspond to electric signals u '(t) and u _ist (t), the spectral functions of which were previously indicated

The relationship between these spectral functions is given in expressions (7), (8), (13), the integration limits are T _c .

где m и k - целые числа, удовлетворяющие соотношениям (15)

Ряд (14) представляет собой чередующиеся друг за другом временные интервалы

. Эти интервалы функциональны: отрезки времени

предназначены для организации режима прохождения сигналов по первой ветви блока обработки сигналов и далее через все элементы тракта звуковоспроизведения до точки прослушивания. В течение этого интервала опорным, "идеальным" сигналом осуществляется зондирование всего тракта звуковоспроизведения и искаженный сигнал наряду с опорным посредством АЦП вводится в ЭВМ. На отрезке

осуществляется процесс вычисления задержек, а также может быть начат процесс статистического анализа параметров сигналов и оптимизация энергетики сигнала.It is possible to adapt the theory of discrete and digital signals to solve the problem formulated above if any time interval T _s between time instants t ₁ and t ₂ (Fig. 1) can be represented as expression (14)

where m and k are integers satisfying relations (15)

Series (14) are alternating time intervals

. These intervals are functional: time spans

designed to organize the mode of signal propagation along the first branch of the signal processing unit and then through all the elements of the sound reproduction path to the listening point. During this interval, the reference, "ideal" signal is used to probe the entire sound path and the distorted signal, along with the reference signal, is input into the computer via the ADC. On the segment

the process of calculating delays is carried out, and the process of statistical analysis of signal parameters and optimization of signal energy can also be started.

предназначены для организации (квазиоптимального) режима воспроизведения по памяти временных параметров сигналов. Этот режим наступает после осуществления всех необходимых вычислений оптимальных временных задержек и переключения управляемого переключателя. В течение отрезка времени

временные параметры сигналов (их задержки друг относительно друга) не изменяются. В это время также осуществляется оптимизация энергетических параметров сигналов и текущая оптимизация при формировании дополнительных сигналов для активного шумопонижения. В течение этого отрезка времени может быть продолжен процесс статистического анализа параметров сигналов и помех.Time segments

are intended for organization of a (quasi-optimal) mode of reproducing from memory the temporal parameters of signals. This mode occurs after all necessary calculations of the optimal time delays and the switching of the controlled switch are performed. Over a period of time

time parameters of signals (their delays relative to each other) are not changed. At this time, the energy parameters of the signals are also being optimized and the current optimization is being carried out while generating additional signals for active noise reduction. During this period of time, the process of statistical analysis of signal and interference parameters can be continued.

 In specialized computers, signal processing can be carried out using several parallel processors.

могут формироваться соответствующим (синхронизирующим) генератором с жестко заданным периодом следования и скважностью импульсов

, который синхронизирует во времени операции по обработке сигналов и определяет состояние управляемого переключателя. В более сложных вариантах реализации программ параметры

могут быть переменными и функционально зависеть от вычисляемых в процессе обработки сигналов динамических изменений параметров сигналов. Например, если в процессе пяти последовательных циклов сбора, анализа параметров сигналов выясняется, что сигналы практически не изменяются, т.е. звуковоспроизведение осуществляется в квазистационарных условиях, то в процессе выполнения программы работы системы параметры

или один из них могут быть изменены.In the simplest version of the implementation of optimal processing programs, the parameters

can be formed by the corresponding (synchronizing) generator with a strictly specified repetition period and pulse duty cycle

, which synchronizes the signal processing operations in time and determines the state of the controlled switch. In more complex program implementation options

can be variables and functionally depend on the dynamic changes of signal parameters calculated during signal processing. For example, if in the course of five consecutive collection cycles, analysis of signal parameters, it turns out that the signals practically do not change, i.e. sound reproduction is carried out in quasi-stationary conditions, then in the process of executing the program of work of the system, the parameters

or one of them is subject to change.

сигнал в точке прослушивания является искаженным: отсутствуют предыскажения информационного сигнала по временным параметрам. Поэтому при выборе параметров

следует руководствоваться следующим.It is important to note that when implementing algorithms for optimizing the time delays of signals, for example, when implementing a full-parameter optimal algorithm, over time

the signal at the listening point is distorted: there is no predistortion of the information signal in time parameters. Therefore, when choosing the parameters

should be guided by the following.

желательно выбрать как можно меньше. Минимальное значение отрезка

определяется нижней частотой звукового диапазона в соответствии с условием (16)

где f_н и T_н - частота и период нижней частоты звукового диапазона. Условие (16) позволяет осуществлять анализ компонентов сигнала на любой частоте в полосе звуковоспроизведения (f_н ≤ f ≤ f_в).Firstly, the value of the parameter

it is advisable to choose as little as possible. The minimum value of the segment

determined by the lower frequency of the sound range in accordance with condition (16)

where f _n and T _n - the frequency and period of the lower frequency of the sound range. Condition (16) allows the analysis of signal components at any frequency in the sound reproduction band (f _n ≤ f ≤ f _in ).

следует выбирать также исходя из возможной удаленности точки прослушивания от громкоговорителя и, как уже отмечалось, исходя из особенностей динамических изменений параметров во времени при звуковоспроизведении в конкретных акустических и шумовых условиях.Parameter

should also be selected on the basis of the possible remoteness of the listening point from the loudspeaker and, as already noted, on the basis of the characteristics of dynamic changes in parameters over time during sound reproduction in specific acoustic and noise conditions.

т.е. увеличивая отрезок реализации сигналов, можно получить более информативный "образ" сигнала и в дальнейшем путем корреляционной обработки этих сигналов более точно находить временные задержки сигналов и их компонентов. Для ввода в систему сигнала, являющегося реакцией системы на сигнал источника, необходимо, чтобы параметр

был больше времени задержки Δt_з сигнала при прохождении всех элементов тракта звуковоспроизведения до точки прослушивания. Более того, это условие целесообразно усилить для повышения точности корреляционной обработки и представить в виде (17)

Полагая, что удаление точки прослушивания от громкоговорителя для подавляющего большинства помещений прослушивания составляет не более 3 - 6 м, что соответствует задержке сигнала не более 0,01 - 0,02 с, можно сделать вывод, что для подобных помещений условие (17) согласуется с условием (16).It is clear that by increasing the parameter

those. by increasing the period of implementation of the signals, it is possible to obtain a more informative "image" of the signal, and later, by correlation processing of these signals, it is more accurate to find the time delays of the signals and their components. To enter the signal into the system, which is the response of the system to the source signal, it is necessary that the parameter

there was a longer delay time Δt _{s of the} signal when all the elements of the sound reproduction path passed to the listening point. Moreover, it is advisable to strengthen this condition to increase the accuracy of correlation processing and present it in the form (17)

Assuming that the distance of the listening point from the loudspeaker for the vast majority of listening rooms is not more than 3 - 6 m, which corresponds to a signal delay of not more than 0.01 - 0.02 s, we can conclude that for such rooms condition (17) is consistent with condition (16).

в интервале, например, [0,05; 1,0] с, а параметр

в интервале, например, [0,1; 120] c.Thus, when compiling the optimal signal processing algorithms, it is advisable to set the parameter

in the range of, for example, [0.05; 1,0] s, and the parameter

in the range of, for example, [0.1; 120] c.

Representation of an audio signal with a duration of T _s in the form (14) allows us to find a general solution to the problem of improving the quality of information transmission in a channel with random parameters.

 The above approach can be illustrated by many examples of the interaction of man and machines in complex, dynamically changing conditions for their use. For example, when driving a car in inclement weather, the driver selects the mode of operation of the windshield wiper in accordance with the parameters of precipitation. In light rain, the driver intuitively sets the mode of operation with a long period of operation of the wiper blades because the driver’s distracting attention and the view of the road interfering with the road are minimized, and the visibility between the wiper cycles is satisfactory. In heavy rain, when it is often falling, large raindrops greatly impair visibility (raindrops and reduced visibility can be interpreted as the appearance of strong noise distorting the perception of information in the optical channel for obtaining this information, similar to acoustic noise during sound reproduction), the driver experimenting with possible the wiper operating modes, switches them to the operation mode with a short period of operation of the wiper blades. And although the interference from the movement of the brushes increases, but at the same time, the distortion of visual information due to raindrops is reduced so that the overall resulting perception of visual information is improved.

После очередного движения щеток очищенное стекло обеспечивает более высокую его способность пропускать через себя без искажений световую информацию аналогично той части отрезка

, на котором должна осуществляться согласованная фильтрация и активное шумопонижение для качественного восприятия звуковой информации.Thus, the period of time during which the wiper blades move along the glass, while giving some deterioration in the perception of information, is similar to the segment

After the next movement of the brushes, the cleaned glass provides a higher ability to pass through it without distortion the light information similar to that part of the segment

at which consistent filtering and active noise reduction should be carried out for a high-quality perception of sound information.

могут изменяться в процессе работы в широких пределах.A significant difference between this example and the above method is that there is no explicit “ideal” (reference) process with which the current parameters of the information transfer process are compared. But in an implicit form, the supporting process is still present in the "man-machine" system - in the form of typical road images (images of roads, cars, road signs, markings, bridges, street lights, cars, traffic lights, landscape samples and etc.) and possible mutual arrangements and dynamic connections between these images or their elements. For example, when rain intensifies, when the perception of road images greatly deteriorates and a person at a subconscious level, performing current correlation processing of visual information, gets low values of correlation coefficients of the analyzed images or their elements with “ideal” images stored in his memory, for example, he does not clearly see the boundaries the roadway or the environment further than 20 meters, then it turns on the wiper in intensive mode, optimizing (maximizing) using this device Signal-to-noise ratio (or RMSE) of visual information. Just as the wiper optimal operation mode depends on weather conditions, the parameters for sound reproduction using complex, intellectually intensive signal processing algorithms

can vary in the course of work over a wide range.

Однако не для всех практических ситуаций увеличение частоты срабатывания щеток стеклоочистителя является оптимальным решением. Например, при увеличении плотности капель дождя вплоть до полного их слияния (например, при падении автомобиля в реку) или в условиях сильного снегопада или тумана работа щеток теряет практический смысл, т.е. с помощью данной машины нельзя повысить качество приема визуальной информации. Аналогично и в звуковоспроизведении при сильных широкополостных или импульсных помехах эффективность работы системы может быть ограничена конкретными условиями звуковоспроизведения. В условиях помех с быстроизменяющимися параметрами, значения которых соизмеримы с временем задержки сигнала в тракте звуковоспроизведения, система может осуществлять лишь частичное подавление помех или вообще не успевать их отслеживать и подавлять. В этой ситуации возможен даже рост СКО. Поэтому при синтезе алгоритмов оптимальной обработки необходимо предусмотреть меры противодействия указанным эффектам, например отключать или дополнительно демпфировать узлы автоматического регулирования параметров сигналов.For example, when the rain ends, the driver turns off the wiper until the next rain, which corresponds to a significant (by orders of magnitude) increase in the parameter value

However, not for all practical situations, increasing the frequency of operation of the wiper blades is the optimal solution. For example, when the density of rain drops increases until they merge completely (for example, when a car falls into a river) or in conditions of heavy snowfall or fog, the operation of the brushes loses its practical meaning, i.e. with the help of this machine it is impossible to improve the quality of reception of visual information. Similarly, in sound reproduction with strong wideband or pulsed noise, the system performance may be limited by the specific conditions of sound reproduction. In conditions of interference with rapidly changing parameters, the values of which are comparable with the delay time of the signal in the sound reproduction path, the system can only partially suppress interference or not have time to track and suppress them at all. In this situation, even an increase in DIS is possible. Therefore, in the synthesis of optimal processing algorithms, it is necessary to provide measures to counteract these effects, for example, disable or additionally damp the nodes of automatic control of signal parameters.

 In the example of interaction between the driver and the car wiper, a similar situation corresponds to a decision by the driver about the need to reduce the driving speed or even completely stop the movement in conditions of poor visibility and, accordingly, turn off the wiper.

 Thus, the representation of time in the form of series (14) in an artificial way allows us to simulate non-ergodic and non-stationary processes in the form of alternating sections of these processes, which are modeled as stationary processes.

 The parameters of these processes during the analysis and processing change insignificantly, so it turns out to be expedient to reproduce the signals in the corrected form based on the previous analysis cycle, i.e. by memory. As a result of such signal processing, it is possible by objective criteria to increase the signal-to-noise ratio (interference) and the accuracy of information transfer to the listener. The specified technique allows you to separate in the time domain the processes of input, processing and output of information without interrupting the perception of the human information flow. In other words, due to these actions, it is possible to break the interdependence of the frequency response and phase response of the optimized signals. The proposed method for adjusting the parameters of reproducible signals has some analogy with the process of interrupted playback of signals from memory in car CD players equipped with a system for reproducing signals during strong vibration and shaking when driving on rough roads.

 In multichannel sound-reproducing systems, it is necessary to provide in the signal processing algorithms the possibility of the appearance of anomalous errors when finding the delays of the signals and their components.

 The indicated errors can appear as a result of erroneous comparison and analysis of signals for different channels, since the signals in the sound reproduction channels are highly correlated and signals having components at the same or close frequencies are present for a significant part of the time. In such a situation, the signals of, for example, the left channel when determining the delays will be an obstacle for the right channel and vice versa. A problem of this kind can be solved by briefly gating one of the channels for the time necessary to reliably find the signal delays in the other channel. Gating can be carried out both in the entire band of sound frequencies, and in separate bands of analysis. Gating of signals can be carried out by microprocessor control devices or software.

 To eliminate anomalous errors due to noise and interference, it is possible to use statistical methods for processing signal parameters. In such methods, the results of measuring the signal-to-noise ratio can be used as a criterion for the reliability of the measured parameter.

 When compiling algorithms for generating additional signals for active noise reduction, it is necessary to take into account that the distance from possible sources of acoustic noise and noise to the listening points, as well as the distance from the speakers to the listening points are generally random parameters. Therefore, it is common for algorithms for generating additional signals to fit a priori unknown parameters of these signals to those required by the following criteria.

 Firstly, it is necessary to determine the signal to noise ratio (interference) in each analysis band. If the signal-to-noise ratio is below the threshold, it is necessary to generate, for example, using an electronic key, an additional signal.

 Secondly, it is necessary to radiate this signal, which has random parameters in the initial phase. This signal, having reached the listening point in the sound reproduction space, in accordance with the principle of signal superposition will lead to a change in the parameters of the total acoustic signal. The signal level will jump either increase (in common-mode signal addition), or decrease. An abrupt change in the level of signals at the listening point, by the way, can be used in algorithms for finding signal delays.

 Thirdly, it is necessary by successive approximations (iteratively) to carry out phase optimization of additional signals, for example, discretely changing the time delay of the additional signal. The criterion for completing the phase optimization process of the additional signal is the minimum level of interference components in the listening signal.

 Fourth, it is necessary through successive approximations to carry out energy optimization of the additional signal. The criterion for the end of the process of optimizing the levels of additional signals is also the minimum level of interference components in the listening signal.

 Obviously, for stationary interference, the parameters of which do not change over time, by choosing the appropriate sampling step in the time and energy domains of parameter optimization, noise reduction can be achieved with the required accuracy, i.e. with acceptable signal to noise ratio and standard deviation.

 Thus, when generating additional signals for active noise reduction, it is necessary to carry out two-parameter signal optimization in the corresponding analysis band.

 In conclusion, the rationale for the proposed method of sound reproduction will comment on the relationship between the functions U (t) and F (t).

In some cases, over some time intervals T ' _c , the functions U (t) and F (t) for the listener can be (within the framework of the mathematical and physical models used to describe sound reproduction) identically equal, i.e. match completely. A practically similar situation is possible, for example, when a visitor to a concert, using a stereo tape recorder and microphones located on his head, listens and simultaneously records magnetic sound information during this concert (T ' _c = t' ₂ -t ' ₁ ). After conducting such a recording, the listener at a convenient time for himself, i.e. already in another time interval T _c = t ₂ -t ₁ and in other acoustic conditions, for example, at home, using the optimal sound reproduction system, in which the same microphones are used as the sounding device as when recording a concert, it can repeatedly listen to the whole ( T ' ₂ = T _c ) or part (T' _c ≠ T _c ) of the record.

 This situation is rare in practice. In practice, the more common option is when a person wants to listen at home, for example, a concert of his favorite rock band, a recording of which was made before his birth or to which the listener could not get for other reasons. In the framework of the proposed method of sound reproduction, the listener can perform high-quality listening to a recording that he had never heard before and, in principle, could never hear. As a storage medium, he should purchase, for example, a personalized CD-ROM corresponding to the size of his head, and elements of an optimal sound reproducing system, calibrated in studio conditions according to universal criteria: with constant frequency response and linear phase response. The listener can perform high-quality sound reproduction of signals repeatedly and in various listening rooms, for example, in a car, in conditions of stronger noise and interference.

можно измерять в процессе обработки сигналов и поэтому данный параметр целесообразно использовать в качестве универсальной интегральной характеристики для описания качества (точности) работы звуковоспроизводящей системы и классификации систем.Obviously, depending on the acoustic properties of the listening room, characteristics and noise intensity, the difference between listening signals from the source signals will be different. In other words, the standard deviation of the listening signal from the source signal is a function of the parameters of the audio reproduction path,

can be measured in the process of signal processing, and therefore this parameter should be used as a universal integral characteristic to describe the quality (accuracy) of the sound reproducing system and the classification of systems.

 Note that the standard deviation as an integral parameter that gives an objective idea of the quality of sound reproduction cannot fundamentally be used for traditional sound reproducing systems because despite the presence of local feedbacks, for example, in low-frequency amplifiers in traditional systems, there is no feedback between input and output of the entire sound reproduction path and, therefore, there is no possibility of comparing signal implementations and determining the standard deviation. In traditional sound reproducing systems, each individual hardware element of the sound reproduction path is described by a whole group of quality indicators, for example, nonlinear distortions, intermodulation distortions, phase distortions, transient distortions, frequency response distortions, etc., which are actually not independent indicators and actually in different physical interpretations show the ability of devices to repeat the shape of the input signal at their output. A large number of quality indicators makes it difficult to integrate the evaluation of devices to convert signals without distortion. Since the difference between the two signals is fully described using the standard deviation, it is advisable to analyze the standard deviation of the entire system as a whole and its individual hardware elements using test signals that simulate all possible forms of real sound signals when classifying sound reproducing systems. The generally accepted normalization of system parameters for several harmonic signals, for example, the nominal level at low, medium, and high frequencies of the audio reproduction band, does not quite correspond to the real sound signals and the features of the operation of devices under the influence of these signals.

The most suitable model of test signals for testing systems and their classification by RMSE values can be, for example, a group of the following signals:
1) a signal of the type "pink noise" with frequency boundaries of the audio reproduction band, with a linearly increasing level from zero of the maximum value; slew rate varies; in a particular case, a pulsed noise signal;
2) a noise-like signal with a stepwise increasing from zero to maximum level.

 The use of such signals is justified by the possibility of combining the analysis of the system in the frequency domain and the dynamic properties of the system, i.e. analysis in the time domain.

 These test signals can be standardized and generated by the most optimal sound reproducing system for testing it during purchase and during operation, for example, for testing the acoustic properties of the listening room, or choosing the most suitable speaker location, or accelerated initial self-adaptation of the system.

 Using this approach will allow the consumer to practice the choice of elements of sound-reproducing systems according to their objective parameters, and not according to random criteria, for example, under the influence of a manufacturer’s advertisement, the declared technical parameters of devices or the seller’s opinion, which argues that a sound-reproducing device or system is superior to a manufacturer’s higher reputation .

 For example, by replacing the components of the equipment under comparison, ceteris paribus, the consumer can quickly choose, for example, ULF, which provides the lowest standard deviation of test signals or at equal standard deviation has a lower cost. Using a similar technique, a consumer can consciously acquire software when implementing a signal processing unit using a computer.

 Consider the options for the implementation of optimal sound reproduction systems and examples of algorithms for optimal signal processing.

 The optimal sound reproduction system (Fig. 2) contains a signal source 1 and a sound reproduction channel, made in the form of a low-frequency amplifier 2 and a loudspeaker 3 connected in series, a sounding device 4, a signal processing unit 5, adapted for coordinated filtering of the signal of the source 1, communication lines 6, while the output of the probing device 4 through the communication line 6 is connected to the first input of the signal processing unit 5. The output of the signal source 1 is connected to the second input of the signal processing unit 5, and the output of the signal processing unit 5 is connected to the input of the low-frequency amplifier 2.

 According to the invention, the signal processing unit 5 is configured to generate additional signals at its output for active noise reduction.

 The system of optimal sound reproduction (Fig. 2) works as follows.

 By means of the receiving element of the probing device 4, the sound signals present at the point of its location are received - direct sound waves from the loudspeaker 3, reflected from the walls and other objects of the room, acoustic waves of the loudspeaker 3, acoustic noise and noise of various origin.

 The received aggregate, acoustic, distorted signal is converted into an electrical listening signal and through the communication line 6 is supplied to the first input of the signal processing unit 5.

 As the communication line 6 can be used cable or wireless communication line 6, for example, communication line 6 using radio waves or infrared waves. The signal distortion in the communication line 6 determines the threshold of accuracy of the system (RMS) and therefore it is preferable to use the communication lines in the form of a closed channel for transmitting information: cable lines, fiber optic lines or information channels that use signal coding methods.

 The second input of signal processing unit 5 receives a signal from the output of source 1.

 In the signal processing unit 5, in accordance with a predetermined hardware or software method, the signal processing is carried out, which consists in performing a coordinated filtering of the signal of the source 1 and generating additional signals for active noise reduction. The output signal of the signal processing unit 5 is the sum of the optimal corrected signal of the source 1 and the additional signals. This signal, passing through the sound reproduction channel: a low-frequency amplifier 2, a loudspeaker 3 and a listening room, at the location of the receiving element of the sounding device 4, repeats the electrical signal of the signal source 1 with the greatest possible accuracy.

 If the specific conditions of sound reproduction do not allow the reproduction at the point of listening to the signal with some accuracy, then sound reproduction is not optimal in the sense of quality guarantees. In such a situation, the system provides a decrease in entropy in the sound reproduction channel, but the required signal-to-noise ratio or standard deviation is not achieved.

 The speed of the optimization process of sound signals at the listening point is determined by the time constants of automatic control of the optimized parameters of the signals, the selected algorithm of the system and the sound conditions.

 The optimal sound reproduction system (figure 2) is optimal for people with one normally functioning hearing organ, i.e. disabled people.

 Signal processing unit 5 can be implemented as a narrow-function device or a computer can be used as unit 5, for example, a standard PC with input devices (two-channel ADC) and output (single-channel DAC) information.

 For listeners having two operable hearing organs, a variant of the functional diagram of the optimal sound reproduction system is shown in FIG. 3.

 The system (figure 3) contains a signal source 1, which is made with at least one additional output for multi-channel audio playback. Additionally introduced according to the number of additional outputs of the signal source 1 at least one additional audio channel. The additional channel is made in the form of an additional low-frequency amplifier 7 and an additional loudspeaker 8 connected in series. The additional channel also includes an additional sounding device 9, an additional signal processing unit 10, configured to filter the signal from the source 1 and generate additional signals for active noise reduction, an additional communication line 11. The output of the additional sounding device 9 through an additional communication line 11 is connected to the first the input of the additional signal processing unit 10. An additional output of the signal source 1 is connected to the second input of the additional signal processing unit 10. The output of the additional signal processing unit 10 is connected to the input of an additional low-frequency amplifier 7.

 In FIG. 3 letters A and B denote a channel and an additional sound reproduction channel, respectively.

 As the signal source 1, any known type of multi-channel signal source 1 can be used using any known primary information carrier, as well as systems and methods for converting and generating multi-channel signals.

 There are various design options for the implementation of probing devices 4 and 9, for example, in the form of headphones, glasses, clothespins. Receiving elements can be placed on the head of the listener, on the headrest of a car or home seat, or in the remote control of the system. The greatest reliability of sound reproduction can be obtained with the location of the receiving elements of the sounding devices 4 and 9 near the hearing organs of the consumer.

 In working condition, the receiving elements of the sounding devices 4 and 9 together with the listener's head form a sounding "microphone-head".

 When using a two-channel (stereo) system, one of the receiving elements, for example, channel A, can be an electronic model of the listener's right ear, and the receiving element of the sounding device 9 of channel B can be an electronic model of the listener's left ear, respectively. With the appropriate (oriented) location on the listener's head of the receiving elements of the sounding devices 4 and 9 during the operation of the system, the output electrical signals of listening will take into account the influence of such random factors as size, shape (due to diffraction of sound vibrations on the head), spatial orientation of the head of a particular listener .

 The principle of operation of channels A and B of the optimal sound reproduction system (Fig. 3) is similar to the principle of operation of the single-channel system described above.

 As a result of the operation of channels A and B at points of space near the receiving elements of the sounding devices 4 and 9 (near the hearing organs), sound vibrations approach in shape to the sound vibrations of the original sources of sounds. Accurate reproduction of the amplitudes and phases (time delays) of the components of sound vibrations allows you to accurately reproduce the spatial perception of the "sound picture" with the localization of the apparent sources of sounds in a specific place relative to the listener's head.

 For the system to work (Fig. 3), it is not necessary to use low frequency amplifiers 2 and 7, loudspeakers 3 and 8 with completely identical characteristics in each of the sound reproduction channels: frequency response, phase response, and efficiency, since the mismatch between these parameters is ideal and to each other is compensated during operation system. It is important that each signal processing unit 5 and the additional signal processing unit 10, as well as the sounding device 4 and the additional sounding device 9, the communication line 6 and the additional communication line 11, are equipped for their sound channel in an unchanged in its parameters unit, which, in in turn, it must be pre-calibrated before operating the system in accordance with condition (12).

 When implementing blocks 5 and 10 of signal processing in the form of a computer with input devices (four-channel ADC) and output (two-channel DAC) with the corresponding software, blocks 5 and 10 do not require preliminary settings, because setting and debugging these blocks is the debugging of the optimal signal processing program. Only devices 4, 9, 6, 11 are subject to calibration. Structurally, the filter adjusting elements for adjusting the frequency response and phase response of these nodes can be implemented inside the cases of the probing devices 4 and 10 in the form of R, C, L elements or equivalent information about the need for correction can be entered in A computer is at the stage of downloading software using, for example, a diskette, magnetic tape, chip, or any other known method of entering information into a computer.

 When configuring these system nodes in an anechoic chamber, a noise generator (including a virtual one implemented in software) can be used as a signal source 1, and as a control and measuring equipment: a multi-channel spectrum analyzer, low-frequency amplifier, microphone head, and loudspeaker studio class. To increase the efficiency of work on tuning and calibration of system elements, computers with the appropriate program can be used.

 The principle of operation of the optimal system can be applied, for example, to a four-channel optimal sound-reproducing system, for example, for two people with normally functioning hearing organs or, for example, for two people with disabilities with one normally functioning hearing organ and one person with two normally functioning hearing organs.

If each output of source 1 corresponds to a signal u _{source k} (t), where k = 1,2 _,. . . K, then this signal can be associated with your listening signal u ' _k (t), which can be optimized in accordance with the proposed method of sound reproduction. According to the principle of reciprocity, each listening signal must have its own source 1 signal and an optimal sound reproduction channel. If the number of signals from source 1 is less than the number of listening points, then optimization of signals in accordance with the proposed method becomes impossible, because the principle of cause and effect is violated. For example, using a single loudspeaker it is not possible to ensure the simultaneous arrival of a direct sound wave for two different points of space along the axis of the loudspeaker. Using two loudspeakers, it becomes possible to realize these signals by selecting the appropriate signal delays and their levels.

Thus, the general solution to the problem of optimal sound reproduction for a group of listeners can be represented as a joint solution to the problems of optimal sound reproduction for each individual listener from this group. The condition for optimal collective sound reproduction can be represented in the form (18)
K = N
where K is the number of outputs of the signal source 1,
N is the number of hearing organs.

 For example, for high-quality sound reproduction in a two-seater car, it is necessary to use four channels of information and the same number of channels of optimal sound reproduction, respectively, in a four-seater car it is necessary to use an eight-channel sound reproduction system, etc.

 It is clear that with this approach, various variations of connecting the outputs of the signal source 1 to the sound channels are possible and, therefore, the optimization of these connection options is possible. As a criterion of the best option for connecting the outputs of the signal source 1 to the sound reproduction channels, there may be a minimum value of the arithmetic mean of the standard deviation for all possible options. This criterion corresponds to the equilibrium process of signal optimization in a variety of information channels.

 The proposed approach can be implemented in the synthesis of optimal home theater systems or a home concert hall with the number of potential visitors up to three to five people. With a larger number of viewers, the proposed method of sound reproduction is difficult to implement practically because of the difficulties in registering multichannel information at the stage of its input, for example, during the scoring of a film using microphone heads.

 In multichannel sound reproduction systems, it is advisable to provide an option for optimizing the combination of activating channels with an incomplete number of listeners.

 In the optimal sound reproduction system, a single-channel (Fig. 4) or multi-channel (Fig. 5) signal generator 12 can be additionally introduced, which is connected by an output to the third input of the signal processing block 5 (10).

 As a signal generator 12, a noise generator can be used.

 The principle of operation of optimal sound reproduction systems (Figs. 4, 5) is similar to the principle of operation of the above-described devices with the only difference being that, to accelerate the system’s adjustment to optimal mode, using the generator 12, auxiliary test signals are formed, for example, before the system starts operation. When implementing blocks 5 and 10 using a computer, the indicated generator 12 can be implemented virtually, programmatically, for example, in the form of the corresponding subroutine of the “random number generator 12” with subsequent “modulation” of this signal by software to obtain, for example, pulse-modulated noise-like test signals. In this example, the third input of processing unit 5 (10) is a virtual, program input that connects the main program with a subroutine, “generator 12”, using the conditional or unconditional jump operator.

 In the constructive implementation of optimal sound reproduction systems, a monoblock version or processing units 5 (10), a communication line 6 (11), a sounding device 4 (9) can be used, functionally, can be combined into a separate device - the “optimal signal processing unit” in order to so that the consumer, already having a stereo system with block configuration, could complement it to the level of an optimal sound reproduction system. As signal processing units, as already noted, one can use ADCs, computers with software, and DACs connected in series.

 If in the optimal sound reproducing system the signal source 1 does not provide high quality of the output signals and information loss and distortion occur in it, for example, due to the imperfect frequency response of the FM tuner receiving path or the scatter of the parameters of the tape recorder heads, it is advisable that the signal source 1 be executed with the ability to control the amplitude-frequency characteristics of the signals at its outputs. The specified adjustment can be implemented using any well-known circuitry solutions timbroblocks or equalizers. This adjustment can also be used by people with hearing impairment. For this, a person with a hearing impairment in the high-frequency region can consult a doctor and take his frequency response for each organ of hearing, and then enter the corresponding correction of the frequency response of signal source 1. The procedure for determining the required correction of signal reproduction in sound channels can be automated using a computer and be part of its software. The dialogue form of communication of the listener with the machine and input into the program of optimal sound reproduction of the indicated information is preferable.

 If the source 1 of the signal is not ideal, it is advisable to perform it with the possibility of noise reduction.

 Since different sources of signals 1 may have different levels of output signals, it is advisable that the source 1 of the signal was made with the possibility of automatic control of the levels of its output signals. For these purposes, you can use technical solutions similar to devices for automatic control of the recording level (ARUZ) of tape recorders. The time constant can be selected on the order of 1 - 20 s.

 In an embodiment of the system, it is advisable to perform a signal source 1 with the possibility of automatic level control and non-automatic regulation of the amplitude-frequency characteristics of the signals at its outputs.

 To ensure volume control at the listening point, the signal source 1 can be configured to control signal levels at its outputs. For the same purpose, the probing device 4 and the additional probing device 9 or the communication line 6 and the additional communication line 11 can be configured to control the transmission coefficient.

 The system of optimal sound reproduction works in accordance with the principle of action already described. If at the first or second inputs of block 5 (10) a signal level changes, including due to the volume control, the system automatically compensates for the inconsistency of the signal energy at the listening point, changing the volume of sound vibrations by the required value.

 If the volume control is performed loudly, including using the results of automated testing of the individual auditory abilities of a particular listener, then in the process, depending on the volume chosen by the listener, the system will automatically carry out frequency predistortions of the signal according to the psychophysiological characteristics of hearing to perceive sounds of different frequencies. In other words, in this case, the system will correct the loss and distortion of sound information occurring in a person due to the deterioration of the hearing organs - the "devices" for inputting sound information into the brain and nervous system ("computer") of a person.

 In systems with block layout of functional units, the volume control can be installed on the first or second inputs of signal processing unit 5 (10) because in classical systems with block layout, the volume control is often installed at the input of the VLF rather than at the output of signal source 1. Volume control with the help of a knob installed at the VLF input will not allow you to change the volume at the listening point, as the attenuation of the signal at the input of the ULF will be compensated in the processing unit 5 (10) by automatically proportionally increasing the signal at its output.

 In embodiments of optimal sound reproduction systems, it is advisable that the signal processing unit 5 and the additional signal processing unit 10 are configured to automatically control the amplitude-frequency characteristics or phase (temporal) -frequency characteristics from the second inputs to the outputs of the blocks to optimize energy or time (phase ) signal parameters, as well as an option in which the signal processing unit 5 and the additional signal processing unit 10 were configured to automatically Cesky adjusting the amplitude-frequency and phase-frequency characteristics to a second input of block outputs for polnoparametricheskoy optimize energy and waveform parameters.

 To clarify the features of the operation of the indicated embodiments of optimal sound reproduction systems, we use the functional diagram of the signal processing unit 5 (10), shown in FIG. 6.

Block 5 (10) of the signal processing (Fig. 6) contains a control device 13, bandpass filters 14 ₁ , 14 ₂ , ..., 14 _R , blocks 15 ₁ , 15 ₂ , ..., 15 _P matched filtering, blocks 16 ₁ , 16 ₂ , ..., 16 _W generating additional signals for active noise reduction, adder 17 connected to each other as shown in FIG. 6.

 As the control device 13, a synchronizing pulse generator or a microprocessor control device can be used. When implementing the block 5 (10) of signal processing using a computer, the control device 13 is implemented programmatically, as part of a software information processing algorithm (program tree).

 As filters 14, any known types of active or passive filters can be used, for example, filters using operational amplifiers and R, C, L elements or filters using SAW technology.

 When implementing a block 5 (10) of signal processing using a computer, the filters 14 are implemented in software, as part of the information processing algorithm. With the digital version, it is advisable to choose the rectangular shape of the frequency response of the filters 14, and place the process of calculating the output signal of the filter 14 in the time domain between the clock samples of the input signals to reduce the time of digital filtering of signals.

 The frequency bands and the shape of the frequency response of the filters 14 at the first input should correspond to the frequency response of the filters 14 at the second input of block 5 (10).

Blocks 15 matched filtering can be made in the form of full-parameter blocks 15 containing blocks 18 ₁ , 18 ₂ , ..., 18 _L optimization of the temporal parameters of the signals and blocks 19 ₁ , 19 ₂ , ..., 19 _Q optimization of the energy parameters of the signals ( Fig. 6), or in the form of non-full-parameter blocks 15, each containing one block 18 or block 19.

 When block 5 (10) is executed, it is possible to use blocks of the same type 15 or blocks of different types 15 in various combinations, for example, part of the blocks at low and high frequencies is performed as one-parameter blocks 15, and part of blocks 15 at the middle frequencies - in the form of full-parameter blocks 15. Given an example of the execution of block 5 (10), it is advisable to use the analog form of execution of blocks 15 to reduce the number of functional nodes of the optimal sound reproduction system and its price. In digital versions of the implementation of blocks 5 (10), it is better to use full-parameter blocks 15, since these blocks can be synthesized programmatically - in the form of the corresponding signal processing routine. This subroutine can be accessed using transition operators any number of times required, as well as varying the parameters of these blocks during their optimization. The speed of software filtering processes of signals is determined by the speed of the computer, the signal processing algorithm, and for sound frequencies approaches the speed of analog filtering. In an embodiment of the signal processing blocks 5 (10) for some frequency bands, for example low, blocks 15 may be excluded as functional devices. In these cases, the equivalent circuit will be a line segment that does not distort the signal, connecting the second input to the output of block 15 or a broadband delay line. However, in order for the operation of block 5 (10) to correspond to the above method of optimal sound reproduction, it is necessary that at least one of the bands of the sound spectrum or the entire spectrum carry out consistent filtering. In the particular case of matched filtering for the entire frequency band of sound reproduction, block 15 is in the form of block 18 and / or 19, the first and second inputs of which are fed from the outputs of the respective adders, the inputs of which are connected to the outputs of the corresponding filters 14.

Число R полосовых фильтров 14 должно быть не менее двух, т.к. при одном полосовом фильтре 14 становится принципиально невозможна обработка сигналов и формирование дополнительных сигналов для активного шумопонижения в рамках предложенного способа звуковоспроизведения. Чем больше число полос анализа R, тем выше разрешающая способность блока обработки 5 (10) в частотной области, т.е. способность более точного компонентного анализа сигналов при формировании дополнительных сигналов для активного шумопонижения.Thus, if the number of blocks 15 is denoted by P, the number of blocks 16 by W, the number of blocks 18 by L, and the number of blocks 19 by Q, in the most general form for any combination of their use, the relationship between these parameters and the number R of filters 14 can be represented as relation (19)

The number R of bandpass filters 14 must be at least two, because with one band-pass filter 14, it becomes fundamentally impossible to process signals and generate additional signals for active noise reduction in the framework of the proposed method of sound reproduction. The larger the number of analysis bands R, the higher the resolution of processing unit 5 (10) in the frequency domain, i.e. the ability of more accurate component analysis of signals during the formation of additional signals for active noise reduction.

 Each of the blocks 18 and 19 has at least two inputs, which receive signals from the outputs of the filters 14, and one output.

 In a fully parametric embodiment of blocks 15, the first inputs of blocks 18 and 19, intended to be connected to filtered listening signals, are connected and are the first input of block 15. Other inputs and outputs of blocks 18 and 19 may have the connections shown in FIG. 6: the second input of block 18 is the second input of block 15, and the output of block 18 is connected to the second input of block 19. The output of block 19 is the output of block 15.

 These relationships can be implemented in the reverse order. The second input of block 15 will be the second input of block 19, the output of which is connected with the second input of block 18, the output of block 18 is the output of block 15.

 Blocks 18 in one-parameter and full-parameter variants of implementation, as well as block 19 in the digital version of its implementation must have additional, control inputs, and in variants, outputs for communication with the control device 13.

Blocks 18 ₁ , 18 ₂ , ..., 18 _{L of} optimizing the time parameters of signals are devices for correlation signal processing and can be performed, for example, in the form of delay lines 20 ₁ , 20 ₂ , ..., 20 _Z , multipliers 21 ₁ , 21 ₂ , ..., 21 _Z , low-pass filters 22 ₁ , 22 ₂ , ..., 22 _Z , analysis circuit 23, switching circuit 24 interconnected as shown in FIG. 6.

Units 19 ₁ , 19 ₂ , ..., 19 _{Q for} optimizing energy parameters generally contain devices for determining the energy of signals during analysis, the signals from the outputs of which are fed to a comparison circuit, the output signal of which is fed to the control input of a controlled amplifier or attenuator. The input of one of the signal energy determination devices is connected to the input of the controlled amplifier (attenuator) and is the second input of block 19. The output of the controlled amplifier is the output of block 19. The input of the other signal energy determination device is the first input of block 19.

 In FIG. 6 illustrates one possible embodiment of block 19 in the form of detectors 25, low-pass filters (LPFs) 26, resistors 27, and a first controlled amplifier 28 connected together as shown in FIG. 6.

Blocks 16 ₁ , 16 ₂ ,. .., 16 _{W the} formation of additional signals for active noise reduction contain two functional units connected in series with each other: a phase shifter made in the form of a controlled phase shifter 29 with a device 30 for controlling the phase shifter 29, and a second controlled amplifier 31.

 The connection order of these nodes can be reversed.

 In FIG. 6 also shows a loudness volume control: R1, C1, C2 installed on the second input of block 5 (10).

2) наличие стационарной помехи при отсутствии сигнала источника

3) наличие нестационарной помехи при отсутствии сигнала источника

4) наличие сигнала источника при отсутствии сигнала прослушивания

5) наличие сигналов источника при слабых помехах

6) наличие сигналов источника при сильных помехах

Если на входах блока 5 (10) сигналы отсутствуют, т.е. u_ист(t) = 0 и u'(t) = 0, то соответственно отсуствуют сигналы и на выходах полосовых фильтров 14. На входах блоков 15 и входах блоков 16 сигналы также отсутствуют. Независимо от сигналов устройства 13 управления, поступающих на блоки 15 и в вариантах на блоки 16, выходные сигналы (предыскаженные информационные сигналы и дополнительные сигналы для активного шумопонижения) также равны нулю. Сигнал на выходе сумматора 17 отсутствует.For the convenience of describing the operation of block 5 (10), we will consider the work with input signals of the following type:
1) there are no signals at the inputs of block 5 (10)

2) the presence of stationary interference in the absence of a source signal

3) the presence of unsteady interference in the absence of a source signal

4) the presence of a source signal in the absence of a listening signal

5) the presence of source signals with low noise

6) the presence of source signals with strong interference

If there are no signals at the inputs of block 5 (10), i.e. u _ist (t) = 0 and u '(t) = 0, then there are accordingly no signals at the outputs of the bandpass filters 14. At the inputs of blocks 15 and the inputs of blocks 16 there are also no signals. Regardless of the signals of the control device 13 arriving at blocks 15 and, alternatively, at blocks 16, the output signals (pre-emphasized information signals and additional signals for active noise reduction) are also equal to zero. The signal at the output of the adder 17 is missing.

 Thus, at the point of listening, the sound signal exactly repeats the signal of the source 1 - silence.

 The considered situation in practice is extremely rare. The most typical situation is when there is no signal, for example, in pauses between phonograms, in the presence of acoustic noise or noise.

If the second input unit 5 (10) No signal source u _ist (t) = 0, and the first input unit 5 (10), the signal is not zero u '(t) ≠ 0, then the signal processing can be represented as the following a set of conditions and actions on signals.

Let for any moment of time, regardless of the parameters of the information signal of the source 1 u _ist (t) and the listening signal u '(t), the control device 13 generates, for example, rectangular pulses of duration Δt ′ and pause Δt ″, which are alternately supplied to blocks 5, 10.

Let these operating pulses determine the operating modes of blocks 15 and 16, for example, in accordance with the following rules:
- during the time Δt ′ in blocks 18 the signal is correlatedly processed - the optimal time delay of the signal of source 1 or the delay of its components is calculated, and for any time interval Δt ′ when passing through the block 18 of the signal of source 1 all its components are delayed by one and the same same time
- after the pulse Δt ′, block 18 does not carry out correlation signal processing, while the signal of source 1, passing through block 18, is delayed by the time corresponding to the optimal signal delay calculated in the previous step Δt ′,
- the operating mode of the blocks 19 is independent of the signals of the control device 13 and is determined only by the parameters of the input signals,
- the output signals of the blocks 16 are determined by the parameters of the input signals and the internal state of the blocks 16.

Consider the operation of block 5 (10), assuming for definiteness that, until time t ₁ , there were no signals, for example, at the inputs of block 5 (10), and starting from the indicated time, a stationary noise appeared, for example, in the form of a hum from a transformer of an household appliance . The indicated initial conditions correspond to a zero signal level at point D (Fig. 6). The zero value of the signal at point D corresponds to a certain transmission coefficient of the first controlled amplifier 28. Let, for example, the midpoint of the control characteristic of the first controlled amplifier 28 correspond to the zero signal level at point D. The parameters of the control characteristic of the first controlled amplifier 28 determine the dynamic range of optimization of the energy parameters of the signals and can be selected, for example, 60 dB, 90 dB or 120 dB.

 In the analogous embodiment of the implementation of the blocks 19, the parameters of the devices are selected for reasons of their physical feasibility and reasonable cost and weight and size parameters. In a digital, software-based method for implementing blocks 5 (10), parameters are selected based on the characteristics of the computer used (memory size, speed), parameters of input and output devices (ADC and DAC), and human parameters. For example, the dynamic range of the first controlled amplifier 28 is selected equal to 120 dB, the quantization step in level in accordance with the quantization step of the ADC.

где

- комплексный коэффициент передачи i-го полосового фильтра 14_i. Эти сигналы поступают на первые входы соответствующих блоков 15_i и 16_i.Relative to the time t ₁ at the outputs of the band-pass filters 14, in the pass-band of which the spectral components of the noise fell, with some delay determined by the pass-band of the filters 14, filtered listening signals appear, the spectrum of which is determined by the expression

Where

- the complex transfer coefficient of the i-th band-pass filter 14 _i . These signals are fed to the first inputs of the corresponding blocks 15 _i and 16 _i .

From time t _{1, the} signals of source 1 do not undergo changes and remain equal to zero. Therefore, at the outputs of blocks 15, the predicted signals of source 1 also remain equal to zero.

The spectral components of the listening signal that fall into the passband of the i-th band-pass filter 14 _i are detected by the detector 25 _i and converted, for example, into negative voltage pulses, which are then smoothed by the low-pass filter 26 _i so that the signal that controls the matched filtering (signal at point D, Fig. 6) begins to change - to decrease.

In proportion to the signal changes at point D, the gain of the first controllable amplifier 28 _i changes. The gear ratio is reduced. Since the signal level of source 1 at the second input of block 19, i.e., at the input of the second detector 25 _i , which has the reverse polarity (positive) of the output pulses with respect to the output signals of the first detector 25 _{i of} block 5 (10), does not change the signal controlling the matched filtering in this particular case is functionally dependent only on the level of the components of the listening signal (interference level).

интегрирования ФНЧ 26_i блока 19_i выбирается одинаковой и определяется низшей частотой полосы анализа сигнала, т.е. низшей частотой полосы пропускания i-го полосового фильтра 14_i согласно условию (20)

где f_н.14i и T_н.14i - соответственно частота и период нижней частоты полосы пропускания i-го полосового фильтра 14_i.Time constant

integrating the low-pass filter 26 _{i of the} block 19 _i is chosen the same and is determined by the lowest frequency of the signal analysis band, i.e. the lowest passband frequency of the i-th bandpass filter 14 _i according to condition (20)

where f _n.14i and T _n.14i are the frequency and period of the lower frequency bandwidth of the i-th bandpass filter 14 _i .

 Thus, the optimization rate of the energy parameter is functionally dependent on the frequency of the processed signal. For example, for a frequency of 100 Hz, the LPF time constant must be selected at least 0.01 s, and for a frequency of 10,000 Hz the time constant can be reduced by two orders of magnitude.

 In highly intelligent signal processing algorithms, this parameter can be optimized by the system itself. The criterion for optimizing the parameter is the minimum value of the standard deviation of the signal from the group or ensemble of signal implementations when the parameter is varied.

In accordance with FIG. 6, the signal controlling the matched filtering also arrives at the control input of the second controlled amplifier 31 _{i of the} block 16 _i generating additional signals for active noise reduction. At a negative level of the control signal, which corresponds to the case under consideration, the transmission coefficient of the second controlled amplifier 31 _i starts to increase proportionally from zero to a certain value.

Thus, from time t _{1 the} output signal of block 16 _i becomes non-zero, the formation of additional signals for active noise reduction begins. These signals come from the outputs of blocks 31 _i to the inputs of the adder 17 and then, passing through all the elements of the sound path to the listening point, through the sounding device 4 (9) and communication line 6 (11) are fed to the first input of block 5 (10).

Since the delay that the signal acquires when passing through all the elements of the audio reproduction path to the listening point is random, it also turns out to be random that the level of the listening signal increases when additional signals are generated for active noise reduction. For example, if an acoustic, additional signal is out of phase with interference, then the level of the resulting listening signal will decrease and, therefore, at the outputs of the bandpass filters 14 _{i the} signal will also decrease. According to these changes, the signal controlling the operation of blocks 28 _i and 31 _i at point D (Fig. 6) will also begin to change.

It is important to note that at the moment of formation of additional signals for active noise reduction, the state of the phase shifter 29 _{i is} also random and therefore it is necessary to find the optimal phase value of the phase shifter 29 _i .

The phase optimization in the phase shifter 29 _i is as follows.

 We will use the functional block diagram of blocks 29 and 30, which is shown in FIG. 7.

The controlled phase shifter comprises: a multi-position controlled switch 32 and first additional delay lines 33 ₁ , 33 ₂ , ..., 33 _S. The outputs of the first additional delay lines 33 are connected to the inputs of the controlled multi-position switch 32. The delay value can be selected with some step, for example, constant. The magnitude of the delay variation step determines the accuracy of the optimization of the phase parameter of the signals.

 The number of first additional delay lines and, accordingly, the value of the delay step is selected based on the psychophysiological abilities of a person to record phase distortions of sounds of certain frequencies.

For example, the phase sampling step can be selected τ ₃₃ ≃ 0.01 T _in (≈ 3.6 ^o ), where T _in is the highest period in the analysis band. With a hardware embodiment of the implementation of the blocks 29, it is possible to reduce the number of delay lines 33 at frequencies of low human susceptibility to phase distortions, as well as the implementation of the phase shifter 29 is possible in the variant with the parametrically controlled R, C or L elemental analog phase shifter 29.

 The phase shifter control device 30 comprises: a second additional delay line 34, additional detectors 35, additional low-pass filters 36, and an additional comparison circuit 37.

 The input of the device 30 is the common input of two branches, in one of which the first additional detector 35 and the first additional low-pass filter 36 are connected in series, in the other - the second additional delay line 34, the second additional detector 35 and the second additional low-pass filter 36. Outputs additional low-pass filters are associated with the inputs of additional comparison circuits 37.

 The phase shifter operates (Fig. 7) as follows.

A multi-position controlled switch 32 sequentially switches contacts in direct relation to delay line numbers or in reverse order. It is advisable to select the switching period longer than the maximum possible signal propagation time through the entire sound reproduction path in order to exclude the occurrence of spurious excitation of the system with active noise reduction, for example, select the value of parameter T _{32 of the} order of 0.02 s.

 The switch 32 is controlled by the signal from the output of the additional comparison circuit 37 according to the rule: when the next signal arrives, the order of switching the delay lines 33 is reversed.

Thus, in the proposed embodiment, the controlled phase shifter 29, the delay value of the signal at its output with respect to the input signal functionally depends on the listening signals shifted in time relative to each other by the delay value τ ₃₄ .
Indeed, in the device 30, the listening signal and the listening signal delayed by τ _{34 are} detected and integrated in different circuits, forming input signals for an additional comparison circuit 37. The output signal of the comparison circuit 37 is generated if the difference of the input signals, i.e. the energies of the listening signals and the delayed listening signal becomes greater than zero. This means that the delay introduced at the next phase optimization step of the additional listening signal has already become non-optimal. In other words, while the difference between the energies of the listening signals and the delayed listening signal is less than zero, a sequential (iterative) approach to the optimal phase (time parameter) of the additional signal for active noise reduction or, alternatively, approaching the extremum point of the listening signal (noise) function. If the indicated difference becomes greater than zero, then this means that in the process of changing the phase of the additional signal for active noise reduction, the extremum point of the function of the energy of the listening signal has already been passed and it is necessary to start again the search for the optimal phase, reversing the process of experimental selection of the phase of the additional signal for active noise reduction .

 In accordance with this algorithm, after the search for the optimal phase, the process of its automatic retention (capture) takes place.

The value of the delay time in the second additional delay line 34, as well as the time constant of the low-pass filter 36, it is advisable to choose less than the period T ₃₂ switching contacts of the multi-position controlled switch 32, and also based on the requirements for the noise reduction level and the maximum achievable resolution of the subtractive node of the comparison circuit 37 to distinguish input levels signals.

 When implementing a controlled phase shifter 29, it is possible to use any known numerical methods for searching for the extremum of a function, for example, using a variable step for changing the phase parameter.

 It is clear that the above method of active noise reduction of sound signals is applicable to lower the level of signals of any physical nature and can be formulated as follows.

 The method of optimal, spatial, active lowering of signals of any physical nature consists in receiving and converting these signals into an electric signal, transmitting the received electric signal to the place of its processing, processing the electric signal and generating an additional electric signal for active lowering of the signal level, its amplification, conversion into a signal of the same physical nature and radiation to a point in the space of signal reception.

 Returning again to the description of the operation of block 5 (10), we note that as a result of two-parameter, multi-band optimization of the phase and energy parameters of the signals, the sound reproduction system suppresses stationary interference with a given accuracy.

 If the interference is unsteady, with varying parameters, then the operation of block 5 (10) is similar to the already described case with the only difference that in block 5 (10) there is a continuous process of tracking the interference parameters.

 In embodiments of blocks 5 (10), it is advisable to control the operation of blocks 28 and 31 not with the help of nodes 25, 26, 27 common to them, but by means of similar additional nodes, but with different values of the automatic control time constant.

 In highly intelligent algorithms, these parameters can be automatically varied, analyzed and set in an optimal way.

 The optimization criterion is the minimum standard deviation for the analysis time.

 A situation in which there is no listening signal in the presence of a signal from source 1 can occur in the event of a system malfunction or when the listener, together with the sounding device 4 (9), is removed from the listening room.

 Under these conditions, the operation of blocks 5 and 10 is reduced to an automatic increase in the signal level of source 1 at the output of block 5 (10) to a maximum or, if there are noises, to an intermediate value. The signal at point D also rises to its maximum value. The rise rate of these signals is determined by the time constant of the low-pass filter 26. Block 31 is closed and the formation of signals for active noise reduction does not occur.

 Consideration of this situation allows us to formulate additional requirements that should be taken into account when synthesizing the signal processing algorithm.

If with an increase in the transmission coefficient of the source signal 1 u _ist (t) there is no proportional increase in the listening signal u '(t), then this means that the optimization of sound reproduction becomes fundamentally impossible, because the principle of cause and effect is violated, and during the operation of the system it is necessary to enter correction, for example, turn it off.

 The case considered can arise in practice, for example, when a wire breakage from the VLF 2 (7) to the loudspeaker 3 (8) occurs in the automobile optimal sound reproducing system or a malfunction occurs in other electrical circuits.

 At the moment of interruption of the transmission of the source signal 1, in block 5 (10) there will be a sharp increase in its transmission coefficient and block 5 (10) will seek to compensate for the decrease in the signal energy at the listening point, and when restoring the broken connection in the electrical part of the sound reproduction path, for example, due to vibration and shaking the car body, the maximum level signal will go to the listening point. A sharp, spasmodic change in the volume level will have a negative psychophysiological effect on the driver (he may be frightened), and also reduce the level of safety of the car.

 When operating traditional sound reproducing systems, a similar situation may also occur. An intermittent signal, even if not at its maximum, but at a nominal volume level, will also annoy and distract the driver from driving.

 Therefore, in order to improve the safety of vehicles, it is advisable to provide in the algorithms for optimal signal processing the process of self-testing and self-monitoring of its operation.

 The considered examples of the operation of unit 5 (10) show that sound-reproducing systems of this type can be installed in car interiors and other vehicles (for example, tractors) as effective noise-reducing systems that increase the ergonomic performance of these machines. In order for the driver to be able to hear external sound signals, a device in the form of an external microphone with a corresponding radiation pattern and a frequency-correcting microphone amplifier can be used as a signal source 1.

 Consider the operation of block 5 (10) in the conditions of sound reproduction of the signals of source 1 under the influence of noise and noise.

When the signals of the source 1 and the listening signal containing interference appear simultaneously, the signals at the outputs of the bandpass filters 14 _i , into the bands of which the components of these signals fall, are fed to blocks 15 _i and 16 _i .

At the output of the side 31 _i and at the corresponding input of the adder 17, the level of the additional signal is determined by the signal level at point D, i.e. signal to noise ratio (interference).

During the time Δt ′ in block 18 _i , correlation signal processing is performed.

Through the output signal of the analysis circuit 23, which is formed relative to the edges of the pulses Δt ′ and Δt ″, the switching circuit 24 _i connects to the output of the block 18 _i delayed by the delay line 20 $\genfrac{}{}{0ex}{}{\text{i}}{\text{η}}$ source signal 1, where in the general case 1 <η <z (Fig. 6).

Indicated nodes (20 $\genfrac{}{}{0ex}{}{\text{i}}{\text{η}}$ 24 _i ) form the signal line described above and a controllable switch of one of the branches of the “white box” (block 5 (10)).

где

среднее время прохождения сигнала через тракт звуковоспроизведения (определяется, в основном, временем прохождения прямой звуковой волны от громкоговорителя 3 (8) до точки прослушивания).It is advisable that the initial value of the delay signal at the output of the delay line 20 $\genfrac{}{}{0ex}{}{\text{i}}{\text{η}}$ and, therefore, at the output of block 18 _i satisfied the condition (22)

Where

average time of the signal passing through the sound reproduction path (determined mainly by the time of the passage of the direct sound wave from the loudspeaker 3 (8) to the listening point).

сигнала следует устанавливать ближайшего значения к уже измеренной на предыдущем этапе ее вычисления задержке. Но поскольку для различных полос анализа указанные задержки вероятней всего будут различными, а для получения истинных значений задержек необходимо, на время Δt′, во всех частотных каналах обработки сигналов установить одну и ту же величину задержки компонентов сигнала источника 1, то после окончания отрезка времени Δt′ следует организовать процесс вычисления оптимального значения задержки сигнала для следующего, очередного периода Δt′ зондирования тракта звуковоспроизведения и нахождения оптимального значения задержки компонентов сигнала.Condition (22) is obtained from the following considerations. Under conditions of random removal of the listener from the speakers 3 (8), it is advisable to set the initial value of the signal delay (at the sensing stage) as the initial conditions of the system operation, corresponding to the most probable distance of the listener from the speaker 3 (8), for example, corresponding to the middle of the maximum calculated distance with in order to minimize the temporal distortion (in the form of a shift) of the signal, which inevitably then occurs when you first switch to the optimal sound playback mode. Subsequently, the specified delay

signal should be set to the closest value to the delay already measured at the previous stage of its calculation. But since these delays are most likely to be different for different analysis bands, and to obtain the true values of the delays, it is necessary, for the time Δt ′, to set the same delay value of the components of the signal of source 1 in all frequency channels of signal processing, after the end of the time interval Δt ′ It is necessary to organize the process of calculating the optimal signal delay value for the next, next period Δt ′ of sounding the sound path and finding the optimal component delay value signal.

 The algorithm for such a calculation of the delays of the signal components can be, for example, an algorithm for finding the arithmetic mean over all delay values or in the form of more complex algorithms - calculating the average delay, where each delay is assigned a weight coefficient in accordance with, for example, a graph of an equal loudness curve, so that when switching the next combination of frequency-corrected delays to minimize the ability of a person to capture distortion of a signal with a relative shift of it components. For example, if in a three-way analysis, the actual delay at a frequency of 100 Hz was 0.015 s, at a frequency of 1000 Hz - 0.0155 s, at a frequency of 10000 Hz - 0.016 s, and the weighting factors for these frequencies are, for example, 1, 5 and 2, then the optimal delay for the time interval Δt ″ will be, respectively: 0.016 s; 0.0155 s and 0.015 s, and for the subsequent interval Δt ′, the optimal delay for all frequencies is 0.0155625 s.

From the output of the switching circuit 24 _{i, the} optimally delayed signal of the source 1 through the first controlled amplifier 28 _i and the adder 17 is fed to the output of block 5 (10) and then, passing through other parts of the sound reproduction path, it reaches the listening point. By means of the probing device 4 (9) and communication line 6 (11), the signal is supplied to the first input of block 5 (10) and from the output of the corresponding band-pass filter 14 _i again to the first input of block 15 _i .

If at this point the stage of correlation signal processing has not yet ended, then, in accordance with the scheme (Fig. 6), the signals of source 1 delayed by the delay lines are multiplied by the listening signals and then filtered by the low-pass filter 22 _i .

At the inputs of the analysis circuit 23, signals are formed whose levels determine the degree of correlation of the signals. In the analysis circuit 23, the maximum signal is found and the corresponding signal delay in the path is calculated, the optimal delay is calculated and the output signal of the circuit 23 is generated, which is fed to the input of the switching unit 24 _i relative to the trailing edge of the pulse Δt ′. From this point in time, the switching units 24 connect the calculated delays to the outputs of the units 18 and upon completion of the Δt ″ segment, a delay that is uniform for all units 18.

During the optimization of time parameters, the operation of the energy parameter optimization block 19 _i can occur according to the algorithm already described, regardless of the state of the block 18 _i . The influence of block 19 _i on the signal parameters at the listening point, in this analysis band, is reduced to amplitude modulation of the corresponding spectral components of the listening signal. This additional modulation of the signal does not affect the results of correlation processing because it is a slow process. Due to the interruption in time of the signal transmission process, its multiband analysis and conversion, it was possible to break the relationship between the frequency response, phase response and the transition characteristics of block 5 (10). In other words, if in traditional signal processing systems (in analog or digital) you start to change, for example, the frequency response, then other indicators, such as the phase response, will automatically start to change, because the time for single-channel devices (four-terminal) is a one-dimensional, monotonically increasing parameter (continuous in the case of analog devices and discrete for digital), and due to the introduction of multiband analysis, signal conversion and time interruption in accordance with (14), i.e., due to the introduction of virtually the second time axis, managed to break the relationship and interdependence of temporary indicators of energy. The above is valid only for signal components falling into different analysis bands. For signal components within the indicated analysis bands, the relationship between the frequency response, phase response, and transient response remains unchanged.

 When narrowing the passband of the band-pass filters 14, the delay time of the signal when passing through the filters 14 increases. Accordingly, the inertia of the operation of unit 5 (10) and the entire sound reproducing system increases.

 Thus, the bandwidth of the bandpass filters 14 should be selected based on the specific conditions of sound reproduction and the dynamics of changes in signal parameters. In the synthesis of highly intelligent algorithms, it is advisable to provide for the possibility of optimizing the indicated parameters of block 5 (10). In the software version of the signal processing process control, this optimization can be carried out using the corresponding subprogram, in which, for example, the passbands (and, accordingly, their number) of the filters 14 are periodically changed, the SDE of the listening signal from the signal of source 1 is calculated, and the optimal values of the bands are found that provide the smallest the value of standard deviation for all experimental data. Subsequently, depending on the chosen algorithm, the number and width of the pass-band of the band-pass filters 14 can be automatically set for some time of quasi-optimal operation of the system, for example, for 120 s. After the specified time, the optimization process of the parameters of block 5 (10) is resumed, and the results, for example, are accumulated in the machine’s memory in the form of statistical data on the corresponding parameter. This data can be used in the extrapolation algorithms of the signal parameters and the parameters of the signal processing unit 5 (10).

Depending on the control voltage of the signal at point D (Fig. 6), which contains information on the signal-to-noise ratio (noise) averaged over a fairly short period of time, the signal is reproduced optimally in the time interval Δt ″. For example, if the noise level is much lower than the signal level of source 1, then in the first controlled amplifier 28 _i the gain varies slightly: a) if the listening point is weak and the signal is in phase, it decreases slightly; b) if at the listening point there is weak interference and the signal is in antiphase, then it slightly increases. With little interference, the second controllable amplifier 31 _{i is} closed and there is no active reduction in noise and interference.

 If the interference level is high and, for example, equal to the source signal level, then: a) if the source signal and the interference signal are in phase, then the signal of source 1 decreases to zero, and at this time the listener perceives (in energy) the corresponding acoustic interference components, how would he perceive the loudspeaker signal or how in the past the signals of the original sources of sounds sounded; b) if the signal of source 1 and the interference signal are in antiphase, then the signal level doubles.

If the noise level is much higher than the signal level, then the signal level at point D begins to decrease and the first controlled amplifier 28 _i closes. At that time, when the voltage at point D becomes negative, the formation of the output of the second controlled amplifier 31 _i and, therefore, the output of the block 16 _{i of} an additional electrical signal for active noise reduction, begins as described above. If the components of the signal of source 1 fall into the analysis band, then in this band of the frequency range the interference suppression is partially carried out.

 Thus, when processing signals in a sound reproduction system, the weighted (for the entire sound reproduction band) signal-to-noise (noise) ratio increases, and the standard deviation of the signal at the listening point (s) decreases.

 Summarizing the examples of interaction between the nodes of the signal processing unit 5 (10) for the case of a dynamic change in the parameters of the signal of the source 1 and the listening signal, we can figuratively compare the operation of the optimal sound reproduction system with the functioning of a rational or intelligent organism according to universal logical rules.

 An example of solving a particular technical problem of improving the quality of transmission of sound information (sound reproduction), as a result of which the optimal methods and systems for their implementation (machines) were found, in accordance with the fundamental cybernetic principle of similarity, should also be optimal solutions (a set of essential features) for other particular cybernetic tasks, for example, solving the problem of improving the quality of visual reproduction of signals, touch signals and signals received by a person in other channels of information perception of reality, a joint decision of all the above tasks for the synthesis of a time machine, and can also be used to improve governance in the public system: the company, industry, country, world, etc.

 Thus, the essential features of any optimal information system of the corresponding type can be formulated as the following method of transmitting information messages.

 A method for optimal transmission of messages of any physical nature in a channel with random parameters, which consists in converting messages to electrical signals of a message source, coordinated filtering of electric signals of a source, their amplification, converting electrical signals into signals of the same physical nature, transmitting these signals through a channel with random parameters to the point of reception, reception and conversion of the signal into a received electrical signal, transmitting the received electrical signal to the place of its processing ki, processing electrical signals of the message source and the received electrical signal, forming control signals by coordinated filtering, characterized in that they generate additional electrical signals for active noise reduction, which amplify, convert into signals of the same physical nature and emit into a channel with random parameters to the receiving point messages.

Industrial applicability
The proposed methods and systems for their implementation can be used in various fields of technology related to information processing, for example, in radio engineering for high-quality sound reproduction of signals, in robotics, in the automotive industry to improve control systems of assemblies and assemblies (for example, to improve active suspension, security systems and ergonomics of the car, "autopilot" systems, etc.).

 The methods can be used to synthesize more advanced weapons systems (targeting systems, tracking systems, etc.).

 The methods can be used in computer technology to create training or entertainment, virtual systems and believable computer games.

Claims (19)

 1. The method of optimal sound reproduction, which consists in coordinated filtering of the electric signal of the source, amplification, conversion of the amplified signal of the source into an audio signal, its emission, reception and transformation at the listening point of the total sound signal - direct and reflected sound waves of the emitted source signal, as well as sound waves interference and noise into an electrical listening signal, transmitting the received electrical signal to the place of its processing, processing the electrical source signals and listening, forming signals coordinated by filtering control, characterized in that they generate additional electrical signals for active noise reduction, which amplify, convert into sound signals and emit to the point of reception of the total sound signal.  2. An optimal sound reproduction system, comprising a signal source and a sound reproduction channel, made in the form of a low-frequency amplifier and a loudspeaker connected in series, a probing device, a signal processing unit, configured to coordinate filtering the source signal, communication line, while the output of the probing device through the line connection is connected to the first input of the signal processing unit, the output of the signal source is connected to the second input of the signal processing unit, and the output of the unit the signal processing is connected to the input of a low-frequency amplifier, characterized in that the signal processing unit is configured to generate additional signals at its output for active noise reduction at the installation point of the probing device.  3. The optimal sound reproduction system according to claim 2, characterized in that the signal source is configured with at least one additional output for multichannel audio reproduction, and at least one additional sound reproduction channel configured as an additional low amplifier is additionally introduced according to the number of additional outputs of the signal source frequency and an additional loudspeaker connected in series, as well as an additional sounding device, an additional processing unit signal ki made with the possibility of coordinated filtering of the source signal and generating additional signals for active noise reduction, an additional communication line, the output of an additional probing device through an additional communication line is connected to the first input of the additional signal processing unit, the additional output of the signal source is connected to the second input of the additional processing unit signals, the output of the additional signal processing unit is connected to the input of the additional amplifier low frequency.  4. The optimal sound reproduction system according to claim 3, characterized in that the signal source is configured to switch output signals to change the order of their connection to the sound reproduction channels.  5. The optimal sound reproduction system according to claim 2, characterized in that an additional signal generator is connected, connected to the third input of the signal processing unit.  6. The optimal sound reproduction system according to claim 3, characterized in that an additional signal generator is connected, connected to the third input of the signal processing unit and to the third input of the additional signal processing unit.  7. The optimal sound reproduction system according to claim 3, characterized in that the signal processing unit, a communication line, a sounding device, an additional signal processing unit, an additional communication line, an additional sounding device are functionally combined in an optimal signal processing unit for a block system configuration.  8. The optimal sound reproduction system according to claims 3, 6 or 7, characterized in that the signal processing unit and the additional signal processing unit are made in the form of a multi-channel analog-to-digital converter, a computer with software and a multi-channel digital-to-analog converter connected in series, while the number the inputs of the analog-to-digital converter are twice as many, and the number of channels of the digital-to-analog converter is equal to the number of sound channels.  9. The optimal sound reproduction system according to claim 2 or 3, characterized in that the signal source is configured to control the amplitude-frequency characteristics of the signals at its outputs.  10. The optimal sound reproduction system according to claim 2 or 3, characterized in that the signal source is made with the possibility of noise reduction.  11. The optimal sound reproduction system according to claim 2 or 3, characterized in that the signal source is configured to automatically adjust the levels of its output signals.  12. The optimal sound reproduction system according to claim 2 or 3, characterized in that the signal source is configured to automatically control levels in non-automatic regulation of the amplitude-frequency characteristics of the signals at its outputs.  13. The optimal sound reproduction system according to claim 2 or 3, characterized in that the signal source is configured to control signal levels at its outputs to control the volume level at the listening point.  14. The optimal sound reproduction system according to claim 3, characterized in that the sounding device and the additional sounding device or communication line and the additional communication line are configured to control the transmission coefficient to control the volume level at the listening point.  15. The system of optimal sound reproduction according to item 13 or 14, characterized in that the regulation of the gear ratios is made finely-compensated.  16. The optimal sound reproduction system according to claim 3 or 8, characterized in that the signal processing unit and the additional signal processing unit are configured to automatically control the amplitude-frequency or phase-frequency characteristics from the second inputs to the outputs of the blocks to optimize energy or time (phase) parameters signals.  17. The optimal sound reproduction system according to claim 3 or 8, characterized in that the signal processing unit and the additional signal processing unit are configured to automatically control the amplitude-frequency and phase-frequency characteristics from the second inputs to the outputs of the blocks for full-parameter optimization of the signals.  18. The method of optimal, spatial, active lowering of signals of any physical nature consists in receiving and converting these signals into an electric signal, transmitting the received electric signal to the place of its processing, processing the electric signal and generating an additional electric signal to actively lower the signal level, amplify it , conversion into a signal of the same physical nature and radiation to a point in the space of signal reception.  19. A method for optimal transmission of messages of any physical nature in a channel with random parameters, which consists in converting messages into electrical signals of a message source, coordinated filtering of electrical signals of a source, their amplification, converting electrical signals into signals of the same physical nature, transmitting these signals through a channel with random parameters to the point of receiving messages, receiving and converting the signal into a received electric signal, transmitting the received electric signal to places processing it, processing the electrical signals of the message source and the received electrical signal, generating control signals by coordinated filtering, characterized in that they generate additional electrical signals for active noise reduction, which amplify, convert into signals of the same physical nature and emit into a channel with random parameters to the point receiving messages.

Images