EP1647972B1 - Intelligibility enhancement of audio signals containing speech - Google Patents

Abstract

The arrangement has a speech detector (200) detecting speech in an audio signal and providing a control signal (226) to control a speech processing device. The device processes the audio signal to determine whether the audio signal includes components which indicate speech. The detector compares a range of detected speech components to a threshold value, and outputs the control signal based on the comparison result. Independent claims are also included for the following: (A) a method for processing audio signals containing speech (B) an audio processing system comprising a speech detector.

Description

The invention relates to a circuit arrangement for improving the intelligibility of speech-containing audio signals with the preamble features of claim 1 and to a method for processing speech-containing audio signals.

Out DE 101 24 699 C1 a circuit arrangement for improving the intelligibility of speech-containing audio signals is known in which frequency and / or amplitude components of the audio signal are changed according to predetermined parameters. In this case, the audio signal is amplified in a processing path by a predetermined factor and passed into a high pass, a corner frequency of the high pass is adjustable so that the amplitude of the audio signal after the processing path is equal to or proportional to the amplitude of the audio signal before the processing path. With this circuit arrangement, the fundamental wave of the speech signal, which contributes relatively little to the intelligibility of the speech components contained, but has the largest energy, to be attenuated, the remaining signal spectrum of the audio signal is raised accordingly. In addition, the amplitude of the vowels having a large amplitude at low frequency can be lowered to a vowel in the transition region from a consonant having a small amplitude at high frequency to reduce a so-called "backward masking". For this, the entire signal is increased by the factor. Ultimately, high-frequency components are raised and the low-frequency fundamental is lowered to the same extent so that the amplitude or energy of the audio signal remains unchanged.

US 5,553,151 describes a "forward masking". Weak consonants are covered in time by previous strong vowels. Proposed is a relatively fast compressor with an "attack time" of about 10 msec and a "release time" of about 75 to 150 msec.

Out US 5,479,560 It is known to divide an audio signal into several frequency bands and relatively strong amplify those frequency bands with high energy and lower the other. This is suggested because language consists of a sequence of phonemes. Phones are made up of a variety of frequencies. These are especially enhanced in the area of the resonance frequencies of the mouth and throat area. A frequency band having such a spectral peak is called a formant. Formants are particularly important for recognizing phonemes and thus speech. One approach to improving speech intelligibility is to amplify the formants of the frequency spectrum of an audio signal and mitigate the errors between them. For an adult male, the fundamental frequency of the speech is about 60 to 250 Hz. The first four associated formants are 500 Hz, 1500 Hz, 2500 Hz and 3500 Hz.

As a further prior art, the US 2003/0044032 , the WO 2004/071130 A1 and the US 2003/0055636 A1 called.

Such circuitry and methods make speech contained in an audio signal more understandable to other components included in the audio signal. At the same time, however, also non-speech-containing signal components are changed or falsified. It is also disadvantageous in the methods or circuit arrangements that they each continuously improve or process rigidly predefined speech components, frequency components or the like. As a result, non-speech signal portions also become changed or falsified at times when the audio signal contains no speech or speech components.

The object of the invention is to improve a circuit arrangement or a method for processing speech-containing audio signals.

This object is achieved by a circuit arrangement for improving the intelligibility of possibly audio-containing audio signals having the features of patent claim 1 or by a method for processing optionally audio-containing audio signals having the features of patent claim 6.

A circuit arrangement is accordingly advantageous for improving the intelligibility of multicomponent audio signals possibly containing speech with an input for inputting such an audio signal. Advantageously, the circuit arrangement is implemented by a speech detector for detecting speech in the input audio signal and for providing a control signal for controlling a speech processing device and / or a speech processing method for processing the audio signal. The speech detector has a correlator for performing cross or autocorrelation of components of the audio signal. Signal components of the audio signal are separated during signal processing by a matrix.

Advantageously, a method is provided for processing multicomponent audio signals containing speech, in which speech or speech components contained in an audio signal are detected and, depending on the result of the detection, a control signal for a speech processing device and / or a speech processing method for speech enhancement is generated and provided ,

The circuit arrangement or the method are thus to be regarded as a precursor to an actual signal processing for improving the intelligibility of speech-containing audio signals. Accordingly, the received audio signal is first examined to see if any speech is included in the audio signal. Depending on the result of the speech detection, a control signal is then output, which is used by an actual speech processing device or an actual speech processing method as a control signal. This makes it possible that in speech processing to improve the speech components in the audio signal relative to other signal components in the audio signal, processing or alteration of the audio signal is performed only if speech or speech components are actually included.

Accordingly, a control signal is provided or output by the circuit arrangement or by the method, which for the actual language improvement z. B. is used as a trigger signal. Thereby, the speech enhancement by means of detection or analysis of a previous audio signal or the like, possibly a time-delayed audio signal can be performed.

The circuit arrangement which generates and provides the control signal can be provided as an independent structural component, but can also be part of a single structural component with the speech processing device or speech enhancement device. In particular, the speech detection circuitry and the speech processing means for enhancing the speech components of the audio signal may be part of an integrated circuit arrangement. Accordingly, the method for detecting speech and the speech processing method for improving speech can also be used Speech components in the audio signal are performed separately from each other. However, a common method which is carried out by means of technical components of a circuit arrangement or by means of a correspondingly running algorithm in a calculation device is particularly preferred.

Advantageous embodiments are the subject of dependent claims.

In particular, a circuit arrangement is preferred in which the speech detector is designed and / or controlled to detect speech components in the audio signal.

In particular, a circuit arrangement is preferred in which the speech detector has a threshold value determination device for comparing a range of detected speech components with a threshold value and for outputting the control signal as a function of the comparison result.

In particular, a circuit arrangement is preferred in which the speech detector has a control input for inputting at least one parameter for variably controlling the detection with regard to a scope of the speech components to be detected and / or with regard to a frequency range of the speech components to be detected.

In particular, a circuit arrangement in which the speech detector for processing a multi-component audio signal, in particular stereo audio signal or multi-channel audio signal, is formed with a plurality of audio signal components and as a processing means for detecting the language based on a comparison or a processing of the components with each other or controlled.

In particular, a circuit arrangement is preferred in which the speech detector has a direction determination device for determining a direction of common signal components of the various components.

In particular, a circuit arrangement is preferred in which the control signal for activating or deactivating the speech enhancement device and / or the speech enhancement method is designed and / or controlled as a function of the speech content of the audio signal.

In particular, a method is preferred in which the control signal is generated as a function of the extent of detected speech components.

In particular, a method is preferred in which the extent of the detected speech components is compared with a threshold value at which a cross or autocorrelation of the audio signal or of components of the audio signal is performed. Signal components of the audio signal are separated during signal processing by a matrix.

In particular, a method is preferred in which the detection is carried out in an adjustable manner with respect to a scope of the speech components to be detected and / or with regard to a frequency range of the speech components to be detected by means of variable parameters.

In particular, a method is preferred in which the audio signal components of a multicomponent audio signal having a plurality of audio signal components are compared with one another or processed with one another to detect the speech. Components are to be understood as meaning signal components from different distances and directions and / or signals of different channels.

In particular, a method in which the audio signal components are compared or processed with regard to common speech components in the various of the audio signal components, in particular for comparing a direction of the common signal components, is preferably compared or processed. Based on different arrival times on, for example, the right and the left channel of a stereo signal and on the basis of specific attenuation of specific frequencies, the distance and direction of the speech component can be determined. As a result, an application of the speech enhancement is applicable in particular only to speech components which are recognized as originating from a person who is close to the microphone. Signal portions or speech portions of more distant persons can thereby be ignored, so that a speech enhancement is activated only when a related person actually speaks.

In particular, a method is preferred in which the control signal is provided for activating or deactivating the speech enhancement device and / or the speech enhancement method.

In particular, a circuit arrangement and / or a method is preferred, wherein a frequency response is determined by means of an FIR or an IIR filter (FIR: finite impulse response, IIR: infinite impulse response).

In particular, a circuit arrangement and / or a method is preferred, wherein matrix coefficients for a matrix are determined via a function dependent on the language component. The function is linear and continuous. Alternatively or additionally, the function has a hysteresis.

The signal components with speech components of the audio signal can be analyzed and detected with regard to various criteria. In addition to a minimum duration, for example, By means of which language is detected as the voice part, it is also possible, for example as a signal component, to focus on the frequency of detectable speech and / or the direction of a speech source of detected speech. The terms signal components and speech components are therefore to be interpreted as general and not restrictive.

Fig. 1

schematisch Verfahrensschritte bzw. Komponenten eines Verfahrens bzw. einer Schaltungsanordnung zum Verarbeiten eines Audiosignals zur Detektion von darin enthaltener Sprache;

Fig. 2

eine beispielhafte Schaltungsanordnung gemäß einer ersten Ausführungsform zur Anwendung einer Korrelation auf Sprachanteile verschiedener Signalkomponenten;

Fig. 3

eine beispielhafte Schaltungsanordnung zur Darstellung einer Matrixberechnung vor einer Durchführung einer Sprachverbesserung des Audiosignals; und

Fig. 4

ein Diagramm zur Veranschaulichung von Kriterien zur Festlegung eines Schwellenwerts.

The invention will be explained in more detail with reference to the drawing. Show it:

Fig. 1

schematically method steps or components of a method or a circuit arrangement for processing an audio signal for the detection of language contained therein;

Fig. 2

an exemplary circuit arrangement according to a first embodiment for applying a correlation to speech components of various signal components;

Fig. 3

an exemplary circuit arrangement for displaying a matrix calculation before performing a speech enhancement of the audio signal; and

Fig. 4

a diagram illustrating criteria for establishing a threshold.

Fig. 1 schematically shows, by way of example, the sequence of a method for detecting speech and / or speech components px in an audio signal i for the optional subsequent or parallel speech enhancement of the speech or of the speech components px, if such are detected, in the audio signal i. An audio signal i is input via an input I of a circuit arrangement for improving the intelligibility of audio signals i.p. which may contain speech or voice components px entered. Depending on the application, the audio signal i may be a single-channel mono signal. However, preference is given to multicomponent audio signals i of a stereo audio signal source or the like, ie a stereo audio signal, a 3D stereo audio signal with additional central component or a surround audio signal with currently usually five components for audio signal components from the right, left, center and from Z. B. two remote sources right and left.

The audio signal i is supplied to a first structural or logical component, which forms a speech detector SD. In the speech detector SD, it is examined whether speech or a speech component px is contained in the audio signal i. According to preferred embodiments, it is checked whether detected speech or speech components px are larger than a correspondingly predetermined threshold value v. Optionally, detection parameters, in particular the threshold value v, can be adapted as needed. In this regard, the illustrated arrangement has an input IV for inputting the threshold v.

If the detection indicates that a sufficient speech component px is contained in the audio signal i, a control signal is set to the value 0, for example. Otherwise, the control signal is set to, for example, the value 1. The control signal s is output from the voice detector SD for further use by a voice processing means.

If the control signal s indicates a speech component px, ie if s = 0 in the present case, the language or speech portions px-improving speech processing is activated. The audio signal i currently input into the speech processing is improved in accordance with known methods or with an otherwise known circuit arrangement. At an output O becomes accordingly, an audio signal o improved with respect to the speech parts is output.

If the speech step px is not detected in the detecting step, that is, if s = 1, the audio signal i input to the speech processing SV is left, that is, output unchanged as the audio signal o.

If there is a time delay of the control signal s applied to the voice processing relative to the currently applied audio signal i due to the speech detection, a delay of the audio signal i input into the circuit arrangement or the method can optionally be performed according to the time delay in the speech detection.

Thus, a circuit arrangement or a method or algorithm is made possible which can only be used for voice enhancement on parts of the audio signal which actually contain speech or which actually contain a specific speech component in the audio signal. Speech detection thus detects speech or separates it from the rest of the signal.

In reality, speech will not be separated mathematically from other signal components of an audio signal. The aim is therefore to provide the best possible estimate. If algorithms or circuit arrangements of embodiments listed below can be mislead by corresponding other signal components, an advantageous improvement of an output audio signal is still achieved according to the first experiments. It is advantageous to ensure that the audio signal i is not falsified too much even in the case of a misdetection in the speech detector SD.

Fig. 2 shows a first embodiment of a speech detector SD. The input consists of two individual inputs for each one audio signal component or an audio signal channel L ', R' of a stereo audio signal. The two audio signal components R ', L' are each fed to a bandpass filter BP for limiting the band. The outputs of the two bandpass filters BP are supplied to a correlator CR for performing a cross-correlation. Each of the two signals output by the bandpass filters BP is respectively multiplied by itself in a multiplier M, ie squared, and then supplied to an adder A. After the addition, a multiplication by the factor 0.5 is optionally carried out in a further multiplier M * in order to reduce the amplitude. The output signal i of the optionally multiplied addition values is fed to a first or second low-pass filter TP.

In addition, each of the output signals of the two bandpass filters BP is fed to an actual circuit for performing the correlation using in particular a further multiplier M. The correlation signal L, * R 'output therefrom is fed to a second low-pass filter TP.

The output signals b, a of the first low-pass filter TP and of the second low-pass filter TP are supplied to a division element DIV for dividing the output signal b of the first low-pass filter TP from the output signal a of the second low-pass filter TP. The division result of the division element DIV is provided as a control signal or as a preliminary stage D1 for the control signal s.

With such a circuit arrangement or a corresponding processing method, a cross-correlation is performed. A conventional stereo audio signal L ', R' is usually the audio signal i from several audio signal components R, L, C, S together. In the case of a multi-channel audio signal, these components can also be provided separately.

$$\mathrm{b}\mathrm{:}\mathrm{Rʹ}\mathrm{=}\mathrm{R}\mathrm{+}\mathrm{C}-\mathrm{S}\mathrm{,}$$

wobei L für eine linke Signalkomponente steht, C für eine zentral von vorne kommende Signalkomponente steht, S für eine Surround-Signalkomponente, d.h, ein rückwärtiges Signalund R für eine rechte Signalkomponente steht.In the case of a stereo audio signal L ', R', the two audio signal channels L ', R' are described by $$\mathrm{a}\mathrm{:}\mathrm{L\text{'}}\mathrm{=}\mathrm{L}\mathrm{+}\mathrm{C}\mathrm{+}\mathrm{S\; resp}\mathrm{,}$$

$$\mathrm{b}\mathrm{:}\mathrm{R\; \text{'}}\mathrm{=}\mathrm{R}\mathrm{+}\mathrm{C}-\mathrm{S}\mathrm{.}$$

where L stands for a left signal component, C stands for a signal component coming from the center, S stands for a surround signal component, ie, a back signal and R stands for a right signal component.

mit RMS als der zeitlich gemittelten Amplitude.Speech or speech components px are located mainly on the central channel or in the central component C. This fact can be used to detect the proportion of speech or speech components px to the remaining signal content of the audio signal i. Can be determined the language included or the included speech portion px in relation to the remaining signal components of the audio signal i according to $$\mathrm{px}\mathrm{=}\mathrm{2}\mathrm{*}\mathrm{RMS}\left(\mathrm{C}\right)\mathrm{/}\left(\left(\mathrm{RMS}\mathrm{/}\mathrm{L\text{'}}\right)\mathrm{/}\mathrm{(}\mathrm{RMS}\mathrm{/}\mathrm{R\; \text{'}}\mathrm{)}\right)$$

with RMS as the time averaged amplitude.

By cross-correlation, the proportion of the central component C can be determined by $$\mathrm{L\text{'}}\mathrm{*}\mathrm{R\; \text{'}}\mathrm{=}\mathrm{2}\mathrm{*}\mathrm{L}\mathrm{*}\mathrm{R}\mathrm{+}\mathrm{L}\mathrm{*}\mathrm{C}\mathrm{+}\mathrm{R}\mathrm{*}\mathrm{C}\mathrm{-}\mathrm{L}\mathrm{*}\mathrm{S}\mathrm{+}\mathrm{R}\mathrm{*}\mathrm{S}\mathrm{+}\mathrm{C}\mathrm{*}\mathrm{C}\mathrm{-}\mathrm{S}\mathrm{*}\mathrm{S}\mathrm{,}$$

On average, for DC-free signals, ie for signal components without a DC voltage component, all non-correlated products become 0. Thus, the criterion for the signal D1 output by the speech detector SD is: $$\begin{array}{ll}\mathrm{D}\mathrm{1}& \mathrm{=}\mathrm{2}\mathrm{*}\mathrm{TP}\left(\mathrm{L\text{'}}\mathrm{*}\mathrm{R\; \text{'}}\right)\mathrm{/}\left(\mathrm{L\text{'}}\mathrm{*}\mathrm{L\text{'}}\mathrm{+}\mathrm{R\; \text{'}}\mathrm{*}\mathrm{R\; \text{'}}\right)\\ & \mathrm{=}\mathrm{2}\mathrm{*}\mathrm{TP}\left(\mathrm{C}\mathrm{*}\mathrm{C}\mathrm{-}\mathrm{S}\mathrm{*}\mathrm{S}\right)\mathrm{/}\mathrm{TP}\mathrm{(}\mathrm{L\text{'}}\mathrm{*}\mathrm{L\text{'}}\mathrm{+}\mathrm{R\; \text{'}}\mathrm{*}\mathrm{R\; \text{'}}\mathrm{)}\mathrm{,}\end{array}$$

This results for the output signal D1, which can be used as a precursor to the control signal s or directly as a control signal s, as a value D1 = 1, if the audio signal i consists exclusively of a central component C. D1 = 0 results if the audio signal i consists exclusively of uncorrelated right and left signal components L, R. D = -1 results if the audio signal i consists exclusively of surround components S. With a mixture of the different components, as given in a real signal, values for D1 are between -1 and +1. The closer the output signal or the output value D1 is to +1, the more central-heavy is the audio signal i or L ', R', so that a correspondingly large speech component px can be closed.

The time constant of the low-pass filter TP can be in the range of approximately 100 ms, if a very fast response to changing signal components is desired. However, the time constant can be extended up to several minutes if a very slow response of the speech detector SD is desired. The time constant of the low-pass filter is therefore an advantageously variable parameter. Before carrying out a detection algorithm, DC components are expediently filtered out by means of a corresponding filter, in particular a DC notch filter (DC Notch). The further band limitation is optional.

Optionally, the illustrated circuit arrangements or methods for providing the control signal s can be followed by a stage in which a threshold value v is set, which of the output signal D1 of the described arrangements or methods to Is exceeded, to switch the control signal s in an active state.

In the case of parallel or subsequent speech signal processing of the audio signal i, the goal is to guide as many signal components as possible, which contain speech or speech components px, through a speech enhancement algorithm and to leave the remaining signal components unchanged, as is also the case Fig. 1 is described. This is advantageously solved by a matrix, as shown Fig. 3 outlined. Matrix coefficients k1, k2,..., K6 are determined as a function of the specific speech component px or dependent on the output value or output signal D1, D2 output by the speech detector SD or are determined as function px = F (D1, D2).
The actual speech enhancement algorithm or actual speech enhancer may be provided in a manner known per se. For example, an in DE 101 24 699 C1 to which full reference is made, described, simple frequency response correction can be performed. However, any other algorithms and devices for improving speech intelligibility can also be used.

At the in Fig. 3 illustrated matrix calculation, the input components or input channels L ', R' of the audio signal i are each multiplied by three factors k1, k3, k5 or k2, k4, k6 and supplied to addition elements. The first adder A is applied the signal of the first channel L 'multiplied by the first coefficient k1 and the signal of the second channel R' multiplied by the second coefficient k2 for addition. The second adder A is applied the signal of the first channel L 'multiplied by the third coefficient k3 and the signal of the second channel R' multiplied by the fourth coefficient k4 for addition. The third Adder A, the signal of the first channel L 'multiplied by the fifth coefficient k5 and the signal of the second channel R' multiplied by the sixth coefficient k6 are applied for addition. The output value of the second adder A is supplied to a speech enhancement circuit VS or a speech enhancement method or algorithm. Its output result is added by means of further addition elements A to the output value or output signal of the first addition element A for providing a first output channel LE and an output value or output signal of the third addition element A by means of a further addition element A for providing a second output channel RE.

$$\mathrm{k}\mathrm{2}\mathrm{=}\mathrm{K}\mathrm{5}\mathrm{=}\mathrm{-}\mathrm{px}\mathrm{/}\mathrm{2}$$

und $$\mathrm{k}\mathrm{3}\mathrm{=}\mathrm{k}\mathrm{4}\mathrm{=}\mathrm{px}\mathrm{/}\mathrm{2.}$$

For the determination of the coefficients, it is considered, for example, that the speech component px can be determined by the described methods by a range of values of, in particular, 0 ≦ P ≦ 1 and as a function of the specific speech components with px = F (D1, D2, D3). According to a simple variant, the coefficients can be determined according to $$\mathrm{k}\mathrm{1}\mathrm{=}\mathrm{k}\mathrm{6}\mathrm{=}\mathrm{1}\mathrm{-}\mathrm{px}\mathrm{/}\mathrm{2}\mathrm{.}$$

$$\mathrm{k}\mathrm{2}\mathrm{=}\mathrm{K}\mathrm{5}\mathrm{=}\mathrm{-}\mathrm{px}\mathrm{/}\mathrm{2}$$

and $$\mathrm{k}\mathrm{3}\mathrm{=}\mathrm{k}\mathrm{4}\mathrm{=}\mathrm{px}\mathrm{/}\mathrm{Second}$$

The two finally output signal channels or components LE, RE correspond to the processed signals which are supplied to the output O for the processed audio signal o.

Fig. 4 exemplifies function F (D1, D2 = 0, D3 = 0). In the case of the first illustrated function F = F1 (D1), the circuit already responds to a small detected speech component. The probability of misdetection is relatively high for small values of D1. However, due to the continuous course of the first function F1 (D1) the Effect of the speech algorithm with a small D1 on the audio signal is relatively low, so that an impairment of the audio signal is barely perceived.

In the case of a second function F2 (D1), the audio signal remains completely unaffected up to a threshold v = Ps2. Thereafter, the effects on the audio signal are greater with changes in the value of P1.

In the case of a third function F = F3 (D1), the algorithm is turned on when a certain threshold value v = Ps31 is exceeded and switched off when another, lower threshold value v = Ps32 is fallen short of. By installing such a hysteresis a constant switching in the transition area is prevented.

Claims (17)

1. A circuit configuration for enhancing the intelligibility of multi -component audio signals (i) possibly containing speech (px), comprising

   - an input (I) for inputting such an audio signal (i),

   - a speech detector (SD) for detecting speech (p x) in the inputted audio signal (i) and for providing a control signal (s) to control a speech processing device (SV) and/or a speech processing method for processing the audio signal (i),

   characterised in that the speech detector (SD) comprises a correlation device (CR) for performing a cross-correlation or an autocorrelation of components of the audio signal, and in that a matrix decoder is provided, which separates signal components of the audio signal by a matrix (MX).
2. A circuit configuration according to Claim 1,
   in which the speech detector (SD) is designed and/or controlled to detect speech components (px) in the audio signal (i).
3. A circuit configuration according to a preceding claim,
   in which the speech detector (SD)

   - is designed to process a multi -component audio signal (i), in particular stereo audio signals (L', R'), 3D stereo audio signals (L, R, C) and/or surround audio signals (L, R, C, S), having several audio signal components (L, R, C, S) and

   - comprises a processing device (CR) for detecting speech by comparing or processing the components (L, R, C, S) with each other.
4. A circuit configuration according to Claim 3,
   in which the speech detector (SD) comprises a direction and/or distance determination device for determining a direction and/or distance of common signal components of the different components (L, R, C, S).
5. A circuit configuration according to a preceding Claim,
   in which the control signal (s) for activating or deactivating the speech enhancement device (SV) and/ or the speech enhancement method is designed and/or controlled in dependence on the speech content of the audio signal (i).
6. A method for processing multi -component audio signals (i) possibly containing speech,
   in which

   - speech or speech components (px) contained in an audio signal (i) are detected and

   - depending on the result of the detection, a control signal ( s) for a speech processing device (SV) and/or a speech processing method for speech enhancement is generated and provided,

   - a cross-correlation or autocorrelation of components (R, L, C, S) of the audio signal (i) is performed and

   - whereby signal components of the audio signal are separated by a matrix.
7. A method according to Claim 6,
   in which the control signal (s) is generated in dependence upon the range of detected speech components (px).
8. A method according to Claim 7,
   in which the range of detected speech components ( px) is compared to a threshold value (v).
9. A method according to one of Claims 6 to 8,
   in which the detection is carried out adjustably with regard to a range of speech components to be detected and/or with regard to a frequency range of the speech components to be detected ( px) by means of variable parameters (v).
10. A method according to one of Claims 6 to 9,
    in which the audio signal components of a multi -component audio signal having several audio signal components (R, L, C, S) are compared to each other or processed with each other for the detection of speech.
11. A method according to Claim 10,
    in which the audio signal components (R, L, C, S) are compared or processed with respect to common speech components in the different audio signal components, especially to determine a direction and/or distance of the common signal components.
12. A method according to one of Claims 6 to 11,
    in which the control signal (s) is provided to activate or deactivate the speech enhancement device (SV) and/or the speech enhancement method.
13. A circuit configuration according to one of Claims 1 to 5 and/or a method according to one of Claims 6 to 12,
    wherein a frequency response is determined by means of an FIR or an IIR filter (FIR: Finite Impulse Response, IIR: Infinite Signal Response).
14. A circuit configuration according to one of Claims 1 to 5 and/or a method according to one of Claims 6 to 12,
    wherein matrix coefficients for a matrix (MX) are determined via a function (P= F(px)) dependent on the speech component (px).
15. A circuit configuration and/or method according to Claim 14,
    wherein the function (P = F(px)) is linear and constant.
16. A circuit configuration and/or a method according to Claim 14, wherein the function (P = F(px)) has a hysteresis.
17. A speech enhancement circuit configuration or method having a circuit configuration , and/or a method according to one of the preceding claims.

Images