EP1456839A2

EP1456839A2 - Method and device for the suppression of periodic interference signals

Info

Publication number: EP1456839A2
Application number: EP02787371A
Authority: EP
Inventors: Stefano Ambrosius Klinke; Christoph Pörschmann
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2001-12-19
Filing date: 2002-11-18
Publication date: 2004-09-15
Anticipated expiration: 2022-11-18
Also published as: CN100380445C; CN1606773A; EP1456839B1; DE50207531D1; US7656933B2; US20050096002A1; DE10162559B4; WO2003052746A2; ES2268123T3; WO2003052746A3; DE10162559A1

Abstract

The invention relates to a method and a device for suppressing periodic interference signals, comprising a unit (1) which is used to provide a period length (T) for the periodic interference signal, an interference detection unit (2) for detecting a signal (S') corresponding to the interference signal (S), and a subtraction unit (3) for subtracting the signal (S') corresponding to the interference signal. The interference detection unit (2) carries out multiple superpositioning of the input signal (E) and scales the multiple superpositioned input signal depending on the period length (T) of the interference signal in order to detect the signal (S') corresponding to the interference signal.

Description

description

Method and device for suppressing periodic interference signals

The present invention relates to a method and an apparatus for the suppression of essentially periodic interference signals and in particular to a method and an apparatus for the suppression of periodic interference in the audio frequency range, which is caused, for example, by a digital telecommunication system during data transmission and in e.g. a mobile telecommunications terminal or an external device such as a hearing aid can be coupled.

In a large number of digital telecommunications systems, data is transmitted between a mobile telecommunications terminal such as B. a mobile phone and an associated base station via a pulsed radio frequency signal with a predetermined carrier frequency. For a so-called GSM telecommunications system (global system for mobile communications), the carrier frequency is 900 MHz and a pulse frequency is approximately 217 Hz. In the case of a DECT telecommunications system, however, the carrier frequency is 1800 MHz and the associated pulse frequency is 100 Hz. One on GSM -based another standard is the DCSl800 standard, which also works at a carrier frequency of 1800 MHz. A large number of carrier frequencies with different pulse frequencies are therefore used in digital telecommunications systems, which is why the terminal device manufacturers are increasingly developing so-called dual-band or triple-band terminals for implementing the various standards.

The pulsed high-frequency signal in particular often causes problems here. The pulsed high-frequency signal is, for example, by the non-linear FET characteristic in the terminal existing microphones demodulates and sometimes causes clearly perceptible interference in the audio frequency range.

Figure 1 shows a simplified temporal representation of a periodically disturbed signal, such as that at the output of a signal source disturbed by a pulsed high-frequency signal, e.g. is output to a microphone.

FIG. 2 shows a simplified time representation of the associated pulsed high-frequency signal or periodic interference signal, as occurs, for example, in GSM or DECT telecommunication systems. In the GSM standard, according to FIG. 2, high-frequency pulses are transmitted at a time interval T of approximately 4.7 milliseconds, which contain the actual information. In the DECT standard, this time interval T is 10 milliseconds and corresponds to a frequency of 100 Hz in contrast to 217 Hz in GSM. These periodic interference signals can now be found in a printed circuit board and in particular on a signal source such as e.g. a microphone can be introduced, resulting in the spikes shown in Figure 1.

Conventional devices and methods for suppressing these periodic interference signals are essentially based on shielding the radio radiation by, for example, a conductive shield housing of the signal source or a conductive microphone housing. Please note that the housing is closed as completely as possible. An optimal effect is usually achieved with a metallic shield. However, the effort for such a shield is particularly important for devices such as a mobile telecommunication terminal and / or a hearing aid is expensive and also space-intensive.

Another option for suppressing these periodic interference signals is usually to eliminate the line-based coupling by filtering. Interference suppression capacitors are usually used here, which are attached spatially close to the field effect transistor (FET) of the microphone in order to attenuate the periodic high-frequency interference signal as much as possible. The selection of the capacitor is particularly critical here, since the influence of parasitic inductances increases sharply at high frequencies.

Consequently, optimal interference suppression is only achieved with a capacitor whose impedance is minimal for the respective frequency of the interference signal. The disadvantage here, however, is that such signal sources or microphones matched with capacitors cost significantly more than conventional standard electret microphones. Furthermore, a new signal source or microphone must be developed for each new telecommunications terminal or cell phone model or also each type of hearing aid, since the hardware environment such as e.g. the PCB layout of the end device or hearing aid influences the properties of the interference suppression capacitor. Another disadvantage is that a respective interference suppression capacitor is required for each carrier frequency, so that signal sources with two interference capacitors are required for a dual-band device and signal sources with even three interference suppression capacitors are necessary for a triple-band device.

The invention is therefore based on the object of providing a method and a device for suppressing essentially periodic interference signals, a simplified and improved interference suppression being made possible.

According to the invention, this object is achieved with regard to the method by the measures of patent claim 1 and with regard to the device by the features of patent claim 20.

In particular by a multiple superposition of the input signal depending on the period of the interference signal and a subsequent normalization of the multiple superimposed input signal, a signal corresponding to the interfering signal can be determined in a particularly simple manner, a subtraction of the signal corresponding to the interfering signal subsequently being carried out from the disturbed input signal resulting in a very well suppressed input signal. Such a method is easy to implement and, moreover, requires very little computing power, with no delays in the input signal, such as an audio signal.

The input signal is preferably temporarily stored as a digitized signal over a large number of period durations, as a result of which a superposition depending on the period duration can be implemented very easily.

The signal corresponding to the interference signal is preferably determined by averaging over a predetermined or changing number of periods, which is easily possible when implemented in software.

In addition, the input signal can be superimposed with different weighting factors and, in particular, a moving averaging can be used, as a result of which a particularly high-quality interference signal suppression is obtained. The weighting factors can be determined as a function of the input signal, which results in a further qualitative improvement in the interference suppression, regardless of a respective input signal level or ratio to the interference signal.

A division is preferably used for the standardization, although further standardization methods are also conceivable in principle in order to convert the superimposed input signal back into its original amplitude range. In the case of an unknown periodic disturbance, the period duration can also be determined from the disturbed input signal, in particular an autocorrelation of a section of the disturbed input signal for determining value maxima and then the period duration being determined from a time interval between the value maxima. In this way, even unknown periodic interference signals can be automatically detected and suppressed. In the same way, interference signals can be detected and suppressed that only have a substantially uniform period and can therefore have slight fluctuations.

Preferably, the method does not use a directly disturbed input signal, but rather an error signal that is dependent on it, for interference suppression, with a signal analysis being carried out on the basis of a disturbed useful signal for outputting the error signal and associated coefficients, and then on the basis for recovering a suppressed useful signal an interference-cleared error signal and the coefficients, a signal synthesis is carried out.

In signal analysis, FIR filtering for outputting a prediction error signal and associated predictor coefficients on the basis of a speech signal is preferably carried out, and in signal synthesis, IIR filtering for recovering the interference-free useful signal on the basis of an interference-free prediction error signal and the predictor coefficients. Thus, the speech estimators used in any case in speech coding in digital telecommunication systems can advantageously be used to further suppress the periodic interference signals. In the same way, such elements known from speech coding or speech estimation can also be used in external devices, such as hearing aids, whereby further miniaturization in particular with further interference suppression compared to the periodic interference signals generated by digital transmission systems.

The particular advantage results in particular from the fact that after the signal analysis carried out, only the error signal contains the periodic interferers, while the associated coefficients remain unaffected.

In the signal analysis, a linear prediction and in particular a short-term prediction is preferably carried out in a time range from 20 to 400 milliseconds. Such linear short-term predictors enable the generation of sufficiently precise error signals and coefficients for further signal processing. The so-called Levinson-Durbin algorithm is particularly suitable for determining the respective coefficients, since this is usually used in particular in mobile end devices for speech coding and is therefore available in any case.

The subtraction is preferably carried out as a function of a signal energy of the disturbed input signal and the suppressed input signal. In this way it is also possible to eliminate interference signals of this type which do not have an interference signal in every frame or after each period T, but rather, for example, skip a period. Such an irregular lack of interference signals within the period is often implied by the telecommunication standards used, so that such a lack of interference signals does not cause an undesired deterioration in interference suppression.

The method for suppressing periodic interference signals, which has been explained above, is preferably carried out during a speech pause of the disturbed input signal, step b) in particular, in which the signal corresponding to the interference signal is determined, should take place during a speech pause. This has the advantage that To determine the signal corresponding to the interference signal, it can be averaged over a comparatively short period of time since the useful data portion is missing in a speech pause. The main advantage, however, is that comb filter effects can be effectively avoided.

The establishment of a speech pause in the disturbed input signal can in principle take place in any way. However, the following methods are preferably used individually or in combination with one another: A speech pause can be determined on the basis of an energy in a current period of the input signal. Alternatively, a speech pause can be determined on the basis of a maximum value in a current period of the input signal. As a further alternative, it is conceivable for a speech pause to be determined on the basis of a change in the input signal in a current period in comparison with a previous period.

These methods are based on the fact that if a

Useful signal basically an energy in a current period and also a maximum value in a current period can be expected to be rather high. With regard to the fixed division of a speech pause based on a change in the input signal from period to period, it is pointed out that the input signal within a speech pause normally differs significantly from the input signal during a speech transmission.

In a preferred embodiment, a noise-reduced input signal can also be used as the input signal, this procedure having the advantage that it is easier to distinguish between the presence and absence of a speech pause, especially in cases where the useful signal is of low intensity. In the event that it is determined that the determination of the signal corresponding to the interference signal has been carried out on the basis of periods in which there was no speech pause by mistake, previous values of the signal corresponding to the interference signal can be used to carry out step c), the For this purpose, the device provided for carrying out the method has a suitable memory for the earlier values of the signal corresponding to the interference signal.

Further advantageous configurations of the invention are characterized in the further claims.

The invention is described below with reference to exemplary embodiments with reference to the drawing.

Show it:

FIG. 1 shows a simplified temporal representation of a periodically disturbed signal generated by a signal source;

FIG. 2 shows a simplified temporal representation of the periodic interference signal;

FIG. 3 shows a simplified block diagram of an overall system with the interference suppression device according to a first exemplary embodiment;

FIG. 4 shows a simplified block diagram of the interference suppression device;

FIG. 5 shows a simplified temporal representation of the signal generated in the interference suppression device and corresponding to the interference signal; and Figure 6 is a simplified block diagram of a subsystem with the interference suppression device according to a second embodiment.

FIG. 7 shows a simplified block diagram of the interference suppression device, combined with a speech pause detection device, according to a third exemplary embodiment,

FIG. 8 shows a simplified block diagram of the interference suppression device, combined with a speech pause detection device, according to a fourth exemplary embodiment, and

FIG. 9 shows a simplified block diagram of the interference suppression device, combined with a speech pause detection device, according to a fifth exemplary embodiment.

First embodiment

FIG. 3 shows a simplified block diagram of a system configuration in which the interference suppression device according to the invention can be used, for example. According to FIG. 3, M denotes a signal source or a microphone for converting an acoustic speech signal into an electrical speech signal or useful signal. As already described above, due to the coupling of interference signals, an actual voice useful signal N can be superimposed with an interference signal S, for example via the printed circuit board or via radio radiation, resulting in a disturbed input signal E. Such a superimposition of a useful signal with a periodic interference signal is generally known, the hum caused by the power network being a typical example.

As already described at the beginning, such interference can, however, also be used in digital telecommunication devices or in the immediate vicinity of these devices happen, in which case the periodic interference signal is caused by the data transmission between the mobile telecommunication terminal and the associated base station. In order to suppress such periodic interference signals, the known measures described at the outset can be carried out, such as, for example, providing a shield for the signal source M and / or providing an interference signal pre-filter, which usually has an interference suppression capacitor and also for reducing the periodic interference signal in the disturbed input signal E is suitable. The initially analog disturbed input signal is converted by an A / D converter W into a digitized disturbed input signal E and then fed to the actual interference signal suppression device U, which is subtracted from the disturbed input signal E by a signal S ^λ corresponding to the interference signal Interference-suppressed input signal E ^{λ is} generated, which is transmitted, for example, via an air interface I or is fed back via a listening path R to implement a necessary echo to a receiver loudspeaker, not shown.

FIG. 4 shows a simplified block diagram of the interference signal suppression device U according to FIG. 3. According to FIG. 4, the digitized, disturbed input signal E output by the converter W, which is composed of the useful signal N and the periodic interference signal S, is supplied to a period duration determination unit 1, for example. which determines a period T of the interference signal S. The period duration T of the interference signal S can be identified in this way in different ways. However, signal maxima are preferably determined by means of autocorrelation in a section of the input signal E or a disturbed audio signal (e.g. shortly after the telephone connection has been set up or in occasional intervals during the call) and from the time intervals between the signal maxima

Autocorrelation function directly determines the period T of the interference signal S. Such a period duration determination can therefore take place once or at predetermined intervals.

In the event that e.g. If no speech signal is transmitted during speech pauses, the period duration can alternatively also be determined directly between two maximum values of the interference signal or the disturbed input signal, as a result of which the period duration T is determined in a particularly simple manner.

Alternatively, the period duration determination unit 1 can also be implemented by a period duration preparation unit, not shown, which outputs the period duration T, for example, when an existing periodic interference signal is known.

In a disturbance signal determination unit 2, a signal S ^Λ corresponding to the disturbance signal S is now determined and then subtracted from the disturbed input signal E by means of a subtraction unit 3, as a result of which the suppressed input signal E ^Λ = N + SS ^λ is obtained, which is consequently im Essentially corresponds to the useful signal N.

In order to determine the signal S ^Λ corresponding to the interference signal S, a multiple superimposition of the input signal E and a subsequent normalization of the multiply superimposed input signal are carried out in the interference signal determination unit 2 as a function of the period T of the periodic interference signal S, as shown in FIG.

In the interference signal determination unit 2, a multiple superimposition of the input signal E is consequently carried out at a time interval of the period T, as a result of which the interference signals located at the same point are increasingly amplified and the statistically distributed useful signal or audio signal N is increasingly canceled. After a standardization, which for example corresponds to a division according to the number of superimpositions speaks, one in turn receives a signal S ^λ corresponding to the disturbed input signal E, which is essentially the same as the interference signal S in the input signal. As a result of the subtraction carried out in the subtractor 3, an interference-cleared input signal E is obtained which essentially corresponds to the useful or audio signal N.

The periodic interference signal is preferably averaged over a number of phases or frames. Since averaging over an infinitely long period of time is not possible, for example, averaging takes place over a predetermined or changing finite number of periods or period durations T. To improve the quality of interference suppression carried out, the introduction of so-called weighting factors has proven to be useful, whereby periods further in the past are to be weighted weaker than a current period in order to obtain a weighted average.

A moving averaging is preferably carried out in the

Interference signal determination unit 2 carried out according to the following scheme:

Average _n = ax Average _n -ι + (1 - a) x _Average current ιι

where n describes the number of the respective periods or frames and a describes a weighting factor.

The weighting factor a can be fixed between 0 and 1. The following values result for a weighting factor of a = 0.8 and a moving averaging over 2 period durations T:

Averages. = 0.2 x input signal ₀

Eingangssignalo) etc.

In this way, the mean value gives a signal S which corresponds very well to the interference signal S and which can subsequently be subtracted from the input signal.

Another possibility is to vary this weighting factor a, i.e. to make the system adaptive. It makes sense to average over a longer period of time if the interference signal is overlaid, for example, by a speaker. To be more precise, for example, the weighting factor a can be chosen to be large depending on the input signal or depending on its signal level (volume). On the other hand, for example during speech breaks, e.g. the signal level of the useful signal or audio signal N is very small, the weighting factor a is chosen to be smaller. In this case, the current phase or the frame or period of the interference signal is weighted more heavily.

This signal ascertained in the interference signal determination unit 2 or the unweighted mean value S ^Λ is then subtracted from the input signal (audio signal) in the current frame or the current period duration, as a result of which the interference signal S can be greatly reduced. If the mean value contains the entire portion of the periodic interference signal, this is even completely calculated out of the input signal.

The quality of the interference suppression device can also be improved by subtraction as a function of a signal energy of the disturbed input signal and the suppressed input signal E. Here, the subtractor 3 is expanded by the following estimate:

If the signal S ^Λ corresponding to the interference signal S is subtracted from the input signal E, a signal energy of Interference-suppressed input signal E is increased in the frame under consideration or the period, the subtraction is dispensed with or the subtraction is carried out with a weighting factor b (less than 1). An increase in the signal energy of the suppressed input signal E compared to the disturbed input signal E due to the subtraction indicates that the interference signal did not (unexpectedly) occur in the frame under consideration and thus a deterioration of the interference suppression would be achieved by the subtraction. Since such a lack of the periodic interference signal, for example in DECT telecommunication systems, is not uncommon and occurs at more or less regular intervals, such a dependent subtraction with possibly adaptive subtraction weighting factors b results in a further improvement in quality.

FIG. 5 shows a simplified time representation of the signal S determined by the interference signal determination unit 2, which essentially corresponds to the interference signal S and is subtracted from the input signal according to FIG. 1. In this way, a method and a device for suppressing periodic interference signals are obtained, as a result of which there is no need for metallic shielding of the microphones, for example. The costs for the microphones or signal sources can thus be reduced. Furthermore, when the input signals or audio signals are routed on a printed circuit board, there is no longer any need to pay attention to HF interference, which considerably simplifies the layout and allows a microphone position to be chosen more freely. Furthermore, the method described above is very easy to implement and, moreover, requires very little computing power, since essentially only two additions and multiplications per sample are necessary. The method also does not cause any additional delays in the audio signal.

The input signal is preferably not in the form of a digitized signal over a plurality of period durations T in one The buffer store shown is stored, so that further processing and in particular the superimposition or averaging described above can be implemented particularly easily.

The method described above was applied directly to the input signal E or the audio signal data according to the first exemplary embodiment. In the same way, however, it can also be applied to error signals or residual signals, such as occur, for example, in speech estimation.

Second embodiment

FIG. 6 shows a simplified block diagram of a subsystem with the interference suppression device according to a second exemplary embodiment. To simplify the following description, it is initially assumed that the optional blocks 4 and 5 are not present in FIG. 6 and that the disturbed useful signal is x (k) = x (k.). The same applies: x * ^Λ (k) = x * (k).

The device for suppressing periodic interference signals in accordance with the second exemplary embodiment essentially consists of a signal analyzer SA for outputting an error signal E (k) and associated coefficients ai on the basis of the disturbed useful signal or an impaired electrical voice signal. On the basis of the error signal E (k) output by the signal analyzer SA, the interference signal suppression device U described above now generates an interference-free error signal E ^Λ (k) with reduced periodic interference signals, which is forwarded to a signal synthesizer SS. The signal synthesizer SS carries out signal synthesis on the basis of the interference-free error signal E ^λ (k) and the coefficients ai generated by the signal analyzer SA in order to recover an interference-free useful signal x * (k) or x * (k). The useful signal quality of the suppressed useful signal x * (k) can therefore be further improved.

The interference suppression device U is preferably used in a mobile telecommunication terminal, e.g. a mobile phone, the elements shown in FIG. 6 already being present, at least in part, for realizing speech coding.

To reduce the amount of data and the susceptibility to interference, so-called speech encoders are used in particular in wireless telecommunication systems, which improve signal quality and immunity to interference, taking into account the human reception possibilities. Here, FIR filters (finite impulse response) or IIR filters for outputting a prediction error signal and associated predictor coefficients are generated on the basis of an applied speech signal as so-called speech estimators. According to the invention, the signal analyzer SA can now use such an FIR filter to output a prediction error signal E (k) and associated predictor coefficients ai on the basis of the perturbed speech signal x (k). As a result, the method used by the interference suppression device U is now not applied directly to the input signal E or the audio signal, but rather to an associated error signal or residual signal. Here, for example, a linear predictor can be used to carry out a linear prediction as a signal analyzer SA, a short-term prediction preferably being carried out in a time range from 20 to 400 milliseconds. Such linear short-term predicators, the so-called Levinson-Durbin algorithm preferably being used to calculate the predictor coefficients ai, are again generally known in speech coding, which is why a detailed description is not given below. The signal analyzer SA accordingly generates a disturbed error signal E (k) and associated coefficients a _± , which do not contain a disturbance.

According to FIG. 6, the actual interference suppression of the periodic interference signal is now carried out in the signal suppression device U described above.

The error signal E (k) generated by the signal analyzer SA essentially consists of the difference between the disturbed useful signal x (k) and an associated estimated value x ^A (k), ie e (k) = x (k) -x ^Λ (k ). Subsequently, the improved synthesized or at least partially interference-error signal E ^Λ (k) associated with the coefficients a _±, yielding x the interference suppression useful signal or original signal * (k) is obtained.

To further improve the coefficient calculation in the signal analyzer SA, a high-pass filter 4 can also be used on the input side according to FIG. 6 for additional high-pass filtering of the disturbed useful signal x (k) and for generating a filtered but still disturbed useful signal x ^Λ (k). Usually a so-called Preenfasys filter is used as HP filter 4, which leads to a further improvement in connection with the signal analyzers used from speech coding. To compensate for the optionally introduced HP filter 4, a TP filter 5 can optionally also be used on the output side for low-pass filtering of the suppressed useful signal x * ^λ (k), which finally outputs the suppressed useful signal x * (k). Such a TP filter usually consists of a so-called Deenfasys filter.

In the same way, the known interference suppression pre-filters and a shielding of the signal source M can optionally be added to the interference signal suppression device according to FIG. 6, which means that inexpensive electret microphones are now used. The interference suppression capacitors would be directly on the connection pins the signal source or the microphone M to attach. The advantage of the second exemplary embodiment described above is consequently that possible artifacts in the useful signal, which can arise due to conventional noise reduction, can be significantly weakened by the signal analysis and signal synthesis.

Third embodiment

In a third exemplary embodiment of the invention, which is illustrated with reference to FIG. 7, an extension is made compared to the exemplary embodiment illustrated in FIG. 4 in that a speech pause detection device β is provided, at the input of which the disturbed input signal E is present. On the basis of features of the disturbed input signal E, the speech pause detection device determines whether there is a speech pause at a current time frame / a current period T of the disturbed input signal E or whether useful voice signals are being transmitted.

The speech pause detection device 6 is connected via a control line 7 to the interference signal determination unit 2, so that the interference signal determination unit 2 is constantly informed as to whether or not there is a speech pause.

As in the embodiment according to FIG. 4, the disturbed input signal E is also present at the interference signal determination unit 2. An update of the mean value formed by the interference signal determination unit in the manner described above now only takes place when the speech pause detection device 6 indicates the presence of a speech pause via the control line 7.

The features on the basis of which the speech pause detection device 6 determines the presence of a speech pause include Ren, for example, a maximum signal value in a current period T or the total energy of the disturbed input signal E within a period T. Also, a comparison between current waveforms of the disturbed input signal E compared to previous waveforms from previous periods can be used to determine whether such a deviation between the Waveforms exist that a pause in speech can be concluded.

Since the useful signal for the detection of the signal S ^λ corresponding to the interference signal S is "disturbing" so to speak, the detection within a speech pause has the advantage that the signal S ^Λ can be determined more quickly with sufficient quality, since fewer averaging steps are required. Also Comb filter effects avoided.

Fourth embodiment

The fourth exemplary embodiment of the invention, which emerges from FIG. 8, differs from the exemplary embodiment according to FIG. 7 in that the speech pause detection device 6 has a further input at which the input signal E is present with reduced interference. For this purpose, the signal S ^Λ corresponding to the interference signal S is fed to a second subtractor 8, at the input of which the disturbed input signal is present and at whose output there is a noise-reduced signal which is fed to the speech pause detection device β. It should be pointed out, however, that the noise-reduced input signal present at the second input of the speech pause detection device 6 is based on an average value for the signal S with respect to its noise reduction, which is due to previous time periods T compared to the currently disturbed input signal E. The exemplary embodiment according to FIG. 8 allows speech pauses to be determined both with the aid of the disturbed input signal E and on the basis of the noise-reduced signal which is present at the second input of the speech pause detection device 6. Especially when the interference component in the disturbed input signal E is very large, it can be difficult to determine the presence of a speech pause solely on the basis of the disturbed input signal E. In this case, it makes sense to record speech pauses based on the noise-reduced input signal. In another

If the interference signal S is exposed to very strong intensity fluctuations or is not present over a period of time, it is more favorable to carry out the detection of speech pauses solely on the basis of the disturbed input signal E.

In the exemplary embodiment from FIG. 8, depending on the signal position, in particular a relationship between the interference signal S and the useful signal N or on the basis of other criteria, it can thus be decided whether to use the interference-reduced input signal E alone, the interference-reduced signal or both to carry out a speech pause detection is. Such a decision can be made in the speech pause detection device 6 with the aid of comparisons of the disturbed input signal E for successive period durations T.

Fifth embodiment

A fifth exemplary embodiment of the invention illustrated with reference to FIG. 9 is based in principle on the exemplary embodiment according to FIG. 7. However, the speech pause detection device 6 is connected via a control line 8 to a memory 9 which contains previous values for the signal S ^λ .

For example, if it is determined that the speech pause detection device 6 is faulty due to a transition from a speech pause to a speech transmission period works, can be used with the help of the memory 9 to the earlier values for the signal S ^Λ corresponding to the interference signal S. In this respect, a more favorable value for the signal S ^Λ supplied to the subtractor 3 can subsequently be found by exchanging incorrect values, for S ^λ which are obtained by averaging, by earlier values originating from a speech pause.

For this purpose, values for the signal S ^Λ , which originate clearly from speech pauses, are transmitted via a signal line 10

Copied memory 9, wherein the presence of uniqueness for a speech pause is conveyed via the signal line 8.

The previous values are copied via a signal line 11 to the interference signal determination unit 2 for the exchange of incorrect values, which have arisen, for example, from a transition from a speech pause to a speech transmission period.

Sixth embodiment

According to a sixth exemplary embodiment, the device according to the invention and the associated method are not integrated into a system generating the periodic interference signal, but rather are implemented as an external device. Such external devices can represent so-called hearing aids, in particular, since they are usually used in the immediate vicinity of a respective mobile telecommunication terminal and are therefore particularly exposed to the coupling of periodic interference signals described above. More precisely, the interference signal suppression device described above with direct or indirect application to the input signal is consequently implemented in a hearing aid device which, for example, includes a behind-the-ear device (BTE), an in-the-ear device (ITE) In the channel device (complete in the canal, CIC), a pocket device, a headset, headphones and / or an implant can be shown. Again- In this way, improved hearing aids can be realized which are essentially insensitive to the periodic interference signals generated by digital telecommunication systems.

The invention has been described above using periodic interference signals in the GSM and DECT telecommunication system. However, it is not limited to this and likewise includes periodic interference signals that are generated by other wireless, wired telecommunication systems or other systems. In the same way, the invention is not limited to mobile telecommunication terminals and hearing aids, but in the same way also includes other devices that are particularly exposed to such periodic interference signals.

Claims

claims

1. A method for suppressing periodic interference signals in a disturbed input signal, comprising the steps of: a) providing a period (T) of the essentially periodic interference signal (S); b) determining a signal (S ^Λ ) corresponding to the interference signal (S); and c) performing a subtraction of the signal (S ^Λ ) corresponding to the interference signal from the disturbed input signal (E) to generate an interference-free input signal (E ^λ ), characterized in that in step b) depending on the period (T) of the interference signal (S) a multiple superposition of the input signal (E) and a subsequent normalization of the multiple superimposed input signal is carried out.

2. The method according to claim 1, so that the input signal (E) is temporarily stored as a digitized signal over a plurality of period durations.

3. The method according to claim 1 or 2, ie, that averaging over a predetermined or changing number of period durations (T) is carried out in step b).

4. The method according to one of claims 1 to 3, so that the input signal (E) is superimposed with different weighting factors (a) in step b).

5. The method according to claim 4, characterized in that a moving averaging is carried out.

6. The method according to claim 4 or 5, so that the weighting factors (a) are determined as a function of the input signal.

7. The method according to any one of claims 1 to β, d a d u r c h g e k e n n z e i c h n e t that the standardization has a division.

8. The method according to any one of claims 1 to 7, d a d u r c h g e k e n n z e i c h n e t that in step a) the period (T) is determined from the disturbed input signal (E).

9. The method according to claim 8, so that a step in an autocorrelation of a section of the disturbed input signal (E) for determining maxima is carried out and the period (T) is determined from a time interval of the maxima.

10. The method according to any one of claims 1 to 9, so that the input signal represents an directly disturbed input signal (E) or an error signal (E (k)) dependent thereon.

11. The method according to claim 10, comprising the steps of: performing a signal analysis for outputting the error signal (E (k)) and associated coefficients (ai) on the basis of a disturbed useful signal (x (k)); and

Performing a signal synthesis to recover an interference-free useful signal (x * (k)) based on an interference-free error signal (E ^Λ (k)) and the coefficients (a _± ).

12. The method according to claim 11, characterized in that in the signal analysis FIR filtering and / or IIR filtering for outputting a prediction error signal (E (k)) and associated predictor coefficients (ai) on the basis of a speech signal (x (k )) and the signal synthesis, an IIR filtering and / or FIR filtering for recovering the suppressed useful signal (x * (k)) on the basis of a suppressed prediction error signal (E ^Λ (k)) and the predictor coefficients (ai) is carried out ,

13. The method according to claim 11 or 12, so that a linear prediction is carried out in the signal analysis.

14. The method according to claim 13, so that the linear prediction represents a short-term prediction in a time range from 20 to 400 milliseconds.

15. The method according to any one of claims 11 to 14, so that the coefficients (ai) are determined by means of the Levinson Durbin algorithm in the signal analysis.

16. The method according to any one of claims 1 to 15, characterized in that in step c) the subtraction is carried out as a function of a signal energy of the disturbed input signal (E) and the suppressed input signal (E ^Λ ).

17. The method according to any one of claims 1 to 16, characterized in that it is carried out in a wireless telecommunications terminal.

18. The method according to any one of claims 1 to 17, d a d u r c h g e k e n n z e i c h n e t that it is carried out in a hearing aid.

19. The method according to any one of claims 1 to 18, that the periodic interference signal is a GSM and / or DECT signal.

20. The method according to any one of claims 1 to 19, characterized in that in step b) the signal (S ^Λ ) corresponding to the interference signal (S) is determined in a speech pause of the disturbed input signal (E).

21. The method according to claim 20, d a d u r c h g e k e n n z e i c h n e t that a

Speech pause is determined based on an energy in a current period (T) of the input signal (E).

22. The method according to any one of claims 20 or 21, d a d u r c h g e k e n n z e i c h n e t that a

Speech pause is determined on the basis of a maximum value in a current period (T) of the input signal (E).

23. The method according to any one of claims 20 to 22, d a d u r c h g e k e n n z e i c h n e t that a

Speech pause is determined based on a change in the input signal (E) in a current period (T) compared to a previous period.

24. The method according to any one of claims 20 to 23, characterized in that a noise-reduced input signal is used as the input signal (E).

25. The method according to any one of claims 1 to 24, characterized in that earlier values of the signal (S ^Λ ) corresponding to the interference signal (S) are used to carry out step c).

26. Device for suppressing periodic interference signals in a disturbed input signal, comprising: a period duration preparation unit (1) for providing a period duration (T) of the essentially periodic interference signal (S); an interference signal determining unit (2) for determining a signal (S) corresponding to the interference signal; and a subtraction unit (3) for subtracting the signal (S ^Λ ) corresponding to the interference signal from a disturbed input signal (E) and for generating an interference-suppressed input signal (E), characterized in that the interference signal determination unit (2) as a function of the period (T) of the periodic interference signal (S) carries out a multiple superimposition of the input signal (E) and a normalization of the multiply superimposed input signal to determine the signal (S ^Λ ) corresponding to the interference signal.

27. The device according to claim 26, so that the interference signal determination unit (2) has a buffer for temporarily storing the input signal (E) as a digitized signal over a plurality of period durations (T).

28. The device according to claim 26 or 27, characterized in that the interference signal determination unit (2) carries out averaging over a predetermined or changing number of periods (T).

29. Device according to one of the claims 26 to 28, so that the interference signal determination unit (2) has a moving averaging unit with different weighting factors (a).

30. Device according to one of the claims 26 to 29, so that the standardization is carried out by a division unit in order to realize a ratio of the superimposed input signals with the number of superimpositions.

31. Device according to one of the claims 26 to 30, characterized by a signal analyzer (SA) for outputting an error signal (E (k)) as an input signal and associated coefficients (a) on the basis of a disturbed useful signal (x (k)); and a signal synthesizer (SS) for recovering an interference-suppressed useful signal (x * (k)) based on an interference-suppressed error signal (E ^Λ (k)) and the coefficients (ai).

32. Device according to claim 31, characterized in that the signal analyzer (SA) is a FIR filter and / or IIR filter for outputting a prediction error signal (E (k)) and associated predictor coefficient (ai) on the basis of a speech signal (x (k )) and the signal synthesizer (SS) has an IIR filter and / or FIR filter for recovering the suppressed useful signal (x * (k)) on the basis of a suppressed prediction error signal (E (k)) and associated predictor coefficients (ai).

33. Device according to claim 31 or 32, characterized in that the signal analyzer (SA) has a linear predictor for performing a linear prediction.

34. Device according to claim 33, so that the linear predictor (SA) carries out a short-term prediction in a time range from 20 to 400 milliseconds.

35. Device according to one of the claims 31 to 34, so that the signal analyzer (SA) determines the coefficients (ai) by means of the Levinson Durbin algorithm.

36. Device according to one of the claims 31 to 35, g e k e n n z e i c h n e t d u r c h an HP filter (4) for filtering the disturbed useful signal (x (k)) and for improving the coefficient calculation in the signal analyzer (SA).

37. Device according to claim 36, so that the HP filter (4) has a Preenf asys filter.

38. Device according to one of claims 36 or 37, g e k e n n z e i c h n e t d u r c h a TP filter (5) for filtering the interference-suppressed useful signal (x * (k)) and for compensating the HP filter (4).

39. Device according to claim 38, so that the TP filter has a Deenf asys filter.

40. Device according to one of the claims 31 to 39, characterized by an interference signal prefilter for reducing the periodic interference signal in the useful signal (x (k)).

41. Device according to one of the claims 26 to 40, so that the disrupted useful signal (x (k)) is generated by an electret microphone.

42. Device according to one of the claims 26 to 41, d a d u r c h g e k e n n z e i c h n e t that it is formed in a wireless telecommunications terminal.

43. Device according to one of the claims 26 to 42, d a d u r c h g e k e n n z e i c h n e t that it is formed in a hearing aid.

44. Device according to claim 43, so that the hearing aid device is a behind-the-ear device, an in-the-ear device, an in-channel device, a pocket device, a headphone and / or an implant.

45. Device according to one of the claims 26 to 44, so that the periodic interference signal represents a GSM and / or DECT signal.

46. Device according to one of the claims 26 to 45, characterized in that the period duration preparation unit has a period duration determination unit (1) which carries out an autocorrelation of a section of the disturbed input signal (E) to determine signal maxima and executes the period duration (T) a time interval between the signal maxima determined.

47. Device according to one of the claims 26 to 46, characterized in that it has a detection device for a speech pause in the disturbed input signal (E), which cooperates with the interference signal determination unit (2).

48. Device according to one of the claims 26 to 46, characterized in that it has a memory for previous values for the signal (S ^λ ) corresponding to the interference signal (S).