JP2005252761A - Method and apparatus for howling suppression - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a howling suppression apparatus for a portable telephone terminal capable of howling suppression, hardly deteriorating an audio volume and audio quality. <P>SOLUTION: In a howling prevention unit 1 in a portable telephone mounting an echo canceller, a first FFT processor 6 performs the FFT processing of a transmission speech signal outputted from the echo canceller 5. A signal processor 7 decreases the bandwidth signal level of a predetermined bandwidth having the highest signal level among the bandwidth of approximately 2 [kHz] or more, on which a first inverse FFT processor 8 performs inverse FFT processing, and outputs it to a line side. Meanwhile, as to a reception speech signal, a second FFT processor 12 performs the FFT processing of the reception speech signal. The signal processor 7 decreases a signal level in the same bandwidth as the bandwidth of which transmission speech signal level is decreased. A second inverse FFT processor 11 then performs inverse FFT processing and outputs if to the echo canceller 5. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method and apparatus for suppressing howling for a mobile phone terminal, and more specifically, a frequency band having a high signal level in a frequency region of about 2 [kHz] or higher that does not significantly affect the speech quality of a transmission / reception signal. The present invention relates to a method and an apparatus for detecting howling and suppressing howling by suppressing the signal level of the frequency band with a bandwidth of about 500 [Hz] that does not greatly affect the call quality.

  In recent years, many mobile phone terminals with a hands-free function have been developed. In a hands-free call using a mobile phone terminal, since the distance between the speaker and the microphone cannot be increased, the sound output from the speaker leaks into the microphone and howling is likely to occur. For this reason, technology for suppressing howling is important for realizing hands-free calling.

  A general method for suppressing howling in a hands-free call using a mobile phone terminal is to mount an echo canceller using an adaptive filter. However, since the cellular phone has a very large change in the surrounding environment such as outdoor use, accurate echo echo generation cannot be performed by the echo canceller, and there is a problem that howling occurs in some cases. In order to solve this problem, for example, “howling suppression method and apparatus” described in Patent Document 1 provides a method for suppressing howling by operating a time-variant notch filter using an error signal of an echo canceller. Yes. However, in this method, when the two people are talking at the same time (double talk), or when the echo path changes frequently and greatly due to external factors such as wind and ambient noise, the transmitted voice is excessive. There was a problem that it was not possible to suppress or howling that occurred.

On the other hand, in Patent Document 2, in a loudspeaker that amplifies the sound picked up by a microphone and outputs it from a speaker, the sound input signal is subjected to frequency analysis by FFT (Fast Fourier Transform) processing, and a certain level or more is exceeded. Further, a band including a frequency component that continues for a certain period is obtained every 20 [Hz] between 20 [Hz] and 10.24 [kHz], and the frequency component of the obtained band is subtracted from the FFT output to correct this. There is described a howling prevention device that suppresses the occurrence of howling by generating a sound reproduction output by performing inverse FFT processing on the subsequent FFT output.
JP 2000-92588 A Japanese Patent No. 2786463

  As described above, in a hands-free call using a mobile phone terminal, it is difficult to reliably remove signal components that cause howling with an echo canceller. Although it is conceivable to combine the echo canceller and the technique described in Patent Document 2, the technique described in Patent Document 2 cannot suppress the signal component that causes the howling until after the howling occurs. It is difficult to prevent howling itself. In addition, in order to suppress the frequency component in any band between 20 [Hz] and 10.24 [kHz] if it is a band that is more than a certain level and continues for a certain period of time, There is a problem that the volume is lowered or the sound quality is deteriorated.

  The present invention has been proposed in view of such circumstances, and an object thereof is to suppress howling without substantially deteriorating sound volume and sound quality.

  The first howling suppression method according to the present invention is a howling suppression method in a cellular phone terminal equipped with an echo canceller. The transmitted signal output from the echo canceller has the highest signal level in a band of about 2 [kHz] or more. In addition to reducing the signal level of a band of a predetermined bandwidth, the signal level of the same band as the previous band is reduced for the received signal and output to the echo canceller.

  The second howling suppression method according to the present invention is a howling suppression method in a cellular phone terminal equipped with an echo canceller. The transmitted signal output from the echo canceller has the highest signal level in a band of about 2 [kHz] or more. In addition to lowering the signal level of the predetermined bandwidth, the received signal is output to the echo canceller by reducing the signal level of the predetermined bandwidth having the highest signal level in the band of about 2 [kHz] or higher. It is characterized by that.

  The third howling suppression method of the present invention is a howling suppression method in a cellular phone terminal equipped with an echo canceller. A) FFT processing is performed on a transmission signal output from the echo canceller, and about 2 a step of reducing the signal level of a predetermined bandwidth having the highest signal level out of a band of [kHz] or more, and performing a reverse FFT process on the processed transmission signal and outputting it; b) FFT processing of the received signal And lowering the signal level of the same band as the band of the signal level of the transmission signal within the band of the reception signal subjected to the FFT processing, and outputting the processed reception signal by performing the inverse FFT , Including.

  The fourth howling suppression method of the present invention is a howling suppression method in a cellular phone terminal equipped with an echo canceller. A) FFT processing is performed on a transmission signal output from the echo canceller, and about 2 a step of reducing the signal level of a predetermined bandwidth having the highest signal level out of a band of [kHz] or more, and performing a reverse FFT process on the processed transmission signal and outputting it; b) FFT processing of the received signal Then, for the received signal subjected to the FFT processing, the signal level of a predetermined bandwidth having the largest signal level in the band of about 2 [kHz] or more is reduced, and the processed received signal is subjected to inverse FFT processing. And a step of outputting.

  A fifth howling suppression method of the present invention is characterized in that, in any of the first to fourth howling suppression methods, the predetermined bandwidth is about 500 [Hz].

  The sixth howling suppression method of the present invention is the fifth howling suppression method, wherein seven bands that overlap each other by about 250 [Hz] between about 2 [kHz] and about 4 [kHz] are suppressed. It is characterized by selecting.

  The seventh howling suppression method of the present invention is characterized in that, in any one of the first to sixth howling suppression methods, the range of suppression amount is about 6 [dB] to about 12 [dB].

  The first howling suppression apparatus of the present invention is a howling suppression apparatus in a cellular phone terminal equipped with an echo canceller. The transmission signal output from the echo canceller has the highest signal level in a band of about 2 [kHz] or more. In addition to reducing the signal level of a band of a predetermined bandwidth, the signal level of the same band as the previous band is reduced for the received signal and output to the echo canceller.

  According to the second howling suppression apparatus of the present invention, in the howling suppression method in the echo canceller-equipped mobile phone terminal, the transmission signal output from the echo canceller has the highest signal level in the band of about 2 [kHz] or more. In addition to lowering the signal level of the predetermined bandwidth, the received signal is output to the echo canceller by reducing the signal level of the predetermined bandwidth having the highest signal level in the band of about 2 [kHz] or higher. It is characterized by that.

  According to a third howling suppression device of the present invention, in a howling suppression device in a mobile phone terminal equipped with an echo canceller, a first FFT processing unit that performs FFT processing on a transmission signal output from the echo canceller and is input. A first inverse FFT processing unit that performs inverse FFT processing on the signal and outputs, a second FFT processing unit that performs FFT processing on the received signal and outputs, and performs an inverse FFT processing on the input signal and outputs the processed signal to the echo canceller. For the transmission signal output from the second inverse FFT processing unit and the first FFT processing unit, the signal level of the band of the predetermined bandwidth having the largest signal level in the band of about 2 [kHz] or more is set. Reduced and output to the first inverse FFT processing unit, the signal level of the transmission signal was reduced within the band of the reception signal output from the second FFT processing unit Characterized in that it comprises a signal processing unit which reduces the signal level of the same band as frequency is output to the second inverse FFT processor.

  According to a fourth howling suppression device of the present invention, in a howling suppression device in a mobile phone terminal equipped with an echo canceller, a first FFT processing unit that performs FFT processing on a transmission signal output from the echo canceller and is input. A first inverse FFT processing unit that performs inverse FFT processing on the signal and outputs, a second FFT processing unit that performs FFT processing on the received signal and outputs, and performs an inverse FFT processing on the input signal and outputs the processed signal to the echo canceller. For the transmission signal output from the second inverse FFT processing unit and the first FFT processing unit, the signal level of the band of the predetermined bandwidth having the largest signal level in the band of about 2 [kHz] or more is set. The received signal output from the second FFT processing unit is lowered and output to the first inverse FFT processing unit, and the signal level having the highest signal level is about 2 [kHz] or more. And a signal processing unit that lowers the signal level of a band of a predetermined bandwidth and outputs the signal level to the second inverse FFT processing unit.

  A fifth howling suppression apparatus according to the present invention is characterized in that, in the first to fourth howling suppression apparatuses, the predetermined bandwidth is about 500 [Hz].

  The sixth howling suppression apparatus according to the present invention is the fifth howling suppression apparatus, and includes seven bands that overlap each other by about 250 [Hz] between about 2 [kHz] and about 4 [kHz]. It is characterized by selecting.

  A seventh howling suppression device of the present invention is characterized in that, in the first to sixth howling suppression devices, the range of suppression amount is about 6 [dB] to about 12 [dB].

  According to the present invention, howling can be suppressed with almost no deterioration in sound volume and sound quality. The reason for this is that the transmission signal output from the echo canceller is a signal of a predetermined bandwidth with the highest signal level in the band of approximately 2 [kHz] or higher that does not have a significant effect on the speech quality of the transmission / reception signal. Decreasing the level and reducing the signal level of the received signal in the same band as the previous band to output to the echo canceller, or a predetermined bandwidth with the highest signal level in the band of about 2 [kHz] or higher This is because the signal level of the band is reduced and output to the echo canceller.

  FIG. 1 is a block diagram of an embodiment of the present invention. The howling prevention apparatus for a mobile phone terminal according to the present invention is used by being incorporated in a mobile phone terminal (not shown). Referring to FIG. 1, a howling prevention device 1 for a mobile phone terminal according to an embodiment of the present invention is arranged on the line side of an echo canceller 5, and includes a first FFT (Fast Fourier Transform) processing unit 6, The processing unit 7 includes a first inverse FFT processing unit 8, a second inverse FFT processing unit 11, and a second FFT processing unit 12.

  The microphone 2 is connected to the A / D converter 3, collects the voice of the speaker, and sends it to the A / D converter 3 as an analog transmission signal. The A / D converter 3 is connected to the microphone 2 and the transmission amplifier 4. The analog transmission signal transmitted from the microphone 2 is sampled at a predetermined frequency, converted into a digital signal, and sent to the transmission amplifier 4. The transmission amplifier 4 is connected to the A / D converter 3 and the echo canceller 5, amplifies the digital signal at each sampling point sent from the A / D converter 3, and sends it to the echo canceller 5. The echo canceller 5 includes an adaptive filter, and is connected to the transmission amplifier 4, the first FFT processing unit 6, the D / A converter 10, and the second inverse FFT converter 11, and the second inverse FFT. A simulated echo signal is generated based on the received signal sent from the processing unit 11, and the echo signal is removed by subtracting the simulated echo signal from the transmitted signal sent from the transmission amplifier 4.

  The first FFT processing unit 6 is connected to the echo canceller 5 and the signal processing unit 7, and transmits the transmission time domain signal after removal of the echo signal sent from the echo canceller 5 by performing 4096-point FFT processing. The signal is converted into a speech frequency domain signal and sent to the signal processing unit 7. The signal processing unit 7 is connected to the first FFT processing unit 6, the first inverse FFT processing unit 8, the second inverse FFT processing unit 11, and the second FFT processing unit 12, and the first FFT From the transmission frequency domain signal sent from the processing unit 6, a frequency band with a maximum bandwidth of 500 [Hz] is detected in the frequency band of 2 to 4 [kHz], and the signal strength of that band is set to 6 [ [dB] and then sent to the first inverse FFT processing unit 8. After that, the signal strength in the same frequency band as the previously detected frequency band is reduced by 6 [dB] with respect to the reception frequency domain signal transmitted from the second FFT processing unit 12, and then the second inverse FFT processing unit 11 is used. Send to. The first inverse FFT processing unit 8 is connected to the signal processing unit 7 and an encoder (not shown). The transmission frequency domain signal sent from the signal processing unit 7 is subjected to inverse FFT processing and converted to a time domain signal. Then, it is sent to an encoder (not shown).

  The speaker 9 is connected to the D / A converter 10 and converts the electrical signal sent from the D / A converter 10 into sound and emits the sound. The D / A converter 10 is connected to the speaker 9 and the echo canceller 5, converts the digital signal sent from the echo canceller 5 into an analog signal, and sends it to the speaker 9. The second inverse FFT processing unit 11 is connected to the echo canceller 5 and the signal processing unit 7. The received frequency domain signal sent from the signal processing unit 7 is inverse FFT processed to convert it into a time domain signal and echo. Send to canceller 5. The second FFT processing unit 12 is connected to the signal processing unit 7 and the reception amplifier 13 and converts the reception time domain signal sent from the reception amplifier 13 into a reception frequency domain signal by performing 4096-point FFT processing. The signal is sent to the signal processing unit 7. The reception amplifier 13 is connected to a decoder (not shown) and the second FFT processing unit 12, amplifies a digital received signal sent from the decoder (not shown), and sends the amplified signal to the second FFT processing unit 12.

  FIG. 2 is a flowchart showing a processing example of the signal processing unit 7. The signal processing unit 7 receives the transmission signal from the first FFT processing unit 6 (step A1), and calculates the signal level of that band for each band of band numbers 1 to 7 shown in FIG. 3 (step A2). . Specifically, the signal level of each frequency is calculated as the sum of the square of the real part and the imaginary part, and the signal level of each frequency is added for each band of band numbers 1 to 7 in FIG. Next, the signal level of each band is compared, and the signal level of each frequency in the band having the highest signal level is reduced by 6 [dB] (step A3).

  Here, as shown in FIG. 3, the reason why the bandwidth of each band is set to 500 [Hz] is that if the suppression is performed in a band wider than 500 [Hz], the volume decreases, and conversely, 500 [Hz]. This is because if it is narrower, a sufficient howling suppression effect cannot be obtained. Adjacent bands are overlapped at half the bandwidth. The reason is to make it possible to reliably detect the howling frequency.

  Thereafter, the processed transmission signal is sent to the first inverse FFT processing unit 8 (step A4). Subsequently, the reception signal is received from the second FFT processing unit 12 (step A5), and the signal level of each frequency in the same band as the band in which the signal level is reduced in step A3 is reduced by 6 [dB] (step A6). After that, the processed reception signal is sent to the second inverse FFT processing unit 11 (step A7). And it returns to step A1 and repeats the process mentioned above.

  According to the present embodiment, the following effects can be obtained.

  The first effect is that howling can be suppressed even when the echo signal removal effect by the echo canceller 5 is insufficient. The reason is that the signal level of each frequency of the highest signal level in the band of 2 [kHz] or more where howling occurs is reduced by 6 [dB] with respect to the transmitted signal, and the same band of the received signal This is also because it decreases 6 [dB].

  The second effect is that the volume of the transmitted voice and the received voice can be maintained while suppressing howling. The reason for this is that the sound energy is concentrated in the low frequency region, and the band of 2 [kHz] or more, which is less energy than the low region, is suppressed only in the bandwidth of 500 [Hz].

  The third effect is that occurrence of howling itself can be prevented. The reason is that the signal level of each frequency of the highest signal level in the band over 2 [kHz] where howling occurs is reduced by 6 [dB], and the same band of the received signal is also reduced by 6 [dB]. This is because the processing to be performed is always performed regardless of the occurrence of howling.

  The fourth effect is that sound quality deterioration of the transmitted voice and the received voice can be prevented while suppressing howling. This is because, unlike the prior art of Patent Document 2, frequency components in a band of 2 [kHz] or less that affect sound quality are not suppressed.

  In the embodiment of FIG. 1, the processing is simplified by uniquely determining which band of the received signal band is to be suppressed from the band determined on the transmission signal side. The band for suppressing the signal level can also be determined by the same method as the transmission signal. A processing example of the signal processing unit 7 in that case is shown in FIG.

  Referring to FIG. 4, steps A8 and A9 are added instead of the process of step A6 in FIG. In step A <b> 8, the signal processing unit 7 calculates the signal level of each band for each band of band numbers 1 to 7 shown in FIG. 3 in the received signal received from the second FFT processing unit 12. Specifically, the signal level of each frequency is calculated as the sum of the square of the real part and the imaginary part, and the signal level of each frequency is added for each band of band numbers 1 to 7 in FIG. In subsequent step A9, the signal levels of the respective bands are compared, and the signal level of each frequency in the band having the highest signal level is reduced by 6 [dB]. Thereby, since the frequency band suppressed by transmission and reception can be controlled independently, occurrence of howling can be prevented more reliably.

  In the above embodiment, the first FFT processing unit 6 and the second FFT processing unit 12 are 4096-point FFT processing. However, the number of FFT points is reduced according to the performance of the arithmetic unit constituting the FFT processing unit. However, it may be increased.

  Further, although the signal level reduction amount is 6 [dB], it may be reduced by 6 [dB] or more. However, since the sound pressure may decrease, the amount of decrease is preferably up to 12 [dB].

  As mentioned above, although the Example of this invention was described, this invention is not limited only to the above embodiment, Various other addition changes are possible. For example, the numerical values of 2 [kHz], 500 [Hz], 4 [kHz], 250 [Hz], 6 [dB], and 12 [dB] shown in the above-described embodiments do not need to be exactly the same, If it is almost the value, it is good.

  As described above, the howling suppression method and apparatus according to the present invention are useful as a method and apparatus for suppressing howling in a mobile phone terminal, and particularly prevent the occurrence of howling itself in a mobile phone terminal with a hands-free function. It is suitable for use in a method and apparatus for suppressing unavoidable howling.

It is a block diagram of one Example of this invention. It is a flowchart which shows the process example of the signal processing part in one Example of this invention. It is a figure which shows seven bands set about 250 [Hz] by overlap between about 2 [kHz]-about 4 [kHz] as a candidate of the band to suppress. It is a flowchart which shows the process example of the signal processing part in another Example of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Howling prevention apparatus 2 for mobile phone terminals ... Microphone 3 ... A / D converter 4 ... Transmission amplifier 5 ... Echo canceller 6 ... First FFT processing unit 7 ... Signal processing unit 8 ... First inverse FFT processing unit 9 ... speakers
10 ... D / A converter
11: Second inverse FFT processing unit
12 ... Second FFT processing unit
13 ... Receiving amplifier

Claims (14)

  In the howling suppression method in a mobile phone terminal equipped with an echo canceller, the signal level of a predetermined bandwidth having the highest signal level is reduced in the band of approximately 2 [kHz] or more for the transmission signal output from the echo canceller. In addition, a howling suppression method, wherein the received signal is output to an echo canceller by reducing the signal level of the same band as the previous band.   In the howling suppression method in a mobile phone terminal equipped with an echo canceller, the signal level of a predetermined bandwidth having the highest signal level is reduced in the band of approximately 2 [kHz] or more for the transmission signal output from the echo canceller. In addition, a howling suppression method, wherein a received signal is output to an echo canceller by reducing a signal level of a predetermined bandwidth having a maximum signal level in a band of about 2 [kHz] or more. In the howling suppression method in the echo canceller-equipped mobile phone terminal,
a) FFT processing is performed on the transmission signal output from the echo canceller, and the signal level of the predetermined bandwidth having the highest signal level is obtained from the FFT processing transmission signal within a band of about 2 [kHz] or more. Reducing and processing the transmission signal after the processing by inverse FFT processing,
b) The received signal is subjected to FFT processing, and the signal level in the same band as the band in which the signal level of the transmitted signal is reduced is reduced in the FFT processed received signal band, and the processed received signal is inverted. FFT processing and outputting,
A method for suppressing howling. In the howling suppression method in the echo canceller-equipped mobile phone terminal,
a) FFT processing is performed on the transmission signal output from the echo canceller, and the signal level of the predetermined bandwidth having the highest signal level is obtained from the FFT processing transmission signal within a band of about 2 [kHz] or more. Reducing and processing the transmission signal after the processing by inverse FFT processing,
b) The received signal is subjected to FFT processing, and the received signal subjected to the FFT processing is reduced in signal level in a predetermined bandwidth having the largest signal level in a band of about 2 [kHz] or more, and the received signal is processed. A step of performing inverse FFT processing on the signal and outputting it,
A method for suppressing howling.   The howling suppression method according to any one of claims 1 to 4, wherein the predetermined bandwidth is about 500 [Hz].   6. The howling suppression method according to claim 5, wherein the band to be suppressed is selected from seven bands that overlap each other by about 250 [Hz] between about 2 [kHz] and about 4 [kHz].   The howling suppression method according to any one of claims 1 to 6, wherein a range of the suppression amount is about 6 [dB] to about 12 [dB].   In a howling suppression device in an echo canceller-equipped mobile phone terminal, the signal level of a predetermined bandwidth having the highest signal level is reduced in the band of approximately 2 [kHz] or more for the transmission signal output from the echo canceller. In addition, a howling suppression apparatus that reduces the signal level of the received signal in the same band as the previous band and outputs the signal to the echo canceller.   In a howling suppression device in an echo canceller-equipped mobile phone terminal, the signal level of a predetermined bandwidth having the highest signal level is reduced in the band of approximately 2 [kHz] or more for the transmission signal output from the echo canceller. In addition, a howling suppression apparatus characterized in that the received signal is output to an echo canceller by reducing the signal level of a predetermined bandwidth having the largest signal level in a band of about 2 [kHz] or more.   In a howling suppression apparatus in an echo canceller-equipped mobile phone terminal, a first FFT processing unit that performs FFT processing on a transmission signal output from the echo canceller, and a first FFT that performs inverse FFT processing on the input signal for output. An inverse FFT processing unit, a second FFT processing unit that performs FFT processing on the received signal and outputs the signal, a second inverse FFT processing unit that performs inverse FFT processing on the input signal and outputs the processed signal to the echo canceller, and For the transmission signal output from one FFT processing unit, the first inverse FFT processing unit is configured by lowering the signal level of a predetermined bandwidth having the largest signal level within a band of about 2 [kHz] or more. The signal level of the same band as the band in which the signal level of the transmission signal is reduced is reduced in the band of the reception signal output from the second FFT processing unit. Howling suppression apparatus, comprising a signal processing unit for outputting to said second inverse FFT processor.   In a howling suppression apparatus in an echo canceller-equipped mobile phone terminal, a first FFT processing section that outputs a speech signal output from the echo canceller by performing FFT processing, and a first FFT processing that outputs the input signal by performing inverse FFT processing. An inverse FFT processing unit, a second FFT processing unit that performs FFT processing on the received signal and outputs the signal, a second inverse FFT processing unit that performs inverse FFT processing on the input signal and outputs the processed signal to the echo canceller, and For the transmission signal output from one FFT processing unit, the first inverse FFT processing unit is configured by lowering the signal level of a predetermined bandwidth having the largest signal level within a band of about 2 [kHz] or more. The received signal output from the second FFT processing unit is reduced in signal level in a predetermined bandwidth having the highest signal level in a band of about 2 [kHz] or more. And a signal processing unit that outputs to the second inverse FFT processing unit.   The howling suppression apparatus according to claim 8, wherein the predetermined bandwidth is about 500 [Hz].   13. The howling suppression apparatus according to claim 12, wherein the band to be suppressed is selected from seven bands that overlap by about 250 [Hz] between about 2 [kHz] and about 4 [kHz].   The howling suppression device according to any one of claims 8 to 13, wherein a range of suppression amount is about 6 [dB] to about 12 [dB].

Images