DE4216911A1 - IMPROVEMENT IN LOSS OF ECHO RETURN - Google Patents

Abstract

A communications transceiver, such as a GSM mobile, is provided, for communicating in frames of encoded audio, comprising an audio input path (10), an audio output path (22), a voice activity detector (VAD 12) for detecting voice on the audio input path, and echo detecting means (25, 26, 27,13, 28) for detecting unwanted echos on the input path resulting from acoustic coupling with the output path. Transmission of encoded audio from the audio input path is inhibited, by means (14, 29), in the presence of voice which is indicated as echo by the echo detecting means. The transceiver preferably has means for generating silence indicator (SID) frames in place of audio frames during periods of detected echo and means (40) Fig. 4 (not shown) for supplementing the silence indicator (SID) frames with comfort noise parameters and transmitting the supplemented frames for decoding as audio frames at a receiver. <IMAGE>

Description

The present invention relates to a Communication system such as a cellular radio system with a transmitter and a receiver and one Device for determining the presence of language in the signal transmitted from the transmitter to the receiver.

In the cellular GSM radio system, speech becomes frames (Frame) of encoded voice data transmitted. The system specifies that each mobile radio unit has one Voice Activity Detector (VAD) for detecting speech on a channel. This is used for operation in a discontinuous Discontinuus Transmission Mode (DTx) used. If the DTx mode is inactive, frames will be of coded speech transmitted continuously, regardless of whether there is language on the channel. If no If there is speech on the channel, the speech encoder encodes the background noise as if it were speech. this will transmitted and reproduced on the receiving end. If DTx mode is active and no language on the channel is the absence of language by the VAD detected and, instead of the transmission of information, how they are necessary for the complete coding of each frame the mobile unit only transmits filter parameters for the speech decoder. A frame of filter parameters is called SID frame (Silence Descriptor Frame). The DTx operation is described in GSM recommendations 06.31 described. The recipient fills in the rest of the frame Random data (so-called comfort noise), as in the GSM recommendations 06.12 is described. Instead of every frame to transmit the spectrum of the background noise only renewed every 24th frame. This reduces the  Total level of radio activity in the system and reduced hence the co-channel interference. It also saves electricity the batteries in the mobile unit.

In this description, the expression "DTx-Mode" for uses the mode (operation) in which SID frames are used Base station during the absence of speech be transmitted.

This mode is used when a command from the Base station is received. The expression "DTx function" is used to change the mobile unit from Voice mode for the transmission of SID frames during the Label DTx modes. The base station recognizes that SID frames in which they contain the excitation information observed. If this is zero, the frame is shown as SID frame considered.

In the early years of commissioning the system is not intended that the DTx mode is used. If the Total message traffic increases over time DTx mode can be activated. All mobile units must however, have the ability in both right from the start Modes to work. It doesn't like in distant areas be necessary to activate the DTx mode. It will therefore there is a situation where the mobile unit, depending on their location, the modes must switch. This Switching between modes is in the GSM recommendations fixed.

The elements of the GSM system as described above are shown in FIG. 4. In this figure, the elements of the mobile radio telephone are shown on the left and the elements of the base station on the right. The mobile unit comprises: a handset, of which a microphone 10 is shown; a speech encoder 11 for encoding speech before it is transmitted; a VAD (Voice Activity Detector) 12 that detects the presence of speech and distinguishes speech from background noise; a VAD hangover element 13 for expanding the VAD indicator, and a discontinuous transmitter DTx / tx 14 for generating SID frames when the VAD flag is cleared. The DTx / tx unit also provides a flag (SP flag) for the radio subsystem to determine whether the frame is a speech frame or a SID frame. The voice SID frames are transmitted to the base station by means of a radio subsystem. The base station has a discontinuous transmission / reception (DTx / rx) element 18 and a speech decoder 19 .

When the DTx mode is active, the VAD 12 detects the absence of speech on the channel and causes the DTx / tx 14 to generate a SID frame. The SID frame contains only the speech coding filter parameters and zeros at the locations of the excitation information. In this state, the mobile unit transmits SID (Silence Descriptor) frames that only have speech coding filter parameters and zeros at the locations of the excitation information. The base station receives the SID frames and replaces the excitation information with "comfort" noise from a comfort noise generator in the DTx / rx element 18 . In this way, the comfort noise generator reconstructs background noise similar to that received by the mobile microphone 10 . This means that the listener at the base station does not hear the disruptive effect of speech and then suddenly no longer hear anything.

A problem with mobile radio units is the acoustic feedback (echo return) from the mobile earpiece (not shown) to the microphone 10. The cellular GSM radio system has strict requirements for the echo return losses. The guidelines require that the echo return losses be less than or equal to 46 dB. With a good mechanical design of the handset, it should be possible for the acoustic coupling from the earpiece to the microphone to reach the 46 dB difference in the directive. However, it is desirable to provide volume control to boost the handset's audio output by 10 dB. If the earpiece output is raised by 10 dB, the echo return losses are only 36 dB. In addition, the maximum maximum echo delay for an acoustic signal that comes from a person connected to the public telephone network (PSTN) and that comes back due to the acoustic coupling between the earpiece and the microphone should be 180 milliseconds. In practice there will be significantly longer delays, depending on the transmission route. For example, a satellite connection can cause delays of 400 milliseconds. This echo interferes with the ear of the recipient connected to the PSTN.

According to the present invention is a Communication transceiver provided with in Frames encoded sound and communicates one Audio input path, an audio output path, a VAD (Voice Activity Detector) for detecting speech the audio input path, an echo detection device for Detection of unwanted echoes on the input path, that from the acoustic coupling with the output path originate, and a facility on the Echo detection device responds to prevent encoded audio information from the audio input path  during the presence of speech that is echoed by the Echo detection device was recognized, is transmitted / having.

The transmitting / receiving device preferably has one Device for generating SID frames instead of Audio frames during periods when echo is detected was, and also has a facility for Supplement the SID frames with Comfort noise parameters and to transfer the supplemented frames so that they are as Audio frames can be decoded in the receiver. To this Echo frames become wise through frames with comfort noise replaced even if the recipient is unable to Decoding of SID frames.

In the following a preferred embodiment of the present invention with reference to the accompanying drawings explained in more detail. The drawings show in single:

Fig. 1 is a Kommunikationssende- / receiving device according to a preferred embodiment of the present invention,

Fig. 2 shows the operation of a preferred embodiment while the DTx mode is active and

Fig. 3 shows the operation of a preferred embodiment of the invention, when the DTX mode is not used,

Fig. 4 element of a mobile transceiver and a transceiver base station according to the prior art.

The conventional arrangement according to FIG. 4 is explained in more detail below. If the VAD 12 detects speech (VAD = 1), coded speech is transmitted, and if no speech is detected (VAD = 0), the radio transmitter, with the exception of short intervals, in which SID frames are transmitted with Comfort noise filter parameters , switched off. The SID frames are sent to the radio subsystem 15 as long as the VAD flag is in the "low" state and are renewed for the base station only during a short interval that follows the speech burst, and only for every 24th frame . The DTx / rx software 18 in the base station receives the SID frames and places random numbers in the places of the missing excitation information, thereby generating comfort noise which simulates the noise on the microphone 10 . The comfort noise frames are decoded by the speech decoder 19 .

The VAD function is based on an adaptive filter to increase the speech-to-noise ratio. The performance of the filtered signal is compared to a threshold. Speech is recognized (VVAD = 1) when the threshold is exceeded. The VAD hangover element 13 is used to eliminate mid-burst and end-burst clipping of the speech.

The VAD filter coefficients are obtained during the speech pauses. In order to renew the filter coefficients during speech, the VAD is correspondingly sensitive and an echo with 45 dB below normal speech level can be detected in a quiet environment, such as an office. This means that the DTx / tx function 12 opens the "WUp-link" (up) radio channel and the echo is transmitted. In a portable "handheld" (handheld device), the transmission of the echo creates another problem. If the radio path is open during both local and long-distance (near-end, far-end talk) calls, the DTx function will not reduce the power consumption of the mobile device, thereby reducing operating time.

To remove the echo, it’s not enough just adjust the VAD sensitivity so that the Echo is not detected. Such an arrangement would the problem with that the VAD its Filter coefficients would renew what during the echo to reduce the ability of the adaptive filter, Remove noise from the received signal would. Furthermore, if the comfort noise parameters (which are renewed every 24th frame) during the In the presence of an echo, this would have to be renewed a very changeable comfort noise, what would disturb the listener at the base station.

In Fig. 1, in addition to the elements of Fig. 4, an earpiece 20 and a speech decoder 21 are provided on the audio output path 22 . The following new elements are also shown: an audio output path feedback line 24 , a weakener 25 , a comparator 26 , a logic gate 27 , another VAD hangover element 28 (in addition to the VAD hangover element 13 ) and a silence descriptor sampler 29 (Silence Descriptor Sampler). The echo cancellation function according to the present invention can be used with or without the DTx function. For the current explanation it should be assumed that the DTx function is active.

The power of the received signal Ptx on the feedback line 24 and the near-end signal (Pvad) from the VAD 12 are compared. (As already described, the near-call signal power is already calculated on the basis of a noise-filtered signal.) The comparison is made in comparator 26 after the Ptx signal in attenuator 25 has been weakened. The echo flag is set accordingly to:

If Ptx-30 dB <Pvad; ECHO = 1 (echo detected)
If Ptx-30 dB <Pvad; ECHO = 0

Since the echo losses in the normal handset are around 45 dB, the size of 30 dB allows a good distance for deviations in the transmitted and received volumes. If there is a volume control 23 , the threshold value of 30 dB can be set in accordance with the set volume. If the volume is increased, the threshold should be lowered.

The VAD flag, which indicates whether encoded speech frames to be sent will be according to the following Modified formula:

VAD = (VVAD * ˆECHO) ++

in which

* = AND
ˆ = logically not
++ = hangover period

mean.  

As can be seen from the table below, the VAD flag is only set when local traffic speaks or double speech.

If only long-distance (far-end) language is available, the VVAD detect or not detect the echo, depending from the noise and the set volume, what through the "don't care" state zero is displayed.

The EVAD flag is set accordingly to:
EVAD = (VVAD ++). This flag is therefore identical to the original VAD flag, as occurs in the prior art. The SID scanning unit 29 reads the echo and the EVAD flags and decides when the comfort noise parameters are renewed.

The comfort noise parameters are only just after one Speech burst and sampled every 24th frame. Between the renewals of the SID frames, the old ones Samples are repeatedly sent to the radio subsystem. This renewal corresponds to the scan as from that Radio subsystem is running, however No renewal of the SID frames for every 24th frame at the same time. The comfort noise parameters are only renewed when eVAD = 0 and ECHO = 0. This avoids  renewing the parameters during an echo and therefore avoids unnecessary disturbances of the Comfort noise at the receiver end. Around Take into account situations where very loud Long-distance call noise causes the echo bit for long periods are set or to situations take into account where there are no breaks between Long-distance and close-range calls will occur additional strategy used to ensure that the best quality of comfort noise is present.

If eVAD = 1 at the time of sampling, this indicates that the echo was detected by the VAD and that the Comfort noise parameters have not been renewed. If eVAD = 0 and echo = 1, an echo was generated by the Echo canceller detected, but not by the VAD. The Echo can be a very quiet speech or a very loud one Be background noise. The comfort noise parameters will not be renewed unless this situation occurs three times in a row and not by any other Combination of the two flags eVAD and Echo interrupted is repeated. This indicates that the echo Background noise was and that the Comfort noise parameters are renewed as long as eVAD = 0 and echo = 1.

Furthermore, the comfort noise parameters only become even renewed for every 24th frame after a speech burst. Between the renewals of the SID frames, the old ones Samples are repeatedly sent to the radio subsystem. This renewal corresponds to the scan as from that Radio subsystem is running, but with the Difference that the renewals of the SID frames for everyone 24. Frame are not at the same time.  

The system described above, in which the DTx mode is active, is described in the block diagram of FIG. 2. The elements have an echo canceller (attenuator 25 ), a comparator 26 , a logic gate 27 and hangover elements 13 , 28 and are all shown as a single block 30 . The transmission path from the microphone horizontally through the circuit diagram to the radio subsystem 15 shows that the speech frames originate from the speech encoder 11 and that speech frames which are subordinate to SID frames originate from the DTx / tx 14 and the SID scanner 29 .

In Fig. 3 is the arrangement of Fig. 2 with a modification according to the invention as used when the DTx mode is inactive (e.g. during the first years of operation of the GSM system or in remote areas) shown.

Since the DTx mode is inactive, the DTx / rx element 18 has been removed from the part of the diagram showing the base station. In practice, this element can be present and only be inactive. According to the invention, a further element is provided in the mobile unit in the form of a DTx / rx element 40 . This element is easy to implement because it only requires reusing the DTx / rx software in the mobile unit. The DTx is permitted both in the "up-link" direction (mobile unit to the base station) and in the "downlink" (base station to the mobile unit) direction. Element 40 is active when the DTx mode is inactive and inactive when a DTx active signal is received by the base station.

The base station informs the mobile unit whether DTx is for the "up-link" direction is permitted. The base station can use DTx in the "downlink" direction, which  DTx / rx unit in the mobile unit is therefore always activated.

The DTx / rx element 40 receives SID frames from the SID scanners 29 and adds random numbers instead of the missing excitation information. In this way, frames with comfort noise are generated, which can be treated as speech frames. For the purpose of operating the system, the frames, as they are emitted by the DTx / rx 40 , are exclusively voice frames and are transmitted by the radio subsystem 15 to the base station for decoding.

For this purpose of implementation, timing coincidence flags as required for the DTx / rx unit 40 are provided in the SID scanner. The DTx / rx unit 40 interpolates the SID frames.

If the echo canceller works without the DTx mode, prevents the DTx / tx unit from opening the radio path, if an echo is incorrectly received by the VAD unit Language is understood. This eliminates the echo and increases the operating time of the portable handheld device. The Echo canceller prevents the DTx / tx function from Comfort noise parameters updated when there is an echo. This advantage lies in both operating modes of the system in front. The quality of the comfort noise is increased.

The described arrangement cancels echoes in situations from which an echo canceller is not suitable per se is because it takes a considerable amount of calculation time would and due to the low level of echo energy. If the mobile unit in "hands-free" mode (speakerphone mode) is operated, a language switchover with  inserted comfort noise as described, in Connection with an echo canceller, used to to ensure that enough echo losses are achieved.  

List of expressions

Abbreviations
DTx / rx
Discontinuous transmission / reception unit. When a speech frame is received, the DTx sends that frame to the speech decoder. When a SID frame is received, the zeros in the frame are padded with random numbers. The old SID frame is used between the renewals of the SID frame. No information is received from the DTx / rx unit between renewals of the SID frame.
DTx / tx
Discontinuous transmission transmission unit. This unit receives an encoded speech frame from the speech encoder and a VAD flag from the VAD unit. When the VAD flag is set, coded speech is sent. When the VAD flag is cleared, the excitation information is set to zero. The resulting frame is called a SID frame. The SID frames are only transmitted for every 24th frame. The transmitter is switched off for the rest of the time.
ECHO
ECHO flag. This flag indicates whether there is only long distance speech. This tells the system that no speech frames should be transmitted since the frames could contain echoes. The flag is set by comparing PVAD and Ptx.
eVAD
This is a copy of the VVAD flag before it is modified by the echo canceller. This flag tells the echo canceller that the comfort noise should not be renewed because the frame contains speech or an echo. A hangover period is added to the VVAD flag.
Frame
Time interval of 20 msec (160 samples of 13 bit uniformly coded samples - 1280 bits), which corresponds to the time segment of the speech transcoder and is also used as a short expression for a traffic frame.
Ptx
Power of long distance (far end) signal. This signal is not noise filtered.
Pvad
Power of the near-end signal. The near call signal is noise filtered and the power is measured.
SID
Silence descriptor frame. Frame that contains only spectral information. The excitation of the filter is set to zero. The DTx / rx unit recognizes the SID frame and fills it with random numbers instead of the zeros. This corresponds to the excitation of the synthesis filter by means of a random generator. The modified SID frame is sent to the speech decoder repeatedly 24 times until a new SID frame is received.
SPflag
Internal flag in the mobile identifier for the radio subsystem, whether the traffic frame is a voice frame or a SID frame.
TAF
Time alignment flag. If the TAF flag is set, the DTx / rx unit looks for SID frames, the SID frames are only renewed at certain times known to the mobile unit and the base station.
Traffic frame
Coded language block. 260 information bits. In principle, the language is encoded as a synthesis filter (spectrum) and as an excitation of the filter.
VAD
The modified VVAD flag with a hangover period at the end of the speech burst.
VVAD
Flag that indicates whether the frame contains speech. This flag can be cleared by the echo canceller if the speech in the frame is an echo.

Claims (6)

1.Communication transceiver, for communicating with frames encoded audio information, with an audio input path, an audio output path, a VAD (Voice Activity Detector) for detecting speech on the audio input path and an echo detection device for detecting unwanted echoes on the input path, which are acoustic Couplings with the output path result, characterized by a device ( 14 , 29 ) which reacts to the echo detection device to inhibit the transmission of coded audio information from the audio input path during the presence of speech which has been displayed as an echo by the echo detection device. 2. Communication transceiver according to claim 1, further comprising a comfort noise generator ( 40 ) for transmitting comfort noise parameters instead of encoded audio information from the audio input path and for decoding as normal speech at the receiver. 3. Communication transceiver according to claim 2, further comprising a DTx mode activation device for receiving and decoding a DTx mode authorization signal, the comfort noise generating device ( 40 ) reacting to the DTx mode activation device and at no such authorization signal is activated. 4. Communication transceiver according to claim 1, operable in a discontinuous transmission (DTx) inactive mode, in which virtually all audio frames that are transmitted are complete frames of encoded audio information and operable in an active DTx mode, in which SID ( Silence Indicator Descriptor) frames for reproducing periods in which there is no language are transmitted, further comprising:
a device ( 14 ) which can be operated in the inactive DTx mode for generating SID (Silence Indicator Descriptor) parameters in response to the detection of an echo by the echo detector,
a device ( 40 ) for supplementing the SID parameters with comfort noise parameters in order to generate complete frames of coded audio information,
means ( 15 ) for transmitting complete frames of encoded audio information and
means ( 29 ) for repeating the frame-to-frame SID parameters during the presence of an echo and the absence of speech. 5. Communication transmitting / receiving device according to at least one of the preceding claims, wherein the echo detection device comprises a comparator device ( 25 , 26 ) for comparing the signal on the audio input path with the signal on the audio output path ( 22 ), if this by a selectable attenuation has been weakened; wherein a volume control device ( 23 ) is provided for controlling the volume on the output path; and wherein the selectable attenuation of the output used for the comparison depends on the setting of the volume control. 6. Communication transceiver according to claim 2 or 3, further comprising a device for generating of comfort noise parameters and for renewing them during renewal periods a number of frames outlast, further comprising a facility for Suppress the renewal of the comfort noise parameters for a prescribed number of renewal periods while a possible echo from the Echo detection device has been detected and none Speech from which VAD was detected and renewing the Parameters after this specified number of frames.