CN104468471A

CN104468471A - Packet acoustic echo cancellation (PAEC) method and equipment

Info

Publication number: CN104468471A
Application number: CN201310419143.1A
Authority: CN
Inventors: 李舟洲; 蔡亦钢
Original assignee: Alcatel Optical Networks Israel Ltd
Current assignee: Alcatel Optical Networks Israel Ltd
Priority date: 2013-09-13
Filing date: 2013-09-13
Publication date: 2015-03-25
Anticipated expiration: 2033-09-13
Also published as: WO2015036857A1; CN104468471B

Abstract

The invention aims to provide a packet acoustic echo cancellation (PAEC) method and equipment. The method comprises the following steps: acquiring source voice packet streams of both conversation ends to be subjected to PAEC by using the echo cancellation equipment; updating target packet streams corresponding to both conversation ends in a target buffer area according to the source voice packet streams; and performing PAEC on the target packet streams according to reference packet streams corresponding to both conversation ends in a corresponding reference buffer area in combination with transmission direction information corresponding to each packet data packet in the target packet streams and reference packet streams in order to obtain echo-cancelled packet streams, and transmitting the echo-cancelled packet streams to corresponding ends in both conversation ends. Compared with the prior art, the method and the equipment have the advantages that bilateral PAEC is realized; the performance of a PAEC channel is multiplied; and the amount of hardware and corresponding maintenance cost are reduced. Meanwhile, call processing and relevant signaling overheads are reduced; any signaling support is not required; and a transparent PAEC function is provided.

Description

A kind of method and apparatus eliminated for the acoustic echo that divides into groups

Technical field

The present invention relates to the communications field, particularly relate to a kind of technology eliminated for the acoustic echo that divides into groups.

Background technology

Acoustic echo in mobile network is because the design of mobile phone or other hand free devices is not good, and the sound that phonetic incepting side's loud speaker sends is delivered to recipient's microphone (and then sending voice transmit leg back to) and caused.Acoustic echo elimination (Acoustic Echo Cancellation, AEC) can remove the echo in signal of communication.It is the core capabilities ensureing sound quality in communication that acoustic echo is eliminated.

In a circuit switched network, traditional AEC technology removes acoustic echo in waveform territory to be done very well.But, in packet network (voice in such as IP network, VoIP), also there is no the approved mode for performing AEC.Some suppliers (if Broadcom(is with reference to US7333447), Samsung, 3Com etc.) invent AEC for packet network, but this kind of AEC needs first (namely stream of packets to be decoded into analog or digital signal, be transformed into waveform territory), use the echo in conventional art erasure signal, then the signal recompile eliminating echo is returned in grouping and (that is, convert back packet domain).Due to coding/decoding repeatedly, result in the decline of sound quality (voice quality, VQ), thus counteract code conversion exempt operation (Transcoder Free Operation, TrFO) getting rid of repeatedly the advantage that Code And Decode obtains.In addition, due to computation complexity and huge buffer requirement, traditional AEC only supports limited tail long delay, and therefore, traditional AEC is very low for efficiency during voip network.

Alcatel-Lucent/Bell Laboratory (Alcatel-Lucent/Bell Labs) has been invented a kind of real packet domain acoustic echo and has been eliminated (Packet Acoustic Echo Cancellation, PAEC) technology, (such as) describe the parameter of waveform in EVRC or EVRC-B only need being used to divide into groups, just can detect the acoustic echo suppressed in stream of packets.Bell Laboratory has 3 relevant patents or patent application in PAEC field:

-US7852792Packet Based Echo Cancellation and Suppression(granted on12/14/2010)by Binshi Cao et al.

-US008144862method and Apparatus for the Detection andSuppression of Echo in Packet based Communication Networks UsingFrame Energy Estimation(granted on3/27/2012)by Binshi Cao et al.

-US2009/0168673Method and Apparatus for Detecting andSuppressing Echo in Packet Networks(published on7/2/2009)byLampros Kalampoukas and Semyon Sosin.

In above-mentioned patent or patent application, compare and prediction by utilizing the waveform characterising parameter of grouping, be grouped in PAEC channel with reference to flow point group and target stream and compare, thus remove (in target stream) similar grouping (being identified as echo), achieve the basic skills eliminating/suppress grouping acoustic echo in packet network.

But the method provided in these patents or patent application only for unidirectional PAEC, and cannot provide two-way PAEC.An audio call relates to two or more correspondent, eliminate the echo that each correspondent produces, will dispose the multiple unidirectional PAEC channel on multiple unidirectional PAEC equipment or single PAEC equipment.From packet-switching performance and capacity analysis, especially in packet switching network inner exchanging scene, the limited capacity of unidirectional PAEC, and industrial quality and performance standard may do not reached.PAEC product with unidirectional grouping echo cancellor may can not meet the user's request in packet switching well.Therefore, dispose for actual industry, these unidirectional PAEC methods all have shortcoming and restriction.

Such as, Fig. 1 illustrates and a kind ofly in US2009/0168673, has described unidirectional grouping acoustic echo to eliminate structure.A unidirectional PAEC channel can only distribute to a correspondent, it need to distinguish the direction of voice flow be " going to " or " from " this correspondent.If go to this correspondent, this voice flow is a reference stream.If from this correspondent, this voice flow is a target stream.Described voice flow or conduct with reference to packet transaction operation, or run as targeted packets process.Key be reference stream processing section different with target stream processing section time parallel running.

The obvious shortcoming of this unidirectional grouping acoustic echo removing method is that efficiency low cost is high.Although have the voice flow that reference packet processing module can cushion other direction, it and the not parallel running of targeted packets processing module, do not eliminate the echo in reference packet yet.Realize two-way grouping echo cancellor, still need to provide two PAEC channels and the signaling doubled and management maintenance expense.In a packet switching network inner exchanging scene, this is undoubtedly the waste to resource.

Summary of the invention

The object of this invention is to provide a kind of method and apparatus eliminated for the acoustic echo that divides into groups.

According to an aspect of the present invention, provide a kind of method eliminated for the acoustic echo that divides into groups, wherein, the method comprises the following steps:

A obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package;

B is according to described source stream of voice packets, and upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream;

C is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream;

Described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by d.

According to a further aspect in the invention, additionally provide a kind of echo cancellation devices eliminated for the acoustic echo that divides into groups, wherein, this equipment comprises:

Acquisition device, for obtaining the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, wherein, described source stream of voice packets comprises one or more packet data package;

Target update device, for according to described source stream of voice packets, upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream;

Cancellation element, for the reference packet stream according to described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream;

Dispensing device, for according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by described echo stream of packets of having eliminated.

Compared with prior art, the present invention by obtaining the source stream of voice packets of the call ends of pending grouping acoustic echo elimination in echo cancellation devices, according to described source stream of voice packets, upgrade the targeted packets stream of corresponding described call ends in destination buffer, according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream, the direction of transfer information corresponding to echo stream of packets has been eliminated described in last basis, described echo stream of packets of having eliminated is sent to corresponding end in described call ends, thus achieve the elimination of two-way grouping acoustic echo, improve the performance of PAEC channel exponentially, decrease hardware quantity and corresponding maintenance cost, decrease call treatment and related signaling expense simultaneously, and then do not need any signaling support, transparent PAEC function is provided.

Accompanying drawing explanation

By reading the detailed description done non-limiting example done with reference to the following drawings, other features, objects and advantages of the present invention will become more obvious:

Fig. 1 illustrates in US2009/0168673, have described unidirectional grouping acoustic echo to eliminate structural representation according to one aspect of the invention a kind of;

Fig. 2 illustrates a kind of echo cancellation devices schematic diagram eliminated for the acoustic echo that divides into groups according to one aspect of the invention;

Fig. 3 illustrates a kind of echo cancellation devices schematic diagram eliminated for the acoustic echo that divides into groups in accordance with a preferred embodiment of the present invention;

Fig. 4 illustrates a kind of method flow diagram eliminated for the acoustic echo that divides into groups according to a further aspect of the present invention;

Fig. 5 illustrates a kind of method flow diagram eliminated for the acoustic echo that divides into groups in accordance with a preferred embodiment of the present invention;

Fig. 6 illustrates that the two-way grouping acoustic echo of one according to a preferred embodiment of the present invention eliminates reference view, and wherein, the packet data package in each direction is as the reference of another direction packet data package;

Fig. 7 illustrates that the two-way grouping acoustic echo of one according to a preferred embodiment of the present invention eliminates reference view, and wherein, the packet data package eliminating echo in each direction is as the reference of the packet data package in another direction;

Fig. 8 buffering that to illustrate according to a preferred embodiment of the present invention a kind of utilizes non-echo cancellor packet data package two-way grouping acoustic echo as a reference to eliminate and compare schematic diagram;

Fig. 9 buffering that to illustrate according to a preferred embodiment of the present invention a kind of utilizes echo cancellor packet data package two-way grouping acoustic echo as a reference to eliminate and compare schematic diagram;

Figure 10 illustrates the comparison of a kind of echo frame for A end according to a preferred embodiment of the present invention and removes algorithm;

Figure 11 illustrates the comparison of a kind of echo frame for B end according to a preferred embodiment of the present invention and removes algorithm.

In accompanying drawing, same or analogous Reference numeral represents same or analogous parts.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail.

Fig. 2 illustrates a kind of echo cancellation devices schematic diagram eliminated for the acoustic echo that divides into groups according to one aspect of the invention; Wherein, described echo cancellation devices comprises acquisition device 1, target update device 2, cancellation element 3, dispensing device 4.Particularly, acquisition device 1 obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package; Target update device 2 is according to described source stream of voice packets, and upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream; Cancellation element 3 is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream; Described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by dispensing device 4.

At this, described echo cancellation devices includes but not limited to according to setting in advance or the instruction stored, can automatically carry out electronic hardware or the software equipment of numerical computations and information processing; Wherein, described hardware device includes but not limited to microprocessor, application-specific integrated circuit (ASIC) (ASIC), programmable gate array (FPGA), digital processing unit (DSP), embedded device etc.Those skilled in the art will be understood that other echo cancellation devices is equally applicable to the present invention, within also should being included in scope, and are contained in this at this with way of reference.

Described echo cancellation devices can be used in any VOIP network, Real Time Communication Network RTC and LTE/EPC network, and above-mentioned network does not also have effectively and the grouping acoustic echo abatement apparatus of generally acknowledging current.

Constant work between above-mentioned each device, at this, it will be understood by those skilled in the art that " continuing " refers to above-mentioned each device respectively in real time, or according to the mode of operation requirement of setting or real-time adjustment, carry out the acquisition of the source stream of voice packets of call ends, targeted packets stream renewal, eliminate the acquisition of echo stream of packets, eliminate the transmission etc. of echo stream of packets, until described echo cancellation devices stops the source stream of voice packets obtaining the call ends that pending grouping acoustic echo is eliminated.

Acquisition device 1 obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package.

Particularly, described acquisition device 1, from the call ends of carrying out conversing (for communicating end A and communicating end B), obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated; Wherein, described source stream of voice packets comprises the source stream of voice packets from communicating end A to communicating end B, also comprises the source stream of voice packets from communicating end B to communicating end A.Wherein, in the stream of voice packets of described source, comprise one or more packet data package (packet), and echo bag in the packet data package of described source stream of voice packets, may be comprised.

Target update device 2 is according to described source stream of voice packets, and upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream.

Particularly, described target update device 2 is according to the source stream of voice packets obtained in described acquisition device 1, by source stream of voice packets is sent to destination buffer, thus utilize described source stream of voice packets to upgrade the targeted packets stream in destination buffer, wherein, due to the stream of voice packets that source stream of voice packets is the call ends that pending grouping acoustic echo is eliminated, therefore, in described targeted packets stream, also comprise the stream of voice packets corresponding to described call ends.At this, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream.

Preferably, described target update device 2 according to described source stream of voice packets, can determine the direction of transfer information corresponding to each packet data package in the stream of voice packets of described source; According to described source stream of voice packets, in conjunction with the direction of transfer information of the packet data package in the stream of voice packets of described source, upgrade the targeted packets stream of corresponding described call ends in destination buffer.

Particularly, described target update device 2 can according to described source stream of voice packets, by according to the source address in the header packet information of each packet in the stream of voice packets of described source and destination address, calculate the direction of transfer information determined corresponding to each packet data package.

Such as, for communicating end A and communicating end B, call ends is described, then described direction of transfer information comprises from A to B or from B to A, if the address of known communicating end A and/or the address of communicating end B, then according to the source address in the header packet information of described packet and destination address, the direction of transfer information corresponding to described packet directly can be determined;

Or, such as, by utilizing predetermined computing function, the source address in the header packet information of described packet and destination address are compared, if source address is greater than destination address, then determine that the direction of transfer of described packet is for from A to B, otherwise, if source address is less than destination address, then determine that the direction of transfer of described packet is for from B to A, if there is other situations, then occur mistake, this packet is dropped.

Described target update device 2 is according to described source stream of voice packets, in conjunction with the direction of transfer information of the packet data package in the stream of voice packets of described source, upgrade the targeted packets stream of corresponding described call ends in destination buffer, therefore, described targeted packets stream comprises the targeted packets stream from A to B and the targeted packets stream from B to A.

Preferably, described target update device 2 according to described source stream of voice packets, can upgrade the targeted packets stream of corresponding described call ends in destination buffer; According to described targeted packets stream, determine the direction of transfer information corresponding to each packet data package in described targeted packets stream.

Particularly, described target update device 2 according to described source stream of voice packets, first can upgrade the targeted packets stream of described call ends corresponding in destination buffer; And then according to described targeted packets stream, by according to the source address in the header packet information of each packet in described targeted packets stream and destination address, calculate the direction of transfer information corresponding to each packet data package determined in described targeted packets stream.At this, described computational methods and described target update device 2 are according to described source stream of voice packets, determine that the method for the direction of transfer information corresponding to each packet data package in the stream of voice packets of described source is same or similar, therefore do not repeat them here, and be contained in this by way of reference.

Cancellation element 3 is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

Particularly, described cancellation element 3 obtains the reference packet stream of corresponding described call ends in the reference buffer district corresponding with described destination buffer, wherein, described reference packet stream can be determined according to the stream of voice packets with echo bag of source stream of voice packets, or the stream of voice packets not comprising echo bag after can carrying out the elimination of grouping acoustic echo according to described source stream of voice packets determined, the direction of transfer information of described cancellation element 3 corresponding to each packet data package in described targeted packets stream and described reference packet stream, the described targeted packets stream of different directions and described reference packet stream are contrasted, such as, by from A end to B end targeted packets stream with hold the reference packet stream held to A to contrast from B, or by from B end to A end targeted packets stream with hold the reference packet stream held to B to contrast from A, detect in described targeted packets stream whether comprise echo bag based on grouping acoustic echo cancellation algorithm (PAEC algorithm), if comprise echo bag, then by deleting described echo bag or utilizing replacement bag to carry out to detected echo bag the mode of replacing etc., echo cancellor is carried out to described targeted packets stream.Particularly, such as, replacement bag is utilized to replace detected echo bag, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.Wherein, described replacement bag includes but not limited to noise bag (such as, comprising the grouping of the noise of certain type, such as white noise, comfort noise etc.), noiseless bag (such as, empty grouping), in targeted packets stream 1/8th rate packet etc. of last buffer memory, and to mix.

At this, the defining method of the direction of transfer information corresponding to each packet data package in described reference packet stream, same or similar with the method for the direction of transfer information corresponding to each packet data package determined in the stream of voice packets of described source, therefore do not repeat them here, and be contained in this by way of reference.

Described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by dispensing device 4.

Particularly, described dispensing device 4 is according to the described direction of transfer information eliminated corresponding to echo stream of packets, such as according to the described destination address information having eliminated echo stream of packets, or according to communicating end information corresponding in described direction of transfer information, described echo stream of packets of having eliminated is sent to and the corresponding end corresponding to the described source having eliminated echo stream of packets.

Such as, the direction of transfer information eliminated if described corresponding to echo stream of packets is that A end is held to B, then described echo stream of packets of having eliminated is sent to B end, and at this, B end is the corresponding end of A end.

Thus, present invention achieves a kind of two-way grouping acoustic echo removing method, the method:

-reduce hardware quantity and corresponding maintenance cost: compared with unidirectional PAEC, two-way PAEC hsrdware requirements reduce by half and have saved relevant maintenance;

-decrease call treatment and signaling consumption: for substantially conversing, only need distribution PAEC channel;

-realizing the PAEC of the implicit expression/transparent supported without any signaling: the gateway in packet voice (transmission) path can integrate two-way PAEC, thinks that side a and b provides the PAEC of implicit expression/transparent.

Preferably, described source stream of voice packets can be sent to described destination buffer and reference buffer district by described target update device 2 respectively, to upgrade the targeted packets stream of corresponding described call ends in described mark buffering area, and the reference packet stream of corresponding described call ends in described reference buffer district, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream, and described reference packet stream comprises the direction of transfer information corresponding to each packet data package in described reference packet stream.

Particularly, described target update device 2 is according to the source stream of voice packets obtained in described acquisition device 1, described source stream of voice packets is sent to described destination buffer and reference buffer district respectively, utilize described source stream of voice packets, the targeted packets stream in described destination buffer and the reference packet stream in reference buffer district are upgraded; Wherein, because source stream of voice packets is the stream of voice packets of call ends that pending grouping acoustic echo is eliminated, therefore, the stream of voice packets corresponding to described call ends is included in described targeted packets stream and reference packet stream.At this, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream, and described reference packet stream comprises the direction of transfer information corresponding to each packet data package in described reference packet stream.

At this, the defining method of the direction of transfer information of the packet data package in described targeted packets stream and reference packet stream, same or similar with the method for the direction of transfer information corresponding to each packet data package determined in the stream of voice packets of described source, therefore do not repeat them here, and be contained in this by way of reference.

Such as, Fig. 6 illustrates that the two-way grouping acoustic echo of one according to a preferred embodiment of the present invention eliminates reference view, and wherein, the packet data package in each direction is as the reference of another direction packet data package.

Particularly, the source stream of voice packets come from A end and/or B end is all sent in reference packet process and targeted packets process by RTP resolver, in the buffering area (destination buffer and reference buffer district) be separated, cushion targeted packets stream and reference packet stream.At this, the source stream of voice packets that RTP resolver sends comprises load and the head of the packet data package of described source stream of voice packets.Wherein, to hold the source stream of voice packets of sending from A or hold echo with B or do not comprise echo, hold the source stream of voice packets of sending or with A end echo or do not comprise echo from B.Because described targeted packets stream determined by source stream of voice packets described in buffer memory, therefore, if comprise echo in the stream of voice packets of described source, then also comprise corresponding echo in described targeted packets stream; If do not comprise echo in the packets of voice of described source, then do not comprise corresponding echo in described targeted packets stream yet.

In described destination buffer and described reference buffer district, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream, and described reference packet stream comprises the direction of transfer information corresponding to each packet data package in described reference packet stream.

In described PAEC algoritic module, the targeted packets stream in a direction in described destination buffer, contrast with the reference packet stream of the other direction prestored in described reference buffer district, as shown in Figure 8, in targeted packets stream, (packet j is to packet j+M in packet set, namely B end is to the targeted packets stream of A extreme direction) corresponding with reference packet stream respectively set 1, set 2, set K(and A end is to the stream of voice packets of B end, for carrying out the reference of B end to A extreme direction) contrast, in targeted packets stream, (packet i is to packet i+N in packet set, namely A end is to the targeted packets stream of B extreme direction) corresponding with reference packet stream respectively set 1, set 2, set Q(and B end is to the stream of voice packets of A end, for carrying out the reference of A end to B extreme direction) contrast, whether to have echo bag in the described targeted packets stream determining different directions.Wherein, corresponding echo bag is comprised in described reference packet stream.

If there is echo bag in described targeted packets stream, then described PAEC algoritic module carries out grouping acoustic echo elimination calculating to it, the echo of the elimination stream of packets after elimination echo is sent to A end respectively and holds with B.

Preferably, described echo cancellation devices also comprises with reference to updating device (not shown), and wherein, described reference updating device can eliminate echo stream of packets according to described, upgrades the reference packet stream in described reference buffer district.

Particularly, described reference updating device can interact with described cancellation element 3, described in obtaining, eliminated echo stream of packets; Then, described reference updating device has eliminated echo stream of packets according to described, upgrades the reference packet stream in described reference buffer district; Thus eliminated the reference packet stream that echo stream of packets is used as comparing with described targeted packets stream described in utilizing, decrease the use to buffering area, have better with reference to effect, thus further increase the accuracy rate of PAEC.

Such as, Fig. 7 illustrates that the two-way grouping acoustic echo of one according to a preferred embodiment of the present invention eliminates reference view, and wherein, the packet data package eliminating echo in each direction is as the reference of the packet data package in another direction.

Particularly, the source stream of voice packets come from A end and/or B end is all sent in targeted packets process by RTP resolver, and described targeted packets stream determined by source stream of voice packets described in buffer memory, therefore, also comprises corresponding echo in described targeted packets stream.At this, the source stream of voice packets that RTP resolver sends comprises load and the head of the packet data package of described source stream of voice packets.Wherein, to hold the source stream of voice packets of sending from A or hold echo with B or do not comprise echo, hold the source stream of voice packets of sending or with A end echo or do not comprise echo from B.

Reference packet process and PAEC algoritic module interact, and to obtain, described PAEC algoritic module is determined has eliminated echo stream of packets, and described echo stream of packets of having eliminated is buffered to described reference buffer district, using as described reference packet stream.

At this, each packet data package in described targeted packets stream and described reference packet stream comprises the direction of transfer information corresponding to it.

In described PAEC algoritic module, the targeted packets stream in a direction in described destination buffer, contrast with the reference packet stream of the other direction prestored in described reference buffer district, as shown in Figure 9, in targeted packets stream, (packet j is to packet j+M in packet set, namely B end is to the targeted packets stream of A extreme direction) corresponding with reference packet stream respectively set 1, set 2, set K(and A end is to the stream of voice packets of B end, for carrying out the reference of B end to A extreme direction) contrast, in targeted packets stream, (packet i is to packet i+N in packet set, namely A end is to the targeted packets stream of B extreme direction) corresponding with reference packet stream respectively set 1, set 2, set Q(and B end is to the stream of voice packets of A end, for carrying out the reference of A end to B extreme direction) contrast, whether to have echo bag in the described targeted packets stream determining different directions.Wherein, no longer comprise corresponding echo bag in described reference packet stream, belong to and eliminate echo stream of packets.

At this, composition graphs 8 or Fig. 9, Figure 10 and Figure 11 respectively illustrate comparing and removing algorithm of a kind of echo frame held with B for A end.

Particularly, in Fig. 10, " N+1 " is the target window size for direction A to B, and " N+Q " is corresponding reference windows size." Q " is by determined according to the echo path delay of B end. represent the N+1(i in the destination buffer from A to B, i+1 ..., i+N) and N+1(q in individual grouping and the reference buffer district from A to B, q+1 ..., q+N) and the comparing result of individual grouping.At this, those skilled in the art will be understood that the described direction of transfer information for the reference packet stream as targeted packets stream A to B should be from B to A. (q=q, q+1 ..., q+Q-1) minimum value will with minimum threshold e _tHcompare, to determine whether there is echo.

E_{i}^{A} = [e_{i, q}^{A}, e_{i, q + 1}^{A}, . . ., e_{i, q + Q - 1}^{A}]

(formula 1)

e_{i, q}^{A} = \sqrt{{Σ_{n = 0}^{N} Σ_{&PartialD; = 1}^{P} (l_{i + n, &PartialD;}^{A} - l_{q + n, &PartialD;}^{B})}^{2}}

(formula 2)

Result minimum value in formula 2 represents the similitude of direction A to B target stream and corresponding reference stream; If the result of formula 2 meets following formula:

\min e_{i, q}^{A} \leq e_{TH}

(formula 3)

Then illustrate in targeted packets stream and reference packet stream to there is similitude, therefore comprise echo in targeted packets stream.

In fig. 11, " M+1 " is the target window size for direction B to A, and " M+K " is corresponding reference windows size." K " is by determined according to the echo path delay of A end. represent the M+1(j in the destination buffer from B to A, j+1 ..., j+M) and M+1(k in individual grouping and the reference buffer district from B to A, k+1 ..., k+M) and the comparing result of individual grouping.At this, those skilled in the art will be understood that the described direction of transfer information for the reference packet stream as targeted packets stream B to A should be from A to B. (k=k, k+1 ..., k+Q-1) minimum value will with minimum threshold e _tHcompare, to determine whether there is echo.

E_{j}^{B} = [e_{j, k}^{B}, e_{j, k + 1}^{B}, . . ., e_{j, k + K - 1}^{B}]

(formula 4)

e_{j, k}^{B} = \sqrt{{Σ_{m = 0}^{M} Σ_{&PartialD; = 1}^{P} (l_{j + m, &PartialD;}^{B} - l_{k + m, &PartialD;}^{A})}^{2}}

(formula 5)

Result minimum value in formula 5 represents the similitude of direction B to A target stream and corresponding reference stream; If the result of formula 5 meets following formula:

\min e_{j, k}^{B} \leq e_{TH}

(formula 6)

At this, P is the value (Line Spectral Pair, line spectrum pair) of LSP.

Fig. 3 illustrates a kind of echo cancellation devices schematic diagram eliminated for the acoustic echo that divides into groups in accordance with a preferred embodiment of the present invention; Wherein, described echo cancellation devices comprises acquisition device 1 ', target update device 2 ', cancellation element 3 ', dispensing device 4 ', and wherein, described cancellation element 3 ' comprises echo determining unit 31 ', echo cancellation unit 32 '.Particularly, acquisition device 1 ' obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package; Target update device 2 ' is according to described source stream of voice packets, and upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream; Echo determining unit 31 ' is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag; Echo cancellation unit 32 ', when comprising echo bag in described targeted packets stream, carries out echo cancellor to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream; Described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by dispensing device 4 '.

Wherein, shown in the acquisition device 1 ' of described echo cancellation devices, target update device 2 ', dispensing device 4 ' with Fig. 2, corresponding intrument is identical or substantially identical, so place repeats no more, and is contained in this by way of reference.

Constant work between above-mentioned each device, at this, it will be understood by those skilled in the art that " continuing " refers to above-mentioned each device respectively in real time, or according to the mode of operation requirement of setting or real-time adjustment, carry out the acquisition of the source stream of voice packets of call ends, targeted packets stream renewal, whether comprise echo bag determination, eliminate the acquisition of echo stream of packets, eliminate the transmission etc. of echo stream of packets, until described echo cancellation devices stops the source stream of voice packets obtaining the call ends that pending grouping acoustic echo is eliminated.

Echo determining unit 31 ' is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag.

Particularly, described echo determining unit 31 ' obtains the reference packet stream of corresponding described call ends in the reference buffer district corresponding with described destination buffer, wherein, described reference packet stream can be determined according to the stream of voice packets with echo bag of source stream of voice packets, or the stream of voice packets not comprising echo bag after can carrying out the elimination of grouping acoustic echo according to described source stream of voice packets determined; The direction of transfer information of described echo determining unit 31 ' corresponding to each packet data package in described targeted packets stream and described reference packet stream, the described targeted packets stream of different directions and described reference packet stream are contrasted, such as, by from A end to B end targeted packets stream with hold the reference packet stream held to A to contrast from B, or by from B end to A end targeted packets stream with hold reference packet stream hold to B to contrast from A, based on divide into groups acoustic echo cancellation algorithm (PAEC algorithm) detect in described targeted packets stream whether comprise echo wrap.

Preferably, described echo determining unit 31 ' is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, and the energy hierarchical information corresponding with multiple consecutive packet data package corresponding in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag.

Particularly, according to Figure 10, Figure 11 and formula 1 to formula 6, the invention provides a kind of based on LSP(Line Spectral Pair, line spectrum pair) carry out the method for the detection of echoes of packet voice, if described echo determining unit 31 ' is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, utilize the method, determine that described targeted packets stream has similar packet to reference packet stream, then described echo determining unit 31 ' can further combined with the described targeted packets stream energy hierarchical information (i.e. all kinds of gain (gain) information) corresponding with multiple consecutive packet data package corresponding in described reference packet stream, judge whether the similar packet in described targeted packets stream exists decay, namely energy hierarchical information is lower than the energy hierarchical information of the described reference packet stream of correspondence, if exist, then prove that described similar packet is echo bag, then comprise echo bag in described targeted packets stream.

This is due in echo, and echo energy generally has decay to a certain degree than original words sound, thus energy level is compared the subsidiary conditions as detecting echo bag.

Echo cancellation unit 32 ', when comprising echo bag in described targeted packets stream, carries out echo cancellor to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

Particularly, if comprise echo bag in described targeted packets stream, echo cancellation unit 32 ' then carries out echo cancellor to described targeted packets stream, such as, utilize to replace to wrap and detected echo bag is replaced, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.Wherein, described replacement bag includes but not limited to noise bag (such as, comprising the grouping of the noise of certain type, such as white noise, comfort noise etc.), noiseless bag (such as, empty grouping), in targeted packets stream 1/8th rate packet etc. of last buffer memory, and to mix.

At this, when utilizing with during to the replacement bag of fixed load, need correspondingly to revise RTP head and other length relevant field and verifications, such as, the specific head of amendment platform, IP head, UDP head, RTP head.

Fig. 4 illustrates a kind of method flow diagram eliminated for the acoustic echo that divides into groups according to a further aspect of the present invention.Particularly, in step s1, echo cancellation devices obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package; In step s2, echo cancellation devices is according to described source stream of voice packets, upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream; In step s3, echo cancellation devices is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream; In step s4, described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by echo cancellation devices.

Constant work between above steps, at this, it will be understood by those skilled in the art that " continuing " refers to that above steps respectively in real time, or according to the mode of operation requirement of setting or real-time adjustment, carry out the acquisition of the source stream of voice packets of call ends, targeted packets stream renewal, eliminate the acquisition of echo stream of packets, eliminate the transmission etc. of echo stream of packets, until described echo cancellation devices stops the source stream of voice packets obtaining the call ends that pending grouping acoustic echo is eliminated.

In step s1, echo cancellation devices obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package.

Particularly, in step s1, echo cancellation devices, from the call ends of carrying out conversing (for communicating end A and communicating end B), obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated; Wherein, described source stream of voice packets comprises the source stream of voice packets from communicating end A to communicating end B, also comprises the source stream of voice packets from communicating end B to communicating end A.Wherein, in the stream of voice packets of described source, comprise one or more packet data package (packet), and echo bag in the packet data package of described source stream of voice packets, may be comprised.

In step s2, echo cancellation devices is according to described source stream of voice packets, upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream.

Particularly, in step s2, echo cancellation devices is according to the source stream of voice packets obtained in described step s1, by source stream of voice packets is sent to destination buffer, thus utilize described source stream of voice packets to upgrade the targeted packets stream in destination buffer, wherein, because source stream of voice packets is the stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, therefore, the stream of voice packets corresponding to described call ends is also comprised in described targeted packets stream.At this, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream.

Preferably, in step s2, echo cancellation devices according to described source stream of voice packets, can determine the direction of transfer information corresponding to each packet data package in the stream of voice packets of described source; According to described source stream of voice packets, in conjunction with the direction of transfer information of the packet data package in the stream of voice packets of described source, upgrade the targeted packets stream of corresponding described call ends in destination buffer.

Particularly, in step s2, echo cancellation devices according to described source stream of voice packets, by according to the source address in the header packet information of each packet in the stream of voice packets of described source and destination address, can calculate the direction of transfer information determined corresponding to each packet data package.

In step s2, echo cancellation devices is according to described source stream of voice packets, in conjunction with the direction of transfer information of the packet data package in the stream of voice packets of described source, upgrade the targeted packets stream of corresponding described call ends in destination buffer, therefore, described targeted packets stream comprises the targeted packets stream from A to B and the targeted packets stream from B to A.

Preferably, in step s2, echo cancellation devices according to described source stream of voice packets, can upgrade the targeted packets stream of corresponding described call ends in destination buffer; According to described targeted packets stream, determine the direction of transfer information corresponding to each packet data package in described targeted packets stream.

Particularly, in step s2, echo cancellation devices according to described source stream of voice packets, first can upgrade the targeted packets stream of described call ends corresponding in destination buffer; And then according to described targeted packets stream, by according to the source address in the header packet information of each packet in described targeted packets stream and destination address, calculate the direction of transfer information corresponding to each packet data package determined in described targeted packets stream.At this, according to described source stream of voice packets in described computational methods and described step s2, determine that the method for the direction of transfer information corresponding to each packet data package in the stream of voice packets of described source is same or similar, therefore do not repeat them here, and be contained in this by way of reference.

In step s3, echo cancellation devices is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

Particularly, in step s3, echo cancellation devices obtains the reference packet stream of corresponding described call ends in the reference buffer district corresponding with described destination buffer, wherein, described reference packet stream can be determined according to the stream of voice packets with echo bag of source stream of voice packets, or the stream of voice packets not comprising echo bag after can carrying out the elimination of grouping acoustic echo according to described source stream of voice packets determined, in step s3, the direction of transfer information of echo cancellation devices corresponding to each packet data package in described targeted packets stream and described reference packet stream, the described targeted packets stream of different directions and described reference packet stream are contrasted, such as, by from A end to B end targeted packets stream with hold the reference packet stream held to A to contrast from B, or by from B end to A end targeted packets stream with hold the reference packet stream held to B to contrast from A, detect in described targeted packets stream whether comprise echo bag based on grouping acoustic echo cancellation algorithm (PAEC algorithm), if comprise echo bag, then by deleting described echo bag or utilizing replacement bag to carry out to detected echo bag the mode of replacing etc., echo cancellor is carried out to described targeted packets stream.Particularly, such as, replacement bag is utilized to replace detected echo bag, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.Wherein, described replacement bag includes but not limited to noise bag (such as, comprising the grouping of the noise of certain type, such as white noise, comfort noise etc.), noiseless bag (such as, empty grouping), in targeted packets stream 1/8th rate packet etc. of last buffer memory, and to mix.

In step s4, described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by echo cancellation devices.

Particularly, in step s4, echo cancellation devices is according to the described direction of transfer information eliminated corresponding to echo stream of packets, such as according to the described destination address information having eliminated echo stream of packets, or according to communicating end information corresponding in described direction of transfer information, described echo stream of packets of having eliminated is sent to and the corresponding end corresponding to the described source having eliminated echo stream of packets.

Preferably, in step s2, described source stream of voice packets can be sent to described destination buffer and reference buffer district by echo cancellation devices respectively, to upgrade the targeted packets stream of corresponding described call ends in described mark buffering area, and the reference packet stream of corresponding described call ends in described reference buffer district, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream, and described reference packet stream comprises the direction of transfer information corresponding to each packet data package in described reference packet stream.

Particularly, in step s2, echo cancellation devices is according to the source stream of voice packets obtained in described step s1, described source stream of voice packets is sent to described destination buffer and reference buffer district respectively, utilize described source stream of voice packets, the targeted packets stream in described destination buffer and the reference packet stream in reference buffer district are upgraded; Wherein, because source stream of voice packets is the stream of voice packets of call ends that pending grouping acoustic echo is eliminated, therefore, the stream of voice packets corresponding to described call ends is included in described targeted packets stream and reference packet stream.At this, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream, and described reference packet stream comprises the direction of transfer information corresponding to each packet data package in described reference packet stream.

Particularly, the source stream of voice packets come from A end and/or B end is sent in reference packet process and targeted packets process by RTP resolver simultaneously, in the buffering area (destination buffer and reference buffer district) be separated, cushion targeted packets stream and reference packet stream.At this, the source stream of voice packets that RTP resolver sends comprises load and the head of the packet data package of described source stream of voice packets.Wherein, to hold the source stream of voice packets of sending from A or hold echo with B or do not comprise echo, hold the source stream of voice packets of sending or with A end echo or do not comprise echo from B.Because described targeted packets stream determined by source stream of voice packets described in buffer memory, therefore, if comprise echo in the stream of voice packets of described source, then also comprise corresponding echo in described targeted packets stream; If do not comprise echo in the packets of voice of described source, then do not comprise corresponding echo in described targeted packets stream yet.

Preferably, it is not shown that described echo cancellation devices also comprises step s5(), wherein, in step s5, echo cancellation devices can eliminate echo stream of packets according to described, upgrades the reference packet stream in described reference buffer district.

Particularly, in step s5, echo cancellation devices can interact with described step s3, described in obtaining, eliminated echo stream of packets; Then, in step s5, echo cancellation devices has eliminated echo stream of packets according to described, upgrades the reference packet stream in described reference buffer district; Thus eliminated the reference packet stream that echo stream of packets is used as comparing with described targeted packets stream described in utilizing, decrease the use to buffering area, have better with reference to effect, thus further increase the accuracy rate of PAEC.

Such as, Fig. 7 illustrates that the two-way grouping acoustic echo of one according to a preferred embodiment of the present invention eliminates reference view, and wherein, the packet data package eliminating echo in each direction is as the reference of the reflective packet data package of another direction band.

Particularly, the source stream of voice packets come from A end and/or B end is sent in targeted packets process by RTP resolver, and described targeted packets stream determined by source stream of voice packets described in buffer memory, therefore, also comprises corresponding echo in described targeted packets stream.At this, the source stream of voice packets that RTP resolver sends comprises load and the head of the packet data package of described source stream of voice packets.Wherein, to hold the source stream of voice packets of sending from A or hold echo with B or do not comprise echo, hold the source stream of voice packets of sending or with A end echo or do not comprise echo from B.

E_{i}^{A} = [e_{i, q}^{A}, e_{i, q + 1}^{A}, . . ., e_{i, q + Q - 1}^{A}]

(formula 7)

e_{i, q}^{A} = \sqrt{{Σ_{n = 0}^{N} Σ_{&PartialD; = 1}^{P} (l_{i + n, &PartialD;}^{A} - l_{q + n, &PartialD;}^{B})}^{2}}

(formula 8)

Result minimum value in formula 8 represents the similitude of direction A to B target stream and corresponding reference stream; If the result of formula 8 meets following formula:

\min e_{i, q}^{A} \leq e_{TH}

(formula 9)

E_{j}^{B} = [e_{j, k}^{B}, e_{j, k + 1}^{B}, . ., e_{j, k + K - 1}^{B}]

(formula 10)

e_{j, k}^{B} = \sqrt{{Σ_{m = 0}^{M} Σ_{&PartialD; = 1}^{P} (l_{j + m, &PartialD;}^{B} - l_{k + m, &PartialD;}^{A})}^{2}}

(formula 11)

Result minimum value in formula 11 represents the similitude of direction B to A target stream and corresponding reference stream; If the result of formula 11 meets following formula:

{\min e}_{j, k}^{B} \leq e_{TH}

(formula 12)

At this, P is the value (Line Spectral Pair, line spectrum pair) of LSP.

Fig. 5 illustrates a kind of method flow diagram eliminated for the acoustic echo that divides into groups in accordance with a preferred embodiment of the present invention.Particularly, in step s1 ', echo cancellation devices obtains the source stream of voice packets of the call ends that pending grouping acoustic echo is eliminated, and wherein, described source stream of voice packets comprises one or more packet data package; In step s2 ', echo cancellation devices is according to described source stream of voice packets, upgrade the targeted packets stream of corresponding described call ends in destination buffer, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream; In step s31 ', echo cancellation devices is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag; In step s32 ', echo cancellation devices, when comprising echo bag in described targeted packets stream, carries out echo cancellor to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream; In step s4 ', described echo stream of packets of having eliminated, according to the described direction of transfer information eliminated corresponding to echo stream of packets, is sent to corresponding end in described call ends by echo cancellation devices.

Wherein, shown in step s1 ', the step s2 ' of described method, step s4 ' with Fig. 4, corresponding step is identical or substantially identical, so place repeats no more, and is contained in this by way of reference.

Constant work between above steps, at this, it will be understood by those skilled in the art that " continuing " refers to that above steps respectively in real time, or according to the mode of operation requirement of setting or real-time adjustment, carry out the acquisition of the source stream of voice packets of call ends, targeted packets stream renewal, whether comprise echo bag determination, eliminate the acquisition of echo stream of packets, eliminate the transmission etc. of echo stream of packets, until described echo cancellation devices stops the source stream of voice packets obtaining the call ends that pending grouping acoustic echo is eliminated.

In step s31 ', echo cancellation devices is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag.

Particularly, in step s31 ', echo cancellation devices obtains the reference packet stream of corresponding described call ends in the reference buffer district corresponding with described destination buffer, wherein, described reference packet stream can be determined according to the stream of voice packets with echo bag of source stream of voice packets, or the stream of voice packets not comprising echo bag after can carrying out the elimination of grouping acoustic echo according to described source stream of voice packets determined, in step s31 ', the direction of transfer information of echo cancellation devices corresponding to each packet data package in described targeted packets stream and described reference packet stream, the described targeted packets stream of different directions and described reference packet stream are contrasted, such as, by from A end to B end targeted packets stream with hold the reference packet stream held to A to contrast from B, or by from B end to A end targeted packets stream with hold the reference packet stream held to B to contrast from A, detect in described targeted packets stream whether comprise echo bag based on grouping acoustic echo cancellation algorithm (PAEC algorithm).

Preferably, in step s31 ', echo cancellation devices is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, and the energy hierarchical information corresponding with multiple consecutive packet data package corresponding in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag.

Particularly, according to Figure 10, Figure 11 and formula 7 to formula 12, the invention provides a kind of based on LSP(Line Spectral Pair, line spectrum pair) carry out the method for the detection of echoes of packet voice, if described echo determining unit 31 ' is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, utilize the method, determine that described targeted packets stream has similar packet to reference packet stream, then described echo determining unit 31 ' can further combined with the described targeted packets stream energy hierarchical information (i.e. all kinds of gain (gain) information) corresponding with multiple consecutive packet data package corresponding in described reference packet stream, judge whether the similar packet in described targeted packets stream exists decay, namely energy hierarchical information is lower than the energy hierarchical information of the described reference packet stream of correspondence, if exist, then prove that described similar packet is echo bag, then comprise echo bag in described targeted packets stream.

In step s32 ', echo cancellation devices, when comprising echo bag in described targeted packets stream, carries out echo cancellor to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

Particularly, if comprise echo bag in described targeted packets stream, in step s32 ', echo cancellation devices then carries out echo cancellor to described targeted packets stream, such as, utilize to replace to wrap and detected echo bag is replaced, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.Wherein, described replacement bag includes but not limited to noise bag (such as, comprising the grouping of the noise of certain type, such as white noise, comfort noise etc.), noiseless bag (such as, empty grouping), in targeted packets stream 1/8th rate packet etc. of last buffer memory, and to mix.

To those skilled in the art, obviously the invention is not restricted to the details of above-mentioned one exemplary embodiment, and when not deviating from spirit of the present invention or essential characteristic, the present invention can be realized in other specific forms.Therefore, no matter from which point, all should embodiment be regarded as exemplary, and be nonrestrictive, scope of the present invention is limited by claims instead of above-mentioned explanation, and all changes be therefore intended in the implication of the equivalency by dropping on claim and scope are included in the present invention.Any Reference numeral in claim should be considered as the claim involved by limiting.In addition, obviously " comprising " one word do not get rid of other unit or step, odd number does not get rid of plural number.Multiple unit of stating in device claim or device also can be realized by software or hardware by a unit or device.First, second word such as grade is used for representing title, and does not represent any specific order.

Claims

1., for the method that the acoustic echo that divides into groups is eliminated, wherein, the method comprises the following steps:

2. method according to claim 1, wherein, described step b comprises following any one:

-according to described source stream of voice packets, determine the direction of transfer information corresponding to each packet data package in the stream of voice packets of described source; According to described source stream of voice packets, in conjunction with the direction of transfer information of the packet data package in the stream of voice packets of described source, upgrade the targeted packets stream of corresponding described call ends in destination buffer;

-according to described source stream of voice packets, upgrade the targeted packets stream of corresponding described call ends in destination buffer; According to described targeted packets stream, determine the direction of transfer information corresponding to each packet data package in described targeted packets stream.

3. method according to claim 1 and 2, wherein, the method also comprises:

-eliminate echo stream of packets according to described, upgrade the reference packet stream in described reference buffer district.

4. method according to claim 1, wherein, described step b comprises:

-described source stream of voice packets is sent to described destination buffer and reference buffer district respectively, to upgrade the targeted packets stream of corresponding described call ends in described mark buffering area, and the reference packet stream of corresponding described call ends in described reference buffer district, wherein, described targeted packets stream comprises the direction of transfer information corresponding to each packet data package in described targeted packets stream, and described reference packet stream comprises the direction of transfer information corresponding to each packet data package in described reference packet stream.

5. method according to any one of claim 1 to 4, wherein, described step c comprises:

C1 is according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag;

C2, when comprising echo bag in described targeted packets stream, carries out echo cancellor to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

6. method according to claim 5, wherein, described step c1 comprises:

-according to the reference packet stream of described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, and the energy hierarchical information corresponding with multiple consecutive packet data package corresponding in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag.

7. the method according to claim 5 or 6, wherein, described step c2 comprises:

-when comprising echo bag in described targeted packets stream, utilize replacement data bag, echo cancellor is carried out to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

8., for the echo cancellation devices that the acoustic echo that divides into groups is eliminated, wherein, this equipment comprises:

9. echo cancellation devices according to claim 8, wherein, described target update device is used for following any one:

10. echo cancellation devices according to claim 8 or claim 9, wherein, this equipment also comprises:

With reference to updating device, for having eliminated echo stream of packets according to described, upgrade the reference packet stream in described reference buffer district.

11. echo cancellation devices according to claim 8, wherein, described target update device is used for:

Echo cancellation devices according to any one of 12. according to Claim 8 to 11, wherein, described cancellation element comprises:

Echo determining unit, for the reference packet stream according to described call ends corresponding in corresponding reference buffer district, in conjunction with the direction of transfer information corresponding to each packet data package in described targeted packets stream and described reference packet stream, determine in described targeted packets stream, whether to comprise echo bag;

Echo cancellation unit, for when comprising echo bag in described targeted packets stream, carries out echo cancellor to described targeted packets stream, to obtain the elimination echo stream of packets corresponding with described targeted packets stream.

13. echo cancellation devices according to claim 12, wherein, described echo determining unit is used for:

14. echo cancellation devices according to claim 12 or 13, wherein, described echo cancellation unit is used for: