CN103347004B - A kind of method improving VBD data transmission quality - Google Patents

Abstract

The invention provides a kind of method improving VBD data transmission quality, including step: (1) is after calling terminal and called end set up Media Stream connection, before entrance VBD data-transmission mode, extracting the data source as assessment network quality of the audio medium stream in this period, data source Continuous plus goes out the jitter value of each RTP bag accordingly；(2) calling terminal and called end are respectively provided with the buffer depth of local terminal static state Jitterbuffer according to jitter value, switch to VBD data-transmission mode afterwards.The present invention judges current network environment quality by the jitter value of the Media Stream after calculating connection establishment, before switching VBD pattern, and then the buffer depth arranging jitterbuffer is cut into VBD data-transmission mode, more can press close to current network environment, on the premise of the integrity ensureing data signal to greatest extent, improve the real-time of data signal as much as possible.

Description

A kind of method improving VBD data transmission quality

Technical field

The present invention relates to VBD (Voice Band Data voice-band data) field of data transmission, particularly relate to one The method improving VBD data transmission quality.

Background technology

In traditional PSTN (Public Switched Telephone Network) the PSTN epoch, In addition to this speech business of phone, also have the data services such as fax, Modem and POS.The integration of three networks makes VoIP The application of (voice over internet protocol) is more and more universal, in order to make VoIP have the function identical with PSTN And Consumer's Experience, FoIP/MoIP(Fax/Modem over internet protocol) just occur in that on the basis of VoIP, The most alternatively FoIP/MoIP is a value-added service of VoIP, although FoIP and MoIP is data service, but they It is slightly different, a kind of pattern that FoIP is many.

FoIP has a both of which in Fax transmits: be respectively VBD (Voice Band Data voice-band data) and T38 computer network facsimile protocol；Mention these they between difference, in addition it is also necessary to first introduce T30 specification: pstn-to- The agreement that pstn is directly directly transmitted in this network environment in the way of facsimile signal.

VBD facsimile mode: T30 facsimile signal is not done any process and be directly packaged in Real-time Transport Protocol, with the side of rtp streaming Formula sends.This mode facsimile signal is same voice flow, so it is more low-loss to need that voice coding modes is consulted into G711 Coded system, to reduce the infringement to facsimile signal.

T38 pattern: T30 facsimile signal is carried out resolves into original facsimile signal and is repackaged into T30 message (peace photograph T30 agreement carry out decomposed signal and be packaged into can be at the message of ip transmission over networks), send (this in the way of T38 message Mode is faxed unrelated with voice so the coded system of voice need not be revised), T30 is equivalent to the agreement of analogue signal, and T38 is just These analog signal figures of T30 words are changed, so T38 can be more a lot, to network environment than T30 improves for anti-transmission interference Require the most lower.But T38 fax has its limitation used, it only supports low speed fax.So only having for high-speed facsimile VBD facsimile mode.Just no longer T38 is described in that patent.

MoIP is also only a kind of VBD pattern: is left intact Modem signal and is directly packaged in Real-time Transport Protocol, The mode of rtp streaming sends.

VBD data-transmission mode is the main flow transmission mode of FoIP/MoIP, and it is to the real-time of network environment, integrity Require higher.If fax, Modem signal occur network delay, shake in transmitting procedure, out of order cause its transmitting terminal to be used Occurring that distorted reception side can not identify in mutual signal, recipient would not send out the response signal of its correspondence, and transmitting terminal is received Signal before will retransmitting less than response signal, if network environment still makes distorted signals, sender continues to retransmit, If it exceeds the sending times in its fax Modem host-host protocol also confiscates the response signal of recipient, then whole transmission Just have failed.This is the Failure Mode of a kind of common VBD data transmission.If voice call occurs network delay, shake, Out of order, sound there will be Caton phenomenon, but what normal talking still can be carried out.As a rule VOIP scheme all can have Jitterbuffer jitter buffer eliminates network delay, shake and out of order via net loss, improves signal quality.This Jitterbuffer is particularly important on the impact of VBD data signal transmission, because data are than the integrity and in real time of voice-to-data Property require higher.

Jitterbuffer wobble buffer is a shared data area, in this data area, every one section Uniform interval, packet can be collected, and stores and is dealt into speech processor.Wrap the change of the time of advent, be referred to as shake, it will Produce due to network congestion, timing wander or routing change.Wobble buffer is put in the receiving terminal that voice connects, and it is wittingly The bag delayed to reach, consequently, it is possible to terminal use will experience one clearly, there is no the connection of what audio distortions.Tremble Dynamic buffer has both of which, static schema and dynamic mode.Static schema: fix and go data to delay every one period of regular time Deposit and go in district to take the packet received, say, that data need to cache one period of regular time, and this regular time is just It it is buffer depth.Dynamic mode: can go to calculate the jitter value of current network at set intervals, adjust according to jitter value Whole how long go to data buffer area to take the packet received, namely adjust buffer depth according to jitter value.Dynamically Jitterbuffer pattern is more relative complex, and it also has the structure that bag compensates in the case of particular network, in order to make sound Continuously, this pattern this be suitable in the case of voice call use, improper under data pattern (such as Fax, Modem) use, Because it can destroy the integrity of data signal.The cycle that under the two pattern, it fetches data is threshold value, it is also possible to is referred to as it and delays Rush the degree of depth in district.Buffer depth is the biggest, its process network jitter, time delay, the via net loss such as out of order ability the biggest, but number The real-time of the number of it is believed that is the poorest.The buffer area degree of depth is the least, and its ability processing via net loss is the least, but the real-time of signal The highest.The integrity of data signal and real-time no less important under VBD data-transmission mode.Thus it is necessary for VBD number Need to provide a kind of effective solution according to transmission optimal balance the two under different network environments.

Summary of the invention

It is an object of the invention to provide a kind of method improving VBD data transmission quality, according to the shake of current network Situation arranges the jitter-buffer degree of depth of Jitterbuffer more accurately, on the premise of ensureing VBD data integrity to the greatest extent Amount improves the real-time of VBD data, thus improves the quality of VBD data transmission.

It is an object of the invention to be achieved through the following technical solutions.

A kind of method improving VBD data transmission quality, including step:

(1) calling terminal and called end set up Media Stream connect after, enter before VBD data-transmission mode, when extracting this section Interior audio medium stream is as the data source of assessment network quality, and data source Continuous plus goes out the jitter of each RTP bag accordingly Value；

(2) calling terminal and called end are respectively provided with the relief area of local terminal static state Jitterbuffer according to described jitter value The degree of depth, switches to VBD data-transmission mode afterwards.

Preferably, in described step (1),

Jitter value is calculated according to formula J (i)=[Ri Si] [R (i-1) S (i-1)]；

Wherein, Si is the time stamp in RTP i bag, and Ri is that this RTP i wraps in the time actually reached under time stamp unit, J (i) Jitter value for RTP i bag.

Preferably, in described step (1), the RTP bag received in front 400ms in described audio medium stream is as assessment The data source of network quality.

Preferably, in described step (1), start received by Continuous plus from setting up after Media Stream connects 80ms The jitter value of RTP bag.

Preferably, in described step (2), take the maximum Jmax in all jitter values, static state is set The buffer depth of Jitterbuffer is Jmax.

Compared with prior art, the embodiment of the present invention has the advantages that

1) present invention is judged currently by the jitter value of the Media Stream after calculating connection establishment, before switching VBD pattern Network environment quality, and then the buffer depth arranging jitterbuffer is cut into VBD data-transmission mode, more can press close to Current network environment, on the premise of the integrity ensureing data signal to greatest extent, improves data signal as much as possible Real-time；

2) either FoIP or MoIP, it enters into VBD data transmission state from voice call state and is required for one Process, at least needs the THP terminal handler processes such as DSP signal detection, signaling negotiation, and this section termination processing procedure is it is generally required to 250ms Left and right, in addition to terminal processes requires time for, often facsimile machine, Modem, POS are also required to a period of time and just can initiate it Signal interaction flow, is properly termed as the front-end processing time during this period of time, and this front-end processing time is indefinite, depending on equipment.Eventually End process time+front-end processing time the most all can be more than 400ms, so present invention can ensure that before VBD data are transmitted There is one section of Media Stream as judging the data source of network environment quality, the accuracy of network environment Quality estimation has been effectively ensured.

Accompanying drawing explanation

Fig. 1 is the networking schematic diagram of FoIP and MoIP.

Fig. 2 is that VBD data transmit Establishing process figure.

Fig. 3 is that Jitter calculates schematic diagram.

Detailed description of the invention

In order to make the purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, right The present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, and It is not used in the restriction present invention.

In order to, on the premise of ensureing data signal integrity to greatest extent, improve the real-time of data signal, Jin Erti High VBD data transmission quality, method provided by the present invention is:

One, calling is initiated in facsimile machine dialing, and far-end facsimile machine response is set up Media Stream and connected, the most also at voice call shape State, has not been entered into VBD data-transmission mode.This state is referred to as voice status S1.

Two, voice status S1 can continue a bit of time, to have the Media Stream of one section of mutually continuous voice during this period, The present invention just with this section of Media Stream as the data source assessing network quality.

RFC 3550(RTP:A Transport Protocol for Real-Time Applications) in have Definition and the computational methods of interarrival jitter are as follows:

D(i-1,i) = [Ri – R(i-1)] – [Si - S(i-1)] = [Ri – Si] – [R(i-1) – S(i-1)](1)

J(i) = J(i-1) + (|D(i-1,i)| - J(i-1))/16(2)

Si in formula (1) is the time stamp in this RTP i bag, and Ri is that this RTP i wraps in and actually reaches under time stamp unit Time, and D (i, i-1) can be understood as the difference of continuous two RTP bag actual times of arrival and the theory difference time of advent；Formula (2) J (i) in is exactly the jitter value of RTP i bag, and its parameter 1/16 is a weighter factor.

Example:

Here having the RTP of three PCMA of continuous print, its RTP information is as follows:

R0 = frame 624: frame.time = Jul 4, 2005 11:56:25.348411000 ；

S0 = frame 624: rtp.timestamp = 1240 ；

R1 = frame 625: frame.time = Jul 4, 2005 11:56:25.418358000 ；

S1 = frame 625: rtp.timestamp = 1400 ；

R2 = frame 626: frame.time = Jul 4, 2005 11:56:25.421891000 ；

S2 = frame 626: rtp.timestamp = 1560 ；

The sample frequency of PCMA is 8000Hz, so the unit of its time stamp is 1/8000 second, and namely 0.000125 second.

In this example, the calculating process of jitter value is as follows:

frame 624:

J(0) = 0 ；

frame 625:

D(0,1) = (R1 - R0) - (S1 - S0) = (.418358000 sec - .348411000 sec) - (1400 * 0.000125sec-1240 * 0.000125 sec)=0.049947 (second)；

J(1) = J(0) + (|D(0,1)| - J(0))/16 = 0 + (|0.049947| - 0)/16 = 0.0031216875 (second)；

frame 626:

D(1,2) = (R2 - R1) - (S2 - S1) = (.421891000 sec - .418358000 sec) - (1560 * 0.000125 sec-1400 * 0.000125 sec)=-0.016467 (second)；

J(2) = J(1) + (|D(1,2)| - J(1))/16= 0.0031216875 + (|-0.016467| - 0.0031216875)/16=0.00395576953125 (second)

Owing to multiplication and division budget is relatively time-consuming, the present invention calculates the time-consuming of J value to reduce, and eliminates and takes advantage of 1/16 weighting The operation of the factor, and do not do accumulative operation, remove+J (i-1), because of without going again to calculate the J defined in RFC 3550 (jitter) value, draws J(jitter) formula:

J(i) = [Ri – R(i-1)] – [Si - S(i-1)] = [Ri – Si] – [R(i-1) – S(i-1)] (3)

J in formula (3) is actually the difference of continuous two RTP bag actual times of arrival and the theory difference time of advent, this Invention describes network jitter with it；And the J of formula (2) is also used to describe network jitter, but it includes accumulative above institute Some network jitter factors, are multiplied by add to make it add up smooth (J that two continuous print bags come out will not saltus step too many) Weight factor 1/16.

The RTP bag that the present invention receives using 400ms before in Media Stream, as data analysis source, has not owing to connecting just foundation Stable situation, so ignoring the bag of above general about 80ms, from 80ms later unwrap its J(jitter of beginning Continuous plus) Value, finally calculates the Jmax of its maximum.Although network environment quality especially mutability, but be the most all stable, Network jitter is analyzed meaningful, so describing current network conditions with the Jmax calculated with the bag of front 400ms Jitter conditions.

Three, facsimile machine is switched to state of faxing, and sends facsimile signal, and the DSP of the VOIP terminal that facsimile machine is connect detects biography True signal, initiates ReInvite VBD facsimile negotiation, and opposite end response has been consulted, and enters S2 VBD data mode, two ends of faxing All the degree of depth of relief area is set according to the value of Jmax and is cut into VBD data-transmission mode.

Method provided by the present invention will be described in detail by example below.In the present embodiment, FoIP's and MoIP Networking mode is as shown in Figure 1.As in figure 2 it is shown, the method flow that the embodiment of the present invention improves VBD data transmission quality includes:

Data equipment (facsimile machine, Modem, the POS) dialing that step 1:VOIP terminal is connected, sends Invite message Initiating calling, this message carries speech coding capacity algorithm, the ptime packaging time length that calling terminal is supported, such as G711A, G722 etc..

Step 2: after called end receives Invite message, returns 100 trying Temporary Responses to calling terminal, and to caller End sends 180 ringing ALERTING messages, so calling and is at ringing condition.

Step 3: called end by obtaining the SDP in Invite message it is known that the code capacity of the other side, link address etc. are believed Cease, and locally-supported ability compares, confirmed to consult, be created as by 200 OK response messages notice callings Merit, this information carries the SDP information such as coding, ptime packaging time length.The most called being also required to records the coding that active calls uses Information, such as G711A.

Step 4: after caller receives 200 OK message, the coding of record negotiation result and packaging time length, to called reply ACK Confirm message, so far enter talking state, Media Stream connection establishment.

Step 5: the packaging time length ptime of negotiation is 20ms, calling terminal, called end all take after Media Stream connects in 400ms The RTP bag that the opposite end received is sent, may the when of just foundation as it is shown on figure 3, connect as the data source of calculating jitter A little unstable, so 4 RTP bags in casting aside above 80ms, the 5th bag (packet 4) starts to take and record in RTP bag Timestamp time stamp is S4, and it is R4 that record receives the present system time of this bag, is accurate to millisecond the most permissible.6th bag (packet 5) too, the timestamp time stamp in its RTP bag is S5, the system time R5 that bag reaches, preceding step 3 and 4 Calling terminal called end all have recorded the coding of negotiation result, here needs timestamp time stamp for calculating this coding Unit, it is all the sample rate of 8KHz that domestic telecommunication commonly uses G711A, G711U and G729 in coding, and G722 is the sample rate of 16KHz. Computing formula according to before:

J (i)=[Ri R (i-1)] [Si-S (i-1)]=[Ri Si] [R (i-1) S (i-1)], J5=(R5-R4) (S5-S4)/8000, by that analogy, calculates the jitter value within these 400ms of J6 ~ J18 respectively, reaches Find out jitter value Jmax of maximum when of 400ms out, be used for describing the jitter conditions of current network.

The data equipment (facsimile machine, Modem, POS) that step 6:VOIP terminal is connect starts to send interactive signal, VOIP The DSP of terminal detects Fax or Modem signal, and reports Call Control Block process.

Step 7: the Call Control Block of called end is initiated ReInvite and re-established calling, and carries a=fax perhaps a= The SDP information that the vbd such as modem consult.

Step 8: after calling terminal receives the ReInvite message that vbd consults, return 100 trying Temporary Responses.

Step 9: calling terminal resolves the ReInvite information received, and has confirmed that VBD consults, and lead to by sending 200 OK Know that called calling re-establishes successfully.

Step 10: called receive 200 OK message after, reply ACK confirm message, so far stage VBD consult success, Next step is accomplished by calling and called and is all cut into VBD data transmission state.

Step 11: according to the Jmax value calculated in step 5, the degree of depth arranged in static Jitterbuffer is Jmax, Namely arranging data in relief area residence time is Jmax, and the result then consulted according to VBD before arranges DSP and connects logical Road is VBD data-transmission mode, is so far put into the state of VBD data transmission.

Step 12:Fax/Modem signal carry out a series of mutual after, complete the transmission of its data, then Fax/Modem leads to Can mourn in silence its data sent out after often completing transmission, the DSP of called end detects that continuing 3 seconds two-way quiet (receives Data and the data oneself sent out) after, just report VBD end of transmission signal.

Step 13: after called end knows VBD DTD, initiates ReInvite message and re-establishes call connection.

Step 14: after calling terminal receives ReInvite message, returns 100 trying Temporary Responses.

Step 15: calling terminal resolves the ReInvite information received, has confirmed that call is consulted, and by sending 200 OK Notice called end calling re-establishes successfully.

Step 16: called receive 200 OK message after, reply ACK confirm message, so far the stage is just from VBD transmission state Switch back to talking state.

Step 17: after switchback call, the behavior of general Fax/Modem equipment all can automatically hang up connection, also having will not be certainly Dynamic hanging up connects by the next manual on-hook of user.In our flow process, calling terminal has hung up connection, by send Bye message to Called end notifies that it terminates call.

Step 18: after called end receives Bye, replys ACK and confirms to calling terminal, and so far Media Stream connects disconnection, whole VBD Transfer process terminates.

Above implementing procedure key point is how to calculate in step 5 in Jmax, and step 11 according to arranging static state The jitterbuffer degree of depth is Jmax, embodies according to current network jitter situation, arranges the Jitterbuffer degree of depth, from And on the premise of ensureing VBD data integrity, improve the real-time of VBD data as far as possible.

The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all essences in the present invention Any amendment, equivalent and the improvement etc. made within god and principle, should be included within the scope of the present invention.

Claims (5)

1. the method improving voice-band data VBD data transmission quality, it is characterised in that the method comprising the steps of:

(1) after calling terminal and called end set up Media Stream connection, before entrance voice-band data VBD data-transmission mode, carry Taking the data source as assessment network quality of the audio medium stream in this period, data source Continuous plus goes out each real-time biography accordingly The jitter value of transmission protocol RTP bag；

(2) calling terminal and called end are respectively provided with local terminal static jitter buffer device Jitterbuffer's according to described jitter value Buffer depth, switches to voice-band data VBD data-transmission mode afterwards.

2. the method improving voice-band data VBD data transmission quality as claimed in claim 1, it is characterised in that described step Suddenly, in (1), jitter value is calculated according to formula J (i)=[Ri Si] [R (i-1) S (i-1)]；

Wherein, Si is the time stamp in realtime transmission protocol RTP i bag, and Ri is that this realtime transmission protocol RTP i wraps in time stamp unit Under time of actually reaching, R (i-1) is that this realtime transmission protocol RTP i-1 wraps in the time actually reached under time stamp unit, S (i-1) being the time stamp in realtime transmission protocol RTP i-1 bag, J (i) is the jitter value of realtime transmission protocol RTP i bag.

3. the method improving voice-band data VBD data transmission quality as claimed in claim 1, it is characterised in that described step Suddenly, in (1), the RTP bag received in front 400ms in described audio medium stream is as the data source of assessment network quality.

4. the method improving voice-band data VBD data transmission quality as claimed in claim 3, it is characterised in that described step Suddenly, in (1), connect start realtime transmission protocol RTP bag received by Continuous plus after 80ms from setting up Media Stream Jitter value.

5. the method improving voice-band data VBD data transmission quality as described in Claims 1-4 is arbitrary, it is characterised in that In described step (2), take the maximum Jmax in all jitter values, the slow of static jitter buffer device Jitterbuffer is set Rushing district's degree of depth is Jmax.