CN1585332A - Speech and data transmission in same frame - Google Patents

Abstract

The method includes following steps: the voice signals that were recorded previously are taken inaudible discrimination in turn according to the depacketizing and the frame to determine the audible period and inaudible period; the audible period and inaudible period are separated, the voice data in audible period are compressed, and the compressed voice data are saved consecutively; the voice data in inaudible period are collected for data transmission. The invention can save ratio threshold of inaudible data in approaching voice and make the network resource get fully utilization.

Description

The same frame transmission method of voice and data

[technical field]: the present invention relates to a kind of multi-media information communication transmission technique field.

[background technology]: enter the information age of 21 century, human information exchanges and carries out the transition to multimedia form from single medium, and multimedia communication has become the theme in these epoch.

Multimedia technology is mainly studied the transmission and the application problem of multimedia messages under the network environment.Asynchronous communication is used for exchanging those and does not need the Still Media data of real-time storage or processing between end system.Real time communication is used for transmitting the continuous media data that needs are handled in real time such as voice, video between end system.Because continuous media communication has real-time, the form with " data flow " realizes continuous media communication end to end usually.The continuous media stream of real time communication transmission has high performance requirement to network, and this also is the problem that multimedia communication technology will solve emphatically.Multimedia data stream is lengthy usually, needs to split into grouping or frame and transmit in network transmission process.In order to keep the integrality of these data itself, need ask for something be proposed to the performance of network.

Multimedia communication comprises multiple transmission of media data and processing, and dissimilar media datas has different requirements to the performance of network with service.For voice, real-time requires very high.After the multi-medium data digitlization, data volume is very big, thereby network has been proposed the requirement of high bandwidth.To high-quality sound, video communication, real-time and bandwidth all have very high requirement.

The time Quito media information relevant with the time, have the characteristics of time continuity and real-time, therefore require the low delay of communication.Voice transfer maximum acceptable time of delay is 250ms, adopts the block exchange mode transmission, and the arrival between the grouping postpones can not be greater than 10ms, otherwise will feel that voice are discontinuous.In real time communication, voice are cores of various medium, under the situation that does not have the output of direct video, are interpersonally normally realized by voice alternately.Sound can transmit in many ways, no matter but be with real-time mode or with the store-and-forward mode transmission, all to guarantee the continuity that plays sound.Guarantee good audio transmission quality, will guarantee enough bandwidth.

People also need to exchange some data messages except the voice and video image information when carrying out long-range interchange, particularly need some data messages as a supplement when carrying out seminar.Simultaneously need some real-time controls in multimedia communication, these control informations are exactly data message.

Voice/data comprehensive transmission technology is a focus of studying in the multimedia communication, also is a direction of multimedia communication development.Because voice signal is had relatively high expectations to the real-time of transmission, generally adopt the circuit exchange mode transmission.But turned to transmission at present, and proposed some solutions based on the IP mode.As document [1] Chaim Ziegler and Gerald Weiss, Mechanisms for Intergrated Voice and Data Conferencin[J] ACM Computer Communicat ion Review, transmit 1990,20 (3) 101～108 pages of disclosed fixed positions that just allow voice and data take respectively in the frame simply; Document [2] J.Vande Lagemeat, J.M.A.Daaemen and I.G.Niemegeers, Voice Transmi ssion In aPriority CSMA/CD LAN:An Efficient Protocol Using Hybrid Switching[J] .ACM Computer Communication Review.1986,16 (3), 102～110 pages of disclosed methods are that frame is divided into voice cycle and cycle data, the voice cycle transferring voice phase, cycle data transmission data; Document [3] Chen Fugui, Hu Xiulin, based on the voice/data comprehensive transmission research [J] of TCP/IP network, computer system application, No.5, P19～21 page .1997 is disclosed then to be real-time from voice, and the priority issues of voice queue has been discussed.

But no matter above-mentioned which kind of method does not all make Internet resources be effectively utilized.

[summary of the invention]: the objective of the invention is to solve the above-mentioned problems in the prior art, provide a kind of can make full use of Internet resources carry out multimedia communication, voice and the same frame transmission method of data.

The same frame transmission method of voice of the present invention and data, finish according to the following steps:

---at first the block of speech signal of having recorded waiting for transmission is carried out silence distinguish successively, determine the sound phase and the silence periods of voice signal;

---the sound phase is separated with silence periods, sound phase voice are compressed, and deposit digitize voice after the compression continuously;

---collect silence periods and be used to transmit data.

The voice/data frame structure that carries out after silence distinguish and compression are handled comprises successively: frame head, efficient voice district, data field and crc field; Wherein frame head is made up of Network Transmission information header field, control information field, sound/noiseless distribution table, Data Identification field and data sequence field etc.

Advantage of the present invention and good effect:

(1) take silence distinguish and the compression treatment measures after, the network bandwidth resources saving is fully utilized Internet resources near noiseless shared rate limit value in the voice.

(2) by formula (1) is chosen silence distinguish threshold value Δ and can be guaranteed voice quality preferably.

(3) voice-and-data is adopted different transmission mechanisms, both can guarantee the real-time of voice transfer, can guarantee the correctness of transfer of data again.

[description of drawings]:

Fig. 1 (a) is the waveform of one section recorded speech;

Fig. 1 (b) is the waveform of silence distinguish after handling;

Fig. 2 (a) shows the distribution situation schematic diagram that a voice recording piece carries out silence distinguish and compression processing sound phase of back and silence periods;

The compressed voice information distribution situation schematic diagram of Fig. 2 (b) for depositing continuously;

Fig. 2 (c) is the frame structure schematic diagram.

[embodiment]:

Embodiment 1:

At first the block of speech signal of recording is carried out silence distinguish, determine the sound phase and the silence periods of voice signal, the sound phase is separated with silence periods (see Fig. 2 a), sound phase voice are compressed, and deposit voice (seeing Fig. 2 b) after the compression continuously.Then, collect silence periods and be used to transmit data.

Fig. 2 (c) is a frame structure, and each Field Definition is as follows in the frame:

1) Network Transmission information header field comprises the required relevant informations of Network Transmission such as source/destination address;

2) control information field contains the control information of making a start to receiving end;

3) sound/noiseless distribution table is the noiseless position table that is drawn by the silence distinguish algorithm, is used for receiving end and recovers former voice;

4) file under the data in the identification field identification frames is used to receive the differentiation of data;

5) position in the transmission data in the data sequence field identification frames is used for the receiving end data sorting;

6) speech region is deposited the digitize voice after the compression;

7) the defeated data of tendency to develop are deposited in the data field;

8) crc field is carried out verification to the data of frame head and data field.

Silence distinguish may produce some influences to voice quality, and still, as long as choose suitable silence distinguish threshold value Δ, its influence can be very little, shown in Fig. 1 (a) and (b).Silence distinguish of the present invention can be obtained satisfied effect by formula (1) decision Δ.

Δ＝E _n*(1+0.2) (1)

In the formula: E _nBe the background noise average energy.

Three kinds of methods introducing in aforementioned background art all voice are not carried out silence distinguish and compression is handled, and just will transmit behind voice and the aggregation of data, and this is conventional transmission method.Saving for network bandwidth resources that the inventive method is described can contrast with them, and existing is that example is analyzed with the disclosed technology of representative document [1].

Suppose that the defeated voice messaging of tendency to develop is the V bit, data message is the L bit, and noiseless proportion is S in the voice, and the speech coding compression ratio is K, and the CRC check code length is the C bit.The frame head that does not carry out silence distinguish is F _HN, the frame head that carries out silence distinguish is F _HSSuppose also that simultaneously the two sends identical voice segments, and the frame number is M.Like this, not carrying out the informational capacity that silence distinguish sent is

T _HN＝VK+L+M(F _HN+C) (2)

Carrying out the informational capacity that silence distinguish sent is

T _HS＝V(1-S)K+L+M(F _HS+C) (3)

This method is saved ratio with respect to the Internet resources of method (1)

$E=[1-\frac{V(1-S)K+L+M(F_{\mathrm{HS}}+C)}{\mathrm{VK}+L+M(F_{\mathrm{HN}}+C)}]$

$=\frac{\mathrm{VKS}+M(F_{\mathrm{HN}}-F_{\mathrm{HS}})}{\mathrm{VK}+L+M(F_{\mathrm{HN}}+C)}---\left(4\right)$

Following formula be in finite time to the two comparison, in order to obtain on 0～∞ time-domain that the comparison of the two is needed the expression-form of change formulas (4), set up the relation with time t, for this reason, order

V＝Pt (5)

Wherein P is the voice recording rate, and unit is b/s, so, send voice Frame number

M＝Pt/R (6)

Wherein R is a packets of voice, and unit is a bit.With formula (5) and (6) substitution formula (4),

$E=\frac{\mathrm{PtKS}+(\mathrm{Pt}/R)(F_{\mathrm{HN}}-F_{\mathrm{HS}})}{\mathrm{PtK}+L+(\mathrm{Pt}/R)(F_{\mathrm{HN}}+C)}---\left(7\right)$

When t → ∞, the limit of E:

$\underset{t\→\∞}{\lim }E=\underset{t\→\∞}{\lim }\frac{\mathrm{PtKS}+(\mathrm{Pt}/R)(F_{\mathrm{HN}}-F_{\mathrm{HS}})}{\mathrm{PtK}+L+(\mathrm{Pt}/R)(F_{\mathrm{HN}}+C)}$

$=\frac{\mathrm{PKS}+F_{\mathrm{HN}}-F_{\mathrm{HS}}}{\mathrm{RK}+F_{\mathrm{HN}}+C}---\left(8\right)$

Following formula is that the inventive method that obtains on the unlimited time is saved limit rate value to the network bandwidth resources of document [1].

Conversation voice for 10 minutes is added up, and the gained data are as follows:

V＝105840000bits；

S＝0.56；

L＝819200bits；

F _HN＝48bits；

F _HS＝56bits；

C＝16bits；

R＝176400bits。

With above-mentioned data substitution formula (8), get K=1: 4 can obtain limit channel saves ratio:

$\underset{t\→\∞}{\lim }E=\frac{\mathrm{RKS}+F_{\mathrm{HN}}-F_{\mathrm{HS}}}{\mathrm{RK}+F_{\mathrm{HN}}+C}\×100\%$

$=\frac{176400\×0.25\×0.56+48-56}{176400\×0.25+48+16}=55.9\%$

This shows, take the silence distinguish measure after, the network bandwidth resources saving is very tangible, near the value of noiseless shared ratio S in the voice.

Claims (2)

1, the same frame transmission method of a kind of voice and data is characterized in that this transmission method finishes according to the following steps:

---at first the block of speech signal of having recorded waiting for transmission is carried out silence distinguish, determine the sound phase and the silence periods of voice signal;

---the sound phase is separated with silence periods, sound phase voice are compressed, and deposit digitize voice after the compression continuously;

---collect silence periods and be used to transmit data.

2, the same frame transmission method of voice according to claim 1 and data is characterized in that the frame structure that carries out after silence distinguish is handled comprises successively: frame head, efficient voice district, data field and crc field; Wherein frame head is made up of Network Transmission information header field, control information field, sound/noiseless distribution table, Data Identification field and data sequence field.

Images