• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method for transmitting voice data between two subscribers (2x, 1w), which are networked via a plurality of redundant communication servers (2a, 2b). The invention provides that voice information is created by the transmitting subscriber (1x) and data packets (X1,... Xn) comprising voice data values (x) are created with a time stamp (t) in the form of a consecutive number, - that the data packets (X1 , ... Xn) are transmitted to the communication servers (2a, 2b), - that from each communication server (2a, 2b) in each case further data packets (A1, ..., An; B1, ..., Bn) are created, each of which at least one of the transmitting subscriber computer (10x) transmitted voice data (x) and a, another time stamp (t ') are assigned, wherein the value of the time stamp (t') the remainder of the division of the time stamp (t) divided by a number N corresponds, and - that the subscriber computer (10w) receives the further data packets (A1, ... An; B1, ... Bn), each of which is assigned the same further time stamp (t '), and a voice data value (w) by selection or summation of the voice data values, each with the same additional time stamp l (t ') determined.
    公开号:AT515441A1
    申请号:T50147/2014
    申请日:2014-02-26
    公开日:2015-09-15
    发明作者:Ernst Dipl Ing Dr Szabo
    申请人:Frequentis Ag；
    IPC主号:
    专利说明:

    The invention relates to a method for the redundant transmission of voice data between a plurality of participants, in particular in the form of a conference call, according to the preamble of claim 1 or 2. Furthermore, the invention relates to a voice communication system according to the preamble of claim 8 and 9. Finally, the invention relates to a disk on which a program for carrying out a method according to the invention is stored.
    A variety of different methods for transmitting voice data and voice switching systems are known from the prior art, which have a plurality of redundant communication server, wherein voice communication data are transmitted via parallel data paths with different transit time characteristic between the communication participants.
    Also known from the prior art conference systems for voice transmission, with which a plurality of conference participants who are connected to each other via a computer system, communicate via communication server voice data, wherein between the individual participants a conference call is switched.
    The disadvantage here is especially that voice signals that are sent by individual participants redundant over the network, arrive due to different maturities in the network at the other communication participants at different times.
    In particular, redundant communication systems with different signal propagation times have the problem that the same voice messages of individual subscribers appear to be delayed, whereby, depending on the propagation delay, addition of these signals can lead to deletions of individual frequency bands.
    In the prior art, there are different approaches to this problem, often methods are used in which the individual signals transmitted via the line are correlated by means of correlation and so the
    Delay differences between the individual signals are determined and compensated. Such a procedure is numerically very complicated, since the execution of a correlation is numerically complex.
    The object of the invention is therefore to provide a method that compensates for differences in runtime of individual redundant signals and avoids a time-shifted superposition of voice data due to different maturities in the transmission in the network.
    The invention solves this problem in a method of the type mentioned above with the characterizing feature of claim 1.
    If the individual signals which originate from the first subscriber computer are superimposed with one another, an extinction which results from different transit time effects of different lines in the network is effectively avoided.
    In addition, it is also possible in a simple manner, in the case of selecting a data packet of a communication server or the prioritization of the data packets of a communication server to effect a switchover of the selection without concrete artifacts occur. If a changeover, e.g. with very poor transmission quality, a higher-priority server made in favor of a lower-priority server, no artefacts arise. This is particularly advantageous when switching very often, e.g. occurs several times per second.
    The invention solves the problem in a method of the type mentioned for communication with two participants with the feature of claim 1, wherein it is provided that the transmitting subscriber voice information is created and this voice information is divided by the respective sending subscriber computer into individual, numeric voice data values, data packets comprising in each case at least one numerical voice data value, in particular a plurality of voice data values recorded immediately after one another, are assigned to each of which at least one time stamp in the form of a consecutive number which determines the order of the individual voice data values to the voice data values of other data packets Data packets of the sending subscriber to be communicated to the transmitting subscriber computer communication server that each of the communication server each additional data packets are created, which in each case at least one voice data value transmitted by the transmitting subscriber computer and at least one further time stamp are assigned, the value of the further time stamp corresponding to the remainder of the division of the time stamp divided by a predetermined constant positive integer N, where N is chosen to be the same for all communication servers, and the receiving subscriber computer receives further data packets transmitted by the communication servers and determines such further data packets to which the same additional time stamp is assigned, and determines a voice data value by selecting or summing the voice data values contained in the further data packets with respectively the same further time stamp and the receiving subscriber available.
    The invention solves the problem in a method of the type mentioned for carrying out a conference with the characterizing features of claim 2. It is provided that - are created by the sending participants voice information and this voice information from each sending subscriber computer divided into individual successive voice data - that the sending subscribers data packets comprising at least one numerical voice data, in particular several immediately consecutively recorded voice data create, each of which at least one timestamp is assigned in the form of a serial number, the order of the individual voice data to the voice data of other data packets of the same subscriber stipulates that - the data packets of each transmitting subscriber are transmitted to the communications server connected to the sending subscriber computer, - that the communication service it receives data packets transmitted to it by the sending subscriber computers and stores the voice data values contained in the data packets, wherein the individual voice data values are each linked to a further time stamp corresponding to the remainder of the division of the time stamp divided by a predetermined positive integer N, in that the communication servers create further data packets to which at least one further time stamp is assigned and to which speech data values are determined, in particular by overlaying the voice data values associated with the same further time stamp, and in that the receiving subscriber computer all of them from the communication servers transmitted data packets, each of which is assigned the same time stamp, superimposed or selected from the transmitted data packets to which the same time stamp is assigned, a data packet and the second Teilneh mer provides.
    In particular, it can be provided that all voice data arriving at the communication servers are assigned at least, in particular precisely, to a further data packet and transmitted to the receiving subscriber.
    A particularly simple embodiment of the invention that is compatible with commercial voice transmission and conferencing systems provides that each data packet is assigned voice data values in the form of a single sample or several samples, and that the voice data is superimposed in the form of sample values by means of addition in the communication servers.
    An advantageous way of avoiding superimposing speech data created at different times on the same sending party is to set N as follows: N > (tdelay.max " tdelay.min) / tsampj - where tdeiay.max denotes the maximum delay time of a connection line between a sending or receiving subscriber's computer and a communication server, - where tdeiay.min is the minimum delay time of a connection line between a sending or receiving subscriber's computer and a communication server, and - where tSamp corresponds to the inverse of the sampling rate.
    In order to avoid transmission errors in the event of failure of one of the communication servers, it can be provided that the communication server exclusively uses data packets from transmitting subscriber computers for the superimposition within a predefined time interval and / or already used for the superimposition
    Data packets for further overlays no longer used and especially rejected.
    In order to avoid transmission errors in the case of failures of one of the subscriber computers, it can be provided that the receiving subscriber computer exclusively uses voice data from communication servers for the overlay within a predetermined time interval and / or no longer uses voice data already used for the overlay for further overlays and discards it in particular.
    Furthermore, the invention solves the problem in a voice communication system of the type mentioned above with the characterizing features of claim 8. It is provided that - the sending subscriber computer has a voice recording unit and a coding unit which subdivides the voice information created by the voice recording unit into individual successive voice data values, and data packets which each have at least one voice data value, in particular a plurality of voice data values recorded immediately after one another, and a time stamp assigned to the voice data values and created by a timer, wherein the time stamp indicates the order of the voice data values, that the transmitting subscriber computer the Data packets transmitted to the communication server connected to the sending subscriber computer, - that the communication server each create additional data packages, which jewe at least one voice data value transmitted by the transmitting subscriber computer and at least one further time stamp are assigned, the value of the further time stamp corresponding to the remainder of the division of the time stamp divided by a predetermined positive integer N, where N is the same for all communication servers, and the receiving subscriber computer receives further data packets transmitted by the communication servers and determines such further data packets to which the same additional time stamp is assigned, and determines a voice data value by selecting or summing the voice data values contained in the further data packets with the same additional time stamp and the receiving subscriber available.
    If the individual signals which originate from the first subscriber computer are superimposed with one another, an extinction which results from different transit time effects of different lines in the network is effectively avoided.
    In addition, it is also possible in a simple manner, in the case of selecting a data packet of a communication server or the prioritization of the data packets of a communication server to effect a switchover of the selection without concrete artifacts occur. If a changeover, e.g. with very poor transmission quality of a higher-priority server in favor of a lower-priority server, no artefacts arise. This is particularly advantageous when switching very often, e.g. occurs several times per second.
    A particularly simple embodiment of the invention that is compatible with commercially available voice transmission and conferencing systems provides that the transmitting users each create voice information and this voice information is subdivided by the respective transmitting subscriber computer into individual successive voice data values such that the transmitting subscriber computers each have a voice recording unit and a coding unit which divides the speech information produced by the speech recording unit into individual successive speech data values, and creates data packets each comprising the speech data values and a timestamp associated with the speech data values and created by a timer, the time stamp representing the order of the timestamps Indicates voice data values to the voice data values of other data packets created by the same subscriber computer, that the transmitting part hmerrechner transmit the data packets created by them to the communication server connected to the respective sending subscriber computer, - that the communication server each receive data packets transmitted by the sending subscriber computers and store the voice data values contained in the data packets in a memory and the individual voice data values each with a further time stamp store linked, which corresponds to the remainder of the division of the time stamp divided by a predetermined positive integer N, - that the communication server create more data packets, which at least one further time stamp is assigned and which each, in particular by superimposing the voice data values with the the same other timestamps are linked, determined, voice data values are assigned and that the receiving subscriber computer all transmitted to him by the communication server data packets, which jewe ils the same time stamp is assigned, superimposed or from the transmitted data packets, which is assigned to the same timestamp, selects a data packet and provides the second participant available.
    An advantageous definition, with which it is avoided that voice data of the same transmitting subscriber created at different times are superimposed, provides that each data packet is respectively assigned voice data in the form of a single sample and that the superimposed units of the communication server or the subscriber computer are superimposed as sums of the form voice data received by them.
    In order to avoid transmission errors in the event of failure of one of the communication servers, provision can be made for N to be set as follows: N > (tdelay.max " tdelay.min) / tsampj - where tdeiay.max denotes the maximum delay time of a connection line between a sending or receiving subscriber's computer and a communication server, - where tdeiay.min is the minimum delay time of a connection line between a sending or receiving subscriber's computer and a communication server, and - where tsamp corresponds to the inverse sampling rate.
    In order to avoid transmission errors in the event of failure of one of the subscriber computers, it may be provided that the communication server exclusively uses voice data from transmitting subscriber computers within a predetermined time period for the overlay and / or no longer uses voice data already used for the overlay for further overlays and rejects it in particular.
    Furthermore, it can be provided that the receiving subscriber computer exclusively uses voice data from communication servers within a predetermined period of time for the overlay and / or no longer uses the voice data already used for the overlay for further overlays and in particular discards it.
    Particularly advantageously, a computer program for carrying out a method according to the invention on a communication server or a computer program for carrying out a method according to the invention can be stored on a data carrier by a subscriber computer.
    The invention is illustrated by a preferred embodiment without limiting the generality. Fig. 1 shows an example of a voice communication system as known from the prior art. Figures 2a and 2b show the disadvantages of the prior art. FIGS. 3a and 3b show a compensation of the delay of the individual audio signals according to the invention in a server. Fig. 4 shows concretely the procedure taken in the communication server. Fig. 5 shows the creation and shipping of the data packets. FIGS. 6a to 6c show an exemplary embodiment of possible delay times in a network. FIGS. 7a to 7c show the procedure in the receiving subscriber computer.
    As already mentioned, in the prior art, shown in FIG. 1, there is a voice communication system comprising four subscribers 1 x, 1 y, 1 z, 1 w each having a subscriber computer 10 x, 10 y, 10 z, 10 w and two communication servers 2 a, 2 b. The individual subscriber computers 10x, 10y, 10z, 10w are interconnected via a network 3 comprising a number of connection lines 3a, 3b, 3c, 3d, 3e, 3f, 3g, 3h. The sending subscriber computers 10x, 10y, 10z have a voice output unit with which the individual voice data superimposed on it are output in the sequence defined by the time stamps τ or other time stamps τ 'and by the order of arrival.
    The voice data of the transmitting participants 1x, 1y, 1z are each created by a voice recording unit in the form of a microphone 12x, 12y, 12z. The transmitting subscriber computers 10x, 10y, 10z each have an encoding unit, not shown, which generates voice data values x, y, z on the basis of the signals received by them from the microphones 12x, 12y, 12z. The receiving subscriber computer 10w has a voice output unit 11 in the form of headphones, speakers may alternatively be used.
    In the present exemplary embodiment, it is shown how the individual voice information delivered by the transmitting subscribers 1 x, 1 y, 1 z are transmitted to the receiving subscriber 1 w. Even if only one communication path from the transmitting subscribers 1x, 1y, 1z to the receiving subscriber 1w is shown here, it is of course also possible for each of the subscribers 1x, 1y, 1z, 1w to exchange the superimposed voice messages of all other subscribers 1x, 1y , 1z, 1w receives received. In the course of a conference call thus the voice information of all participants 1x, 1y, 1z, 1w are transmitted to all other participants 1x, 1y, 1z, 1w.
    So that faults in individual lines or the failure of a communication server do not lead to interruptions or to the cutting off of individual subscribers 1 x, 1 y, 1z, 1 w from the communication with the other subscribers 1 x, 1 y, 1 z, 1 w, are two identically structured redundant communication servers 2a, 2b provided, which may also be provided even more communication server.
    As shown by way of example in Fig. 1, a connection line 3e is interrupted so that the signal output from the transmitting party 1z does not arrive at the communication server 2a. Due to the redundant design of the communication servers 2a, 2b, however, it is possible to transmit the voice signal from the transmitting subscriber 1z via the connection line 3f and the communication server 2b to the receiving subscriber 1w. Such a procedure is basically known from the prior art. However, the redundant design of the individual communication servers 2a, 2b and the use of different connection lines 3a to 3h is accompanied by the disadvantage described above that due to different line delays, voice signals of individual transmitting subscribers 1x, 1y, 1z arrive at the communication servers 2a, 2b at different times , The communication servers 2 a, 2 b receive and superimpose the signals emitted by the transmitting subscribers 1 x, 1 y, 1 z each with a specific time offset and forward these mixed voice communication signals to the receiving subscriber 1 w via the data lines 3 g and 3 h. Due to the different time delays on the lines 3a, 3b, 3c, 3d, 3e, 3f and the different time delays on the lines 3g, 3h individual transmitted voice information is transmitted multiple times via different lines each with different time offset to the receiving party 1w.
    FIGS. 2a and 2b show in more detail the block diagrams of a communication system with two redundant data paths P1, P2 between the transmitting subscriber 1x and the receiving subscriber 1w, the upper path P1 the data lines 3a and 3g, the lower path P2 the data lines 3b and 3h. The speech data x (t) are passed over the paths, each path P1, P2 having a certain predetermined time delay ΔΤ. As also shown in FIG. 1, there are now sometimes different data paths P1, P2 for the transmission of voice data signals from one subscriber to another subscriber 1w. In Fig. 1, for example, data from the transmitting subscriber 1x, 1y, 1z to the receiving subscriber 1w on the one hand via the lines 3a and 3g for the subscriber 1x, the lines 3c and 3g for the subscriber 1y and the lines 3e and 3g for the subscriber 1z on the other hand are transmitted over the lines 3b and 3h for the subscriber 1x, the lines 3d and 3h for the subscriber 1y and the lines 3f and 3h for the subscriber 1z. Both paths P1, P2 have a different time delay, the first path P1 has a time delay T1, the second path P2 has a time delay T2 = T1 + ΔΤ, which is greater than the time delay T1 by a time offset. At the receiving subscriber, these two voice data information x1 (t), x2 (t) transmitted via the separate paths P1, P2 can be added (FIG. 2a), amplified in an amplifier K and transmitted to the respective receiving subscriber, for example via a loudspeaker and output played to the receiving participant 1w. This results in the initially described unpleasant overlay or cancellation effect, due to the different time offset T1, T2 via two separate paths P1, P2 in the network. If errors occur, it may be necessary to switch between the individual paths. When switching between the paths P1 and P2 (Fig. 2b), there are interferences in the form of artifacts, which have a negative effect on the transmission quality.
    As shown in FIGS. 3a and 3b, a method is known from the prior art in which a compensation unit 11 is provided at the receiving subscriber computer 10w, that is to say at the computer of the receiving subscriber 1w, which determines by means of correlation whether the signals x1; x2 are correlated and, if so, with which time delay ΔΤ the two superposed partial signals Xi (t), x2 (t) at
    Arrive at the recipient's party. If this time delay ΔΤ known, the respective signal can be compensated. Hereby one obtains the compensated signals xS (t), x'2 (t). These can be added together (FIG. 3 a), the sum obtained being amplified by an amplifier K. As a result, one obtains a result r (t) in which no cancellations are heard. However, such a compensation is relatively resource-consuming and can only be carried out with great effort in real time, so that in this way is to be expected with much greater delay times. When switching between the paths P1 and P2 (FIG. 3b), there are no disturbances in the form of artifacts.
    However, such a compensation is no longer possible in the network illustrated in FIG. 1, in which the signals x, y, z of the transmitting stations 1x, 1y, 1z have already been superposed in the servers 2a, 2b and different delays in the individual are already mutually exclusive superimposed signals exist.
    In Fig. 4, the procedure is shown in more detail in an embodiment of the invention. Each transmitting subscriber 1x, 1y, 1z respectively generates voice information x (t), y (t), z (t) via a microphone 12a, 12b, 12c assigned to its subscriber computer 10x, 10y, 10z. This voice information x (t), y (t), z (t) is converted by the respective transmitting subscriber computer 10x, 10y, 10z into individual successive voice data Xi ... xn, y ·, ... yn "ζΛ ... zn , divided. In the present case, the speech data are sample values x (L), x (t2),... X (tn), which are recorded at individual, equidistant sampling instants L... Tn. Each sampling time ti ... tn is indicated by a continuous time stamp τ1; ... τη, where the discretized samples are referred to below as x ^] = x (T). These sampled values x (ti), x (t2),... X (tn) are assigned their respective time stamps τ1; τ2, ... τη to data packets X ·, ... Xn, ... Yn, Ζλ ... Zn summarized (Fig. 5). Each data packet Xi ... Xn, Yi ... Yn, Ζλ ... Zn comprises at least one sample χ [τ-1], y [x-i], zfa] and a time stamp τ1; which is created in the present case by the sending user computer 10x, 10y, 10z and in the form of a continuous integer τ1; ... τη is awarded. It is also possible that a data packet Xi ... Xn, Yi ... Yn, Zi ... Zn contains a plurality of samples, as long as the temporal succession of all samples x, y, z is at least implicitly detectable. Thus, for example, a plurality of samples x, y, z, which were determined with a known sampling rate, can be transmitted in a common data packet Xi... Xn, Yi... Yn, .omega..times..Zn, wherein only the sampling instant t.sub.i of the first sample value is specified and sampling times t2, ... tn of the remaining samples with knowledge of the sampling frequency can be determined. At each sampling time t1; ... tn or when creating the respective first sample value of the data packet, a counter is thereby increased by a predetermined value, in particular 1, and used as a time stamp. The timestamp τ1; ... τη corresponds to the respective counter value at the time of recording t1; ... tn, that is an integer corresponding to the time for sampling or recognition of the respective sample.
    The timestamp τ1; ... τη are not synchronized between the individual subscriber computers 10x, 10y, 10z and have different values. However, the sampling rate is the same for all sending subscriber computers 10x, 10y, 10z.
    The data packets Xi... Xn, Yi... Yn, Zi... Zn of each transmitting subscriber computer 10x, 10y, 10z are transmitted to the communication server 2a, 2b connected to the transmitting subscriber computers 10x, 10y, 10z. each
    Communication line 3a, 3b, 3c, 3d, 3e, 3f in this case has a different delay time Ta, ... T, so that the individual created by the sending subscriber computers 10x, 10y, 10z data packets Xi ... Xn, Yi ... Yn, ΖΛ ... Zn arrive at respectively different times at the communication servers 2a, 2b (FIG. 5). The communication lines 3g, 3h between the communication servers 2a, 2b and the receiving subscriber computer 10x have different delay times 3g, 3h. Each of the sending subscribers 1x, 1y, 1z transmits to the communication servers 2a, 2b the data packets Xi ... Xn, Yi ... Yn, Ζλ ... Zn, which in addition to the voice information also each have a timestamp τ, which the respective Time t of the creation of the respective language information indicates, so that the total
    Order of creating the language information for the individual communication server 2a, 2b is known. Each of the communication servers 2a, 2b determines for each incoming data packet Xi ... Xn, Yi ... Yn, Zi ... Zn of a sending user computer 10x, 10y, 10z respectively the remainder of the division of the time stamp τ divided by a predetermined positive integer N. The remainder of the remainder is referred to below as τ% Ν. This number N is the same for all the communication servers 2a, 2b. In this preferred form of embodiment of the invention, N is set to a value for which:
    In the present embodiment, the sampling rate is 16 kHz, so that the number of samples per time is 16,000. Each data packet Xi ... Xn, Yi ... Yn, ~ Z - ... Zn contains 80 samples, so that 200 data packets Xi ... Xn, Yi ... Yn, Zi ... Zn be created by each sending participant. The sampling time tsamp corresponds to the inverse number of the sampling frequency and due to the line delays in the transmission network 3, different maximum delay times, which are taken into account for the determination of the number N result. N is constant and is determined offline due to the expected minimum and maximum delay times. The time period tdeiay, max gives the maximum delay time among the connection lines 3a-3h between a sending user computer 10x, 10y, 10z, 10w and a communication server 2a, 2b. The time period tdeiay, min indicates the minimum delay time among the connection lines 3a-3h between a subscriber computer 10x, 10y, 10z, 10w and a communication server 2a, 2b. In the present embodiment, N is set as the smallest integer satisfying the above inequality. For a simple representation, N is set to the value 8 in the present embodiment.
    As shown in FIGS. 6a to 6c, the individual data packets Xi..Xn, Yi... Yn, .about.Z-... Zn which arrive at each communication server 2a, 2b are stored in a memory 211, 212, 213, each of which has N memory locations for each of the sending subscriber computers 10x, 10y, 10z connected to the communication server 2a, 2b. For each of the timestamps τ in a data packet ... Xn, Yi ... Yn, ~ Z- ... Zn, a further time stamp τ '= τ% Ν is determined, which divides the remainder of the division of the time stamp τ by the number N corresponds. The language information χ [τ], y [x], ζ [τ] contained in the data packet Xi ... Xn, Yi ... Yn, Zi ... Zn are stored in the memory 211, 212, 213 of the respective communication server 2a, 2b in the memory space of the respective sending
    Subscriber computer 10x, 10y, 10z and the respective other time stamp τ 'corresponds to stored.
    After all transmitting subscribers 1 x, 1 y, 1 z have transmitted to the communication server 2 a, 2 b respectively voice data with a time stamp τ, this superimposes the voice data to a superimposed voice data and transmits a data packet A1 containing voice data; ... to, B1; ... Bn to the receiving subscriber computer 10w connected to it. In the present case superimposed speech data values determined as the sum of the incoming speech data. All voice data x, y, z arriving at the communication servers 2a, 2b belong to a further data packet A1; ... to, B1; Bn are assigned and transmitted to the receiving subscriber 1w. The servers 2a, 2b each execute the steps shown below and act independently of each other.
    In Fig. 6a, the superposition of the speech signals at a first time at the first communication server 2a is shown. In the first communication server 2a, the data packets Xi, Yi, Z1 extend; one. In each case, the division remainder of the division of the transmitted time stamp Xi is determined by N and the speech data values xi [xi], yi [xi] and z ^ Xi] are written to the relevant locations in a memory 211, 212, 213 provided for this purpose. The memory positions described in each case are marked with IN. The communication server 2 a provides an index OUT, which is incremented with each execution step and is reset to the value 0 when the memory size is exceeded. All values in the memories 211, 212, 213, which are stored at the memory position τ '= 5, are superimposed and stored in a further memory 220 at the position τ' = 5. Overlaying results in the value a [OUT] = x [OUT] + y [OUT] + z [OUT] = 215 + 104 + 44 = 363. At the output of the communication server 2a, there is a data packet Ai which equals the value of the index OUT = 5 as a time stamp τ and the overlay value 363 contains. Similarly, the second communication server 2b creates data packets B1 independently of the communication server 2a; B2, B3 and transmits them to the receiving subscriber computer 10w.
    In Fig. 6b, the superposition of the speech signals x, y, z is shown at a second time at the first communication server 2a. At the first communication server 2a, the data packets X2, Y2, Z2 are received. In each case, the remainder of the division of the transmitted time stamp ^ is determined by N and the speech data values x2 [x2], y2 [x2] and z2 [x2] are written to the relevant locations in a memory 211, 212, 213 provided for this purpose. The memory positions described in each case are marked with IN.
    In the server 2a increases the index OUT, which now has the value OUT = 6. All values in the memories 211, 212, 213, which are stored at the memory position τ '= 6, are superimposed and stored in a further memory 220 at the position τ' = 6. Overlaying results in the value a [OUT]
    = 225 + 116 + 21 = 362. At the output of the
    Communication server 2a is applied to a data packet A2, which contains the value of the index OUT = 6 as a time stamp τ and the overlay value 362. Similarly, the second communication server 2b creates data packets B1 independently of the communication server 2a; B2, B3 and transmits them to the receiving subscriber computer 10w.
    In Fig. 6c, the superposition of the speech signals x, y, z is shown at a third time at the first communication server 2a. At the first communication server 2a, the data packets X3, Y3, Z3 are received. In each case, the division remainder of the division of the transmitted time stamp τ! are determined by N and the voice data values χ3 [τ3], γ3 [τ3] and ζ3 [τ3] are written to the respective locations in a designated memory 211, 212, 213. The memory positions described in each case are marked with IN. In the server 2a increases the index OUT, which now has the value OUT = 7 .. All values in the memories 211, 212, 213, which are stored at the memory position τ '= 7, are superimposed and in another memory 220 at the Position τ '= 7 stored. Overlaying results in the value a [OUT]
    = 17 + 139 + 271 = 427. At the exit of the
    Communication server 2a is a data packet A3, which contains the value of the index OUT = 7 as a timestamp τ and the overlay value 427. Similarly, the second communication server 2b creates data packets B1 independently of the communication server 2a; B2, B3 and transmits them to the receiving subscriber computer 10w.
    The receiving subscriber computer 10w in turn is connected to the communication servers 2a, 2b and each receive data packets A, B, to which a time stamp τ 'and voice data a, b are assigned. Like the communication server 2a, 2b, the respective subscriber computer 10w superimposes all the voice data received from it on data packets A1 transmitted to it; A2, A3, B1; B2, B3, which are each assigned the same time stamp τ '. As with the communication servers 2a, 2b, an addition may be used for superimposing the voice data a, b at the receiving subscriber computers 10w.
    The receiving subscriber computer 10w also has memory 111, 112 in which a memory space is available for all the communication servers 2a, 2b connected to it as well as for all values 0... N-1. The individual in the data packets A1; A2, A3, B1; B2, B3 speech data stored in memory 111, 112 at the point which is denoted by an index associated with the respective communication server 2a, 2b and the time stamp τ 'associated with the respective data packet Ai, A2, A3, Bi, B2, B3. If, for a time stamp τ ', voice information is available from all subscribers or a predefined time-out time span has elapsed, the superimposed voice data is transmitted to the respective receiving subscriber 10w, for example output via a loudspeaker. In order to avoid that the receiving subscriber computer 10w exclusively uses those voice data of communication servers 2a, 2b within a predetermined period of time for the overlay, which have been created in the predetermined period of time, voice data afc], bfc] once used for the overlaying are not used for further overlays more used and therefore discarded.
    In Fig. 7a, the superposition of the speech signals is shown at a first time. At the receiving subscriber computer 10w, the data packets A1 extend; B- a. In each case, the division remainder of the division of the transmitted time stamp τ 'is determined by N and the voice data values A1; are described to the respective positions in the respective memory 111, 112. The memory positions described in each case are marked with IN. In the subscriber computer 10w, an index OUT counts the numbers from 0 to N-1 and in the present case has the value OUT = 3. All values in the memories 111, 112, which are stored at the memory position τ '= 3, are superimposed. The incoming value 122 of the voice data packet Bi is written to the memory 112 in the location indicated by the data packet Bi at τ = 4. The incoming value 363 of the voice data packet Ai is written to the memory 111 in the location indicated by the data packet Ai at τ = 5. In the case of determining the output signal by superimposing the voice data a, b, the value f [OUT] = a [OUT] + b [OUT] = 108 + 54 = 162. The data value f [OUT at the output of the receiving subscriber computer 1w ], which is brought to the receiving party 10w note. Alternatively, the output signal at the subscriber computer 10w can also be determined by selecting one of the two voice data a or b.
    In Fig. 7b, the superposition of the speech signals a, b is shown at a second time. At the receiving subscriber computer 10, the data packets A2, B2 arrive. In each case, the division remainder of the division of the transmitted time stamp τ 'is determined by N and the voice data values A2, B2 are written to the relevant positions in the respective memory 111, 112. The memory positions described in each case are marked with IN. In the subscriber computer 10w, an index OUT counts the numbers from 0 to N-1 and in the present case has the value OUT = 4. All values in the memories 111, 112, which are stored at the memory position τ '= 4, are superimposed. The incoming value 102 of the voice data packet B2 is written to the memory 112 in the location indicated by the data packet B2 at τ = 5. The incoming value 362 of the voice data packet A2 is written to the memory 111 in the location indicated by the data packet A2 at τ = 6. In the case of determining the output signal by superimposing the voice data a, b, the value f [OUT] = a [OUT] + b [OUT] = 259 results. At the output of the receiving subscriber computer 10y, the data value f [OUT] is present the receiving subscriber 10w is brought to the knowledge. Alternatively, the output signal at the subscriber computer 10w can also be determined by selecting one of the two voice data a or b.
    In Fig. 7c, the superposition of the speech signals a, b is shown at a second time, in the receiving subscriber computer 10 long the data packets A2, B2 a. In each case, the division remainder of the division of the transmitted time stamp τ 'is determined by N and the voice data values A2, B2 are written to the relevant positions in the respective memory 111, 112. The memory positions described in each case are marked with IN. In the subscriber computer 10w, an index counts the numbers from 0 to N-1 and in the present case has the value OUT = 5. All values in the memories 111, 112 which are stored at the memory position τ '= 5 are. The incoming value 84 of the voice data packet B3 is written to the memory 112 in the location indicated by the data packet B3 at τ = 6. The incoming value 427 of the voice data packet A3 is written to the memory 111 in the location indicated by the data packet A3 at τ = 7. In the case of determining the output signal by superimposing the voice data a, b, the value f [OUT] = a [OUT] + b [OUT] = 363 + 102 = 465 results. At the output of the receiving subscriber computer 10w, the data value f [OUT ], which is brought to the receiving subscriber computer 1w note. Alternatively, the output signal at the subscriber computer 10w can also be determined by selecting one of the two voice data a or b.
    Instead of superimposing the voice data in the receiving subscriber computer 10w, it is also possible for the receiving subscriber computer 10w to make a selection among the available data packets Ai, An, Bi, Bn, each with the same further time stamp τ '. Thus, for example, it can be specified that the first communication server 2a is prioritized and data packets Ai,..., An originating from it are compared with the data packets B1; originating from the second communication server 2b; ..., Bn are preferred. If no data packets with a specific further time stamp τ 'are available from the prioritized communication server 2a, selection can be made on the data packets B1; ..., Bndes second communication server 2b are used. Due to the knowledge of the exact recording time arise despite the different data path of the data packets B1; ..., Bn opposite the data packets A1; ..., no artifacts at the first communication server 2a when switching.
    Of course, it is of course possible that each of the subscriber computers 10w ... 10z simultaneously transmits voice data and receives the voice data of the other subscriber computers 10w ... 10z, i. at the same time as sender and receiver
    Participant computer acts. In this case, all voice data in the servers 2a, 2b are superimposed on one another and data packets A, B created by superposition are transmitted to all subscriber computers 10w... 10z.
    Moreover, it is also possible to achieve a voice communication only between two subscriber computers 10x, 10w, without a conference call would be required. Such an embodiment of the invention has at least two communication servers 2a, 2b, which are connected to the two subscriber computers 10x, 10w. The sending subscriber computer 10x respectively generates voice data which, as shown in the previous exemplary embodiment, is transmitted to the communication servers 2a, 2b. Whereas in the previous exemplary embodiment summation of the individual voice data values x, y, z is performed in the individual communication servers 2a, 2b, in this embodiment of the invention, only an additional time stamp is assigned in the same way as in the first embodiment of the invention. As in the first exemplary embodiment of the invention, the voice data is transmitted to the receiving subscriber computer 10w, which transmits the data in the data packets A1; ..., An, B1; ..., Bn speech data either summed or the data packets A1; ..., An, B1; ..., Bn selects one of the communication servers 2a, 2b. The selected or summed voice data is output via the headphone 11. In this exemplary embodiment, too, both subscriber computers 10w, 10x can be designed as sending and receiving subscriber computers.
    权利要求:
    Claims (14)
    [1]
    1. A method for transmitting voice data from a transmitting subscriber (1x) to a receiving subscriber (1w), wherein the sending subscriber (1x) each a transmitting subscriber computer (10x) and the receiving subscriber (1w) a receiving subscriber computer (1w ) and the two subscriber computers (10x, 10w) are networked via a plurality of redundant communication servers (2a, 2b), characterized in that voice information is created by the transmitting subscriber (1x) and this voice information is transmitted from the respective sending subscriber computer (1 Ox ) are divided into individual, numerical voice data values (x), - data packets (X1; ... Xn) comprising at least one numerical voice data value (x), in particular a plurality of voice data values recorded directly after one another, are generated by the transmitting subscriber (1 x), each time assigned at least one time stamp (τ) in the form of a consecutive number n, which determines the order of the individual voice data values (x) relative to the voice data values of other data packets, - that the data packets (Xi, ... Xn) of the transmitting subscriber (1 x) communicate with the communication server (1 Ox) connected to the transmitting subscriber computer (1 Ox) ( 2a, 2b), - that from each communication server (2a, 2b) in each case further data packets (A1; ..., At; B1; ..., Bn) are created, to which at least one of the transmitting subscriber computer (1 Ox) transmitted voice data (x) and at least one, further time stamp (τ ') are assigned, wherein the value of the further time stamp (τ') the rest the division of the time stamp (τ) divided by a predetermined constant positive integer N, where N is chosen to be the same for all communication servers (2a, 2b), and - that the receiving subscriber computer (10w) receives it from the communication servers (2a, 2b) transmitted further data packets (A1; ... An; B1; ... Bn) and determines such further data packets (A1; ... An; B1; ... Bn) which are each assigned the same further time stamp (τ ') and a voice data value (w) is determined by selecting or summing the voice data values contained in the further data packets (A1; ... An; B1; ... Bn) with respectively the same further time stamp (τ ') and the receiving subscriber (1w ) is available.
    [2]
    2. A method, in particular according to claim 1, for transmitting voice data between a plurality of transmitting participants (1x, 1y, 1z) to at least one receiving subscriber (1w) in the form of a conference call, each transmitting subscriber (1x, 1y, 1z) via a transmitting subscriber computers (10x, 10y, 10z) and each receiving subscriber (1w) has a receiving subscriber computer (10w) and these are networked via a plurality of redundant communication servers (2a, 2b), characterized in that - by the sending subscribers (1x, 1y, 1z) voice information is created and this voice information from each transmitting subscriber computer (10x, 10y, 10z) into individual successive voice data values (x, y, z) are divided, - that the sending participants (1 x, 1y, 1z) data packets (Xi, ... Xn, Yi, ... Yn, Z1, ... Zn) each comprising at least one numerical voice data value (x, y, z), in particular a plurality of directly after one another a Recorded voice data values, each of which is assigned at least one time stamp (τ) in the form of a consecutive number which determines the order of the individual voice data values (x, y, z) to the voice data values of other data packets of the same user (1x, 1y, 1z), - That the data packets (Xi, ... Xn, Yi, ... Yn, ~ Ζ-λ, ... Zn) of each transmitting subscriber (1 x, 1y, 1z) to the transmitting subscriber computer (10x, 10y , 10z) are communicated to the communication server (2a, 2b), - that the communication server (2a, 2b) transmit data packets (Xi, ... Xn, Yi, ...) transmitted by the sending user computers (10x, 10y, 10z) to each of them. Yn, Z1; ... Zn) and store the voice data values (x, y, z) contained in the data packets, wherein the individual voice data values (x, y, z) are respectively linked to a further time stamp (τ ') associated with the remainder of the Division of the time stamp (τ) divided by a predetermined positive integer N corresponds, - that the communication server (2a, 2b) superimpose voice data values (x, y, z) to voice data values (a, b), each with the same further time stamp ( τ '), - that the communication server (2a, 2b) from the voice data (a, b) create further data packets (A ^ A, Β ^ .Βη), where they each the result of the superposition of the same further time stamp (τ ') assigned voice data values (x, yz) and the respective time stamp (τ'), and - that the receiving subscriber computer (10w) all of him from the communication servers (2a, 2b) transmitted data packets (A1; ... An, B1; ... Bn), each of which the same time stamp (τ ') is assigned, superimposed or from the transmitted data packets (A1; ... to, B1; ... Bn) to which the same time stamp (τ ') is assigned, selects a data packet and makes it available to the second user (1w).
    [3]
    3. The method according to claim 1 or 2, characterized in that all of the communication servers (2a, 2b) incoming voice data (x, y, z) at least, in particular, another data packet (Ai, ... An, Bi,. .. Bn) and transmitted to the receiving subscriber (1w).
    [4]
    4. The method according to any one of the preceding claims, characterized in that each data packet (X1, ... Xn, Yi, ... Yn, Zu ... Zn) each voice data values (x, y, z) in the form of a single sample or multiple samples and that the speech data (x, y, z) are superimposed as samples in the communication servers (2a, 2b).
    [5]
    5. The method according to any one of the preceding claims, characterized in that N is determined as follows:

    where tdeiay.max denotes the maximum delay time of a connection line between a sending or receiving subscriber computer (10x, 10y, 10z, 10w) and a communication server (2a, 2b), where tdeiay.min is the minimum delay time of a connection line between a transmitting or receiving one Subscriber computer (10x, 10y, 10z, 10w) and a communication server (2a, 2b), and - where tsamp corresponds to the inverse of the sampling rate.
    [6]
    6. The method according to any one of the preceding claims, characterized in that the communication server (2a, 2b) exclusively data packets (X1, ... Xn, Yi, ... Yn, Z1; ... Zn) of sending subscriber computers (10x, 10y, 10z) within a predetermined period of time for the superposition and / or already used for the overlay data packets (X1, ... Xn, Yi, ... Yn, Z1; ... Zn) for further overlays no longer uses and in particular, rejects.
    [7]
    7. The method according to any one of the preceding claims, characterized in that the receiving subscriber computer (10w) exclusively uses voice data (a, b) of communication servers (2a, 2b) within a predetermined time period for the overlay and / or already used for the overlay language data (a, b) is no longer used for further overlays and in particular rejected.
    [8]
    8. A voice switching system for transmitting voice data from a transmitting subscriber (1x) to a receiving subscriber (1w), the transmitting subscriber (1x) each having a transmitting subscriber computer (1 Ox) and the receiving subscriber (1w) receiving a subscriber computer (10w) and the subscriber computers (10x, 10w) are networked via a plurality of redundant communication servers (2a, 2b), characterized in that the transmitting subscriber computer (10x) has a voice recording unit and an encoding unit which converts the voice information created by the voice recording unit into subdivides individual consecutive voice data values (x), and creates data packets (X1; ... Xn), each of which comprises at least one voice data value (x), in particular several voice data values (x) recorded immediately after one another and one associated with the voice data values (x) Timestamp (τ) created in a timer have sequential number, wherein the time stamp (τ) indicates the order of the voice data values (x), - that the transmitting subscriber computer (1 Ox) the data packets (Xi, ... Xn) connected to the transmitting subscriber computer (10x) communication server (2a , 2b), - that the communication server (2a, 2b) in each case further data packets (A1; ..., At; B1; ..., Bn), to which at least one voice data value (x) transmitted by the transmitting subscriber computer (1 Ox) and at least one further time stamp (τ ') are assigned, the value of the further time stamp (τ') being assigned to the rest of the Division of the time stamp (τ) divided by a predetermined positive integer N corresponds, where N is the same for all communication servers (2a, 2b), and - that the receiving subscriber computer (10w) him another of the communication servers (2a, 2b) transmitted Receives data packets (A1; ... An; B1; ... Bn) and determines such further data packets (A1; ... An; B1; ... Bn) to which the same further time stamp (τ ') is assigned in each case, and a voice data value (w) is determined by selecting or summing the voice data values contained in the further data packets (A1; ... An; B1; ... Bn) with respectively the same further time stamp (τ ') and the receiving subscriber (1w) Available.
    [9]
    9. Voice switching system, in particular according to claim 1, for the transmission of voice data between a plurality of transmitting participants (1x, 1y, 1z) to at least one receiving party (1w) in the form of a conference, wherein sending participants (1x, 1y, 1z) via a sending Subscriber computer (10x, 10y, 10z) and each receiving subscriber (1w) has a receiving subscriber computer (10w), and these are networked via a plurality of redundant communication servers (2a, 2b), characterized in that - the transmitting subscriber (1x , 1y, 1z) in each case generate voice information and this voice information is subdivided by the respective transmitting subscriber computer (10x, 10y, 10z) into individual successive voice data values (x, y, z), - that the transmitting subscriber computers (10x, 10y, 10z) respectively have a voice recorder unit and a coding unit that contain the voice information created by the voice recorder unit divided into individual successive speech data values (x, y, z), and data packets (X1; ... Xn, Yi, ... Ynj Zu ... Zn), each of the voice data values (x, y, z) and a timestamp (τ) assigned to the voice data values (x) and created by a timer in the form of a wherein the time stamp (τ) indicates the order of the speech data values (x, y, z) to the speech data values (x, y, z) of other data packets (X1, ...) generated by the same subscriber computer (10x, 10y, 10z). Xn, Yi, ... YRj Ζλ, ... Zn), - that the sending subscriber computers (10x, 10y, 10z), the data packets created by them (Xi, ... Xn, Yi, ... Yn, Ζλ, ... Zn) to the communication server (2a, 2b) connected to the respective transmitting subscriber computer (10x, 10y, 10z), - that the communication server (2a, 2b) sends them respectively from the sending subscriber computers (10x, 10y, 10z) receive transmitted data packets (X1; ... Xn, Yi, ... Yn, Zu ... Zn) and store the voice data values (x, y, z) contained in the data packets in a memory and the individual voice data values (x, y, z) are each linked to a further time stamp (τ '), which corresponds to the remainder of the division of the time stamp (τ) divided by a predetermined positive integer N, - the communication servers (2a , 2b) voice data values (x, y, z) superpose each other to voice data values (a, b), which are each associated with the same further time stamp (τ '), - that the communication server (2a, 2b) from the voice data values (a, b ) create further data packets (A ^ A, B ^ A) to which they respectively assign the result of the superimposition of the same further time stamp (τ ') associated voice data values (x, y, z) and the respective time stamp (τ'), and the receiving subscriber computer (10w) transmits all the data packets (A1; ... to, B1; ... Bn), to each of which the same time stamp (τ ') is assigned, superimposed or from the transmitted data packets (A1; ... An, B1; ... Bn) to which the same time stamp (τ') is assigned Selects data packet and makes it available to the second subscriber (1w).
    [10]
    10. The voice switching system according to claim 8 or 9, characterized in that each data packet (X1, ... Xn, Yi, ... Yn, Zu ... Zn) each voice data (x, y, z) in the form of a single sample and that the overlay units of the communication server (2a, 2b) or the subscriber computer (10x, 10y, 10z, 10w) each superimposed as sums of the incoming them voice data (x, y, z).
    [11]
    The voice switching system according to any one of claims 8 to 10, characterized in that N is set as follows:

    where tdeiay.max denotes the maximum delay time of a connection line between a sending or receiving subscriber computer (10x, 10y, 10z, 10w) and a communication server (2a, 2b), where tdeiay.min is the minimum delay time of a connection line between a transmitting or receiving one Subscriber computer (10x, 10y, 10z, 10w) and a communication server (2a, 2b), and - where tsamp corresponds to the inverse sampling rate.
    [12]
    12. Voice switching system according to one of claims 8 to 11, characterized in that the communication server (2a, 2b) exclusively voice data (x, y, z) of sending subscriber computers (10x, 10y, 10z) within a predetermined time period for the superimposition and uses / or already used for the overlay language data (x, y, z) for further overlays no longer uses and in particular rejects.
    [13]
    13. Voice-switching system according to one of claims 8 to 12, characterized in that the receiving subscriber computer (10w) exclusively uses voice data (a, b) from communication servers (2a, 2b) within a predetermined time interval for the superimposition and / or already for the superimposition used language data (a, b) is no longer used for further overlays and in particular discarded.
    [14]
    14. A data carrier on which a computer program for carrying out a method according to one of claims 1 to 7 on a communication server (2a, 2b) and / or subscriber computer (10x, 10y, 10z, 10w) is stored.
    类似技术:
    公开号 | 公开日 | 专利标题
    DE60104353T2|2005-07-21|Method and device for reducing the delay jitter in data transmission
    DE60130665T2|2008-06-26|AUDIO DATA PROCESSING
    DE60209637T2|2006-12-21|Controlling a conference call
    US7571210B2|2009-08-04|Methods and systems for linking virtual meeting attendees over a network
    EP1283611A2|2003-02-12|Method for synchronization of a communication system via a packet oriented data network
    EP1496680B1|2016-11-02|Conference system and method applying HRTF-filter
    AT515441B1|2016-01-15|Voice transmission in redundant systems
    EP1855437A1|2007-11-14|Method for setting up a Push-to-Talk communication connection
    DE102014105211A1|2015-10-15|Method for operating a subscriber of a communication network
    DE102010003248A1|2011-09-29|Method and device for processing data in a network of a vehicle
    DE112018006602T5|2020-09-10|VIRTUAL WINDOW FOR TELEPHONE CONFERENCES
    AT516870B1|2020-11-15|Communication system for voice communication
    EP3327699B1|2019-12-25|Method for transferring voice signals
    WO2016128145A1|2016-08-18|Method for transmitting data in a multimedia system, and software product and device for controlling the transmission of data in a multimedia system
    DE3222453C2|1987-10-08|
    EP3267621B1|2021-11-03|Method for generating an overall room noise for passing to a real endpoint, use of said method and teleconferencing system
    DE202019106741U1|2019-12-16|Interaction system
    AT518561B1|2020-09-15|Communication system for voice communication
    DE19819551C2|2001-06-28|Process for the transmission of electrical audio signals in real time
    DE102008048880B4|2010-07-08|A method, apparatus and computer program product for outputting communication contributions of a conference and method, apparatus and computer program product for generating a message with a communication contribution of a conference
    EP1142234B1|2004-11-03|Transmission system with echo cancellation
    DE102004041884A1|2006-03-09|Communication terminal and method for controlling a communication terminal employs messages containing user identity information
    DE102014116610A1|2016-05-19|Digital telephone conference system, subscriber access and switching device, working method and computer program product
    DE102009043804B4|2013-11-28|Providing a summary communication channel in a telecommunication system
    DE2104263C2|1977-05-05|Circuit arrangement for a telephone exchange for setting up television conference connections
    同族专利:
    公开号 | 公开日
    EP2913946B1|2016-06-15|
    AT515441B1|2016-01-15|
    EP2913946A1|2015-09-02|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    EP0661881A2|1993-12-24|1995-07-05|Sharp Kabushiki Kaisha|Image-information storage regenerating device|
    US20060007943A1|2004-07-07|2006-01-12|Fellman Ronald D|Method and system for providing site independent real-time multimedia transport over packet-switched networks|
    JP2007300498A|2006-05-01|2007-11-15|Matsushita Electric Ind Co Ltd|Repeating apparatus and repeating method|
    US20100034212A1|2008-08-11|2010-02-11|Qualcomm Incorporated|Methods and apparatus for providing modified timestamps in a communication system|
    US6327276B1|1998-12-22|2001-12-04|Nortel Networks Limited|Conferencing over LAN/WAN using a hybrid client/server configuration|
    WO2008147272A1|2007-06-01|2008-12-04|Telefonaktiebolaget Lm Ericsson |A conference bridge and a method for managing packets arriving therein|CN110867185A|2019-11-04|2020-03-06|江苏税软软件科技有限公司|Real-time voice-to-text input system and method with isolated internal and external networks|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    ATA50147/2014A|AT515441B1|2014-02-26|2014-02-26|Voice transmission in redundant systems|ATA50147/2014A| AT515441B1|2014-02-26|2014-02-26|Voice transmission in redundant systems|
    EP15455002.4A| EP2913946B1|2014-02-26|2015-02-13|Voice transmission in redundant systems|
    [返回顶部]