Method and circuit for establishing a conference connection

The present invention relates to a method and circuit for conducting teleconferences in a digital telephone switching system, where non-linearly encoded voice samples are transformed into linearly encoded voice samples that are then assembled and then retransformed into non-linearly encoded voice samples.

: It is known that because the voice signals of the subscribers in digital telephone switching systems, especially in the systems with modulation in pulse code - PCM - are encoded in digital values ​​according to a nonlinear - approximately logarithmic - and at the receiving end, decoded feature. in analog voice signals according to an appropriate feature, in order to improve the transmission quality, a teleconference, that is, a connection between more than two subscribers, cannot be established by simply summing up the digitized voice samples of the subscribers.

Thus, according to a known method, the voice samples, after being encoded in PCM words consisting of 8 bits according to a nonlinear characteristic, are converted into PCM words consisting of, sixteen bits in a non-linear / linear code converter. The de-realization of cold teleconferences remains an unoccupied domain between the positive and negative domains of the vocal samples, the allowed amplitude field should not be exceeded when subtotals are formed during the assembly of several vocal samples having relatively large, unidirectional amplitudes. Distortions of voice signals, however, can only be avoided if the total sample10 of the voice waves of all teleconferencing participants does not exceed the allowed amplitude range, that is, is not within the unoccupied residual domain of the extended domain per15 set for subtotals. If the total of the vocal samples is within the unoccupied residual domain, and the allowed amplitude range has been exceeded, the vocal sample will be limited to the maximum positive or negative value20 allowed during the linear / nonlinear code conversation. This leads to a distortion of the analog voice signal to be reconstituted before the distortion at reception25.

The method, according to the invention, eliminates the above disadvantages by realizing the teleconferences by reducing the distortion of the voice signals to the mini30 mum, foresees that the voice samples initiated by the individual participants in the teleconference will be integrated in time and

LEI PRICE 10.68 rr compared, the voice sample with the highest intensity being added unintentionally while the remaining voice samples are attenuated with a predetermined value before being assembled, and 5 the circuit for applying the method in a digital telephone switching system. which includes a first code converter, which converts non-linearly encoded voice samples into linearly encoded voice samples, a summator, which summarizes the transformed voice samples initiated by the participants at a teleconference, and a second code converter, which retransforms the values ​​of sum 15 into non-linearly encoded voice samples, and the voice samples of the individual participants are integrated in time and compared in an evaluation circuit.

Following is an example of an embodiment of the invention in connection with FIG. 1 and 2, representing:

FIG. 1 is a block diagram of a circuit position according to the invention, FIG. 2, the schematic of a potential regulator sec- 25 for the circuit of fig. 1.

The conference circuit, according to the embodiment of the invention is connected - iig. 1 - to the switching network of a PCM 30 telephone exchange through a connection unit (not shown). Through an RX reception line, a serial bitstream is transferred from the conference circuit switching network. From the conference circuit, 35 bit stream in the output series is transferred to the switching network via a TX transmission line.

As usual in multiplexing switching systems with distribution in ⁴⁰In PCM time, the input bit current is divided into thirty-two channels or time positions per frame, and each frame contains two hundred and fifty-six bits. Each speech sample, that is, whatever sample value of the voice signal is encoded in a PCM word consisting of eight bits, which gives two hundred and fifty-six quantification steps. The sampling frequency is 8 KHz. 50 Thirty-two channels are message channels, while channel number zero serves to frame synchronization, and channel number sixteen is available for signaling information 55. In one embodiment of the circuit, according to the invention, it is capable of forming six independent conference groups each consisting of up to five participants. 60

Each participant receives the total of the voice samples of the other participants, the PCM voice samples being signed in a linear coded form. If in person, attend a conference, circ⁵¹ 4 the conference cult produces 11 different summaries of nl of voice samples.

The vocal samples are evaluated by taking the time integral from them in a way that will be described later and the participant's vocal samples considered to have the highest sound level are added unintentionally, while the vocal levels of all other participants are added afterwards. were attenuated by 12 dB.

The serial bitstream arriving through the RX receiving line is fed into an SPC serial-parallel converter and from there into the form of groups of eight parallel bits, in a SIC printing / switching circuit. The two protocol bits representing the bit format of each channel are transferred from the SPC converter to the SIC circuit, which detects the unoccupied channels by checking the bit formats and suppresses such channels for further processing operations.

From the SIC circuit, the eight-bit words - each representing a vocal sample - are introduced, on the one hand representing a vocal sample - they are introduced, on the one hand, into an SSM vocal sample intermediate circuit, and on the other hand. , in an EVC evaluation circuit. In a first CFS voice sample comparator, all the voice samples of a conference are compared in groups. A CNT counter is assigned to each channel and, consequently, to each participant (only one is shown in the figure). After each comparison of the voice samples of a conference group, the counter associated with the participant with the highest sound level is increased with an impulse, while the counters associated with the other participants are reduced by an impulse. CNT meters register from 0 to 63 and thus thirty-two frames - ie 8X125 ps = -4 ms - at most until a meter associated with the participant with the loudest intensity has exceeded the meter associated with another participant that was before the speaker with the highest sound level. In this limit case, a participant must thus deliver the highest intensity vocal samples during thirty-two frames,

To determine the participant whose voice samples have the highest intensity, that is, the participant with the highest sound level, the associated meter impulses in the participants are compared in groups in a second CFS comparator. Based on this comparison, the CFC comparator provides control bits, so-called attenuation bits, to an LFL attenuation regulator, where they are temporarily stored. From there, the attenuation bits are transferred to a first CCL code converter, which performs a dual function. On the one hand, it converts non-linearly encoded vocal samples - produced by encoding analog voice signal through a technique commonly called "compressed encoding" - into linear encoding vocal samples as they are required for non-distortion-free assembly. Each non-linear coded PCM word is eight bits long, as mentioned above, while linearly encoded speech samples consist of each of thirteen bits. This is shown in FIG. 1 by the number of pairs of threads that form the lines connecting the various sub-circuits. From the SSM intermediate circuit of the voice sample, the compressed voice samples are transferred to the first CCL code converter through a line consisting of eight pairs of wires, while they exit this code converter as linear samples encoded linearly through thirteen pairs. wires. The attenuation bit is provided to the CCL converter by a pair of wires.

On the other hand, the CCL code converter attenuates all the vocal samples, except the one of the participant with the highest sound level. In one version, a 12 dB attenuation was chosen, but it is also easy to introduce other attenuations. The CCL converter is a module so it is easy to set a different attenuation value. Which voice samples are attenuated and which are not attenuated is determined by the attenuation bit, which is fixed in the LFL attenuation regulator.

The totalization of the results of the vocal sample comparisons performed in the first CFS comparator, which is performed in the CNT meters associated with the individual participants, and the model of these counters gives a first time integration of the vocal samples to be evaluated. In order to prevent any too frequent change of the attenuation as a reaction to instantaneous values, the impulses are compared in the second CFC comparator not in each frame, but only in each fourth frame. This gives a second integration of voice samples over time.

From the first CCL converter, the successive voice samples of the participants and the associated attenuation bits are transferred to an ADD adder, in which they are summed. After each single meeting, they are temporarily stored in an ACC memory.

After assembly, the voice samples are transferred to a second LCC code converter, where they are again transformed into linearly encoded voice samples, which are transferred to a PSC series-parallel converter through eight pairs of wires. From the PSC converter, the words PCM leave the conference circuit, according to the invention, in series form through the TX line; they are transferred through the connection unit associated with the switching network and from there to the participants.

Between the PCM channels received in the conference circuit and the starting channels, a time shift of 6 channels, ie approximately 24 μs is obtained.

The conference circuit, according to the invention, also makes it possible to put the circuit on busy lines. In a private secondary telephone exchange, for example, the calling party may circuit the busy line of the calling party, or an operator may enter the existing connection to announce a call from the public network.

By protocol bits, which are evaluated in the SIC circuit, as mentioned above, further information can be transferred and evaluated. An example is the information that a telephone subscriber line is very long, which is known in the telephone exchange; this information is used to subject the voice samples of the interested participant to a lesser amount of attenuation.

in FIG. 2, a ROC sequential regulator is presented, which receives 4 MHz and 8 MHz signals and FSYNC frame synchronization signals and emits CS control signals, which are applied to the individual sub-circuits of the conference circuit, according to the invention, to control the time sequence. of the operations performed there. Individual command lines from the ROC sequential controller to the various sub-circuits are not shown in order to simplify the drawing.

Because all but one of the vocal samples are attenuated in the conference circuit, according to the invention, before being summed, the probability that, and the amount with which, the amplitude of the vocal signal sum to be exceeded are substantially reduced.

The main advantage of the invention is that an improvement in signal-to-noise performance is achieved. In addition, the circular trigger attenuation required in a four-wire circuit with vocal signal amplification is maintained without increasing insertion attenuation. The stability and the echo conditions, determined by 84971 born of hybrids that close the four-wire circuits, are improved, and the distortion factor is reduced.