NO803527L - PROCEDURE FOR CUSTOMIZING A DIGITAL TONE SIGNAL TO A DATA FLOW - Google Patents

Abstract

1. Method for matching the clock rate of a voice frequency signal occurring in digital form to a data flow which is organised in pulse frames and superframes, serves for digital speech transmission of telephone quality and is provided with signalling bits for signalling transmission, and whose clock rate is plesiochronous to the clock rate of the digital voice frequency signal, characterized in that clock rate matching in respect of each superframe is carried out word-wise, i.e. with one entire code word of the digital voice frequency signal, that, in respect of each superframe in those positions (Z1, Z2 and Z3 in fig. 1) of the signalling bits (ZK 16) which have become free during the transmission of the digital voice frequency signal over a plurality of telephone channels (K1, K2 and K3 in Fig. 1) and which are assigned to the latter, there is transmitted one positive or one negative clock rate matching signal (++ or -- in Fig. 2 to 4 in R1 and R2), that, in the event of the identity of the clock rates of the data flow out of the digital voice frequency signals (fig.2), the clock rate matching signal (++ or --) is alternately positive and negative from one superframe to the next, that in the event of a positive clock rate matching (Fig. 3), the clock rate matching signals (++) of two consecutive superframes are both positive and within the second superframe - which contains the second of the directly consecutive positive clock rate signals - at an agreed position a blank word (LW) is inserted between two code words of the digital voice frequency signal (Tn6 and Tn7), and that, in the event of negative clock rate matching (fig. 4), the clock rate matching signals (--) of two consecutive superframes are both negative and the released positions of the signalling bits in the superframe which follows the second negative clock rate matching signal is filled with a code word gated out from the coded voice frequency signal or with gated out components of code words (Tn5, Tn7 and Tn9 in fig. 4).

Description

Procedure for timing adaptation of a digital tone signal to a data stream.

The present invention relates to a method for timing adjustment of a tone signal that exists in digital form, to a data stream organized in pulse frames and superframes, for digital speech transmission of telephony quality and with characteristic bits for characteristic transmission, whose rate is plessynchronous to the digital tone signal's rate, i.e. that the two bars--4 -7

Quenser has an insignificant deviation of e.g. 10 - 10

If the transformation of an analogue tone broadcast signal into digital form is controlled e.g. with the rate of a digital signal connection for 2-Mbit/s signals (DSV2, until now PCM30), the insertion of the resulting data stream into the pulse frames in DSV2 and the use of any subsequent further digital hierarchy steps are unproblematic. However, use cases can be imagined where the analog/digital conversion (A/D conversion) takes place spatially at a distance from the devices in DSV2 and the utilization of their rates is not possible for various reasons. For A/D conversion, only a locally sourced clock that is plesiochronous to the clock in DSV2 can then be used. In these cases, a tempo adjustment is necessary.

For summarizing the data streams in several digitalé "subsystems" with mutually plesiochronous clocks to a "supersystem", several clock adaptation procedures are known. These methods can be divided into those with and those without information loss. If, for the aforementioned use case, one wanted to take into account a beat adjustment with loss of information, then this, in order not to lose word synchronicity, would mean the omission or repetition of an entire code word from time to time. Investigations carried out by the applicant have shown that the result of these precautions was a quality deterioration of the tone signal unacceptable to the listener when the beat deviation is greater than 10— 7 Beat matching methods without information loss during bitwise beat matching are known from "Nachrichtentechnischen Fachberichten Nr. 42, 1972", PCM -technique, VDE-Verlag GmbH Berlin-Charlottenburg, pages 235 - 244 and from the journal

"Elektronik", booklet 6/1978, pages 78-83. In this connection, at an agreed place in the frame - mostly in an "supersystem" - blank bits are added (positive beat adjustment) or information bits are sorted out (negative beat adjustment).

This sequence must be fed to the receiver with the help of additional capacity that is available in the pulse frame, which reverses the manipulation that was carried out on the transmitter side, and again produces the original beat. Alongside the methods of positive clock adaptation or negative clock adaptation, where the clock frequency in the "subsystem" in the first method is lower and in the second method is higher than the clock frequency of the available data stream, there is a combination of the two possibilities as positive/negative beat adjustment. In this connection, relative tempo deviations can be processed in a positive and negative direction.

The multiplex system known from the "Taschenbuch

der Fernmeldepraxis 1979", pages 13 - 41, is suitable both for digital transmission of tone signals and for combined transmission of tone and telephony signals in the pulse frame in DSV2.

Fig. 1 shows the known structure of the pulse frame and the upper frame in a DSV2. This pulse frame with a duration of 125^is contains 256 bits, which corresponds to a data stream of

2048 kbit/s. The pulse frame is divided into 32 time channels ZKO - ZK31. The first time channel ZKO alternately serves to transmit a message and a frame characteristic word (RSy).

The seventeenth time channel ZK16 is mainly arranged for the transmission of the selection characteristics Zl - Z15, Z17 - Z31 for the thirty telephone channels Kl - K15, K17 - K31. Sixteen successively transmitted pulse frames RO - R15 form a superframe. The synchronization of the upper frame takes place by an upper frame synchronization word UR - Sy in the seventeenth time channel ZK16 in the first pulse frame RO. For the sake of overview is

the 16 pulse frames in an upper frame in fig. 1 drawn below each other.

The proposal known from the "Taschenbuch der Fernmeldepraxis" allows mixed coverage of DSV2 with telephony and tone channels. The first time channels ZKO and the seventeenth time channels ZK16 are thus not affected by the tone transmission.

With this digital tone transmission, high-quality channels with a 15 kHz bandwidth and medium-quality channels with a 7 kHz bandwidth are used. In the former case the sensing frequency is 32 kHz and in the latter case 16 kHz. Both are multiples of the sensing frequency of 8 kHz for telephony signals. Each sensing value with the protection bits against the bit errors that occur on the transmission path occupies a word of 12 bits in length. In a pulse frame, four 12-bit code words must therefore be arranged for a 15-kHz tone channel and two 12-bit code words for a 7-kHz tone channel. When transmitting a 15-kHz tone channel, six time channels ZK1 - ZK3, ZK17 - ZK19 are therefore covered per eight bits, and in the transmission of a 7-kHz tone signal three time channels ZK1 - ZK3. If additional tone channels are to be transmitted, additional time channels ZK4 - ZK6, ZK20 - ZK22 resp. ZK4 - ZK6. The bits in two consecutive code words are also shuffled together to protect against double errors on the transmission path. A 15-kHz tone channel thus has an information flow of 384 kbit/s, and there is an interface between the encoder/decode device of the tone signal and DSV2 with a data transfer rate of 384 kbit/s (for a 7-kHz tone channel: 192 kbit/s).

The known method has the disadvantage that it only requires a synchronous introduction of the information in coded tone channels in the frame in a DSV2. However, this is not always possible, as for example: at a digital cut point, between a broadcast studio and the Post Office's digital network when the rates have a permissible deviation of ^ 10 from each other,

in national digital networks, at the points of intersection between such parts of this network which do not, or have not yet,

is operated synchronously, and where the clock deviation is ^> 10 , and

in the exchange of digital tone signals between international networks which are indeed synchronous in themselves, but in relation to each other are plesiosynchronous with a time deviation > 10~<7>.

The present invention is based on the task of specifying a method for beat matching in which the beginning of the first code word of the digital tone signal within a pulse frame largely coincides with the beginning of the first time channel reserved for the tone transmission, i.e. that the the first bit in the digital tone signals within an upper frame is almost always the most significant bit in the code word.

This task is solved by the invention that is defined

in the main requirement.

The sub-claims relate to advantageous further developments of the invention.

The invention provides the advantage that a multiplex system which is arranged for the transmission of synchronously coded tone channels, without changing the frame structure, can also be used in connection with plesiochronous mode of operation of the clock generator in question. The beat matching according to the invention makes it possible to leave the boundaries of the code words in the digital tone signal unchanged in the pulse frame, and thereby carry out a rapid synchronization on the receiving side in the event of a transmission error.

A further advantage achieved by the fact that parts of a synchronous network within an inherently plesiochronously driven transmission network can be realized for digital tone signal transmission with a bit transmission rate of e.g. 192 kbit/s or multiples of this, whereby time-multiplex transmission or cross-linking is enabled.

The invention will be described in more detail under reference

to figures 2-4.

Fig. 2 shows the structure of the data stream in a DSV2 at

a plesiochronous 7-kHz tone signal transmission in synchronous mode, Fig. 3 shows a possible structure of the data stream in the case of a positive, and Fig. 4 in the case of a negative clock adaptation of the digital 7-kHz tone signal.

In fig. 2 - 4, three and three upper frames are drawn below each other, which follow each other in time, and each of which consists of 16 pulse frames RO - R15, which in turn are made up of 32 time channels ZKO - ZK31 of eight bits.

In fig. 2, in each upper frame, for example, the time channels ZKl - ZK3 are coated with 32 tone signal words or code words Tnl - Tn32 a 12 bits in a 7-kHz tone signal. When covering these three time channels ZKl - ZK3 with the coded tone signal, the 3 times 4 bit locations released in the time channels ZK16 (originally the selection characteristics Zl - Z3 a 4 bits in the pulse frames RI, R2 and R3) can be used for the clock adjustment. The first 8 bits in the pulse frames RI and R2 are alternated per upper frame incurred with a positive and a negative clock adaptation signal ++ resp. —. The rate adaptation signal ++ resp. — is thus distributed over the pulse frames Ri and R2. The remaining 4 bit positions in the pulse frame R3 require a signal word S such as can be used for the transmission of alarms or for signaling purposes.

The synchronous state "no timing adjustment required" need not be transmitted separately, as the regular sequence ++—++— etc. of the alternating positive and negative timing signals does not trigger any reaction in the receiver.

In fig. 3 shows the case of a positive clock deviation, i.e. the clock frequency of the tone signal is lower than the proportional clock frequency in DSV2. If the time shift of the two clocks is 12 clock periods, according to the invention, positive clock matching signals are transmitted in two successive superframes, and in the second superframe there is a place agreed with the receiver, here the place of the first code word after the frame characteristic word RSy in the pulse frame R3, inserted an information-free blank word LW. In the pulse frame R3, only one code word of the 7-kHz tone signal is thereby transmitted, and within this overframe the number of transmitted codewords is reduced by one to 31. The synchronism is thereby restored, and in the subsequent overframes positive and negative timing signals with the negative timing signal first..

By that in fig. 4 shows negative clock deviations, the clock frequency of the tone signal is greater than the proportional clock frequency in DSV2. In order to signal this state, negative clock matching signals are transmitted in two consecutive superframes. Since a code word sorted out from the tone signal must also be transmitted within an overframe, it appears that the overframe (here third overframe) that follows the overframe with the second negative beat matching signal contains 33 codewords. The bit locations within the seventeenth time channel ZK16 already used during a tone transmission for the clock matching signal and signal word S are covered in the pulse frame Ri with the first four bits of the code word Tn5, in the pulse frame R2 with the fifth to eighth bits of the code word Tn7 and in the pulse frame R3 with the last four bits of the code word Tn9. A thereby conditioned displacement of the code word boundaries for the code words Tn5 - Tn9 within the pulse frames R2 and R3 is canceled in the fifth pulse frame not shown, and a tenth code word begins at the first bit position of the time channel ZKl in this pulse frame.

Here too, positive and negative (with a positive start) beat adjustment signals are transmitted alternately after the beat adjustment in the subsequent overframes.

There is also the possibility of adding 3 times 4 blank bits between the code words within an upper frame (e.g. in three successive pulse frames of 4 blank bits) with the positive clock adaptation in order to thereby achieve a less stepwise clock adaptation than with the addition of a 12-bit blank words.

In the same way, it is possible with the negative clock adjustment to transmit a code word resolved into 3 times 4 bits and sorted out on the transmitter side within the time channel ZK16 and only after the complete transmission to insert this code word in its entirety on the receiver side at the agreed place in the tone signal. In this connection, no shifting of the code word boundaries within the pulse frame takes place.

Assuming that there is a relative rate deviation of e.g. 5 ' 10 ^, then in the worst case a clock adaptation must take place only every 0.316 s (a distance of 157 overframes), and in the case of negative clock adaptation the transmission of the signal word S must be interrupted only after the passage of 157 overframes.

In the event of a further possibility for the negative beat adjustment, the transmission of the code word sorted out on the sending side by the digital tone signal is waived,

and the timing of the clock adjustment communicated to the receiver only by means of the two consecutive negative clock adjustment signals. The recipient replaces the sorted out code word with an estimated value which is either the same as the preceding code word, an extrapolation of two or more preceding code words or the mean value (geometric, arithmetic, etc.) of the code word that borders the sorted out code word. In this case, there is no interruption in the transmission of the signal word S.

When transmitting a 15-kHz tone signal, e.g. in the six time channels ZK1, ZK2, ZK3, ZK17, ZK18 and ZK19, a further 12 bits are free at the locations of the selection characteristics Z17 - Z19. In addition to the signal word S, these can take over other tasks, such as e.g. transmission of remote control signals.

A negative clock adaptation of the 15-kHz tone signal according to the example in FIG. 4 also causes a shift of the code word boundary within time channels ZK17 - ZK19. Because in the pulse frame Ri, 4 bits of the code word that should normally be transmitted in the time channel ZK17 and in the first half of the time channel ZK18 are already transmitted in the time channel ZK16. This code word offset increases in the pulse frame R2 after the time channel ZK16 by 8 bits and in the pulse frame R3 after the time channel ZK16 increases by 12 bits, so that the beginning of the code word in the time channels ZK17 and ZK18 corresponds to the beginning of the time channel ZK17.

The rate adaptation signals ++ or — are in the embodiment 8-position codewords which are complementary to each other. Their signal distance is thus 8. In this connection, up to three errors contained in these code words can be corrected.

In connection technology, the method according to the invention can be realized with known variable bearings (Ålinger) for sorting out and inserting bit groups. The recovery of the original beat for the coded tone signal is possible in the control circuit by means of a known PLL switching circuit with a low time constant.

The procedure can be used for beat adjustment for

all coded analog signals, but especially for image and sound signals, regardless of the frame structure of the transmission system, only the frame exhibits sufficient information capacity for signaling the clock matching signals.

Claims (4)

1. Procedure for timing adjustment of a tone signal that exists in digital form, to a data stream organized in pulse frames and superframes, for digital voice transmission of telephony quality and with characteristic bits for characteristic transmission, whose rate is plesiochronous to the digital tone signal's rate, characterized by that the rate adjustment per upper frame takes place word by word, i.e. with an entire code word of the digital tone signal, and that there in the places (Zl, Z2 and Z3 in Fig. 1) of the characteristic bits in (ZK16) which have become vacant during the transmission of the digital tone signal over several telephone channels (Kl, K2 and K3 in fig. 1) and is assigned to these, per upper frame is transferred a positive resp. a negative clock matching signal (++ or — in Fig. 2 - 4 in RI and R2), - that the beat matching signal when the beat of the data stream and the digital tone signal (fig. 2) is (++ or —) alternately positive and negative from overframe to overframe, that the clock matching signals (++) for two follow each other ging overframes in the case of a positive beat adjustment (Fig. 3) both are positive, and - that the clock adjustment signals (—) for two successive overframes in the case of a negative clock adjustment (Fig. 4) are both negative. 2. Method as stated in claim 1, characterized in that the one with a positive beat adjustment (fig. 3) within the second upper frame - which contains the second of the directly following positive beat adjustment signals - is added at the agreed place a blank word (LW) between two code words for the digital tone signal (Tn6 and Tn7). 3. Method as stated in claim 1, characterized in that the freed places of the characteristic bits in the upper frame that follows the second negative beat adjustment signal, in the case of negative beat adjustment (fig. 4), are coated with a code word sorted out from the coded tone signal, resp. . with sorted parts of code words (Tn5, Tn7 and Tn6 in fig. 4). 4. Method as set forth in claim 1, characterized in that a code word at the agreed location during a negative beat adjustment is sorted out from the digital tone signal on the transmitter side and replaced on the receiver side with an estimate value formation, e.g. mean value, based on the adjacent code words.