ITU-T J 41-1988 CHARACTERISTICS OF EQUIPMENT FOR THE CODING OF ANALOGUE HIGH QUALITY SOUND PROGRAMME SIGNALS FOR TRANSMISSION ON 384 kbit s CHANNELS《用于编码在384kbit s通道中传输的高质量模拟声音节目信号.pdf

1、INTERNATIONAL TELECOMMUNICATION UNION)45G134 * TELECOMMUNICATIONSTANDARDIZATION SECTOROF ITU4%,%6)3)/.G0G0!.$G0G03/5.$G0G042!.3-)33)/.#(!2!#4%2)34)#3G0G0/ amended at Melbourne, 1988)1 General1.1 This Recommendation gives the characteristics of equipment for the coding of 15 kHz monophonic analogueso

2、und-programme signals into a digital signal of 384 kbit/s. For stereophonic operation, two monophonic digital codecscan be utilized. Two monophonic digital signals that form a stereophonic signal should be routed together over thesame transmission systems (path) to avoid any difference in transmissi

3、on delay.1.2 Equipment for coding of analogue sound-programme signals, as specified in this Recommendation, can be:a) A stand-alone encoder/decoder with a digital interface at 384 kbit/s. The encoder operation and the decoderoperation may be performed in two separate equipments or in the same equipm

4、ent.b) A combined encoder-multiplex/decoder-demultiplex with a digital interface at 1544 or 2048 kbit/s. Theencoder-multiplex operation and the decoder-demultiplex operation may be performed in two separateequipments or in the same equipment.In case b), it is not mandatory to provide an external dig

5、ital sound programme access port at 384 kbit/s.1.3 Two methods of encoding have been specified by the CMTT 1 and these form the basis for thisRecommendation.2 Transmission performanceThe transmission performance per encoder/decoder pair shall be such that the limits specified inRecommendation J.21 (

6、CCIR Recommendation 505) are not exceeded by three encoder/decoder pairs connected intandem at audio frequencies.Note When transmitting stereophonic sound programme signals, it is necessary that the encoder and decoderare designed such that they will meet the specified requirements for phase differe

7、nce.In order to avoid any unnecessary complexity, the sampling of channels A and B should be performedsimultaneously.3 Method of encoding3.1 The recommended encoding laws are as specified in 1.3.2 These encoding laws are based on a uniformly quantized 14-bit per sample PCM technique with compandinga

8、nd employ either:a) eleven-segment 14- to 11-bit instantaneous A-law companding, orb) five-range 14- to 10-bit near instantaneous companding.For provisional rules for through connection between the two companding methods, see Note 4 in 1.3.3 Other coding techniques which may be used by bilateral agr

9、eement of the Administrations concerned are alsolisted in Annex A. However, these techniques do not form part of this Recommendation.2 Fascicle III.6 - Rec. J.413.4 Equipment characteristics common to both methods of encoding are:Nominal audio bandwidth: 0.04 to 15 kHz.Audio interface: see Recommend

10、ation J.21, 2.Sampling frequency (CCIR Recommendation 606): 32 (1 5 105) kHz.Pre/de-emphasis: Recommendation J.17 with 6.5 dBattenuation at 800 Hz.Note Pre-emphasis and de-emphasis are not used by the Administrations of Canada, Japan and the UnitedStates on their national circuits and on internation

11、al circuits between each other, but are used on international circuits toother countries.4 Equipment using instantaneous companding4.1 Coding table4.1.1 The coding law is specified in Table 1/J.41.4.1.2 The allocation of character signals (PCM code words) is also given in Table 1/J.41. Two variants

12、(A and B) ofcharacter signals are allowed.Note In the case of digital interconnection between variants A and B, the conversion from one set of charactersignals to the other in Table 1/J.41 can be implemented without any performance degradation. In the case of analogueinterconnection, a small reducti

13、on in the S/N ratio, in the order of 3 dB, is expected.4.2 Bit ratesNominal source coding bit rate (32 kHz 11 bits/sample) 352 kbit/sError protection 32 kbit/sTransmission bit rate 384 kbit/s4.3 Overload levelThe overload level for a sine-wave signal at zero dB insertion loss frequency (2.1 kHz) of

14、the pre-emphasis is+ 15 dBm0s.4.4 Digital signal formatThe character signal bit sequences for variants A and B are shown in Figure 1/J.41.4.5 Bit error protectionOne parity bit is added to each 11-bit character signal.Fascicle III.6 - Rec. J.41 3TABLE 1/J.4111 segment, 14 to 11 bit instantaneous com

15、panding A-law PCM for sound-programme signals (positive half only)a)11 bit codingAllocation of character signalsNormalizedanalogueinputNormalizedanalogueoutputCompresseddigitalcodeSegmentNo.Effectiveresolution (bits)1 2 3 4Variant Ab)5 6 7 8 9 1011 S X Y ZVariant B c)A B C D E F G8160 to 81924096 to

16、 4128817641128957681 9 0 1 1 11 1 1 1 1 10 0 0 0 0 0100 1 1 01 1 1 1 1 1 10 0 0 0 0 0 04080 to 40962048 to 2064408820567676402 10 0 1 1 01 1 1 1 1 10 0 0 0 0 0100 1 0 11 1 1 1 1 1 10 0 0 0 0 0 02040 to 20481024 to 1032204410286395123 11 0 1 0 11 1 1 1 1 10 0 0 0 0 0100 1 0 01 1 1 1 1 1 10 0 0 0 0 0

17、01020 to 1024512 to 51610225145113844 12 0 1 0 01 1 1 1 1 10 0 0 0 0 0100 0 1 11 1 1 1 1 1 10 0 0 0 0 0 0510 to 512256 to 2585112573832565 13 0 0 1 11 1 1 1 1 10 0 0 0 0 0100 0 1 01 1 1 1 1 1 10 0 0 0 0 0 0255 to 256128 to 129255.5128.52551280 0 1 01 1 1 1 1 10 0 0 0 0 0100 0 0 11 1 1 1 1 1 10 0 0 0

18、 0 0 0127 to 1280 to 1127.50.51270614010001 1 1 1 1 10 0 0 0 0 0X00 0 0 01 1 1 1 1 1 10 0 0 0 0 0 0X 11th bit freely available in variant A.a)Character signals for the negative half are the same as those for the positive half except that the sign bits (bit 1 and S for variants A and B respe ctively)

19、 are inverted.b)Variant A is presently used with digital equipment based on a 2048 kbit/s hierarchy. After coding and before the parity bit is inserted, bits 1 to 5 are inverted.c)Variant B is presently used with digital equipment based on a 1544 kbit/s digital hierarchy. All bits, including the par

20、ity bit, are inverted and reformatted before transmission (see Figure 1/J.41).4 Fascicle III.6 - Rec. J.414.5.1 Variant AThe five most significant bits of each sample are protected against errors by means of a parity bit. In the converter ofthe transmitting part, the parity bit is added as the 12th

21、bit to each code word. Its value is fixed so that the 6 bit parity blockalways contains only an odd number of “one” values. In order that even bit error structures can also result in parityviolations, the protected and unprotected bits of each code word are interleaved in ascending and descending se

22、quence, asshown in Figure 1/J.41.4.5.2 Variant BThe added parity bit shall be based on the 7 most significant bits of the 11-bit PCM word. These are bits S, X, Y, Z,A, B, C. The parity of “ones” bit shall be even. Since the chord bits (X, Y, Z) always contain a one, the minimum number ofones per sam

23、ple is 2, resulting in a minimum ones density of 1/6.4.5.3 Error concealmentIf a parity violation is detected, an error concealment technique should be applied (for instance, replacement byinterpolation, extrapolation or repetition. For multiple parity violation (error bursts), a muting technique sh

24、ould be applied.4.6 Digital interface at 384 kbit/sUnder study (see Recommendations G.735 and G.737).Fascicle III.6 - Rec. J.41 54.7 SynchronizationThe coding equipment operates in synchronism with the clock of subsequent multiplex equipment or the networkclock. In cases where the digital interface

25、is provided, bit and byte (24 bit, as shown in Figure 1/J.41) timing information isrequired.Variant A: A solution for synchronous access is given in Recommendations G.735 and G.737.Variant B: The solution for synchronous access is under study.4.8 Fault condition and consequent actions4.8.1 Variant A

26、Where a 384 kbit/s digital interface is provided, the same principles for fault conditions and subsequent actions asthose outlined in Recommendation G.732 should be followed.4.8.2 Variant BUnder study.5 Equipment using near-instantaneous companding5.1 IntroductionThe equipment described in this sect

27、ion uses the near-instantaneous method of companding in the coding of highquality sound-programme signals into digital form.A two-stage process is used in the encoding equipment:a) Conversion of a 15 kHz channel into a 338 kbit/s stream.Note The value of 338 kbit/s has been chosen to allow for the p

28、ossible multiplexing of 6 channels into a2048 kbit/s dedicated frame format.b) Asynchronous insertion of the 338 kbit/s stream into a 384 kbit/s stream.Note The asynchronous insertion of the 338 kbit/s stream into a 384 kbit/s stream allows the use, at theencoder location, of a clock not necessarily

29、 synchronous to the network clock. It can be advantageous when theencoder equipment and the insertion equipment (see Recommendations G.735 and G.737) are located indifferent places, and when the transmission link between them is unidirectional,and the reverse processes in the decoding equipment.5.2

30、Conversion from 15 kHz to 338 kbit/s5.2.1 Overload levelThe overload level for a sine-wave signal at the zero dB insertion loss frequency (2.1 kHz) of the pre-emphasiscircuit, is + 12 dBm0s.5.2.2 CompandingNear-instantaneous companding is used to achieve a data rate reduction from 14 bits/sample to

31、10 bits/sample. Thesystem codes a block of 32 samples into one of 5 gain ranges, according to the highest value sample in the block. Thecompanding characteristic is shown diagramatically in Figure 2/J.41 and the parameters are specified in Table 2/J.41.6 Fascicle III.6 - Rec. J.415.2.3 Range coding

32、and protectionInformation defining the range used is transmitted over 3 successive blocks as a 7-bit word, increasing to 11 bits in aHamming 7, 11 single error correcting code and distributed throughout the 3 blocks as follows:The five possible values for each of the 3 range codes (one range code fo

33、r each block in the 3 ms frame; seeFigure 3/J.41), are:Range 4 highest signal levelRange 3Range 2Range 1Range 0 lowest signal levelRange codes generated in this way from three successive blocks are designated Ra, Rb and Rc. They are then used tocompute a single 7-bit range code, R, as follows:R = 25

34、Ra + 5Rb + Rc + 1R1 to R7 form the unsigned binary representation of this code which is transmitted LSB first (R1 to R7), followedby 4 protection bits R8 to R11 made up as follows:R8 = ( R3 + R2 + R1) MOD 2R9 = ( R6 + R5 + R4 ) MOD 2R10 = (R7 + R5 + R4 + R2 + R1) MOD 2R11 = (R7 + R6 + R4 + R3 + R1)

35、MOD 2Fascicle III.6 - Rec. J.41 78Fascicle III.6 - Rec. J.41Fascicle III.6 - Rec. J.41 95.2.4 Sample error protection32 bits per frame are used for sample error detection on the basis of 1 parity bit per 3 samples. Odd parity isemployed, i.e., the total number of data bits set to state 1, in the pro

36、tected samples, plus the parity bit is always an oddnumber. The distribution of the parity bits within the frame and the allocation of the parity bits to the samples is shown inFigure 3/J.41 and Table 3/J.41, respectively. Only the 5 most significant bits of the samples are protected. In order to en

37、surethat, if two sequential bits are corrupted, the error can still be detected by the parity checking process, the protected andunprotected bits of each sample are interleaved in descending and ascending order, respectively: 1, 10, 2, 9, 3, 8, 4, 7, 5, 6.LSB is transmitted first and the bits underl

38、ined are those protected by the parity check. Error concealment should be usedand can be achieved, for example, by replacing an erroneous sample value by a sample value calculated by linearinterpolation between adjacent correct samples, or by extrapolation of the previous sample if the following sam

39、ple is itself inerror.TABLE 3/J.41Allocation of parity bits to the samplesParity bit Protects samples Parity bit Protects samples123456789101112131415163, 35, 668, 39, 7112, 44, 7517, 48, 7921, 53, 8426, 57, 8831, 62, 9219, 51, 8224, 55, 8628, 60, 9032, 64, 942, 37, 696, 42, 7311, 46, 774, 36, 679,

40、41, 721718192021222324252627282930313214, 47, 7818, 52, 8323, 58, 8927, 63, 9515, 50, 8022, 56, 8529, 61, 910, 34, 655, 40, 7010, 45, 747, 33, 6813, 38, 7616, 43, 8120, 49, 8725, 54, 931, 30, 59This order has been chosen:a) to spread each group of 3 protected samples as widely as possible;b) to spre

41、ad the 18 or 21 samples protected by each housekeeping word, with the maximum number of other samples between them.5.2.5 Single channel frame formatThree 32 sample blocks, together with various housekeeping bits, form a single channel frame having a bit rate of338 kbit/s and a duration of 3 ms. The

42、number of bits per frame is therefore 3338 = 1014 bits, and these have been allocatedas shown in Table 4/J.41. Figure 3/J.41 illustrates the frame arrangement for a single channel. Two frames are shown inFigure 3/J.41 and this format is referred to as a multiframe. Framing information is reversed, i

43、.e. alternate bits in each frameof the multiframe.10 Fascicle III.6 - Rec. J.415.2.6 Two channels (stereo-pair) formatTwo separate 338 kbit/s streams are used to form a stereo-pair. Each of these bit streams is arranged as shown inFigure 3/J.41. The coders of the stereo-pair must be in synchronizati

44、on. Care must be taken at the receiving end tocompensate for any phase difference between the 2 channels.5.2.7 Synchronization of the 338-kbit/s streamThe 338 kbit/s stream is synchronized to the coder sampling frequency.TABLE 4/J.41Bit allocation in the frameFrame allocation(bits/frame)Bit rate per

45、 channel(kbit/s)Sample wordsRange coding(including error protection)Sample word error protectionSignallingFrame alignmentTotal9601132471014320.03.610.61.32.3338.05.2.8 Loss and recovery of frame alignmentOne of the following strategies is used:a) Loss of single channel frame alignment shall occur if

46、 two or more consecutive frame alignment words arereceived incorrectly (for this purpose, bits F1 to F7, Frame 0, and bits F8 to F14, Frame 1, are both consideredas frame alignment words: see Figure 3/J.41). An incorrect frame alignment signal is defined as one in whichtwo or more bits are in error.

47、 Realignment shall be achieved when a single frame alignment signal is receivedcorrectly. If this word is a spurious code, a second attempt at realignment shall be made.b) Only bits 1 to 10 of the 14 bit frame alignment word, derived from Frame 0 and Frame 1 (see Figure 3/J.41), aretaken into accoun

48、t at the receiving end. Loss of frame alignment is assumed to have occurred when threeconsecutive frame alignment signals are received incorrectly in their predicted position. When frame alignmentis assumed to have been lost, the automatic frame alignment recovery device will decide that alignment h

49、as beenrecovered when it registers two consecutive correct frame alignment signals.5.3 Conversion from 338 kbit/s to 384 kbit/s5.3.1 Frame structureThe frame structure (see Figure 4/J.41) with a nominal bit rate of 384 kbit/s and 613 bits in length is composed of: data input of 338 kbit/s; 63 redundancy bits for single error correction; bits for justification (J) and for identification of justification (IJ); the frame alignment (FA) signal,The frame is arranged in 4 sections.Fascicle III.