Biblio5

432 TELEVISION TRANSMISSION

mation, the unique identification of each packet or packet type, and appropriate methods of multiplexing video bit stream packets, audio bit stream packets, and ancillary data bit stream packets into a single data stream. A prime consideration in the system transport design was interoperability among the digital media, such as terrestrial broadcasting, cable distribution, satellite distribution, recording media, and computer interfaces. MPEG-2 transport stream syntax was developed for applications where channel bandwidth is limited and the requirement for efficient channel transport was overriding. Another aspect of the design was interoperability with ATM transport systems (see Chapter 18).

RF/ transmission deals with channel coding and modulation. The input to the channel coder is the multiplexed data stream from the service multiplex unit. The coder adds overhead to be used by the far-end receiver to reconstruct the data from the received signal. At the receiver we can expect that this signal has been corrupted by channel impairments. The resulting bit stream out of the coder modulates the transmitted signal. One of two modes can be used by the modulator: 8-VSB for the terrestrial broadcast mode and 16-VSB for the high data rate mode.

14.9.4.1 Video Compression. The ATSC standard is based on a specific subset of MPEG-2 algorithmic elements and supports its Main Profile. The Main Profile includes three types of frames for prediction (I-frames, P-frames, and B-frames), and an organization of luminance and chrominance samples (designated 4 : 2 : 0) within the frame. The Main Profile is limited to compressed data of no more than 80 Mbps. Figure 14.8 is a simplified block diagram of signal flow for the ATSC system.

Video preprocessing converts the analog input signals to digital samples in such a form needed for later compression. The analog input signals are red (R), green (G), and blue (B).

Figure 14.8 Video coding in relation to the ATSC system. (From Ref. 18, Figure 5.1. Reprinted with permission.)

Source: Ref. 18, Table 5.1. (Reprinted with permission.)

Table 14.4 lists the compression formats covered by the ATSC standard. The following explains some of the items in the table. Vertical lines refers to the number of active lines in the picture. Pixels are the number of pixels during the active line. Apect ratio, of course, refers to the picture aspect ratio. Picture rate gives the number of frames or fields per second. Regarding picture rate values, P refers to progressive scanning and I refers to interlaced scanning. It should be noted that both the 60.00-Hz and 59.95-Hz (i.e., 60 × 1000/ 1001) picture rates are allowed. Dual rates are permitted at 30 Hz and 24 Hz.

Sampling rates. Three active line formats are considered: 1080, 720, and 483. Table 14.5 summarizes the sampling rates.

For the 480-line format, there may be 704 or 640 pixels in an active line. If the input is based on ITU-R Rec. BT.601-4 (Ref. 19), it will have 483 active lines with 720 pixels in each active line. Only 480 of the 483 active lines are used for encoding. Only 704 of the 720 pixels are used for encoding: the first eight and the last eight are dropped. The 480-line, 640-pixel format corresponds only to the IBM VGA graphics format and may be used with ITU-R Rec. BT601-4 (Ref. 19) sources by employing appropriate resampling techniques.

Sampling precision is based on the 8-bit sample.

Colorimetry means the combination of color primaries, transfer characteristics, and matrix coefficients. The standard accepts colorimetry that conforms to SMPTE.9 Video inputs corresponding to ITU-R Rec. BT.601-4 may have SMPTE 274M or 170M colorimetry.

The input video consists of the RGB components that are matrixed into luminance

(Y) and chrominance (Cb and Cr) components using a linear transformation by means of a 3 × 3 matrix. Of course, the luminance carries picture intensity information (black- and-white) and the chrominance components contain the color. There is a high degree of correlation of the original RGB components, whereas the resulting Y, Cb, and Cr have less correlation and can be coded efficiently.

In the coding process, advantage is taken of the differences in the ways humans perceive luminance and chrominance. The human visual system is less sensitive to the high frequencies in the chrominance components than to the high frequencies in the

Table 14.5 Sampling Rate Summary

a fps c frames per second.

9SMPTE stands for Society of Motion Picture and Television Engineers.

434 TELEVISION TRANSMISSION

luminance component. To exploit these characteristics the chrominance components are low-pass filtered and subsampled by a factor of 2 along with the horizontal and vertical dimensions, thus producing chrominance components that are one-fourth the spatial resolution of the luminance components.

14.10 CONFERENCE TELEVISION

14.10.1 Introduction

Video conferencing (conference television) systems have seen phenomenal growth since 1990. Many of the world’s corporations have branches and subsidiaries that are widely dispersed. Rather than pay travel expenses to send executives to periodic meetings at one central location, video conferencing is used, saving money on the travel budget.

The video and telecommunications technology has matured in the intervening period to make video conferencing cost effective. Among these developments we include:

•Video compression techniques;

•Eroding cost of digital processing; and

•Arrival of the all-digital network.

Proprietary video conferencing systems normally use lower line rates than conventional broadcast TV. Whereas conventional broadcast TV systems have line rates at 525/ 480 lines (NTSC countries) or 625/ 580 for PAL/ SECAM countries, proprietary video conferencing systems use either 256/ 240 or 352/ 288 lines. For the common applications of conference television (e.g., meetings and demonstrations), the reduced resolution is basically unnoticeable.

One of the compression schemes widely used for video conference systems is based on ITU-T Rec. H.261 (Ref. 20), titled “Video Codec for Audiovisual Services at pX64 kbps,” which is described in the following.

14.10.2 pX64 kbps Codec

The pX64 codec has been designed for use with some of the common ISDN data rates, specifically the B-channel (64 kbps), H0 channel (384 kbps), and the H11/ H12 channels (1.536/ 1.920 Mbps) for the equivalent DS1/ E1 data rates. A functional block diagram of the codec (coder/ decoder) is shown in Figure 14.9. However, the pX64 system uses standard line rate (i.e., 525/ 625 lines) rather than the reduced line rate structure mentioned earlier. The reduced line rate techniques are employed in proprietary systems produced by such firms as Compression Labs Inc. and Picture Tel.

One of the most popular data rates for conference television is 384 kbps, which is six DS0 or E0 channels. However, it is not unusual to find numerous systems operating at 64/ 56 kbps.

14.10.2.1 pX64 Compression Overview

Sampling frequency. Pictures are sampled at an integer multiple of the video line rate. A sampling clock and network clock are asynchronous.

Source coding algorithm. Compression is based on interpicture prediction to utilize temporal redundancy, and transform coding of the remaining signal to reduce spatial redundancy. The decoder has motion compensation capability, allowing optional incor-

436 TELEVISION TRANSMISSION

Figure 14.10 Positioning of luminance and chrominance samples. (From Figure 2/ H.261, ITU-T Rec. H.261, [Ref. 20].)

line sources at 6.75 MHz or 3.375 MHz, respectively. These frequencies have a simple relationship with those in ITU-R Rec. BT.601 (Ref. 19).

The second format, called quarter-CIF or QCIF, has half the number of pels and half the number of lines stated of the CIF format. All codecs must be able to operate using QCIF.

A means is provided to restrict the maximum picture rate of encoders by having at least 0, 1, 2, or 3 nontransmitted pictures between transmitted pictures. Both CIF/ QCIF and this minimum number of nontransmitted frames shall be selectable externally.

Video source coding algorithm. A block diagram of the coder is illustrated in Figure 14.11. The principal functions are prediction, block transformation, and quantization. The picture error (INTER mode) or the input picture (INTRA mode) is subdivided into 8-pel-by-8-pel line blocks, which are segmented as transmitted or nontransmitted. Furthermore, four luminance blocks and two spatially corresponding color-difference blocks are combined to form a macroblock. Transmitted blocks are transformed and the resulting coefficients are quantized and then variable length coded.

Motion compensation is optional. The decoder will accept one vector per macroblock. The components, both horizontal and vertical, of these motion vectors have integer values not exceeding ±15. The vector is used for all four luminance blocks in the macroblock. The motion vector for both color-difference blocks is derived by halving the component values of the macroblock vector and truncating the magnitude parts toward zero to yield integer components.

A positive value of the horizontal or vertical component of the motion vector signifies that the prediction is formed from pels in the previous picture, which are spatially to

Figure 14.11 Functional block diagram of the source coder. (From Figure 3/ H.261, ITU-T Rec. H.261 [Ref. 20].)

the right or below the pels being predicted. Motion vectors are restricted such that all pels referenced by them are within the coded picture area.

Loop filter. A two-dimensional spatial filter may be used in the prediction process. The filter operates on pels within a predicted 8-by-8 block. It is separable into onedimensional horizontal and vertical functions. Both are nonrecursive carrying coefficients of 14 , 12 , 14 except at block edges, where one of the taps would fall outside the block. In this case the one-dimensional filter is changed to have coefficients of 0, 1, 0. There is rounding to 8-bit integer values at the two-dimensional filter output and full arithmetic precision is retained. Rounding upward is used where values whose fractional part is one-half. The filter is switched on/ off for all six blocks in a macroblock according to the macroblock type. There are ten types of macroblocks such as INTRA, INTER, INTER + MC (motion compensation), and INTER + MC + FIL (filter).

Discrete cosine transform. The transmitted blocks are first processed by a separable two-dimensional discrete cosine transform, which is 8 by 8 in size. There is an output range of the inverse transform from −256 to +255 after clipping to be represented by 9 bits.

Quantization. There are 31 quantizers for all other coefficients except the INTRA dc coefficient, which has just 1. The decision levels are not defined in CCIT-Rec. H.261. The INTRA dc coefficient is nominally the transform value linearly quantized with a step size of 8 and no dead-zone. The other 31 quantizers are nominally linear but with a central deadzone around zero and with a step size of an even value in the range of 2–62.

438 TELEVISION TRANSMISSION

Clipping and reconstructed picture. Clipping functions are inserted to prevent quantization distortion of transform coefficient amplitudes causing arithmetic overflow in the encoder and decoder loops. The clipping function is applied to the reconstructed picture. This picture is formed by summing the prediction and the prediction error as modified by the coding process. When resulting pel values are less than 0 or greater than 255, the clipper changes them to 0 and 255, respectively.

Coding control. To control the rate of generation of coded video data, several parameters may be varied. These parameters include processing prior to source coder, the quantizer, block significance criterion, and temporal subsampling. When invoked, temporal subsampling is performed by discarding complete pictures.

Forced updating. Forced updating is achieved by forcing the use of the INTRA mode of the coding algorithm. Recommendation H.261 does not define the update pattern. For the control of accumulation of inverse transform mismatch error, a macroblock should be updated forcibly at least once per every 132 times it is transmitted.

14.10.2.3 Video Multiplex Coder. The video multiplex is arranged in a hierarchical structure with four layers. From top to bottom these layers are:

1. Picture;

2. Group of blocks (GOB);

3. Macroblock; and

4. Block.

For further description of these layers, consult CCITT Rec. H.261 (Ref. 20).

14.11 BRIEF OVERVIEW OF FRAME TRANSPORT FOR VIDEO

CONFERENCING

This section briefly reviews one method of transporting on the PSTN digital network the video conferencing signals developed in Section 14.10. It is based on ITU-T Rec. H.221 (Ref. 22).

14.11.1 Basic Principle

An overall transmission channel of 64 kbps–1920 kbps is dynamically subdivided into lower rates suitable for transport of audio, video, and data for telematic purposes. The transmission channel is derived by synchronizing and ordering transmission over from one to six B-channels, from one to five H0 channels or an H11 or H12 channel.11 The initial connection is the first connection established and it carries the initial channel in each direction. The additional connections carry the necessary additional channels. The total rate of transmitted information is called the transfer rate. The transfer rate can be less than the capacity of the overall transmission channel.

A single 64-kbps channel is structured into octets transmitted at an 8-kHz rate.12 Each bit position of the octets may be regarded as a subchannel of 8 kbps. A frame

11Note the use of ISDN terminology for channel types (e.g., B-channels). Turn back to Section 12.4.2.1 for a description of ISDN user channels.

12This is the standard Nyquist sampling rate discussed in Chapter 6.

Figure 14.12 Frame structure for a 64-kbps channel (i.e., a B-channel). (From Figure 1/ H.221, ITU-T Rec. H.221 [Ref. 22].)

reflecting this concept is illustrated in Figure 14.12. The service channel (SC) resides in the eighth subchannel. It carries the frame alignment signal (FAS), a bit-rate allocation signal (BAS), and an encryption control signal (ECS).

We can regard an H0, H11, H12 channel as consisting of a number of 64 kbps time slots (TS). The lowest numbered time slot is structured exactly as just described for a 64-kbps channel, whereas the other time slots have no such structure. All channels have a frame structure in the case of multiple B and H0 channels; the initial channel controls most functions across the overall transmission, while the frame structure in the additional channels is used for synchronization, channel numbering, and related controls. The term I-channel is applied to the initial or only B-channel, to time slot one of the initial or only H0 channels, and to time slot one on H11 and H12 channels.

REVIEW EXERCISES

1. What four factors must be dealt with by a color video transmission system transmitting images of moving objects?

2. Describe scanning, horizontally and vertically.

3. Define a pel or pixel (besides the translation of the acronym).

4. If the aspect ratio of a television system is 4 : 3 and the width of a television screen is 12 inches, what is its height?

REFERENCES

1. IEEE Standard Dictionary of Electrical and Electronics Terms, 6th ed., IEEE Std. 100-1996, IEEE, New York, 1996.

2. Fundamentals of Television Transmission, Bell System Practices, Section AB 96.100, American Telephone & Telegraph Co., New York, 1954.

3. Television Systems Descriptive Information—General Television Signal Analysis, Bell System Practices, Section 318-015-100, No. 3, American Telephone & Telegraph Co., New York, 1963.

REFERENCES 441

4. A. F. Inglis and Arch C. Luther, Video Engineering, 2nd ed., McGraw-Hill, New York, 1996. 5. K. Simons, Technical Handbook for CATV Systems, 3rd ed., General Instrument–Jerrold

Electronics Corp., Hatboro, PA, 1980.

6. Electrical Performance for Television Transmission Systems, ANSI/ EIA/ TIA-250C, EIA/ TIA Washington, DC, 1990.

7. Television Systems, ITU-R Rec. BT.470-3, 1994 BT Series, ITU, Geneva, 1994.

8. Reference Data for Engineers: Radio, Electronics, Computer & Communications, 8th ed, Sams–Prentice-Hall, Carmel, IN, 1993.

9. The Preferred Characteristics of a Single Sound Channel Transmitted with a Television Signal on an Analogue Radio-Relay System, CCIR Rec. 402-2, Part 1, Vol. IX, ITU, Geneva, 1990.

10. The Preferred Characteristics of Simultaneous Transmission of Television and a Maximum of Four Sound Channels on Analogue Radio-Relay Systems, CCIR Rep. 289-4, Part 1, Vol. IX, CCIR, Dubrovnik, 1986.

11. Pre-emphasis Characteristics for Frequency Modulation Radio-Relay Systems for Television, CCIR Rec. 405-1, Vol. IX, Part 1, ITU, Geneva, 1990.

12. Transmission Performance of Television Circuits Designed for Use in International Connections, CCIR Rec. 567-3, Vol. XII, ITU, Geneva, 1990.

13. Transmission Performance of Television Circuits over Systems in the Fixed Satellite Service, CCIR Rep. 965-1, Annex to Vol. XII, ITU, Geneva, 1990.

14. Digital or Mixed Analogue-and-Digital Transmission of Television Signals, CCIR Rep. 646-4, Annex to Vol. XII, ITU, Geneva, 1990.

15. Digital Transmission of Component-Coded Television Signals at 30-34 Mbps and 45 Mbps, CCIR Rep. 1235, Annex to Vol. XII, ITU, Geneva, 1990.

16. A Compilation of Advanced Television Systems Committee Standards, Advanced Television Systems Committee (ATSC), Washington, DC, 1996.

17. ATSC Digital Television Standard, Doc. A/ 53, ATSC, Washington, DC, 1995.

18. Guide to the Use of the ATSC Digital Transmission Standard, Doc. A/ 54, ATSC, Washington, DC, 1995.

19. Encoding Parameters of Digital Television for Studios, ITU-R Rec. BT.601-4, 1994 BT Series Volume, ITU, Geneva, 1994.

20. Video Codec for Audiovisual Services at pX64 kbps, CCITT Rec. H.261, ITU, Geneva, 1990. 21. R. L. Freeman, Radio System Design for Telecommunications, 2nd ed., Wiley, New York,

1997.

22. Frame Structure of a 64 to 1920 kbps Channel in Audiovisual Teleservices, ITU-T Rec. H.221, ITU, Geneva, 1993.