University of Canberra Advanced Communications Topics

University of Canberra Advanced Communications Topics

University of Canberra Advanced Communications Topics Television Broadcasting into the Digital Era Lecture 1 Television Fundamentals, Analog TV and Formats 1 by: Neil Pickford Overview of Topics 1 - Fundamentals of Television Systems, Digital Video Sampling & Standards (2hr) 2 - Digital Audio/Video Stream Compression (2hr) 3 - Digital Modulation Systems for DTV 4 - Transmission System Error Protection 5 - Digital System Parameters, Planning and SI 6 - DTV Hardware 8 Hours total Fri 08:30-10:30 & Thu 12:30-13:30 1/2 Hour Multipart Question on Examination 2 Digital Media

First media systems were Analog Most media are converting to digital 3 Computer storage Music (LP-CD) Telecommunications Multimedia Internet Networking (TCPIP) Radio (DAB) Television (DTTB) What is Television Images - Black and White Shades of Grey Colour - Hue & Saturation

Sound - Audio Information Data - Teletext & Other Data Synchronisation - Specifies the Timing Transport System - Gets the Above to your TV 4 History - Ferdinand Braun - CRT 5

1890 Ferdinand Braun developed the Cathode Ray Tube. 1897 developed the Cathode Ray Oscillograph, the precursor to the radar screen and the television tube 1907 First use of cathode ray tube to produce the rudiments of television images. He shared the Nobel Prize for physics in 1909 with Guglielmo Marconi for his contributions to the development of wireless telegraphy. John Logie Baird - Basic TV 6 Oct 1923 John Logie Baird was the first person anywhere in the world to demonstrate true television in the form of recognisable images, instantaneous movement and correct gradations in light and shade. Scanning was done mechanically with a Nipkow disc. The first 30 line picture transmitted was a Maltese cross.

1927 he also demonstrated video recording 1928 transatlantic television 1937 the broadcast of high definition colour pictures 1941 stereoscopic television in colour 1944 the multi-gun colour television tube, the forerunner of the type used in most homes today. Early Mechanical Approach to TV Mechanical Nipkow discs were used to scan the image and reconstitute the image at the receiver. PE cells were used to capture the image. The problem was synchronising the disks. 7 30 Line Mechanical TV 8 Electronic Television - Farnsworth 9 In 1922 at Age 14 Philo Farnsworth had the idea of how to make Electronic Television possible. Sept. 7, 1927, Farnsworth painted a square of glass black and scratched a straight line on the centre.

The slide was dropped between the Image Dissector (the camera tube that Farnsworth had invented earlier that year) and a hot, bright, carbon arc lamp. On the receiver they saw the straight-line image and then, as the slide was turned 90 degrees, they saw it move. This was the first all-electronic television picture ever transmitted. Vladimir Zworykin - Iconoscope 1 In 1923 Vladimir Zworykin of RCA made a patent application for a camera device, and by 1933 had developed a camera tube he called an Iconoscope. Although Zworykin submitted his patent application first after many years of legal battle Farnsworth was acknowledged as the inventor of electronic television. By the end of 1923 he had also produced a picture display tube,

the "Kinescope" Significant Television Inventions 1 These inventions were the underlying basis of the development of Television as we know it today Aspect ratio First TV displays were Round Rectangular Rasters easier to Generate Television Developed using a 4:3 Aspect Ratio Cinematic formats are much wider World now moving to 16:9 Aspect Ratio 4:3 (12:9) 1 16:9

Film 1 Has been the highest Resolution storage format. Various frame sizes used. 16mm, 35mm & 70mm Difficult to produce, store, handle and display. Easily degraded due to contamination and scratches. Generally recorded at 24 fps. Generally displayed at 72 fps (each frame 3x) to reduce flicker Use a device called a Telecine to convert to television formats The Video Signal First Television Pictures were Black & White Referred to as Luminance

Video refers to the linear base-band signal that contains the image information White 700 mV Stripe Front Back Porch Grey Background 0 mV -300 mV 1 Sync Pulse Black Stripe Sync Pulse Video Timing

SDTV 64 us for each line (15.625 kHz) 52 us Active Picture Area 12 us Blanking and Synchronisation Two level sync pulse 300 mV below blanking Active Picture 52 us Sync 4.7 us 1 Line = 64 us 1 Line Blanking 12 us Frame Rate

1 A Frame represents a complete TV picture Our analog TV Frame consists of 625 lines. A Frame is usually comprised of 2 Fields each containing 1/2 the picture information Our system has a Frame rate of 25 Hz The Field rate is 50 Hz Pictures displayed at 25 Hz exhibit obvious flicker Interleaving the Fields reduces flicker. Flicker and Judder Flicker and Judder are terms used to describe visual interruptions between successive fields of a displayed image. It affects both Film & TV. If the update rate is too low, persistence of vision is unable to give illusion of continuous motion.

Flicker is caused by: 1 Slow update of motion Information Refresh rate of the Display device Phosphor persistence Vs Motion Blur Judder usually results from Aliasing between Sampling rates, Display rates and Scene motion Interlace 1 To reduce the perceived screen flicker (25 Hz) on a television, a technique called 'interlacing' is employed.

Interlacing divides each video frame into two fields; the first field consists of the odd scan lines of the image, and the second field of each frame consists even scan lines. Interlace was also used to decrease the requirement for video bandwidth. It is a form of Compression Interlaced Vs Progressive Scan 1 Interlaced pictures. - 1/2 the lines presented each scan 1,3,5,7,9,11,13...............623,625 field 1 2,4,6,8,10,12,14.............622,624 field 2 Because the fields are recorded at separate times this leads to picture twitter & judder Progressive pictures - all the lines sent in the one scan. 1,2,3,4,5,6,7,8................623,624,625 picture No twitter or judder. But twice the information rate.

Progressive Scan 2 Simplifies the interpolation and filtering of images Allows MPEG-2 compression to work more efficiently by processing complete pictures Direct processing of progressively-scanned sources 24 frame/second progressive film mode can be provided. Assists video conversions with different: Progressive numbers of scan lines Doubles numbers of samples per line Raw Data Requirement temporal sampling (i.e., picture rate) Resolution

The number of picture elements resolved on the display Resolution in TV is limited by: Capture device Sampling Rate Transmission System / Bandwidth Display Device 2 Dot Pitch, Phosphor Focus & Convergence Viewing distance / Display size

Human Eye Typical SDTV systems attempt to transfer 720 pixels per line Colour Equations for PAL For B&W only had to transmit Luminance (Y) A Colour Image has Red, Green & Blue Components which need to be transmitted. We already have the Y signal. To remain compatible with Monochrome sets use Y, U & V to represent the Full Colour Picture Y = 0.299 R + 0.587 G + 0.114 B Colour U = 0.564 (B - Y) Difference Signals V = 0.713 (R - Y) 2 A Compatible Colour System Y V U 2

Y R G B Colour Sub Carrier 2 Colour Sub-Carrier is added at 4.43361875 MHz Frequency selected to interleave colour information spectra with Luma spectrum More efficient use of spectrum. Adding Colour to B&W Video First TV signals were only Luminance In 1975 we added PAL Colour System

A Colour Reference Burst on Back Porch And IQ modulated Colour Information 2 Television Modulation - AM 100% 0% 100% 2 TV uses Negative AM Modulation Amplitude Modulation 2 RF Carrier Wave Modulation Information Amplitude Modulation Amplitude Modulation (Min Carrier 20%) TV Modulation - AM Min 20%

100% 76% 20% 0% 20% 76% 100% 2 Peak White 20% Black 76% Syncs 100% TV Modulation - PAL AM 100% 76% 20% 0% 20% 76%

100% 2 Headroom prevents Colour Over/Under Modulating Frequency Modulation Modulation Information RF Carrier Wave Frequency Modulation 3 Intercarrier Sound A FM subcarrier is added to the AM picture to carry the Audio information FM Deviation 50 kHz used with 50 us Emphasis PAL-B uses 5.5 MHz Sound subcarrier (L+R)

2nd Sound subcarrier for Stereo (R) 3 -10 dB wrt Vision for mono single carrier mode -13 dB wrt Vision for Stereo & Dual mode 5.7421875 MHz (242.1875 kHz above main sound) -20 dB wrt Vision carrier 54.7 kHz Subcarrier Pilot tone added to indicate: Stereo (117.5 Hz) or Dual mode (274.1 Hz) FM Sound Emphasis dB 50 us Emphasis 30 25 20 Emphasis 15 10 5 0

10 3 100 1000 Frequency (Hz) 10000 100000 TV Modulation - Sound 100% 76% 20% 0% 20% 76% 100% 3 FM Sound Subcarriers Superimpose over the AM

Vestigial Side Band - VSB AM Modulation gives a Double Side Band signal To conserve spectrum Analog TV uses VSB 3 Each sideband contains identical information 5 MHz of information means required BW > 10 MHz Only one sideband is required for demodulation Only 1.25 MHz of the lower sideband is retained VSB truncates the high frequency part of the lower sideband. To implement Analog TV in 1950s with no lower sideband would have been very expensive because of the filtering required.

PAL-B Spectrum 0 dB -13 dB Sound -20 dB Vision Carrier Truncated Lower Sideband Chroma -1.25 -2 3 +5.75 -1 0 1 2 3 4 5 Relative Frequency (MHz) 4.433

6 Frequencies Used 3 Australia uses 7 MHz Channels VHF Band I Ch 0-2 45 - 70 MHz VHF Band III Ch 6-12 174 - 230 MHz UHF Band IV Ch 27-35 520 - 582 MHz UHF Band V Ch 36-69 582 - 820 MHz World TV Standards NTSC PAL SECAM PAL/SECAM Unknown 3 Australia is PAL

NTSC 3 National Television Systems Committee (NTSC) First world wide Colour system Adopted (1966) Generally used in 60 Hz countries Predominantly 525 line TV systems AM modulation of Luma & Syncs (4.2 MHz) U & V Chroma AM Quadrature Modulated (IQ) Chroma Subcarrier 3.579545 MHz FM or Digital subcarrier modulation of Sound SECAM

3 Sequentiel Couleur Avec Memoire (SECAM) Developed by France before PAL 625 Line 50 Hz Colour system Uses AM modulation for Luminance & Sync Line sequentially sends U & V Chroma components on alternate lines Receiver requires a 1H chroma delay line Uses FM for Colour subcarrier 4.43361875 MHz Uses FM for sound subcarrier PAL 4 Phase Alternation Line-rate (PAL) Colour System Developed in Europe after NTSC & SECAM Generally associated with 50 Hz Countries

Predominantly 625 Line system AM modulation of Luma & Syncs (5 MHz) U & V Chroma AM Quadrature Modulated with V (R-Y) component inverted on alternate lines Chroma Subcarrier 4.43361875 MHz FM or Digital subcarrier modulation of Sound Transmission Bandwidth - VHF 6 MHz 7 MHz 8 MHz Not in Use Australia is one of a few countries with 7 MHz VHF TV 4 Transmission Bandwidth - UHF 6 MHz 7 MHz 8 MHz Not in Use 4 Australia is Alone using 7 MHz on UHF U & V Components

Y = 0.299 R + 0.587 G + 0.114 B B-Y = -0.299R - 0.587G + 0.866B U = B-Y R-Y = 0.701R - 0.587G + 0.114B 4 V = R-Y Y, B-Y & R-Y Values B-Y = -0.299R - 0.587G + 0.866B Condition White Black Red Green Blue Yellow Cyan Magenta 4 R 1 0 1 0 0

1 0 1 G 1 0 0 1 0 1 1 0 B 1 0 0 0 1 0 1 1 Y 1 0 0.299 0.587

0.114 0.886 0.701 0.413 B-Y Range is too large B -Y 0 0 -0.299 -0.587 0.886 -0.886 0.299 0.587 R -Y 0 0 0.701 -0.587 -0.114 0.114 -0.701 0.587 What makes a Colour Bar - RGB Red

Green 4 Blue Colour Bar Component Colour Bar - YUV 4 Colour Bar U Y V Y, B-Y & R-Y Values R-Y = 0.701R - 0.587G + 0.114B Condition White Black

Red Green Blue Yellow Cyan Magenta 4 R 1 0 1 0 0 1 0 1 G 1 0 0 1 0 1 1 0 B

1 0 0 0 1 0 1 1 Y 1 0 0.299 0.587 0.114 0.886 0.701 0.413 R-Y Range is too large B-Y 0 0 -0.299 -0.587 0.886 -0.886 0.299

0.587 R-Y 0 0 0.701 -0.587 -0.114 0.114 -0.701 0.587 Y, U & V Values U = 0.564 (B-Y) V = 0.713 (R-Y) Condition White Black Red Green Blue Yellow Cyan Magenta 4 R 1 0

1 0 0 1 0 1 G 1 0 0 1 0 1 1 0 B 1 0 0 0 1 0 1 1 Y 1

0 0.299 0.587 0.114 0.886 0.701 0.413 U 0 0 -0.169 -0.331 0.500 -0.500 0.169 0.331 V 0 0 0.500 -0.419 -0.081 0.081 -0.500 0.419 Component Video

700 mV 350 mV 0 mV -300 mV4 Video distributed as separate Y U V Components Y signal is 700 mV for Video Black-White Y Signal carries Sync at -300 mV U & V signals are 700 mV pk-pk. 350 mV at 0 Y U V Coax

5 Video Signals are transmitted on Coaxial Cable 75 Ohm Coax - RG-59 or RG-178 Video is usually 1 Volt Peak to Peak Terminated with 75 Ohms at end of run High impedance loop through taps are used To split video must us a Distribution Amplifier For Component signals all coaxes must be the same length otherwise mistiming of the video components will occur Standard Definition Television SDTV The current television display system 4:3 aspect ratio picture, interlace scan Australia/Europe USA/Japan

5 625 lines - 720 pixels x 576 lines displayed 50 frames/sec 25 pictures/sec 414720 pixels total 525 lines - 704 pixels x 480 lines displayed 60 frames/sec 30 pictures/sec 337920 pixels total Enhanced Definition Television EDTV 5 Intermediate step to HDTV Doubled scan rate - reduce flicker Double lines on picture - calculated

Image processing - ghost cancelling Wider aspect ratio - 16:9 Multi-channel sound High Definition Television - HDTV 5 Not exactly defined - number of systems System with a higher picture resolution Greater than 1000 lines resolution Picture with less artefacts or distortions Bigger picture to give a viewing experience Wider aspect ratio to use peripheral vision Progressive instead of interlaced pictures HDTV Parameters - AS 4599

HDTV Defined as a MPEG-2 stream which is compliant with [email protected] encoding. HDTV sample rate: 5 Less than 62 668 800 samples per second Greater than 10 368 000 samples per second Systems with less than 10 368 000 samples per second are defined as SDTV HDTV Have We Heard This Before? 5

The first TV system had just 32 lines When the 405 line system was introduced it was called HDTV! When 625 line black & white came along it was called HDTV! When the PAL colour system was introduced it was called HDTV by some people. Now we have 1000+ line systems and digital television - guess what? Its called HDTV! Do You Use A PC? All Current Generation PCs use Progressive Scan and display Pictures which match or exceed HDTV resolutions although the pixel pitch, aspect ratio and colorimetry are not correct. 5 HDTV Video Formats - SDTV - 50 Hz Pixel x Line Pixels/Picture Bitrate Mb/s

704 x 576 405,504 5.0 - 15.0 544 x 576 313,344 3.5 - 8.0 352 x 576 202,752 2.5 - 4.0 544 x 288 156,672 1.5 - 3.0 352 x 288 101,376 1.0 - 2.5 All these formats are Interlaced 5 Video Formats - HDTV - 50 Hz Pixel x Line Pixels/Picture Bitrate Mb/s 1920 x 1080 I 2,073,600 19 - 25 1920 x 1035 I 1,987,200 18 - 25 1440 x 1152 I 1,658,880

15 - 20 1280 x 720 I/P 921,600 8 - 11 720 x 576 I/P 414,720 6 - 15 720 x 480 I/P 345,600 5 - 12 5 HD Video Formats 720 0 1280 1440 1920 345,600 480 576 720 414,720 921,600

1,552,200 1080 1152 5 2,073,600 1,658,880 Common Image Format CIF 6 1920 pixels x 1080 lines is now the world CIF. All HDTV systems support this image format and then allow conversion to any other display formats that are supported by the equipment. In Australia we have adopted the CIF for our HDTV production format. The Recommended Video format is 1920 x 1080 Interlaced at 50 Hz with a total line count of 1125 lines. Chromaticity

SDTV needs compatibility with legacy displays, so default SDTV chromaticity in DVB is: HDTV has unified world-wide chromaticity and no legacy displays 6 same as PAL for 25Hz same as NTSC for 30Hz default is BT.709 for both 25Hz and 30Hz simulcast allows mixture of legacy chromaticity for SDTV and BT.709 for HDTV BT-709 Colorimetry HDTV uses a different colour space to SDTV

HDTV display Phosphors not same as SDTV BT-709 defines the parameter values for HDTV HDTV has a slightly different colour equation Y = 0.2126 R + 0.7152 G + 0.0722 B Colour U = 0.539 (B Y) Difference Signals V = 0.635 (R - Y) 6 Digital Television Why digital? To Overcome Limitations of Analog Television Noise free pictures Higher resolution images Widescreen / HDTV No Ghosting Multi-channel, Enhanced Sound Services Other Data services. 6

Digital Television - Types Satellite (DBS) DVB-S Program interchange Direct view / pay TV SMATV Uplink 6 Downlink Digital Television - Types Cable HFC - pay TV

MATV DVB-C / 16-VSB Fibre Main Coax Tap 6 Spur Tee Digital Television - Types Terrestrial (DTTB) 6 DVB-T / 8-VSB Free to air TV (broadcasting) Narrowcasting/value added services Untethered - portable reception

Digital Terrestrial Television Broadcasting - DTTB 6 Regional free to air television Replacement of current analog PAL broadcast television services Operating in adjacent unused taboo channels to analog PAL service Carries a range of services HDTV, SDTV, audio, teletext, data Providing an un-tethered portable service Enabling Technologies Source digitisation (Rec 601 digital studio) Compression technology (MPEG, AC-3) Data multiplexing (MPEG) Transmission technology (modulation)

6 Digitising Video - Rec BT-601 6 Output 27 MHz - Y Cr Y Cb Y Cr Y Cb .. 10 bit x 27 MHz = 270 Mbit/s Rec BT-601 - Sampling 7 Nyquist Rate for SDTV 11 MHz 13.5 MHz base sampling rate. Chrominance sample rate 6.75 MHz 8 or 10 bit

component samples Parallel BT-656 7 1st Rec 656 connection format used. Uses 110 Ohm twisted pairs for data and clock ECL level signalling @ 27 MHz Width: 10 bits NRZ data + 1 clock pair Uses standard DB-25 Female on Equipment All cables are DB-25 Male to Male pin for pin All cables have overall shield to prevent EMI Max length without a DA 50 m, with EQ 200 m SDI - Serial BT-656

7 Serial Data Interface - Current version of 656 Uses standard 75 Ohm video coax Cabling 1300 nm Optical fibre interface also defined 270 Mb/s Serial data stream of 10 bit data X9+X4+1 scrambling used for data protection Encoding polarity free NRZI 800 mV pk-pk 4 channel Audio can be encoded into ancillary data areas during the blanking period Sampling 7 Digital video requires sampling of the Analog image information. Highest quality achieved when sampling

Component video signals. For SDTV a basic luminance sampling frequency of 13.5 MHz has been adopted. Various methods exist to sample the complete colour image information 4:2:2 4:1:1 4:4:4 4:2:0 4:4:4 & 4:2:2 Sampling YUV YUV YUV Sampling Points 13.5 MHz 4:4:4 YUV Y Only 4:2:2 7

4:1:1 & 4:2:0 MPEG-1 Sampling YUV Y Only Y Only Y Only YUV Sampling Points 13.5 MHz 4:1:1 Y V Y JPEG/JFIF H.261 MPEG-1 7 4:2:0 Y U

Y 4:1:1 & 4:2:0 MPEG-2 Sampling YUV Y Only Y Only Y Only YUV Sampling Points 13.5 MHz 4:1:1 YV Y Only Co-sited Sampling MPEG-2 7 4:2:0 YU Y Only Rec BT-601/656

235 128 16 0 & 2557 Digital Standard for Component Video 27 MHz stream of 8 / 10 bit 4:2:2 Samples 8 bit range 219 levels black to white (16-235) Sync/Blanking replaced by SAV & EAV signals Ancillary data can be sent during Blanking Y U V Decoding Rec BT-601 7

Rec BT-601 - Filtering Multiple A/D and D/A conversion generations should be avoided 7 Enabling Technologies Source digitisation (Rec 601 digital studio) Compression technology (MPEG, AC-3) Data multiplexing (MPEG) Transmission technology (modulation) 8 Video Bitrate - HDTV 2 M pixels * 25 pictures * 3 colours * 8 bits = 1.24416 G bits / sec for Interlace Scan or = 2.4833 G bits / sec for Progressive We need to Compress this a bit! 8 Compression Technology

8 When low bandwidth analog information is digitised the result is high amounts of digital information. 5 MHz bandwidth analog TV picture 170 - 270 Mb/s digital data stream. 270 Mb/s would require a bandwidth of at least 140 MHz to transport Compression of the information is required Compression - Types Two types of compression available Loss-less compression 2 to 5 times Lossy compression 5 to 250 times 8

Compression - Loss-less Types Picture differences - temporal Run length data coding - GIF 101000100010001001101 = 1 + 4x0100 + 1101 01 11 31 31 31 21 01 11 8 21 symbols source = 16 symbols compressed Huffman coding - PKZIP

21 bits source = 12 bits compressed Short codes for common blocks Longer codes for uncommon blocks Lookup tables Compression - Lossy Types Quantisation - rounding Motion vectors Prediction & interpolation Fractal coding Discrete cosine transform (DCT) 8 Approaches to Image Compression Intraframe compression treats each frame of an image sequence as a still image. Interframe compression employs temporal predictions and thus aims to reduce temporal as well as spatial redundancies, increasing the

efficiency of data compression. 8 Intraframe compression, when applied to image sequences, reduces only the spatial redundancies present in an image sequence. Example: Temporal motion-compensated predictive compression. MPEG-1: General Remarks - 1 MPEG-1 standard simultaneously supports both interframe and intraframe compression modes. MPEG-1 standard considers: 8

Progressive-format video only: Luminance and two chroma channels representation where chroma channels are subsampled by a factor of 2 in both directions; 8 bit/pixel video Otherwise, appropriate pre- and post- processing steps should be carried out. MPEG-1: General Remarks - 2 MPEG-1 standardises a syntax for the representation of encoded bit-stream and a method of decoding. The standard syntax supports the operations of: 8 Discrete Cosine Transformation (DCT), Motion-compensated prediction, Quantisation, and Variable Length Coding (VLC).

MPEG-1 - I, P & B Frames Uncompressed SDTV Digital Video Stream - 170 Mb/s Picture 830kBytes I Frame 100 kBytes Picture 830kBytes Picture 830kBytes B Frame B Frame 12-30 kBytes 12-30 kBytes Picture 830kBytes P Frame 33-50 kBytes MPEG-2 Compressed SDTV Digital Video Stream - 3.9 Mb/s I - intra picture coded without reference to other pictures. Compressed using spatial redundancy only

P - predictive picture coded using motion compensated prediction from past I or P frames B - bi-directionally predictive picture using both past and future I or P frames 8 I Frames I I I I I I I

I I I I I Intraframe Compression 9 I Frames marked by (I) denote the frames that are strictly intraframe compressed. The purpose of these frames, called the "I pictures", is to serve as random access points to the sequence. P Frames Forward Prediction I

P P I P Frames use motion-compensated forward predictive compression on a block basis. 9 P Motion vectors and prediction errors are coded. Predicting blocks from closest (most recently decoded) I and P pictures are utilised. B Frames Forward Prediction I B B

P B B P B B P B B I Bi-Directional Prediction B frames use motion-compensated bi-directional predictive compression on a block basis.

9 Motion vectors and prediction errors are coded. Predicting blocks from closest (most recently decoded) I and P pictures are utilised. In Case of Poor Predictions Forward Prediction I B B P B B P B B P

B B I Bi-Directional Prediction 9 In both P and B pictures, the blocks are allowed to be intra compressed if the motion prediction is deemed to be poor. Group of Pictures GoP = 12 9 I B

B P B B P B B P B B 1 2 3 4 5

6 7 8 9 10 11 12 1 I Relative number of (I), (P), and (B) pictures can be arbitrary. Group of Pictures (GoP) is the Distance from one I frame to the next I frame Some Other Frame Patterns An I picture is mandatory at least once in a sequence of 132 frames (period_max= 132) GoP = 6 I B B

P B B I B B P B B I GoP = 2 I B I

B I B I B I B I B I GoP = 2 I P I P

I I I P I P I I I 9 Frame Transmission Sequence Source and Display Order 1 2 3

4 5 6 7 8 9 10 11 12 1 I1 B1 B2 P1 B3 B4 P2 B5 B6 P3 B7 B8 I2 I1 P1 B1 B2 P2 B3 B4 P3 B5 B6 I2 B7 B8 Transmission Order 9 MPEG Typical Frame Size GoP = 15 9 Compression - DCT 8x8 Pixels 9

Steps of Intra Frame Compression Original Image Segment DCT Quantise f(n,m) F(u,v) Image Transform (Reduce number into 8x8 (Efficient of symbols) Representation) Pixel Blocks F*(u,v) Lossless Lossy Symbol Coding (Minimise average length of symbols) 9 Compressed Bit Stream Data

Discrete Cosine Transformation (DCT) DCT can be applied to various sample block sizes For MPEG DCT is applied to 8 x 8 Blocks of Luminance and Chrominance data. and 1 DCT - Original Spatial Pixels m=0 Spatial 8x8 Pixel Values f(m,n) 1 1 2

3 4 5 6 7 n = 0 55 55 109 109 109 109 109 109 n = 1 55 55 109 109 109 109 109 109 n = 2 55 55 109 109 109 109 109 109 n = 3 55 55 109 109 109 109 109 109 n = 4 55 55 55

55 55 55 55 55 n = 5 55 55 55 55 55 55 55 55 n = 6 55 55 55 55 55

55 55 55 n = 7 55 55 55 55 55 55 55 55 DCT - Raw Values Frequency Domain 8x8 Transform Values NINT[ F(u,v)] u=0 3

4 6 7 v = 0 602 -69 -50 -24 0 16 21 14 v = 1 147 -63 -45 -22 0 15 19 12 v=2 0 5 0

0 0 0 0 v = 3 -52 22 16 8 0 -5 -7 -4 v=4 0 0

0 0 0 0 5 0 3 4 3 0 v=6 0 v = 7 -29 0 2

0 v = 5 34 1 1 0 -15 -11 0 0 0 0 0 0 0 12 9

4 0 -3 -4 -2 NINT = Nearest INteger Truncation Why use transform Coding? The purpose of transformation is to convert the data into a form where compression is easier Transformation yields energy compaction The transform coefficients can now be quantised according to their statistical properties.

1 Facilitates reduction of irrelevant information This transformation will reduce the correlation between the pixels (decorrelate X, the transform coefficients are assumed to be completely decorrelated (Redundancy Reduction). How Do Transforms Work? Basic Fourier Analysis The component frequencies are the basis of the original waveform. 1 Waveform is composed of simpler sinusoidal functions Providing enough of the waveform is sampled, component frequencies can be determined that

approximate the original waveform Basis waveforms change with each type of transform. 2D DCT Basis Function 1 1D DCT Basis Function For Simplicity the 2D Basis function can be reduced to a 1D function that is applied in both x and y dimensions nx = cos X 1 x=0 x=4 x=1 x=5 x=2 x=6

x=3 x=7 4 x 4 - DCT Basis Block Pattern u=0 u=1 u=2 u=3 v=0 v=1 v=2 v=3 1 Diagram Simplified by 1 bit Quantising the pattern DCT Block Scan Sequence v=0 v=1

v=2 v=3 u=0 1 u=1 u=2 u=3 4 x 4 DCT Patterns 1 Quantisation - DC Coefficient The DCT coefficients are uniformly quantised. DC and AC Coefficients are treated differently. The DC Coefficient

1 The DC coefficient is divided by 8, and the result is truncated to the nearest integer in [-256 255] range. F*(0,0) = NINT[F(0,0)/8] Quantisation - AC Coefficients 1 Each AC coefficient, F(U, V) is first multiplied by 16 and the result is divided by a weight. w(u, v). times the quantiser_scale. F*(u,v) = NINT[16 * F(u,v)/w(u,v) * quantiser_scale]. The result is then truncated to [-256,255] range. The 8 x 8 array of weights, w(u,v), is called the

quantisation matrix. The parameter quantiser_scale facilitates adaptive quantisation. MPEG-1 Quantisation Matrix w(u,v) 1 8 16 19 22 26 27 29 34 16 16 22 24 27

29 34 37 19 22 26 27 29 34 34 38 22 22 26 27 29 34 37 Default 40 Matrix

22 26 27 29 32 35 40 48 26 26 29 32 35 40 48 58 26 27 29

34 38 46 56 69 27 29 35 38 46 56 69 83 Weights are 8 bit integers Matrix can be downloaded DCT Example - Original Image 1

55 55 109 109 109 109 109 109 55 55 109 109 109 109 109 109 55 55 109 109 109 109 109 109 55 55 109 109 109 109 109 109 55 55 55 55 55 55 55 55

55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55

55 55 55 55 Example - Raw DCT Coefficients 602 -69 -50 -24 0 16 21 14 147 -63 -45 -22 0 15 19 12 0 -52 0 34

1 0 0 0 0 0 0 0 22 16 8 0 -5 -7 -4 0

0 0 0 0 0 5 0 3 4 3 0 -15 -11 0 0

0 0 0 0 0 0 -29 12 9 4 0 -3 -4 -2 Example - Quantised DCT - QS=2

1 75 -35 -21 -9 0 5 6 3 74 -32 -16 -7 0 4 4

3 0 0 0 0 0 0 0 0 -19 8 5 2 0 -1

-2 -1 0 0 0 0 0 0 0 0 10 -4 -3 1

0 1 1 0 0 0 0 0 0 0 0 0 -9 3 2

1 0 0 0 0 Quantiser_scale = 2 Example - Quantised DCT - QS=7 1 75 -10 -6 -2 0 1 2

1 21 -9 -5 -2 0 1 1 1 0 0 0 0 0 0

0 0 -5 2 1 1 0 0 0 0 0 0 0 0

0 0 0 0 3 -1 -1 0 0 0 0 0 0 0 0

0 0 0 0 0 -2 1 1 0 0 0 0 0 Quantiser_scale = 7

Example - 8 x 8 Scan Sequence 1 75 -10 -6 -2 0 1 2 1 21 -9 -5 -2 0 1

1 1 0 0 0 0 0 0 0 0 -5 2 1 1

0 0 0 0 0 0 0 0 0 0 0 0 3 -1 -1

0 0 0 0 0 0 0 0 0 0 0 0 0 -2

1 1 0 0 0 0 0 Quantiser_scale = 7 Example - Inverse DCT - Result Received 8x8 pixel block at the Decoder 1 59

59 105 107 107 110 107 107 56 59 105 108 107 108 106 107 56 62 105 110 107 107 106 107 57 64 100 108 107 106 103 102 54 50 61 57 55 56 59 60 55 52 56

58 54 55 57 55 55 55 56 56 54 54 56 55 52 55 56 59 54

53 56 55 Quantiser_scale = 7 Assignment - Draw accurately the Full 8x8 DCT Basis block set Simplified Diagram by 1 bit Quantising the pattern 1 Quantised Data Stream Quantiser_scale = 2 75 -35 74 0 -32 -21 -9 -16 0 -19 0 8 0 -7 0 5 0 0 5 0 10 0 -4 0 2 0 4 6 3 4 0 0 0 -3 0 -9 3 0 1 0 -1 0 3 0 -2 0 0 0 2 1 0 1 0 -1 0 1 0 0000000 Quantiser_scale = 4 75 -17 37 0 -16 -11 -4 -8 0 -9 0 4 0 -4 0 2 0 0 2 0 5 0 -2 0 1 0 2 3 2 2 0 0 0 -2 0 -4 2 0 1 0 -1 0 1 0 -1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Quantiser_scale = 7 75 -10 21 0 -9 -6 -2 -5 0 -5 0 2 0 -2 0 1 0 0 1 0 3 0 -1 0 1 0 1 2 1 1 0 0 -1 0 -2 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Quantiser_scale = 16 75 -4 9 0 -4 -3 -1 -2 0 -2 0 1 0 -1 0 1 0 0 1 0 1 0 -1 0 0 0 1 1 0 1

0 0 0 0 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 Spatially-Adaptive Quantisation Spatially-adaptive quantisation is implemented by the quantiser_scale, that scales the w(u,v) values 1 The quantiser_scale is allowed to vary from one "macroblock to another within a picture to adaptively adjust the quantisation on a macroblock basis. The quantiser_scale is chosen from a specified set of values on the basis of spatial activity of the block (e.g., macroblocks containing busy, textured areas are quantised relatively coarsely), and on the basis of buffer fullness in constant bitrate applications. Coding: AC Coefficients

Coding is based on the fact that most of the quantised coefficients are zero and hence it is more efficient to represent the data by location and value of the non-zero coefficients. The quantised AC coefficients are scanned in a zigzag fashion and ordered into symbol = [Run, level] pairs and then coded using variable length (Huffman) codes (VLC) (longer codes for less frequent pairs and vice versa). 1 (The VLC tables are standardised.) Example - Run Level Coding Level: is the value of a non-zero coefficient; Run: is the number of zero coefficients preceding it. Run Level 0 -10 0 21 1 -9

0 -6 0 -2 0 -5 1 -5 1 2 1 Run Level 1 -2 1 1 2 1 1 3 1 -1 1 1 1 1 0

2 Run Level 0 1 0 1 2 -1 1 -2 0 1 5 1 5 1 EOB Quantiser_scale = 7 75 -10 21 0 -9 -6 -2 -5 0 -5 0 2 0 -2 0 1 0 0 1 0 3 0 -1 0 1 0 1 2 1 1 0 0 -1 0 -2 1 0 0 0 0 01000001000000000000000 63 DCT coefficients represented by 47 symbols

Coding: DC Coefficients Redundancy among quantised DC coefficients of 8 x 8 blocks is reduced via differential pulse coded modulation (DPCM). The resultant differential signal ([-255, 255] range) is coded using variable length codes. 1 Standard VLC tables are specified. These tables are the only standard tables in MPEG-1 that make a distinction between luminance and chrominance components of the data.) MPEG-1 Bit Stream Hierarchy 1 MPEG Encoder A typical MPEG encoder includes modules for:

1 Motion estimation Motion-compensated prediction (predictors and framestores) Quantisation and de-quantisation DCT and IDCT Variable length coding a Multiplexer a Memory buffer a Buffer regulator Simplified MPEG Encoder Digital Video DCT

Q VLC IQ IDCT MC Pred Motion Vectors 1 Store Side Info Mux Bit Stream Audio, System & Other Data

MPEG Decoder The decoder basically reverses the operations of the encoder. The incoming bit stream (with a standard syntax) is demultiplexed into: 1 DCT coefficients Side information Displacement vectors Quantisation parameter, etc. In the case of B pictures, two reference frames are used to decode the frame. MPEG Decoder Image Data VLC

IQ IDCT Digital Video Store MC Pred DeMux Program Stream 1 Motion Vectors + Side Information MPEG-2 - Formats ML & HL MPEG-2 defines profiles & levels

They describe sets of compression tools DTTB uses main profile. Choice of levels Higher levels include lower levels Level resolution Low level (LL) 360 by 288 Main level (ML) 720 by 576 High level (HL) 1920 by 1152 1 SIF SDTV HDTV MPEG Profiles and Levels [email protected] MAX. BITRATE 300 Mbit/s [email protected]

100 Mbit/s 80 Mbit/s 60 Mbit/s 40 Mbit/s [email protected] [email protected] [email protected] [email protected] [email protected] 20 Mbit/s [email protected] HIGH [email protected] [email protected] HIGH-1440 LEVELS 1 [email protected] 4:2:2 [email protected]

MAIN [email protected] LOW SIMPLE MAIN SNR SCALABLE SPATIALLY SCALABLE HIGH PROFILES [email protected] [email protected] It is preferable that all decoders sold in Australia be [email protected] capable allowing all viewers access to HD resolution when it becomes commonly available 1

Digital Audio - Multichannel Two sound coding systems exist for Digital TV Cover a wide variety of Audio Applications 1 MPEG 1 & 2 Dolby AC-3 DVB VCD and S-VCD DAB, DBS, DVD Cinema (Film) Computer Operating Systems (Windows) Professional (ISDN codecs, tapeless studio, .)

Multichannel Sound TV L Ls 1 C LFE R Rs Masking Both use perceptual audio coding that exploits a psychoacoustic effect known as masking masker Inaudible (masked) tone Level (dB)

Masked Threshold 500 1000 2000 4000 8000 Frequency (Hz) 1 Multichannel Sound - MPEG 1/2 MPEG Audio Layer II was developed in conjunction with the European DVB technology 1

Uses Musicam Compression with 32 sub bands MPEG 1 is basic Stereo 2 channel mode MPEG 2 adds enhancement information to allow 5.1 or 7.1 channels with full backwards compatibility with the simple MPEG 1 decoders MPEG 1 is compatible with Pro-Logic processing. Bitrate 224 kb/s MPEG 1 Bitrate 480 - 512 kb/s MPEG 2 5.1 MPEG Audio Encoder 32 Sub-bands Audio In 2 x 768 kb/s Subband Filter PsychoAcoustic Model 1 Audio Bit Stream O/P

Quantiser & Coder Frame Packer Bit Allocation Coding of Side Information 2x 32-192 kb/s MPEG Audio Decoder Audio Frame Bit Unpacker Stream 2 x 32-192 kb/s De-Quantiser

Decoding of Side Information 1 Inverse Subband Filter Audio Out 2 x 768 kb/s Multichannel Sound - Dolby AC-3 Dolby AC-3 was developed as a 5.1 channel surround sound system from the beginning.

1 Compression Filter bank is 8 x greater than MPEG 2 (256) Must always send full 5.1 channel mix One bitstream serves everyone Decoder provides down-mix for Mono, Stereo or Pro-Logic Listener controls the dynamic range, Audio is sent clean Bitrate 384 kb/s or 448 kb/s Dialogue level passed in bit-stream AC-3 Multichannel Coder L R C LS RS 5.1-ch Encoder 5.1-ch Decoder

LFE LFE Encoder 1 L R C LS RS Decoder AC-3 Stereo Decoder L R C LS RS 5.1-ch Encoder Matrix

LFE LFE Encoder 1 5.1-ch Decoder L R C LS RS 2-channel Decoder Lo Ro MPEG-2 Multichannel Coder concept Lo Ro L

R C LS RS MPEG-1 Encoder MPEG-1 Decoder Re matrix Down mix Extension Encoder LFE MPEG-2 Encoder 1 Lo Ro Extension

Decoder MPEG-2 Decoder L R C LS RS LFE Low cost 2-channel decoder L R C LS RS LFE Down mix Lo Ro MPEG-1 Encoder T2

T3 T4 LFE Extension Encoder MPEG-2 Encoder Low cost 2-channel decoder 1 MPEG-1 Decoder 2-channel Decoder Lo Ro Widely Available 1

All major MPEG-2 Video decoders incorporate 2-channel or 5.1 channel MPEG-2 Audio Several dedicated MPEG-2 multichannel decoders More than 100 Million decoders world-wide Enabling Technologies Source digitisation (Rec 601 digital studio) Compression technology (MPEG, AC-3) Data multiplexing (MPEG) Transmission technology (modulation) 1 MPEG-2 1 Compresses source video, audio & data Segments video into I, P & B frames

Generates system control data Packetises elements into data stream Multiplexes program elements - services Multiplexes services - transport stream Organises transport stream data into 188 byte packets Digital Terrestrial TV - Layers . . . provide clean interface points. . . . 1920 x 1080 1280 x 720 50,25, 24 Hz Picture Layer Video Compression Layer Data Headers Motion Vectors Multiple MultiplePicture PictureFormats

Formats and Frame Rates and Frame Rates MPEG-2 compression syntax [email protected] or [email protected] Chroma and Luma DCT Coefficients Variable Length Codes Transport Layer Transmission Layer 1 Packet Headers Video packet Audio packet

Flexible Flexibledelivery deliveryof ofdata data Video packet VHF/UHF TV Channel 7 MHz Aux data MPEG-2 packets COFDM / 8-VSB Digital Television Encode Layers Control Data Video Picture Coding Control Data MPEG-2

Data Data Coding Sound Audio Coding MPEG Transport Stream Mux Program 1 Multiplexer Program 2 Other Data Control Data MPEG-2 or AC-3 Program 3 Service Mux Bouquet Multiplexer MPEG Transport Data Stream 188 byte packets Control Data

Modulator & Transmitter Delivery 1 System Error Protection Digital Television Decode Layers MPEG-2 Transport Stream Mon Data Picture Decoder Data Decoder MPEG Transport Stream De-Multiplexer

Demodulator & Receiver Delivery 1 System Speakers Audio Decoder MPEG DeMux Error Control MPEG or AC-3 Set top Box (STB) - Interfacing

1 Domestic and Professional interfaces still to be defined Transport Stream via IEEE 1394 (Firewire) Baseband Audio & RGB/YUV Video signals. STB can convert between line standards so you do not have to have a HD display. Display and transmitted information must be at same Frame/Field rate. (25/50) DTTB - Content & Services DTTB was designed to carry video, audio and program data for television DTTB can carry much more than just TV 1

Electronic program guide, teletext Broadband multimedia data, news, weather Best of internet service Interactive services Software updates, games Services can be dynamically reconfigured DVB Data Containers MPEG Transport Stream is used to provide DVB data containers which may contain a flexible mixture of: Streams with variable data rate requirements can be Statistically Multiplexed together. 1 Video Audio

Data services Allows Six 2 Mb/s programs to be placed in a 8 Mb/s channel Examples of DVB Data Containers Channel bandwidth can be used in different ways: 1 SDTV 1 SDTV 2 SDTV 3 SDTV 4 SDTV 5 HDTV 1 Multiple SDTV programs Single HDTV program HDTV 1 SDTV 1 Simulcast

HDTV & SDTV Video Program Capacity For a payload of around 19 Mb/s 1 HDTV service - sport & high action 2 HDTV services - both film material 1 HDTV + 1 or 2 SDTV non action/sport 3 SDTV for high action & sport video 6 SDTV for film, news & soap operas However you do not get more for nothing. More services means less quality 1 Fixed Bit Rate Multiplexing 1

Most early digital services used fixed data rates for each of the component streams. The fixed rate had to allow for a high Quality of Service for demanding material. Fixed Data Rate was set to a high value for QoS Less demanding material is sent at a higher quality level. Works well with systems having similar material on the transport channels. Spare Data Capacity 1 Spare data capacity is available even on a fully loaded channel. Opportunistic use of spare data capacity when available can provide other non real time data services. Example: 51 second BMW commercial

The Commercial was shown using 1080 Lines Interlaced. 60 Mb of data was transferred during it. In the Final 3 seconds the BMW Logo was displayed allowing 3 Phone Books of data to be transmitted. Statistical Multiplexing - 1 1 Increases efficiency of a multi-channel digital television transmission multiplex by varying the bit-rate of each of its channels to take only that share of the total multiplex bit-rate it needs at any one time. The share apportioned to each channel is predicted statistically with reference to its current and recent-past demands. Data rate control fed back to the encoders from the multiplexer.

Statistical Multiplexing - 2 1 More demanding material can request a higher data rate to maintain Quality of Service. More channels can be multiplexed together than an equivalent fixed rate system. Relies on demand peaks on only a few channels while other channels idle at a lower demand.

Recently Viewed Presentations

  • Overcoming Bad Habits Through Control

    Overcoming Bad Habits Through Control

    Controlling Our Habits. Christians . don't habitually sin "No . one who is born of God [deliberately, knowingly, and habitually] practices sin, because God's seed [His principle of life, the essence of His righteous character] remains [permanently] in him [who...
  • Similar Polygons - Teachers.Henrico Webserver

    Similar Polygons - Teachers.Henrico Webserver

    Similar Polygons Informal Definition of Similar Figures Two figures are similar if they have the same shape. (They do not necessarily have the same size.) Formal Definition of Similar Polygons Two polygons are similar iff their corresponding angles are congruent,and...
  • Frank-Wolfe optimization insights in machine learning

    Frank-Wolfe optimization insights in machine learning

    Frank-Wolfe optimization insights in machine learning. Simon . Lacoste-Julien. INRIA / École Normale Supérieure. SIERRA Project Team. SMILE- November 4th 2013
  • How is CQC ensuring safe, high quality and

    How is CQC ensuring safe, high quality and

    Alan Rosenbach, Special Policy Lead, CQC [email protected] * Who we are and what we do * * Our model is ongoing and dynamic. Here is the cycle of registration and ongoing monitoring of compliance. Our regulatory judgements - and any...
  • 4 Continuous Random Variables and Probability Distributions Copyright

    4 Continuous Random Variables and Probability Distributions Copyright

    The continuous analog of a difference is a derivative. The following result is a consequence of the Fundamental Theorem of Calculus. Proposition. Example 8.8. When . X . has a uniform distribution, ... 2 is facilitated by using the same...
  • Electronic Journal Site Licenses: A Boon for Whom?

    Electronic Journal Site Licenses: A Boon for Whom?

    A---Elsevier. Of course they use it to maximize their profits, but you can use it to maximize your distribution. Enough for now. A Publisher's View "So, we should have models where we make a deal with the university, the consortia...
  • Career Clusters & Pathways - Clinton County High School

    Career Clusters & Pathways - Clinton County High School

    Career Clusters & Pathways. ... Oxyfuel Systems. Basic Welding 2. Shielded Metal Arc Welding. Gas Tungsten Arc Welding. Cas Metal Arc Welding. SMAW Groove (Industry Cert.) GMAW Groove (Industry Cert.) * Career Major = 4 Credits in a Career Pathway...
  • Soil Conservation - cpb-us-e1.wpmucdn.com

    Soil Conservation - cpb-us-e1.wpmucdn.com

    contour plowing. 3. terracing. 4. no-till farming. 5. cover crop. 6. crop rotation (1) Wind Breakers slow down the wind (2) Contour Plowingslows down water erosion (for gentle hills) Direction of water flow (3) Terracingfor steephills (makes STEPS to slow...