The Open Collections website will be unavailable July 27 from 2100-2200 PST ahead of planned usability and performance enhancements on July 28. More information here.

Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A real-time source encoder for tv bandwidth compression. Lui, Hung Lam 1972

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


831-UBC_1972_A7 L83.pdf [ 3.13MB ]
JSON: 831-1.0101671.json
JSON-LD: 831-1.0101671-ld.json
RDF/XML (Pretty): 831-1.0101671-rdf.xml
RDF/JSON: 831-1.0101671-rdf.json
Turtle: 831-1.0101671-turtle.txt
N-Triples: 831-1.0101671-rdf-ntriples.txt
Original Record: 831-1.0101671-source.json
Full Text

Full Text

REAL-TIME SOURCE ENCODER FOR TV BANDWIDTH COMPRESSION by HUNG LAM LUI B.E., Tokyo I n s t i t u t e o f T e c h n o l o g y , 1970 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n t he Department o f E l e c t r i c a l E n g i n e e r i n g We a c c e p t t h i s t h e s i s as c o n f o r m i n g to t h e r e q u i r e d s t a n d a r d THE UNIVERSITY OF BRITISH COLUMBIA August 1972 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I ag r ee tha t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r ag ree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y pu rposes may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . It i s u n d e r s t o o d tha t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f The U n i v e r s i t y o f B r i t i s h Co lumb ia Vancouve r 8, Canada Date ABSTRACT A TV bandwidth compression scheme based on psychophysical approach had been studied with an optical real-time processor. This thesis re-ports the design and instrumentation of an electronic version of the op-t i c a l -p.rocess.or.. The electronic. ..processor . is very close .to an actual TV system. The processor requires relatively simple c i r c u i t s , and perfor-mance checks are described. Subjective evaluation of the output picture, showed an improve-ment in picture quality for low compression r a t i o s $ less than 2. Above this ratio, the compression introduces excessive distortions. i TABLE OF CONTENTS Page ABSTRACT • . i TABLE OF CONTENTS i i LIST OF ILLUSTRATIONS . . . . . i i i ACKNOWLEDGEMENT i v 1. INTRODUCTION 1 2. THE PROPOSED SCHEME AND BASIC SYSTEM DESIGN . . 4 2.1 The Proposed Scheme 4 2.2. Compression Ratio Calculations 9 2.3 Basic System Design 10 3. INSTRUMENTATION • . • 12 3.1 The Dual-Mode Processor 12 3.1.1 The Buffering Amplifiers . 14 3.1.2 The Low-pass Filters and the Delay Lines . . . . 14 3.1.3 The Video Switches 16 3.2 The Logic Unit 19 3.2.1 Replenishment Pattern Generator . 20 3.2.2 The Synchronizing Generator 20 3.3 Computer Interface 23 3.4 Test of the Instrumentation 24 4. SUBJECTIVE EVALUATION 28 4.1 S t i l l Picture Evaluation . . . 28 4.2 Estimation of the Point of Subjective Equality 31 4.3 Motion Picture Evaluation 32 4.4 Results and Discussion 34 REFERENCES 37 i i LIST OF ILLUSTRATIONS Figure Page 1- 1 Block Diagram of the Realr-time Source Encoder 3 2- 1 Replenishment Patterns Based on a Frame Basis 6 2- 2 Replenishment Patterns Based on a Line Basis . 7 3- 1 The Dual-mode Processor • 13 '3-2 The 'Buffering Amplifier 15 3-3 Delay lines Switching 15 3-4 The Video Switches and Driver . . . . . 18 3-5 The Logic Unit ..' 21 3-6 Timing Chart of the Replenishment Pattern Generator . . . 22 3-7 System Waveforms for a Frame Basis (R=l) 25 3-8 System Waveforms for a Line Basis (R=l) 26 3- 9 System Waveforms for a Line Basis 27 4- 1 Subjective Evaluation Set-up 29 4-2 The Test Pictures 30 4-3 The Staircase Comparison Method 33 TABLE 4-1 35 i i i ACKNOWLEDGEMENT Acknowledgement i s due to many persons who have helped during the course of this research project. In particular I would like to thank my supervisor, Dr. M.P. Beddoes for guidance and encouragement. I would also like to thank Prof. F.K. Bowers for reading the manuscript. Further thanks go to Mr. W. Walters, Mr. Herb Black, Mrs. Gra-ciela Sutter and Miss Norma Duggan for their professional assistance. I would like to thank my colleagues who participated in the sub-jective tests. Finally I would l i k e to thank MRC and NRC for the financial support.of this work. i v 1. INTRODUCTION It i s generally supposed that a large amount of redundancy i s present i n today's video transmission procedures. At the same time, the appearance of video-phone, long-distance video r e l a y l i n k s and g l o b a l video communication v i a s a t e l l i t e s demand that the data rate or the band-width of video signals be reduced while preserving p i c t u r e q u a l i t y . Numerous techniques for compressing the bandwidth of video s i g n a l s have been developed during the l a s t two decades. They may be c l a s s i f i e d into the following three categories (1), ( 2 ) . ( i ) Pure s t a t i s t i c a l methods that make use of the s t a t i s t i c a l properties of the source. ( i i ) Pure psychophysical methods that make use of the li.mita-tions of v i s u a l perception. ( i i i ) A combination of the above two. Most of the work i n t h i s f i e l d has been investigated by- computer simulation. Computer simulation i s v e r s a t i l e and economical. Real-time studies are p r o h i b i t i v e because i t often takes the f a s t e s t computer se-v e r a l seconds to process a s i n g l e p i c t u r e frame, so that i t i s v i r t u a l -l y impossible to study s i t u a t i o n s that involves motion. I t has been reported (3) that real-time studies revealed unpredictable e f f e c t s which might not be noticeable i n s t a t i c simulations. Thus real-time studies should be made to study the p r a c t i c a l case c l o s e l y before f i n a l conclu-sions are drawn. A bandwidth compression scheme r e s u l t i n g from f l i c k e r e x p e r i -ments have been proposed and studies by Meier (4), (5) and Chu (6). An 1 2 optical-processing system with coherent light was used for real-time simulations. The eye's sensitivity to fli c k e r frequency was found to diffe r according to the highest spatial frequency present in the pic-ture. The television source encoder could be designed to match more closely the fli c k e r sensitivity of the eye. In the present work, the design and instrumentation of an elec-tronic encoder that performs real-time encoding in synchronization with a closed circuit TV system is reported. It i s the electronic version of Meier's optical processing system (4), (5) and is much closer to an actual TV system. Figure 1-1 shows the. block diagram of the encoder. The design principles and instrumentation can be found in Chapter 2 and 3. Sub-jective test results can be found in Chapter 4. TV CAMERA VI!>EG SYHC P>UAL-MO"D£ PROCESSOR CHANNEL CHANNEL | I REPLENISHMENT PATTERN £r£NERA70fL T (-BNEfcATofC PDP-1Z F i g . 1-1 B l o c k D i a g r a m o f t h e R e a l - t i m e E n c o d e r T l / MONITOR 4 2. The Proposed Scheme and Basic System Design. 2.1 The Proposed Scheme The poss i b i l i t y of compressing the bandwidth of TV signals by alternate presentations of high-quality and low-quality versions of the same picture arises from the result of Meier's spatial frequency f l i c k e r experiments reported in 1968 (4), (5). We define a picture to be a high-quality one or a low-quality one by the highest spatial frequency present, which in turn affects the subjective quality we conceive from i t . With an optical data processor, Meier measured "the response of the human eye to spatial frequencies affected by the fact that the picture i s presented only part of the time rather than continuously, or in other words, the c r i t i c a l f l i c k e r frequency as a function of spatial frequen-cy" (4). Following this, Chu reported a series of subjective evaluation of picture quality on half-tone pictures by using an improved version of the same optical data processor (6). From the results of their work, the human visual system's sen-s i t i v i t y to f l i c k e r frequency i s found to diffe r according to the high-est spatial frequency present in a picture. The visual system tends to exhibit lower c r i t i c a l f l i c k e r frequency for high spatial frequencies. A TV picture may be compressed by presenting the low spatial frequency components at the regular rate and the higher spatial frequency compo-nents at a reduced rate, while the observer might not notice any speci-f i c f l i c k e r or subjective degradation. Subjective picture quality tests, however, showed that the picture degradation could increase very rapid-ly as the compression ratio i s raised. Before going over to the basic system design, consider what re-quirements the scheme would impose on. the encoder when applied to com-mercial TV. Commercial TV with 2-fold line-interlace scans 60 half-pictures (fields) per second, which is equivalent to 30 complete pic-tures (frames) per second. Each picture element i s replenished with brightness information at a frequency of 30 Hz. According to the proposed scheme, for each picture element, low spatial frequency components would be replenished at the normal rate of 30 Hz, while high spatial frequency components could be replenished at any rate between 30 Hz and 0 Hz, as long as subjective f l i c k e r sen-sation i s not evoked. When picture elements are replenished only with low spatial frequency information, we get a low-quality picture. When pi ture elements are replenished with a l l spatial frequency information available, we get a high-quality picture. Consider the case in which picture elements are replenished a l -ternatively in the low-quality mode and in the high-quality mode on a frame basis. The temporal replenishment ratio, R is defined as the ratio between the time of replenishment occupied by the low-quality mode and that of the high-quality mode , during the active scan of the TV pic-ture. For R=l, the compression scheme would have one picture frame re-plenished entirely in the low-quality mode, followed by the next frame replenished entirely in the high-quality mode, and vice versa. The tran-sitions from one mode to the other take place during the blanking periods Figure 2-1 illustrates the replenishment patterns based on a frame basis for R=l, 2 and 3. The more general replenishment pattern can be derived from a line basis. Figure 2-2 summarizes the replenishment patterns based on a line basis for R= 1, 2 and 3. For sake of c l a r i t y , a frame simply con-6 ?ICJURE FRAMES FROM TV CAMERA F3 h J J L.;f 1 VuAL - MODE TJCTUZE FRAMES FROM TROCZSSOR. - j—1\ — HQ. La Ha La na La HO. La EvERy Z FFAMES MA KB ONE COHPRESSiofJ CYCLE — Ha La La HQ, \ / u Evt7f?y 3 FRMFS Ha La La - j j -La La LGL. •IV La Ha La LCI La I—> EVZRy 4 rRAM£$ MAKE OME CoMPRE£Sl6tJ CYCLE * /EACH FRAME CONTAit/S Z FlEU>S OF THE SAMS auALIjy F i g . 2-1 R e p l e n i s h m e n t P a t t e r n s B a s e d on a Frame B a s i s 7 PICTURE FRAMES, FROM TV CAMERA 1.1 tJe 1 1 3 4 S F, Fz F3 F4 II-1 z 1 4 / PROCESSOR. R=1 ^ V Z Z Z Z PICTURE FRAMES FI?OM P&CESSORS/ -—II / V////A : A LINE REP/JNI^EP //V //£ MODE ta MODE zzzzz 7777. V777 / / / / zzZZ, V y - i f -/ // // '// / .-• '•//,' / / //// ' / / / / / / / / / /// /A L-»• EVERy 2 FRAMES MAK& ONE COMPRESSION1 CYCLE 1 4 _ S //// y / / / / ///•/.• //'/// 777777 '/•/// ///'// ////, 7 / / // V y z z z 7.///  77777 W77 ////-. 77277. 272 fy^ey 3 FRAME? M/tfce cvf CDMPZESSIOM CYCLE 1 5 V ////A z z z^. '/ / / A/ 777. y -if-7,vv,' . ' / / / / '//// // //, ////V |V //V ///// ///// 7/77, ///// / / // . '////' ////../ / / / / •'//// 1—*- £ V « / 4 FRAMES M4KE ONE COMPRESS/ti^ CYCLE * /F/l^fV CONTAINS Z F/eLOS WITH IDENTICAL REPlBf/isHME*JJ PATTERN" F i g . 2-2 Replenishment Pattern Based on a Line Basis 8 sists of 5 lines. By repeating the same pattern twice for each frame, the problem that each frame consists of 2 interlaced fields i s solved. The replenishment patterns for higher values of R or for the practical 525 lines system can be derived by following the examples shown. One can also derive the replenishment patterns based on n-line basis, with the upper limit being that of a frame basis. To reduce the complexity of the flexible memory (7) required to convert the irregular data rate in the encoder to uniform rate, mode-transitions should be made as often as possible. Considering synchronization problem in the decoding end, the highest rate i s taken to be that of a line basis. The line-sync pulses can thenbe ut i l i z e d for synchronizing the transitions, which are completed during the blanking periods of the scan. It w i l l be clear in Chapter 3 that this choice also eases the design considerations of the video switches. The highest mode-transition rate is half of the horizontal scanning frequency. For commerical TV the highest mode-transition rate is 7875 Hz, corresponding to replenishment based on a line basis with R = 1. For replenishment based on a frame basis, the highest mode-tran-si t i o n rate i s 15 Hz (R = 1), half of the frame frequency. Other mode-transition ratio can be easily calculated since they are merely fractions of the horizontal-line or frame frequency. 9 2.2 Compression Ratio Calculations By relating the spatial frequency contents of a TV picture to the bandwidth of the corresponding TV signal, the compression ratio, C, can be expressed as a function of the temporal replenishment ratio,R, and the bandwidths of the high-quality and low-quality picture signals. The highest spatial frequency present in a TV picture can be expressed by the horizontal resolution: n lines/picture width, and the bandwidth (BW) of the corresponding signal is given by BV = - 4 - * — [Hz] ( 1 ) 2 T H J_ where i s the horizontal scanning period. The compressed picture, generated from high-quality pictures of bandwidth (BW) and low-quality pictures of bandwidth (BW) has an equivalent composite bandwidth (^)rjoMP °^ ( B W ) 0 ( B W >HQ 'THQ + ( B W )LQ • TLQ COMP T H Q + T L Q (BW)HQ + (BW)LQ . R (2) where TJJQ1s t^-e time during which the replenishment i s done i n the high-quality mode and T i s the time during which the replenishment i s done in the low-quality mode, for one complete cycle of replenishment, and 10 T • A L ^ R = — by definition. HQ Defining the compression ratio C as A ( B W ) H O c k S2 (3) we have 1 + R c 5 5 B ^ ( 4 ) 1 + R ^ <BW>HQ For example, with (BW) = 3 MHz, (BW)T_ = 1 MHz and R = 1, the HQ tiQ compression ratio is C - ^ =1.5 i + r -4-2.3 Basic System Design Based on the above discussions, a real-time encoder was designed, instrumented with electronic components and tested. The block diagram of the encoder and i t s relevant peripherals is shown in Figure 1 - 1 . A closed circuit TV system was used as the source and the receiving end of the system. The TV camera supplies the video input while the moni-tor displays the encoded output. The camera also provides the main clock pulses from which a l l timing pulses of the encoder are derived. The encoder can be divided into two subsystems. One of the 11 s y s t e m s , t h e d u a l - m o d e p r o c e s s o r , p r o c e s s e s t h e i n c o m i n g a n a l o g v i d e o s i g n a l i n e i t h e r t h e h i g h - q u a l i t y mode o r t h e l o w - q u a l i t y mode. I t i s c o n t r o l l e d b y t h e o t h e r s u b s y s t e m , a d i g i t a l l o g i c u n i t t h a t t a k e s c a r e o f s y n c h r o n i z a t i o n a n d t i m i n g . A p a r t f r o m c o n t r o l l i n g t h e t r a n s i t i o n s o f t h e e n c o d i n g mode, t h e l o g i c u n i t a l s o e n a b l e s t h e s e l e c t i o n o f v a r i o u s v a l u e s o f R. The s e l e c t i o n o f r e p l e n i s h m e n t p a t t e r n s c a n b e b a s e d on e i t h e r a f r a m e b a s i s o r a l i n e b a s i s . E x t e r n a l s y n c h r o n i z a -t i o n i s u s e d o n t h e d i s p l a y u n i t t o e n s u r e t h a t s t e a d y s y n c h r o n o u s s c a n -n i n g a c t i o n w i l l b e o b t a i n e d r e g a r d l e s s o f t h e e n c o d i n g p r o c e s s . 12 3. Instrumentation 3.1 The Dual-mode Processor The dual-mode processor as mentioned in Figure 1-1 processes the incoming video signal in either the high-quality mode or the low-quality mode under the control of the logic unit. The block diagram of the processor is shown in Figure 3-1. The video signal from the TV camera i s bandlimited to 3 MHz, corresponding to a horizontal resolu-tion of approximately 280 lines/picture width. Two identical buffering amplifiers convey the 3 MHz signal simultaneously to the high-quality (HQ) and the low-quality (LQ) processing channels. In the LQ processing channel, low-quality picture signals are generated by further bandlimiting the 3 MHz signal with low-pass f i l t e r s . The HQ processing channel i s es-sentially an all-pass network, in which delay lines are inserted to com-pensate for the delay in the low-quality picture signal caused by low-pass f i l t e r i n g . The correct amount of delay given to the high-quality signal ensures the spatial registration of the superimposed high-quality and low-quality pictures on the display unit. The selection of HQ or LQ mode is performed by 2 FET switches, represented by the switch "S" in Figure 3-1. A third buffering amplifier then isolates and drives the TV monitor. With the switch closed for trans-mission through path H, the output signal i s in high-quality mode. With the switch closed for transmission through path L, the output i s in low-quality mode. No effort is made to convert or process the picture signal i n di g i t a l form. The dual-mode processing is done purely i n analog format to ensure the experimental results are free from d i g i t a l quantization errors. TV CAMERA IFF PRiUTS LDHOOSO To LoCjiC , UNIJ sync EMITTER. F0LL0WE.R. BUFF£RIM6T AM?UFl£R. TV MONITOR VIDEO —-CoURAC KNB1 Lo^rc UNIT F i g . 3-1 The Dual-mode Processor 14 3.1.1. The Buffering Amplifiers The buffering amplifiers are identically b u i l t . Each consists of a common-emitter stage for amplification followed by a common-collec-tor stage for impedance transformation, as shown in Figure 3-2. The frequency response is f l a t within 3db from 50 Hz to 8 MHz, and the high-frequency r o l l - o f f i s smooth. A single stage of emitter follower was used for driving the 3 MHz p r e - f l i t e r at the input of the processor. 3.1.2. The Low-pass F i l t e r s and the Delay Lines Low-pass f i l t e r s are required in the LQ processing channel. As w i l l be discussed in Chapter 4, in order to evaluate the subjective quality of the compressed pictures, a third processing channel with several low-pass f i l t e r s to generate picture standards is required. It is essential for the low-pass f i l t e r s in both channels to have similar characteristics. A set of 6 pairs of f i l t e r s , each pair having a different cut-off frequency, had been constructed by Farr for f i l t e r i n g video signals (8). The f i l t e r s were checked and found to be very handy for the pre-sent purpose. The f i l t e r s are third-order Transitional Butterworth-Thomson f i l t e r s (9), having a calculated overshoot of 3.9% for a step input and an attenuation of 15.4 db/octave. The f i l t e r s , with cut-off frequen-cies of 2.0, 1.5, 1.0, 0.75, 0.5, 0.25 MHz are considered to have spanned suitably the region of interest in which the compression scheme is to be studied. The selection of the low-pass f i l t e r s was done by mecha-nical rotary switches. G-MD 15 i-'iSv (rND /777m Ti Tz • 2N4 124 Pig. 3-2 The Buffering Amplifier Module /A/ o — PFSLAY LitJBS »1 P o— o— o—i 3 o OUT - o /7J77 Pig. 3 - 3 Delay Lines Switching 16-The second stage of the buffering amplifier, which i s an emitter follower, acts as a voltage source for driving the low-pass f i l t e r s each terminated in 75 ohms. To compensate for the delay in the low-quality picture signal, delay lines were used in the HQ processing channel to give the high-quality picture signal the appropriate amount, of delay. Although i t is possible to calculate the amount of delay from the characteristic of low-pass f i l t e r s , experimental compensation by t r i a l and error was found to be most effective for f i n a l visual matching of the pictures. The approximate amount of delay required in the high-quality picture signal was 100 ns, 150 ns, 250 ns,:350 ns to compensate for-the delay of the low-quality signal bandlimited to 1.5 Mhz, 1 MHz, 0.75 MHz and 0.5 MHz respectively. Commercially available delay lines of 50 ns per module were used. A cascade of 15 sections of such modules has a delay time to rise time ratio of 10. Defining the rise time (tr) as the time required for the output to rise from 10% to 90% of i t s f i n a l value, the equiva-lent bandwidth B is related to tr as B = 0.44tr \ Thus the bandwidth of 15 sections cascaded delay lines exceeds 5 MHz, while giving a delay time of 750 ns. For shorter cascaded lines the bandwidths are wider than 5 MHz, which should be more than adequate for the present purpose. To minimize the number of sections required, the delay lines were cas-caded in groups and switched as shown in Figure 3-3 to give the appro-priate amount of delay. 3.1.3. The Video Switches To multiplex the high-quality picture signal and the low-quality 17 picture signal, electronic switches were used. The dual-input single-ended multiplexer using JFET as switches i s shown in Figure 3-4. A switch-driving circuit with complementary outputs was used to turn the switches on and off. Amelco 2N4303 N-channel JFETs have been used. Their pinch-off voltage i s rated at 6 volts maximum . To turn the JFET on, the diode connected to the gate of the JFET must be reverse - biased so that the voltage at the gate can follow that at the source, satisfying the con-dition V =0. The magnitude of the video signal to be switched i s 2 volts peak-to-peak maximum. A control voltage more positive than 1 volt w i l l suffice. To turn the JFET off,the gate must be held more negative than -7 volts. Allowing for a forward voltage drop of one volt across the diode, a control voltage of -8 volts or more w i l l be required. The switch-driving ci r c u i t in Figure 3-4 delivers control voltages well over + 10 volts to the JFETs. The choice of + 15 volts as power supplies was made since the same power supplies were used by the buffering amplifiers. The choice of having the mode-transitions completed during the blanking periods of scan eases many design considerations on the switches. The horizontal blanking period i s approximately 10 us long. The mode-transitions are made at the leading edge of the sync pulses. 10 us w i l l be more than adequate to allow for the settling of a step input signal. Any AC signal pickup (spikes) from the switch-driving ci r c u i t w i l l not interfere with the display since a l l transitions are blanked out during that 10 us period. Mode-transitions completed during the vert i c a l blank-ing periods offer no problem too since the vertical blanking periods are even longer than 10 us. s-vv CONTROL IN -AAA linn Fig. 3-4 The Video Switches and Driver 19 The source-drain capacitance of the JFETs and stray wiring capaci-tance are responsible for the input signal feedthrough from the "off" switch to the output. Signal feedthrough is directly proportional to the frequency and amplitude of the signal applied to the "off" switch and the signal source impedance of the "on" switch. Measurements i n -dicated that'the actual signal feedthrough level in the multiplexer is better than -38 db at 2 MHz, with 2 volts peak-to-peak signal applied to the "off" switch. This was considered adequate for our purposes. The passband of the multiplexer i s checked to be f l a t from DC to 8 MHz minimum. 3.2 The Logic Unit The logic unit has the following functions: (1) permits the selection of replenishment patterns based on a frame basis or a line basis, (2) permits the selection of different values of R, (3) generates the appropriate replenishment pattern control signal, (4) generates composite synchronization pulses to the display unit, and (5) interfaces with a computer i f desired. The original design of the logic unit had been tedious and complicated. To encode in different values of R required considerable amount of ROM (Read-Only-Memory) to store the replenishment patterns for each value of R. Then i t was j u s t i f i e d to put the encoding action under the control of a computer for any value of R greater than 1. The logic unit became simple and the replenishment patterns can be modified simply by changing the control program. The f i n a l version of the logic 20. unit i s shown i n Figure...3-5. 3.2.1. Replenishment Pattern Generator A replenishment pattern generator was used to generate the re-plenishment patterns for R=l. It contains a MCFF (Mode-control F l i p Flop) which, being triggered by the leading edges of selected sync pulses, stores the mode-control information and holds the video switches to the appropriate mode before the next sync pulse occurs. A 2-stage counter (T 4) is used to generate the replenishment pattern based on a frame basis from the vertical sync pulses. Then i t i s triggered into the MCFF. The extra gating action by the NOR gate i s redundant in this case. The timing chart for replenishment based on a frame basis for R= 1 is shown in Figure 3-6, (a). The same 2-stage counter can be used to i n i t i a l i z e the status of the MCFF for the f i r s t line of each f i e l d when the encoder i s work-ing on a line basis. The MCFF i s then triggered by the leading edge of every horizontal sync pulse, reversing i t s state by every line. At the beginning of each f i e l d , the proper status i s i n i t i a l i z e d by setting the output of the counter into the MCFF just one horizontal period be-fore the active scan starts. Two NOR gates check that no further set-ting of the MCFF is made during the active scan period, allowing the MCFF to reverse i t s state properly. The timing chart for replenishment based on a line basis for R = 1 is shown in Figure 3-6 (b). 3.2.2. The Synchronizing Generator Sync signals from the TV camera are injected Into the sync gene-rator through two level converters, labelled LI, L2 in Figure 3-5. Hori-" fzoM TV CAMERA FROM Ls2> of BUFFER 3f SyucHRoNiZtMCr GrENERAToR To to Fig. 3-5 The Logic Unit Fl?oM T V CMBKh: COMPOSITE VIDEO Siett/AL - - I FRAME / FRAME ZFF2 TRIGGER MCFF I FlEEDi I FlELPz FlELPi (HoR\.sytk) I II MilII I! II MCFF--JU|J1 U LI u F»£i-2>4 FIB LP 5 ' F/£U>* | FRAML-4 FlELVj I M J l I U l R R j J FRAME BASIS LINE BASIS ONE REPLEHIZH'MEHT cyct£ (R=1) * Horizontal axis not to scale Fig. 3-6 Timing Chart of the Replenishment Pattern Generator 23 zontal sync pulses (5ys in width) are regenerated by a mono-stable multi-vibrator MM2. Composite sync signal are generated by combining these 5us pulses with the vertical sync signal. Monstable multivibrator MM1 provides the gating pulses required by the replenishment pattern gene-rator. Two further level converters, L3, L4, convey the gating pulses from MM1 and the horizontal sync pulses from MM2 to the sense lines of the PDP-12. The mode-control signal generated by the computer i s strobed into a f l i p flop at the beginning of each horizontal sync pe-riod. The output of the f l i p flop can then be used to control the video switches. 3.3. Computer Interface Computer interface a b i l i t y i s not only necessary for the pre-sent format of replenishment, i t is necessary for further investigations such as adaptive type replenishment studies. Programs can be easily written and modified to provide different versions of replenishment format. A DEC PDP-12 d i g i t a l computer was chosen for the interface. The PDP-12 has 12 d i g i t a l Sense-lines that can be individually tested with a Line mode instruction. It also has 2 I/O Bus Transfer Registers that can move data in or out of the computer's accumulator, with PDP-8 mode instructions (10). Two control programs were written, one for replenishment on a frame basis, the other for replenishment on a line basis. Both programs are able to generate replenishment pat-terns for R=l, 2, 3 and 5. Horizontal sync signal and ver t i c a l sync signal are fed into sense-line 1 and 2 respectively. The mode-control signal is taken out from the LSD of Buffer 31, one of the I/O Bus Transfer Registers. 24 3.4 Test of the Instrumentation To test the system's operation, waveforms at various points were monitored and recorded. The input to the HQ processing channel was purposely grounded when taking these photographs. Figure 3-7 shows the waveforms of the system working on a frame basis (R = 1). The switching action of the dual-mode processor can be seen in waveform (c) determined by the replenishment pattern control signal, waveform (b), from the logic unit. Figure 3-8 shows the waveforms obtained when the system is work-ing on a line basis (R=l). The upper photo shows the f i e l d i n i t i a l i z a -tion and the lower photo shows the replenishment pattern control signal by expanding the time axis of the upper photo. The switching action of the dual-mode processor working on a line basis i s shown in Figure 3-9. These waveforms were obtained while interfacing with the PDP-12, so that replenishment patterns for R= 1, 2, and 3 could be generated. The upper photo shows the replenishment pattern control signals and the lower photo shows the output of the dual-mode processor for each value of R. Comparing the above waveforms with the timing chart (Figure 3-6), the logic unit and the video switches were found to function in the correct time sequence. Tests were also run on the analog processor based on the appea-rance of the Marconi Resolution Chart No. 1 (Figure 4-2,a). The 4 l e -vels of grey in the pattern were reproduced, and the highest resolution corresponded to ~280 lines/picture width. Hori.: 20 ms/cm Vert.: 2 v./cm Fig. 3-7 System Waveforms for a frame basis (R=l) (a) video input from TV camera (b) replenishment pattern control signal (c) video output from encoder (HQ processing channel grounded) 26 J _ ~ • -U-J-« ^ 7 - ^ -1 1 +^++ - r f r f - nTTt! 11 E -7-7+7 1 J J . _ - Hori.: - 12 ms/cm (uncalib.) Vert.: 2 v/cm Hori.: 50 us/cm Vert.: 2 v/cm Fig. 3-8 System Waveforms for a Line Basis (R=l) (a) video input from TV camera (b) f i e l d i n i t i a l i z a t i o n (c) same as (a), except viewed with different time-base on scope (d) replenishment pattern control signal 27 H o r i : 50 ys/cm Vert: - 2 v/cm (uncalib.) (e) (f) (g) r Mb* i? n i Mr - H 1 til 1 1 1 1 i1 1.»IMH 1 j ™ 1 1 a f u 1 t >^1 ' 11 v 1 1 r* I • I' ' a 1 I F imp f ¥ if i » • i 1 R = 1 R = 2 R = 3 Hori: 50 ys/cm Vert: 2 v/cm F i g . 3-9 System Waveforms for a Line Basis (a) video input from TV camera (b) replenishment pattern control s i g n a l , f o r R=l (c) replenishment " 11 " , f o r R=2 (d) " " " " , f o r R=3 (e) video output from encoder, f o r R = 1 (f) (g) , f o r R = 2 , f o r R = 3 28 4. Subjective Evaluation 4.1 S t i l l Picture Evaluation The following work was done to answer an important question. Is the compressed p i c t u r e superior to the normal p i c t u r e at the same band-width? Figure 4-1 shows the complete set-up for the subjective evalua-t i o n . A t h i r d processing channel had been added to the system. I t i s i d e n t i c a l to the LQ channel i n the dual-mode processor, with an extra b u f f e r i n g a m p l i f i e r added to drive the monitor. They were taken as standards to which the compressed pic t u r e s were compared. There were a l l together 7 standards, with bandwidths of 3, 2, 1.5, 1, 0.75, 0.5, 0.25 MHz. A Conrac I I , KNB9 TV monitor modified to operate with external sync was used. It displays a 7-inch x 5-inch r a s t e r . The amplitude-frequency response of the video a m p l i f i e r i s down 3db at 8 MHz. The subject was seated at a distance of about 4 p i c t u r e heights (20 inches) from the screen of the monitor. He was allowed to adjust the brightness and contrast of the d i s p l a y as he desired. The room i l l u m i n a t i o n was held constantly at 55 lm/m2. Three s t i l l p i ctures were used i n the t e s t s . As shown i n Figure 4-2, they are selected to represent a wide range of pi c t u r e s that.might be found on the d a i l y TV screen. The o r a l i n s t r u c t i o n s given to each of the i n d i v i d u a l l y tested subjects were: " You w i l l be given p i c t u r e sets to compare. Each set contains 2 p i c t u r e s . You can e i t h e r show your preference f o r one of them, or decide that they are of the same q u a l i t y . With the switch provided you can s e l e c t p i c t u r e A or p i c t u r e B to have i t displayed on the monitor. TEST PATTER t*J TV CAMERA SYNC PHILIPS Ll>H0050 V/DEO VIPEO THE REAL- T'ME ENCODER. (REFER, TO FfCr- 3 - f j 3-5) E.F. 3 MHz, A 59 THE THlRP CHANNEL GtENERATtUt* STANPKDS SYNC TV MONITOR, VIDEO COA/EAC KN&9 L _J SUBJECT Co/vreol B.AC BUFFERING AMPLIFIER. _ Low-PASS FILTER. F i g . 4-1 Subjective Picture Quality Evaluation Set-up 30 (a) Marconi Resolution Chart (c) Cyclist Fig. 4-2 The Test Pictures (Bandwidth = 3 MHz) 31 You may switch back and forth as often as you wish and for as long as you want." " Please announce your decision once you have made up your mind." The subject was presented the next pair of pictures immediately after he had reached a decision. The same procedure continued un t i l the evaluation of a test picture was completed. The subject was then given 2 to 5 minutes of recess. Evaluation of a new test picture was repeated in the same manner. 4.2. Estimation of the Point of Subjective Equality With the subjects indicating their preference among the picture pairs, i t seemed possible to obtain an idea of not only which picture was better, but how much better. To estimate the point of subjective equality (PSE), the Staircase Comparison Method was used (11). A com-pressed picture was presented along with a standard of arbitrary band-width. Based on the subject's preference, the bandwidth of the standard was changed by a predetermined amount, in such a way as to induce the subject to change his preference in the next comparison. For example, i f the preference was for the standard picture, the bandwidth of the standard would be decreased for the next comparison, and i f i t was for the compressed picture, the bandwidth would be increased. The time series of the bandwidth of the standards followed an up-and-down course converging on the PSE as shown in Figure 4-3. The PSE was estimated by averaging the peaks and valleys of the up-and-down sequence, shown as A, B, C, D, E in Figure 4-3. / • - • . . . 32 4.3 Motion P i c t u r e Evaluation It i s important i n bandwidth compression studies to include s i -tuations i n v o l v i n g motion i n the d i p l a y . The real-time encoder permits the study of motion p i c t u r e s . A man s i t t i n g i n a sound-proofed- room reading a book was imaged by the TV camera. The t e s t scene might be considered a simulation of face-to-face conversation. Movement of the l i p s and eyes could be no-t i c e d most of the time, while movement of the head and shoulder could be observed quite frequently. The model was i n s t r u c t e d to read at a constant speed and to t r y h i s best to repeat the same movements for each t e s t . The same tes t conditions and procedures as used i n the s t i l l p i c t u r e test were adopted. 4.4 Results and Discussion The r e s u l t s of the subjective evaluation are summarized i n Table 4-1. Only pi c t u r e s compressed on a l i n e basis with. R = 1 were evaluated i n t e n s i v e l y and reported here. Attempts to evaluate p i c t u r e s compressed on a frame b a s i s , or on a l i n e b a s i s with R > 1 revealed that d i s t o r t i o n s brought about by compression made i t d i f f i c u l t to compare such p i c t u r e s with the standards. Based on complaints from subjects, the evaluation of such pictures was terminated. When compressed on a frame b a s i s , the p i c t u r e was found to e x h i b i t l o c a l - f l i c k e r along v e r t i c a l l y oriented edges. The l o c a l - f l i c k e r sensa-t i o n increased as the compression r a t i o was r a i s e d . Subjects agreed that i t was d i f f i c u l t to associate the q u a l i t y of such p i c t u r e s with the n o n - f l i c k e r i n g standards. When compressed on a l i n e b a s i s , the p i c t u r e was found to e x h i b i t X SOBWSCT ? REFERS STANDARD Fig. 4-3 The Staircase Comparison Method (JO 34 line crawling effect along verti c a l l y oriented edges. When R > 1, the line crawling effect was further complicated by appearance of lo c a l -f l i c k e r sensation. Evaluation of such pictures was rejected in a pre-liminary viewing. When R = 1, the line crawling effect was found to be negligible i f the difference in bandwidths of the HQ picture and the LQ picture was small. The results summarized in Table 4-1 had the difference kept constantly at 1.5 MHz. The mean of each PSE was obtained by averaging the estimated PSEs of seven subjects. They were mostly graduate students who had l i t t l e or no experience i n picture quality evaluation. Assuming that the data sets were taken from a population with a normal cumulative distribution, the 95% confidence interval of each PSE was calculated using the Student's t-test. Table 4-1 shows that results obtained with the four different test pictures were very similar: there were no significant differences. Hence the PSE's for the four pictures could be averaged to obtain a mean PSE, designated PSE, with higher accuracy. Since the PSE is the hypothetical bandwidth required to transmit a picture of equivalent quality without the use of compression, i t is instructive to compare PSE with BW , the bandwidth needed to transmit the comp actual compressed picture. If PSE is less than BW , there i s no point i n comp constructing the compression terminals, better results could be obtained by simple f i l t e r i n g . If PSE exceeds BW , then there i s some improvement in r ° comp picture quality through the use of compression techniques, when compared with' simple f i l t e r i n g using the same transmission bandwidth. We therefore define an Improvement Factor, I, as I = PSE/ BW , comp COMPRESSED PICTURE (LINE BASIS , fc= 1 ) ? 5 £ P S £ (BW) , V /COM/" 6 TEST PICTURE CA) cc; _ 4 A PS£ 65H^) COMP 3 0 2-25" f-33 2-28 ± o-24 2-5o ± 0-26 ±o-22 ±0-24 ±0-1 z ± o- o5~ 2-5" /•o 1-75 f-43 2 1 4 ±0-2Z 1-13 ±o-1i /• 73 2- 12 2-o3 ± O-O7 ±o-o5 2-25 0-75"- 1-5 1-5 1-5-7 ±o-37 1-64 ±o-18 ± 0-23 l-6o ±O-13 1-o7 ±o-o? 2-0 o - J 1-ZS f-6 ± 0-2S f. oS to- IJ 1-0 ±0-11 1-11 to-24 /•o? ± 0 - /o 0-85 ± o- 08 Mr/2 MHz Mr/? . MHz MHz -~~ TABLE 4-I RESUL" TEST PICTURE (a) REZoLUTiorJ CHART (b) FAcE Cc ) cycLis.T i d ) "MOTION" 36 th i s i s tabulated i n the l a s t column of Table 4-1 for the various com-pression r a t i o s . I t appears that compression r a t i o s above 1,5 are useless, since I drops below unity at C = 1.6. For smaller compression r a t i o s , there i s some improvement through the use of compression, but i t i s a very .modest one. Thus at a compression r a t i o of 1.43, I = 1.16, i n d i c a t i n g a 16% improvement i n the e f f i c i e n t use of bandwidth, when compared with simple f i l t e r i n g . This r e s u l t was obtained with tests using one p a r t i c u l a r combin-ation of high-quality and low-quality bandwidths, and further tests would be needed to v e r i f y that other bandwidths leading to the same compression r a t i o w i l l have a s i m i l a r improvement f a c t o r . I t may also be worthwhile i n v e s t i g a t i n g further methods of overcoming the f l i c k e r and other d i s t o r t i o n e f f e c t s which have greatly l i m i t e d the scope of the compression scheme. At present i t appears that the compression scheme used by i t s e l f o f f e r s only very l i m i t e d savings i n transmission bandwidth, - perhaps i t should be used as a supplement to other s t a t i s t i c a l encoding schemes ( D , ( 2 ) . 37 REFERENCES 1. "Redundancy Reduction", Proc. IEEE (special issue), vol. 55, Mar., 1967. 2. A.H. Frei et a l . , "An Adaptive Dual-mode Coder/Decoder for Television Signals", IEEE Trans, on Communication Tech., vol. Com-19, no. 6, pp. 933-944, Dec., 1971. 3. R.C. Brainard et a l . , "Low-Resolution TV: Subjetive Effect of Frame Repetition and Picture Replenishment", B.S.T.J., vol. XLVI, no. 1, pp. 261-272, Jan., 1967. 4. 0. Meier, "Television Picture Transmission and Optical Signal Pro-cessing", M.A.Sc. Thesis, Dept. of E l e c t r i c a l Engineering, U.B.C., July, 1968. 5. M.P. Beddoes and 0. Meier, "Flicker Effect and Television Compression", IEEE Trans, on Info. Theory, vol. IT-16, no. 2, pp. 214-218, Mar., 1970. 6. T.K. Chu, "Optical Signal Processing and Television Bandwidth Com-pression", M.A.Sc. Thesis, Dept. of El e c t r i c a l Engineering, U.B.C. Sept., 1970. 7. A.H. Robinson and C. Cherry, "Results of a Prototype TV Bandwidth Com-pression Scheme" , Proc. IEEE (special issue), vol ..5.5., p.p. 356-364, Mar. . 1967. 8. J.P. Farr, "Some Television Bandwidth-Compression System Using Edge Coding", M.A.Sc. Thesis, Dept. of El e c t r i c a l Engineering, U.B.C, April , 1966. 9. Y. Peless and T. Murakami, "Analysis and Synthesis of Transitional Butterworth-Thomson Filters and Bandpass Amplifiers", R.C.A. Review, 18, pp. 60, 1957. 10. Digital Equipment Corporation, "PDP-12 User Handbook", Maryland, Massachusetts. 11. K.A. Brownlee et a l . , "The Up-ancVDown Method with Small Samples", J. Amer. Statis. Ass., 48, pp. 262-277, 1953. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items