Digital Video on PC’s What advantage does having video on PCs in digital form have over the various other formats?

Digital Video - Advantages Can be used with very little extra hardware –just larger hard disks –video from tape can be shown on the PC monitor but quality interface cards to do this are costly - plus cost of video equipment Almost random access –videotape is sequential slow to edit impractical for interactive applications

Digital Video – Advantages cont Quality Editing and Duplication –copies are exact - tape copies are imperfect –frame by frame editing –only degradation of edited material is due to lossy compression Storage Flexibility –long life – Hard disk 30,000 hrs - professional tape 3 hrs Ease of use

Digital Video Disadvantages Requires lots of storage space PC performance is constrained –getting the images from disk to screen quickly enough to avoid jerky motion Large image size only practical on powerful machines –240  180 not uncommon –320  240 practical –640  480 with hardware support Achievable frame rates go down as image size goes up

Common Digital Video Terminology Full frame: 640 x 480 Half frame: 320 x 240 –MPEG 1 –is actually one-quarter screen Quarter frame: 160 x 120 –Streaming/webcam “postage stamp” – is actually one-sixteenth screen Full motion: –30 fps (NTSC) –25 fps (PAL) Full colour: 24-bits - 16 million colours

Standard Digital Video Resolution Terminology MPEG terms –SIF Standard Image Format H.261 (ITU-R) video conferencing terms –CIF Common Intermediate Format –QCIF Quarter Common Intermediate Format

Square pixel formats Square pixels PAL: –full-screen x 576 –SIF – 352 x 288 –CIF x 288 – quarter screen –QCIF x 144 –sub QCIF x 96 Square pixels NTSC: – full-screen x 480 –CIF x 240– quarter screen –QCIF x 120 –sub QCIF x 80

CCIR pixel formats CCIR 601 pixels PAL: –full-screen x 576 –SIF – 360 x 288 –CIF x 288– quarter screen –QCIF x 144 –sub QCIF x 96 CCIR 601 pixels NTSC: –full-screen x 480 –SIF –360 x 240 –CIF x 240– quarter screen –QCIF x 120 –sub QCIF x 80

MPEG MPEG-1 for computer screens – 320 x 240 MPEG-2 also allows for larger formats: 1440 x 1152; 1920 x 1080 (16:9)

Analogue Capture UK TV images and PC screen sizes do not match –PAL images are (768)  588; NTSC (720)  480 composite signals –professional sampling rates standards give: PAL (864)  588; NTSC (858)  490 component signals but several of the lines sampled and parts of each line do not form part of the picture

TV cropping pixels can be cut off the image edges (top, bottom, left, right). For many applications, these areas are not visible and removing them saves on the amount of data captured with no visible degradation in quality e.g., Square pixels PAL: –full-screen x 540 –half-screen – 360 x 540 –CIF x 270 – quarter screen still different from screen resolution

TV frame rates –NTSC 30 fps, PAL/SECAM 25 fps are different from: –arbitrary capture rates 30 fps OK for capture of US material; not so for UK –pc screen refresh rates; 72Hz+ Non- Interlaced typically –not usually related in any way to video frame rates

Digital Video (Camera) Capture Light hits CCDs (charge coupled devices) and generates electric currents Electronics encodes the electrical signal into three parts: –Luminance –Red difference –Blue difference The signals are sampled and quantised

Digital Video (Camera) Capture 2 The digital information is then compressed in three stages: –DCT (Discrete Cosine Transform) using 8 x 8 pixel blocks –AQ (Adaptive Quantisation) reduces complexity of DCT output (if necessary) to improve VLE effectiveness - lossy –VLE (Variable Length Encoding)

Digital Video

Video – Television Standards What are the names of the TV standards in: USA, Japan, UK, France, Australia, and Russia? How do these standards differ? Can one TV receive all the standards? See vision_systems vision_systems vision_systems

NTSC – National Television Standards Committee USA, Canada, Central America, Japan, etc. 525 line 60Hz interlaced – 60 fields/s, 30 frames/sec –interlaced 525/59.94/2:1 uses YIQ model ( a conversion of RGB colour) NTSC 4.43 (NTSC video signal, PAL colour carrier) 525/29.97/2:1

PAL – Phase Alternate Line UK, Western Europe (except France), Australia, South Africa, India, China, etc. 625 line 50Hz interlaced – 50 fields/s, 25 frames/s 625/50/2:1 uses YUV model PAL M (e.g. Brazil) 525/29.97/2:1 PAL N (e.g. Argentina) 625/50/2:1 Note: TV picture does not use all lines

SECAM – Système Electronique pour Coleur avec Mémoire France, Eastern Europe, Russia, parts of Africa and the Middle East 625 line 50Hz interlaced – 50 fields/s, 25 frames/s

TV images sizes in pixels In reality these are samples of an analogue signal PAL images are (768)  576; NTSC (720)  480 actually: PAL (864)  625; NTSC (858)  525 –but several of the lines sampled and parts of each line do not form part of the picture

Aspect Ratio Aspect Ratio – horizontal to vertical, nominally, 4:3 on TV defines the ratios 768  576 & 720  480 so video capture of analogue signals is normally done at (maximum) these figures: 768  576 & 720  :9 on HDTV (1920  1080)and on widescreen

Note: Film runs at 24fps when played on 50Hz TV (telecine) it runs at 25fps i.e. 50 fields/s Movies run a little faster on TV! when played on 60Hz TV in order to get 60 fields/s the second field of every other frame is broadcast twice ‘3-2 pulldown’ –i.e. the first frame of the film is recorded for three fields, the second for two, the third for three, and so on...

Video Tape Formats What are the names of the main video tape formats? How do they differ? If you have a video tape from the US in your video player’s tape format can you be sure it will play OK?

Analogue Tape Formats VHS (Video Home System ) - ½” tape and Video8 - 8mm tape –suitable for normal TV recording –1 helical track for video (composite) 1 straight track for sound; (C chroma) –resolve 240 and 230 lines

Betamax - ½” tape suitable for normal TV recording composite but superior to VHS almost obsolete

Super VHS (S-VHS) - ½” tape & High-Band 8mm (Hi-8) superior to VHS Y/C component using S-Video connections to keep the signals separate suitable for in-house semi-pro work resolve 400 lines

S-Video is not a tape format: Separate Video Betacam SP (Sony) - ½” tape –industry standard pro. format - Component YUV –Separate channels for Y, U and V –20 minutes recording on two-hour Betamax tape –three video signal tracks and four audio tracks –very high image quality –resolve 550 lines

CCIR 601 Most digital formats conform to ITU-R Recommendation BT 601 (commonly known as CCIR 601) See CCIR 601 is a family of formats - each written l:m:n – that define sampling for digital video as YUV component video Defines sampling rates relative to a basic rate of 3.375MHz

CCIR cont l is the no. times base rate that luminance is sampled m and n. represent the relative manner in which colour differences are sampled horizontally and vertically – chrominance sub-sampling

e.g., 4:2:2 means luminance is sampled at 13.5MHz; both colour differences are sample for every other pixel –4:2:2 means that 720 luminance samples and two lots of 360 colour difference samples per line –4:2:0 means that 720 luminance samples per line and two lots of 360 colour difference samples every other line –4:1:1 means that 720 luminance samples and two lots of 180 colour difference samples per line the no. of lines sampled is 576 for PAL and 480 for NTSC

Are Pixels Round or Square? The assumption (as a matter of convenience) that TV screens are 4:3 implies that for square pixels 4/3 times as many values must be recorded for each line as there are lines making up the picture To sample an analogue video frame so as to obtain square pixels requires that the sampling rate is different for PAL and NTSC

Round and square pixels –sampling all 288 PAL lines in each field takes 1/50s as opposed to 1/60s for the 240 NTSC lines, so each line takes exactly the same time to sample –i.e., a PAL line’s samples, 768, are made in the same time as an NTSC line’s samples, 640, so the PAL sampling rate is higher. –this produces video suitable for playback on a computer monitor using a 4:3 resolution, i.e., 640 x 480, 1024 x768 but not 1280 x 1024

Capturing analogue video vs DV Capture of analogue video is normally done using square pixels Recording using a Digital video camera is normally done using CCIR 601 pixels Some video capture cards will allow you to capture analogue video with nominally CCIR 601 pixels See

CCIR non square pixels CCIR 601 pixels are not square. They are the result of sampling both PAL and NTSC lines at the same rate taking 720 samples for each –i.e. a PAL line’s samples are a smaller proportion of the nominal resolution of PAL analogue video (768  576) than an NTSC line’s (720  480) sample are –thus CCIR 601 pixels for PAL are slightly wider and for NTSC are slightly narrower than square

NTSC/PAL v CCIR –Ratio is NTSC 10:11 and PAL is 59:54 –this produces video suitable for playback on the appropriate TV system, it appears narrower (PAL) or wider (NTSC) by about 10% when played back on a computer monitor using a 4:3 resolution

Digital Tape Formats D-1 (Sony, BTS) - ¾” (19mm) tape –720  480 – 168Mb/s –component video; YUV –CCIR 601; 4:2:2 –8-bits per component –broadcast Standard D-2 (Ampex, Sony) - ¾” tape –Y/C digital - 8-bit D-3 (Panasonic) - ½” (12.7mm) tape –Y/C digital - 8-bit

Digital Tape Formats 2 D-5 (Panasonic) - ½” tape –YUV - derived from D bits per component ; 2:1 compressed –CCIR 601 4:2:2 –broadcast Standard DCT (Digital Component Technology - Ampex) - ¾”tape –YUV - derived from D bits per component; 2:1 compression –CCIR 601 4:2:2 –broadcast Standard

Digital Tape Formats 3 DVCPRO (Panasonic) –DV with wider faster-running tape holding additional analogue recording and control track DVCAM (Sony) –DV with high faster-running quality tape Digital Betacam (Sony, BTS) - ½” tape –YUV - 10-bits per component; 2:1 compression –CCIR 601 4:2:2 –broadcast Standard

Digital Tape Formats 4 Digital8 (Sony) –DV using Video8 or Hi8 tapes DVCPRO50 (Panasonic) –DV with high quality tape –4:2:2 –uses two DV codecs in parallel –50 Mbps (video) compressed about 3.3:1 –visually lossless and high quality D-9 formerly Digital-S (JVC) –Variant of DV –4:2:2 –uses two DV codecs in parallel –50 Mbps (video) compressed about 3.3:1 –visually lossless and high quality

Digital Tape Formats 5 DVC (Digital Video Cassette) or DV ¼” (6.35mm) tape –nominal 25 Mbit/s data stream (now also 50 and 100Mbps) –can record up to three hours of video in SP (standard play) mode on cassettes which measure 125x78x14.6mm –using MiniDV format cassettes - 1/12th the size of a standard VHS tape can record an hour in standard format or up to 90 minutes of lower quality output in LP

Other Video Formats DVC - DV cont –PAL DV - CCIR 601 4:2:0 - 8-bits per component; 5:1 compressed –DV compression is directly done by the DV camcorder using the DV codec

VideoCD –A competitor to VHS tapes – successful in the Asia –Format defined 1993 – quite complex –Use MPEG-1 compressed video –Can accommodate multiple sound tracks

Other video formats –Usually utilise proprietary formats Sony MicroMV format Half the size of a MiniDV cassette Uses MPEG 2 recording so lower quality than DV Recently abandoned by Sony –Some tapeless camcorders use flash memory usually using MPEG4 Compression

Other formats Apple QuickTime – Movies –interleaved format – (.MOV) Frames of video and sound are interspersed Microsoft's Video for Windows –Audio Visual Interleave – (.AVI) –Being superseded by Active Streaming Format – (.ASF) and windows media video (.WMV)