Presentation on theme: "www.smartphonetech.org Contents l Forms of Motion Picture l Multimedia products with Motion Picture Content l Motion Video Technology l Source for Motion."— Presentation transcript:
Contents l Forms of Motion Picture l Multimedia products with Motion Picture Content l Motion Video Technology l Source for Motion Video Objects l Playing Video on PC l Video Formats : lines, frames, fields & aspect Ratio l Analog Video Signal Waveform l T.V. Broadcast Standards l Color of Video l Analog Video Camera Principle
Contents (cont.) l Luma & Chroma l Hue & Saturation l Chroma Component Options l Video Signals l Composite Video Signal Wavwform :Color l Digitisation Basics l Effect of Low Pass Filters l Magnitude Domain Discretisation l Color Bitmap Encoding & CLUT l Digitised Video Standards
Contents (cont.) l Video Compression l Motion Compensation l Artifacts due to digitisation & compression l Video Compression Techniques & Standards l Data Rates of Media l Analog Video Recording : Tape Formats l Digital Video Camera l Digital Video Recording : Tape Formats l Video File Formats l Steps to PC Video
Contents (cont.) l Video Capture : Basics & Platform l Video Capture Cards l Video Capture Considerations & Procedures l Video editing : Primer & concepts l Transitions & Superimpositions l Non-linear editing l Video Editing Tools l Shooting Tips
Motion Video Technology for Multimedia Learning Objectives n Appreciation of Technical Aspects of lFull Motion Video lDigital Video Storage and Compression lVideo Editing lMultimedia Video Standards n Obtaining necessary background for acquiring and processing Full Motion Video Objects for embedding in Multimedia Contents.
Many Forms of Motion Picture n Films n35 mm, 70mm, 16mm,8mm n Broadcast TV l Ordinary l DBS TV l HDTV ( coming up) n Laser Disk (Players) n DVD (Players) n Video in Cassettes and Tapes n Video in computer l Multimedia component l Stand alone CD Video l Web Content
Why use Motion Video in Multimedia? n Sin qua non in True Multimedia n Most effective communication mode n Entertains and educates with least audience initiative l Brings realism l Multiple views (vista and zoom) l Being there - presence l People are already sensitised to movies l Shows changes - as and when they happen
The downside of Video in Multimedia n Takes two much disk space n Needs large bandwidth when transmitted n Costly to produce n Quality is difficult to achieve and highly dependent on target machine n May easily become the weakest link in the product
Content is the King n Every king needs subjects - so does video n what to show in video lDrama lDances linterviews lLarge artwork or mural larchitecture lRecord of events (village fair) and social functions (marriage) lHow to do things wOperation wInstallation wShow How is know-how lInternals of a machine n You get the idea
Multimedia Products with Motion Picture content n As embedded object. lWithin an interactive multimedia product (say, web page, encyclopedia entry). lShort length snippets (one or many) lSmall format (1/4th- 1/16th). n As the primary content lWithin an interactive multimedia presentation. lGood quality, with voice and music lAt least several minutes long. lRest of contents organised around it. lMust be at least 1/4th screen n Stand-alone object lLibrary/depository object lGood quality n Stand-alone digital video product lGood quality, Longer duration, Full screen preferred
Before you begin To incorporate video in multimedia, ask the following questions: n Could the video be substituted by still ? l If not, why, what is the USP? n Could Animation be better than a video ? n What would be the role of the video in the product? ( e.g. embedded object, album ) n Who would provide the footage ? n What would be the duration of the video clip(s)? n Delivery medium (CD-ROM, DVD,Internet, LAN) n Who would be editing and integrating it (skill level) and where (facilities)? n Does the producer know about the nuances of video technology?
Studying Motion Video Technology n All skills are to be converged n Many techniques of Multi Media borrowed from full motion video production methodology. n Possibility of auxiliary services like Desktop Video Editing. n Close kin of Animation. n Most complex and jargon ridden
Principle of Motion Picture: Frame Rates n Successive images (frames) of a moving object are captured in a movie camera at a fixed rate (frame rate) n The frames, when played (projected) at the same frame rate gives an illusion of moving object. n Frame rate above 20/sec is adequate. n For slow moving objects even 15 frame/sec would do. Below this rate, jerks are perceived. n Not only object motion but camera -optical motion (pan, zoom ) is to be considered. n Cine uses 24 frames per second. n Higher frame rates used for special effects.
Source for Motion Video Objects n Computer Files l Only form usable in Multimedia. l Many formats. l Usually compressed. n Analog Video Camera l Needs frame grabber for direct storage as a computer file. l Equipment technology mature and low cost. n Analog Video Cassettes/Tapes l Camera to tapes/cassettes. l Different qualities and standards. l Generation loss of quality. l Needs frame grabber again.
Source for Motion Video Objects (contd.) n Digital Video Camera l May be stored in tapes/cassettes or directly stored as computer files. l Technology still evolving. l Equipment Costly. n Digital Video Cassettes/Tapes/ disks l Surprisingly predates digital video cameras. l Old Digital Video Cassettes/Tapes/ disks are not readable by computers. l New Computer readable (compressed) formats are now available. l No generation loss. l Best suited for editing.
Playing Video on PC n Pentium and Power PC based personal computers are capable of playing full motion video of different qualities. n From l Hard Disk (AV, SCSI-II F&W preferable) l CD ROM (32x) l Intranet/Internet l DVD l TV antenna or Cable TV (need special cards) l VCP (you really do not need a PC) n Quality varies: l Full speed, full screen, full colour l Small quarter size (Fixed, variable and movable) windows l Smaller frame rates l Video Overlay
Playing Video on PC ( Contd.. ) n Configuration: l Pentium-II, MHz, 32MB Ram l SVGA monitor and adapter with required resolution and colour depth l Sound Card and speakers l Software l Graphic accelerator ( desirable for full speed full screen) l CD-ROM n Configuration: additinally may require l DVD (if source is DVD) l Hardware assist (For playing MPEG-II) l TV tuner Card
Video Formats: Lines, Frames and Fields n Brodcast and VCR video are generally 25/30 frames /sec n Each frame is composed of a large number of horizontal lines. n Each frame is divided into two fields. l The even field contains only even lines of a frame. l The odd field contains only odd lines of a frame. l The two filelds are displayed one after the other : thus covering the whole frame in two installments. l Each field takes half the time of a frame. n This is known as interlacing.
Video Formats: Lines, Frames and Fields
Video Formats: Lines, Frames and Fields (contd) n Interlacing l Reduces Broadcast bandwidth (by transmitting only half the lines per field) l Reduces flicker as fields are updated 50/60 fields /sec. n To synchronise the frames and fields, the source supplies vertical synchronisation signal, once per field. n Loss of vertical sync results in jumpy vertical scroll. n To synchronise the lines in a field, the source supplies horizontal synchronisation signal, once per line. n Loss of horizontal sync results in loss of picture. n Computer Video is progressive i.e non-interlaced.
Video Formats: Aspect Ratio n Films l 35 mm 1.85:1(16:9) to 2.35:1 (~19:8) l 70mm 2.35 : 1 l 16mm/8mm 1.85:1 n Standard TV Monitors: 5:3 n HDTV - 16:9 n Computer Monitors : 5:3 (1.67:1) n Video in computer (multimedia component) height -h width-w AR=w:h
Analog Video Signal Waveform: Monochrome n TBD
Television Broadcast Standards
Color Components of Video n The information of a point in a colour picture may be described in terms of intensities of three colour signals RED, GREEN and BLUE. n Almost all perceivable colours can be described by varying the proportions of the three components. n If we keep the proportions same but change the absolute values of the RGB signals, the colour will remain the same but brightness will change. n However, If we keep the the absolute values of two signals same but change the proportions by changing the third signal, both the colour and brightness will change.
Analog Video Camera Principle n Two basic types of video cameras l Tube (e.g videocon, orthicon) l CCD (semiconductor) n In a monochrome video camera the external image is focussed on the active surface of the tube or the CCD array. n An electron beam scans the active surface horizontally and vertically just like in a TV monitor. n The scanning action generates the video signal proportional to the luminosity of the points scanned.
Analog Video Camera Principle
Analog Video Camera Principle (monochrome) n The scans are synchronised with V and H sync signals generated from a stable source. n The stable source may be the camera or another camera or a studio master sync generator (gen- locking). n The above video signal is composited with the V and H sync signals to create the composite video signal.
Analog Video Camera Principle (colour) n In a colour camera l The scanning mechanism is identical to that of the monochrome. l An image is optically seperated into R G and B components using filters and prisms. l The three components are focussed into three different active surfaces which generate corresponding (R,G,B) intensity signals. n In a single CCD colour camera l There is only one active surface but containing three different kinds of active elements, one each for R,G and B. l Each type of active element responds only to its corresponding colour lights.
Analog Video Camera Principle(colour) (contd.) n Depending on the output type desired l The R,G,B signals and the sync signals are separately ouput l The signals are composited and output.
Luma and Chroma n The Luminance or brightness of a small region in a colour picture is given by:- l Y=0.299R+.0587G+0.114B n In black and white TV, only the Luma component is required. n When the Luminance is known, we need two more quantities (say proportions) do describe the colour. These signals are known as chrominance components of the signal. n There are several ways to describe the Luma. n Eye is most sensitive to spatial and temporal variation of the Luma component.
Hue and Saturation n If we (optically) add white colour to another pure colour say blue, what do we get ? Does the colour change (say to pink or yellow)? n Something (tint) remains constant and something (say shade) changes. n What remains constant is the tint or hue. n What changes is technically called saturation. n If a pure colour does not contain any white component it is called fully saturated. n An image with saturated components looks bright. n Unsaturated components create pastel shades.
CHROMA component options n Naive: Y and any two ( not used) n CCIR 601 l U= (B-Y): -yellow, V= (R-Y) : - cyan n PAL (actual) l U= 0.492(B-Y), V=0.877(R-Y) l Also known as Y-U-V n NTSC (Y-I-Q) l I=0.596R-0.275G-0.321B (orange to blue) l Q=.212R-.523G+.311B (purple to green) l Eye is least sensitive to Q n JPEG/MPEG l C b =(B-Y)/2+0.5 l C r = (R-Y)/2+0.5 l Range is always one.
Video Signals n Component Video l RGB and +Hsync (every line) +Vsync (every line) : 3+2 or 3+1 (combined sync) l R,G+ combined sync,B n TV signal and composite VHS l Picture info +Hsync (every line) +Vsync (every line) +chroma burst (every line) l Picture info= Luma +chroma l Luma + chroma on colour carrier as a combined signal n SVHS (camera and deck) l Luma and chroma on colour carrier as two separate signals
Composite Video Signal Waveform: colour
Digitisation Basics n Lines and frames are first steps of discretisation - already available, even in analog video. n Two dimensions for video data within a line : l Time domain sampling, l Magnitude domain discretisation. n Both these degrade picture quality to some extent. n Without sacrificing quality l How often to sample ? l How many bits to discretise the data?.
Spatial Resolution and Bandwidth n Consider the figure below which shows an alternate patterns of black and white and the corresponding video signal. n The number of pulses would roughly corresponding to the number of dark bands N. n If the horizontal display line time is T h, then the frequency is f= N/ T h n The frequency content (bandwidth) of the video signal is therefore directly proportional to the horizontal details (resolution) and vice versa. n Thus the bandwidth can be expressed as dots per line (spatially) or alternatively in MHz ( time equivalent) for a given horizontal clock.
Spatial Resolution and Bandwidth (contd) If processing, transmission or display bandwidth of the video signal is much below the video signal f generated due to horizontal details, nthe video signal amplitude would decrease nand hence the contrast would decrease nand hence details ( bands ) cannot be distinguished ni.e resolution would be lost. n Similar effects may be seen after employing low pass spatial filter in still images. n Due to the limitation of transmission (and display) system, the bandwidth of video signal is (artificially) limited.
Effect of Low-pass filters (spatial or temporal)
Resolution and Bandwidth (contd) n Limiting bandwidth of video signal limits resolution of broadcast / VCR generated images but l allows broadcast, reception and processing using inexpensive equipment l allows slower sampling rate. n The vertical resolution is limited by the number of lines. n However, the actual resolution perceived by human eye is a bit different than what the bandwidth calculation would predict. n Due to interlacing, a single pixel width line is displayed only in alternate frames and would tend to blink. n In SECAM a thin line may show two differnt colours (as Cr and Cb are sent alternately) n A projected film has much higher resolution in both vertical and horizontal directions.
Sampling and Nyquist n A theorem due to Shannon, attempted to answer the question “How often to sample ?” Like many other question the answer is “it depends”. The best sampling rate depends on how quickly the signal itself is changing. n The agility of a signal or the quickness of its change is measured by its Bandwidth. For example in broadcast, the TV picture signal bandwidth is less than 6 MHz (6 million times/second). n The sampling rate should be at least twice the bandwidth or a part of the information would be lost and signal may not be reconstructed from its sampled version. This rate is called the Nyquist Frequency. n Nyquist frequency for TV picture signal is 12 MHz.
Sampling and Nyquist (contd.)
Sampling RGB or Composite? n Sampling of composite signals are to be avoided as : l Resultant digital signal is difficult to manipulate. l The colour subcarrier is also sampled. l Artifacts (defects) are introduced. n Sampling of S-Video signal i.e. Luma and combined chroma is slightly better. n RGB is simplest and easiest to manipulate - but wastes bandwidth. n YUV (Luma and separate chroma) are the best as optimal sampling rates and compression are possible.
Luma and Chroma Sampling n The Luma signal changes more rapidly and eye is most sensitive to it. It has a bandwidth of about 4 MHz in broadcast TV. n Luma signal needs to be known for every pixel. n Chroma signal changes more slowly and can be sampled less frequently. n Good quality may be obtained by sampling the colour components at half the frequency. (4:2:2). n Reasonable colour rendering is obtained by 4:1:1 sampling. n Colour signals may also be sampled in alternate lines (4:2:0)
Luma and Chroma Sampling(illustration)
Magnitude Domain Discretisation n For Digitisation the sampled magnitudes or values need be converted into binary integers. n For a black and white image, we would require only 1 bit (0 or 1) per pixel. n For grey scale image, the number of bits required would depend on the number of grey shades desired. For example, l with 6-bits, 64 different shades may be distinguished. l with 8-bits the shade count would be 64
Magnitude Domain Discretisation (contd.) n The total number of bits employed in RGB component video would determine the varieties of shades. n For RGB component video, the number of bits employed are usually identical for each channel. n For Chroma-Luma component video,the Luma channel may use more bits than the Chroma n Employing more number of bits per pixel would l increase the fidelity of the picture, but l increase file size and play banndwidth requirement. n Some lossy compression techniques may reduce the bit depth
Colour Bitmap Encoding n Colour of a point (pixel ) may be described by the quantities of the R-G-B components.For computer,the quantities would be described by digital numbers- say integers. n So we need three integers for every point to describe the image. n For the image, the sets of numbers for all points is called the Colour Bit Map. n For true colour representation, each such numbers should be of 8-bit size. So we need 24 bits, giving Millions of shades. n We may not need such huge varieties of shades for every single picture.
Bitmap Encoding and CLUT (contd.) n If our requirement for shades is modest, say 200 shades only, we may reduce the number of bits required. Example: night scene, sky, buildings... n This is done by using a palette or a colour look-up table (CLUT). n A colour look-up table contains entries of RGB components against shade numbers. (we may throw- in a colour sample strip along, just like the shade cards available in paint shops) n As there are less than 256 numbers, an 8-bit integer would be adequate. n We may then represent each point with just one 8-bit number. n Of course we knew this while studying still image and graphics. n What we did not know is that for a video, the 256 shade CLUT limitation may be applicable for the whole clip.
Rationale for Compression n Example : 640x240 (full screen), 24 bit full colour, 30fps l 221 M bits/s (~ Indian Internet BW) l 28 M Bytes/s l 15 s file takes 420 MByte space (about 1 CD) n Storage n Streaming rates n Distribution n Transmission
Data rates/BW of Medium
Compression Basics n Redundancies Exist l Spatial (subsampling, subband) w neighbouring pixels tend to have very similar brightness and colour, except for the edges. l Temporal (predictive) w Successive frames have many pixels identical to those of previous. w Intraframe Vs. Interframe l Spectral (subband) w Not all shades of colour appear in a clip n lossy l Compression may destroy information n Also called source encoding !
Lossless Encoding & Compression n No loss of information n Compression depends on content. n Low compression n fast encoding and decoding n Run length encoding (RLE) l Simplest form of compression l Usually intraframe, also used in fax, zip etc. l Example: l aaaaabbbccccd=5a3b4cd n Huffman Encoding l Determine the most frequently occurring pattern, give it a short code and so on. l Dictionary is required. n Transform Encoding (some)
Lossy compression n Enforces ‘almost similar ‘ things to be similar and then compress. n Large compression ratios may be obtained. n Quality may suffer n Trade-off may be done between compression ratio and quality. n When applied spatially, reduces levels of l contrast l Hue n Almost similar successive frames are treated as similar n May be carried out without notice (Backdoor Compression) l When you reduce size l When you reduce colour depth l When you reduce frame rates
Transform Encoding: The basic idea n Similarity is not always obvious. n By transforming the original data, the similarities are revealed. n In the above picture, the horizontal lines are not similar. n If we rotate the thick line (i.e. mathematically transform it) to be horizontal, similarity is revealed.
Transform Encoding: The basic idea (continued) n A picture may be transformed into components which may be easily and compactly expressed. n A picture may produce very large number of components. n One may discard the insignificant components. n Discarding component implies loss of information- but this may not be perceptible. n By varying the number of components retained, one may trade off between quality and compactness. n The trick is to find the right transformation which would produce a small number of significant components.
Transform Encoding(contd.) n Available Transforms l Direct Cosine Transform (DCT) l Fractal transform n Fractal Transform l Highly asymmetrical l Needs huge time during compression l proprietary n Discrete Cosine Transform l Popular (MPEG, JPEG) l Well understood l Good trade off l 1:200 compression feasible
1-D Cosine Transform Example n The following waveform represents the luminosity distribution along a 180 pixel line. n A cosine transform of this would need ~ 180 components
1-D Cosine Transform Example (contd): total Vs 30
1-D Cosine Transform Example (contd): 40 Vs component Approximation
1-D Cosine Transform Example (contd): total Vs 20
1-D Cosine Transform Example (contd): 30 Vs 20
1-D Cosine Transform Example (conclusion) n If we retain just 50 components there is almost no loss of information, except some noise. n If we retain just 40 components, we get a reasonable replica n We may try going down to 30 or even 20 components. One may observe the rapid deterioration of details. n We see the trade-off between compactness and loss of detail. n The above concept may be extended to 2-D as well.
8x8 Discrete Cosine Transform n Popular 2-D transform with 64 coefficients
Interframe Compression n Temporal redundancy is eliminated by interframe compression. n Consider a title page, which have to be shown for 3 seconds (i.e frames) l Each such frame is absolutely identical l Sending one frame and N-1 ditto’s are good enough and non-lossy. n In general, successive frames may be very similer. n Compression could be obtained by sending a reference frame and then sending only the changes or differneces. l In order that the differences do not become too big, reference frames may be sent periodically
Motion Compensation n A special case of interframe compression n Applicable with camera pan or moving subject n A part of the picture remain same but move as a group (translate) n By supplying only the motion vector (direction and amount) the pixels concerned may be defined. n Macroblocks (16x16) between successive frames are compared for close matches. n Differences are computed after considering the motion vector.
Why filtering is required n With spatial filters we may limit the details and levels of contrast n Limiting the details reduces bandwith and sampling at lower frequency is possible without violating Nyquist limits. n Bandwidth and contrast limited pictures compress well. n Lossy compression will reduce bandwidth and contrast in any case n Artifacts may be introduced if filtering is not performed.
Artifacts due to Digitisation and Compression n Pixelisation n Jagged edges n Moiré fringe n Contouring n Colour seepage and smear n Palette Flash n Jerky movement n Salt and pepper
Video Compression Techniques and Standards n JPEG & M-JPEG (intra frame) n H.261 (intra frame) n MPEG-1 n MPEG-2 n MPEG-4
JPEG and M-JPEG n JPEG (Joint Photography Expert Group) l Recommended standard for digitisation and compression of still images l May be used for frame by frame compression of MP. n JPEG Modes: l lossless l sequential l progressive l hierarchical n JPEG lossless mode : pixel-wise predictor encoding using neighbours n Sequential Mode l each image component encoded in a single left to right, top to bottom fashion.
JPEG and M-JPEG (continued) n M-JPEG l JPEG extended for motion picture and animation, still intraframe l Sequential JPEG applied to each frame n JPEG progressive : l image quality improves as more data arrives l Suitable for web images n JPEG hierarchical l Code and compress a down-sampled image l Upsample it and take the difference with original, l send the down sampled and difference l may be repaeated multiple times
MPEG n MPEG (Motion Picture Expert Group) l Recommended encoding and compression standard. l Accepted by ITU and ISO n Three different standards l MPEG-1 (for PC Video, near VHS) l MPEG-2 (for HDTV and Broadcast studio) l MPEG-4 (primarily for video telephony) n MPEG-1 and MPEG-2 are already available and popular
MPEG standards Overview n Three layers or streams l System l Audio l Video n System = Info regarding l synchronisation l random access, l management of buffers l time stamp n State of the art compression n Flexibility for optimisation n Software Decoders available l Quicktime l Xing l Mediamatics with windows Media player
MPEG-1 n 1991 Vintage n For CD-ROM video and Audio n Audio is near studio quality n Video is near VHS n Targeted frame rates ~ 1.5 Mbits/s n lossy & asymmetric compression l Intra frame spatial redundancy crunch using DCT (6:1) l Interframe,temporal redundancy elimination, using P (predictive) difference l A number of RLE and compaction steps n Software coding and decoding possible.
MPEG-2 n For higher data rates (4 to 9 Mbits/s) n For better picture quality and HDTV (US and Europe) n Would play MPEG-1 objects n Used in DVD video and DBS (hughes) n Interframe,temporal redundancy elimination, using l P (predictive) and B (bi-directional) frames l Seq: I-P-P-P-B-I-P-P-B-I n 4:2:2 Chroma subsampling
MPEG-2 (continued) n 10-bit DCT DC precision n non linear qualtisation n Other compression steps n Optional time stamp n error and transport control n Hardware codec recommended n Don’t expect MPEG-2 to be better than MPEG- 1 at lower bit rate.
MPEG-2 Standards and levels
MPEG-4 n Expected to be formalized in Nov 1998 n Targeted bit rate 4.8 to px 64 Kbps (p <80) n Suitable for l Video phones l Video conference l Internet video l Mobile video l Computer generated animation n Object based encoding
Data Rates of Media
Analog Video Recording:Tape Formats n VHS (1/2”: Consumer, composite) n 8 mm ( Consumer camcorder, composite signal) n S-VHS ( improved VHS, S-Video, midrange) n Hi-8 ( improved 8 mm, S-Video/RGB,midrange) n Betacam 1/2”(high end, studio, RGB and other) n Betacam-sp (high end, studio, RGB and other) n U-matic-3/4’’ - now obsolete n Tapes (1’’) : Broadcast and studio l Type -B l Type-C
Digital Video Cameras
Media For Storage and Distribution n Considerations nCapacity nStreaming rates nerasability ntransportability n Types nHard Disk nZip disks (between stations or for back up) nTape (between stations or for back up) nCD-ROM l Writable for small volume l Mass produced nDVD
Digital Video Recording:Tape Formats
Media For Storage:DVD n DVD = Digital Video/Versatile Disks n The Latest Storage Medium, capable of storing l multiple full length feature films. l Songs (multiple albums) l Data (4 GB- 16 GB) n Standards still evolving n Different Capacities l Single/double layer l Single/double side(may require manual flip) l Typical capacity of single side-single layer 4.7 GB
DVD(contd.) n Different Versions n DVD-Video n DVD -Audio n DVD-ROM n DVD-R (write once-by user) n DVD-RAM (rewritable) n DVD -Video and DVD -Audio may be played in dedicated players. n Computer may play all. n DVD-R : Best suited for small volume multimedia content distribution. n DVD-RAM : Useful for object capture and edit n Data rates 2-5 Mbytes /s (video) l Supports MPEG-II
Video File Formats n Video For Windows AVI (Audio- Video interleaved) n QUICKTIME l A very popular PC-Video format from Apple for Mac,PC, Internet l Mature and open standard. l Play backs most forms of files (CODECs). l Suitable for virtually all forms of multimedia objects like video, sound animation stills, MIDI,VR. l Scaleable - quality adjusted automatically to suit platform. l Latest :- real-time support and Vector graphics in QT3. n Special Animation files (FLI, FLC)
CODECS n Cinepack l Most widely used and supported. l Software and hardware versions available. l Best for 15 fps,180 KBps l 10:1 achievable compression l Shortcomings : pixelisation, chunky, poor colour saturation. n Indeo l From Intel (Hardware and software) l 3x version good for talking heads, poor colour saturation. l 4x and 5x versions comparable to MPEG-1 l Realtime adjustment of brightness and contrast
Steps to PC Video n Plan Content n Decide Target platform, run length and quality. n Shoot and Grab or otherwise Acquire (vide Video tips) n Prepare Edit Decision List (EDL) n Edit and Compose n Dub n Clean and Compress n Store in CD or Distribute via Net
Video Capture: Basics n Required for analog sources like. l VCP l VTP l Laser-Disk players l Camera nRequires Special Hardware (frame Grabber Cards) which l Samples the video signal l Digitise the chroma and Luma components l Stores them in disk files l Optionally compresses the data l Displays while grabbing
Video Capture Platforms. n Should be the highest performance machine affordable. n Would serve as video editing and compressing stations as well. n Windows NT n Pentium-II, 128 MB memory, SCSI-Disk 4GB AV, RAID is better. n Separate disk for software and data storage recommended. n Zip drive for transportation of large files or 100 Mbps fast Eathernet n and of course the frame grabber card
Video Capture Cards n Also called frame grabbers n Some popular cards l Megamotion (JPEG) l Broadway (MPEG-1) l Video Blaster (Creative) l Video Spigot (Creative) l Matrox Illuminator
Video Capture Cards :check list n Input Signal types l Composite, S-Video, RGB l Whether Audio channels supported l NTSC/PAL/SECAM n Sampling rates supported n Colour resolution (8/16/24) n Whether full screen frame grab possible. n Maximum frame rates n Video Overlay, whether needs special video cards. n Whether saves in raw or compressed mode. n Software Provided
Video Capture Cards :Typical Specs Data Translation- Broadway n 32 bit PCI master n NTSC or PAL n capture Frame Size: 352x288 (240) n Colour: 16bit,YUV 4:2:2 n SVideo and Composite n Preview (15 fps typical ) n Overlay (full fps) n Capture to AVI (uncompressed) n Capture to MPEG (compressed) n 16 bit audio n Bundled nEdit Software (Ulead Media Studio VE) n and Compress Software (MPEG-1)
Video Capture Cards :Connection
Video Capture- considerations. n In house or outsource? l Budget l Type of capture l CODEC l Quality n Equipment availability l High end camera/VTP/VCP l Video capture platform, card with appropriate CODEC. Example QUICKTIME l Time required to perform the job. n Asymmetric Encoding n Operator skill n Do not use lossy compression if editing is envisaged(MJPEG-ok)
Video Capture- considerations(contd.). n Two Approaches èBest Capture èJust Capture n Best Capture wCapture with the highest screen size, colour depth and frame rates wSame for audio wBetter safe than sorry : you can always reduce information wHuge file sizes wRequires the most expensive boards n Just Capture wWhy bother for full frame and true colour when the target machines cannot display them? wUse the target screen size, colour depth and frame rates wSave time and money
Video Capture procedures n Estimate the file size. n Create a blank file and defragment the disk. n Check the settings n Trial run nElectrical nSettings nPalette
Video Editing Primer n Contents: lConcepts lTools lAdobe Premiere lStory board
Video Editing Concepts n Edit Decision List n Linear and non linear editing lEdit like serial tape or Edit Like disks n File edit lJust cut and paste n True Edit lFile edit and more n Transition n Conversion n Superposition n Picture edit lFilters, Colour correction &c
Video Editing: Conversions (continued) n Cropping n Resizing n Pan & Scan lFor 70 mm format lPerformed often at telecine stage. n Frame Rate Conversion lMore : duplicate frames lLess: discard frames lCAUTION ! n Pace lSlow motion lFast Play n Colour Depth
Transitions n Dissolve lGradual superposition of the incoming clip on the outgoing lMay ustilise other subtechniques also. n Fade-out lThe ongoing clip gradually turns into black n Fade-in lThe incoming clip gradually turns from black to normal n Wipe lA line divides the frame into the outgoing and incoming clip areas. lThe line moves to increase the incoming clip area. n Many more
Superposition n Whatisit lMany layers of picture with transparent and opaque regions, superposed together n Kinds lPicture +Text lMatte + Text lVideo +Video lPicture + Animation lVideo +Animation (Spilberg) n Keying lChroma Keying lAlpha Channels: pixel by pixel transparency control
Traditional Non-linear Editing! n Tradition goes back to lMusic Channels. l “Jurassic Park” l“Who framed Roger Rabbit” n Used lSilicon Graphics High End Work stations lComposite or RGB digitised lRandom Access Abekas Disks and Exabyte digital tape drives lDigital Edit and special effects including animation lTape to film n Not been a cup of tea for the PC’s n But stay tuned
Non Linear Editing using Model n Instead of an “edit decision list” on paper, one or more story boards may be created using compressed (lower quality) clips. n Such “dummy” products are called models of the final. We would call them prototypes. n Models may be iteratively improved. n Compressed, smaller screen, slower frame rate and lower resolution content allows quick prototyping and improvement.
Video Editing Tools n VidEdit lComes Free with Video for Windows and many frame grabber cards. lFrom the stable of Microsoft. n Adobe Premiere lMost popular
VidEdit Features n Basic Video File Edit Tool. n Input may be.AVI file, Autodesk Animator.FLI or.FLC file. n Cut, Paste and Insert of Video and Audio components. n Frame rate convert n Colour Depth convert n Crop and Resizing n Palette building n compression lRLE, Indeo, Microsoft video-1 l Custom settings based on target (CD-ROM etc.)
Adobe Premiere: Features n Fairly Comprehensive Video Clip Editing n Wide selection of picture edit n Conversion applied on whole clip n Drag and Drop n Wide selection of Transitions n Film Strip metaphor n Super n Audio edit n Compression
Adobe Premiere Windows
Adobe Premiere Construction Window
Media Merge n Simpler video editor n Uses story board metaphor n Sequence of clip and transition boxes n Text superposition n Drag and drop n Three components lStory board editor lScene editor lAudio Editor
Video Compression Techniques and Standards n JPEG & M-JPEG (intra frame) n H.261 (intra frame) n MPEG-1 n MPEG-2 n MPEG-4
Tips for Video in multimedia n Shoot or Acquire n Lights more Lights n Shooting Ancillaries n More tips
Shoot or Acquire n Don’t shoot if good quality “footage” is available. n Employ a professional for shoots if fund permits. n Choose the best available camera (SVHS and High-8 recommended). n Tripod is equally important. n Lighting can make or break. n Choose a simple background. n Check if direct storage to PC is feasible. n See ‘tips’ at the end.
Lights... more lights Eye is more forgiving than camera, which needs better lighting. n At least three light sources l(including natural light). lThe main light from front but a bit left (or right). lThe key light from up and behind. lA compensatory light to reduce stark shadows from the main n Don’t neglect the background: lThe background and the foreground should have difference in illumination. lFor bright background, exposure should be adjusted or the subject would look too dark.
Shooting Ancillaries n Reflectors l Thermocol on masonite backing l Aluminium foil l White board n Halogen main light n Halogen “table lamps” and stand n Turn table (for machine parts and small objects) n Back screen l Remember chroma key n Prompting stand n Mike boom
More Tips for Video in multimedia n While shooting landscapes and large buildings, use slow pans. It would look good, compress well. n Use shooting notes - detailing location, date, event, shot number, identity of persons etc.. Do not rely on memory. n Always capture audio along with video. It would back up your shooting notes. n Take shots that may be used as the back drop of titles for long clips.
More Tips n For indoor objects experiment using short clips with l back ground l lighting l shooting techniques l filtering l rate conversions l compression. n This extra trouble would be rewarded. n Cuts take less disk space than transition. Use transitions judiciously. n Use fonts specially designed for video (e.g. sans serif faunts) in titles.
The Future : n DVD video’s at Indian homes and libraries. n CD-ROM social function video a must for Indian middle class. n DVD-video of museums n Indian Cartoon films (Amar Chitrakatha) on CD-ROM’s and DVD. n Amature film directors. n Spielberg from India n Desktop Editing of Broadcast Quality Software n Video on demand
Desktop Editing of Broadcast Quality Content n Favourable Factors: lDigital Camera, DVD Players and HDTV lIncreased Processing Power at the Desktop at affordable price. lCost 1/100 th of present analog edit station cost lEmergence of MPEG-II n Special Equipment would still be required n Media lTape.. Computer.. Tape lAnalog Tape or Digital ? lDVD.. or Tape ? lCamera to Computer ? n India- the destination ? lLabour intensive lSpecial Care, Closer control