Inside DVD-Video/Glossary


 * Aspect Ratio : a somewhat ambiguous term, sometimes used as a synonym for pixel density. In this book, I will take it to mean the relative dimensions of the video image as viewed on a TV screen, measured in units of distance, not pixels. Thus, narrowscreen video has an aspect ratio of 4:3, while widescreen video has an aspect ratio of 16:9.


 * BCD : short for Binary-Coded Decimal. BCD numbers are used for the playback time field in the PGC, and for the cell elapsed time in each PCI packet. Presumably this format is used, not for calculation purposes by the player, but for easy conversion to a human-readable form in the player’s display.


 * B-Frame : a video frame which is encoded as the difference between two other reference frames, one temporally preceding and one temporally following.


 * Button : a defined rectangular area on the screen during the display of a menu or title (yes, titles, not just menus, may have buttons). May be selectively highlighted by the user pressing the up/down/left/right arrow keys on the player remote to navigate around the available buttons; when the OK/Enter button is pressed, the highlighted button (depending on the player) may be briefly shown in the selected state, after which the VM instruction attached to the button is executed. A button may also be an auto-action button, which means it is selected (and the instruction executed) by the mere act of highlighting it.


 * Cell : a unit consisting of one or more VOBUs, possibly with a single VM instruction attached to be executed when it finishes playing. The smallest unit that can be targeted by a VM jump instruction.


 * Domain : one of the major divisions in the DVD-Video disc structure; either a VMG or a VTS.


 * DSI : Data Search Information. A packet that contains information needed for scanning/stepping forward and backward in trick-play modes.


 * DTS : Decoder Timestamp. This value needs to be a fraction of a second in advance of the PTS in order to give the decoder machinery time to process the packet. I’m not sure why this needs to be present at all; since the necessary delay would presumably vary from one decoder design to another, so it should be up to the decoder implementation to deal with any necessary timing compensation, this information has no business being in the MPEG stream.


 * DVD : an optical disc format which was a development from the earlier Compact Disc format. Originally supposed to stand for “Digital Video Disc”, then “Digital Versatile Disc”, then finally it was decreed that it stood for nothing at all, so that it could be trademarked.


 * DVD-Video : a specification for using DVDs to hold video and movies together with options for interactive menus for accessing them; the subject of this book.


 * Field : half of a frame (either all the odd-numbered scanlines, or all the even-numbered ones), for interlacing.


 * FPC : First-Play PGC. A special PGC in the VMG that gets control when the disc is inserted in the player. Would contain instructions to bring up the main menu, but this is often preceded by distributor logos, copyright warnings, and sometimes even advertisements that can’t be skipped.


 * Frame : a video frame is a single still video image. A succession of frames displayed one after another at a sufficiently high rate gives the illusion of a moving image. Cinema movies normally use a rate of 24 frames per second, which is adequate for most, but not all, motion. However, flashing 24 images per second onto a screen, separated by darkness, leads to very discernible and annoying flicker. So each frame is actually projected twice in succession, giving 48 flashes per second, which most people don’t notice. Broadcast TV solved the flicker problem in a different way, by using interlacing.


 * GOP : Group Of Pictures. An I-Frame, together with all dependent P- and B-Frames. In DVD-Video, there must be one I-Frame every 18 frames (36 fields) for NTSC, and one every 15 frames (30 fields) for PAL.


 * Header : a unit of information in an MPEG file. Begins with 2 bytes of 0, then one byte containing 1, then another byte containing an ID code which indicates what (if any) further information might be following. Some codes indicate the presence of a packet.


 * I-Frame : a video frame which is compressed in a self-contained way, and can be decoded without reference to any other frame. See also P-Frame and B-Frame.


 * Interlacing : a bandwidth-saving hack that was originally introduced for analog broadcast TV, and for some reason kept for digital TV. Instead of showing each frame in its entirety, and using the double-projection trick to mask the flicker, each frame is split into two fields, which are shown in succession. Undoing the interlacing, by combining pairs of fields into frames, is not so simple, because video cameras record the fields at different times, so any movement of the objects in the picture can lead to strange artifacts.


 * Letterbox : a way of displaying widescreen footage on a narrowscreen TV, by shrinking the image to fit the width of the screen, and padding out the unused parts of the screen above and below the picture with a black background. Compare Pan &amp; Scan.


 * MPEG : stands for Motion Picture Experts Group. The name for a family of standards to do with digital video, of which the only ones relevant to DVD-Video are the earliest two: MPEG-1 and MPEG-2.


 * Narrowscreen : the original 4:3 aspect ratio for broadcast TV, adopted from early cinema movies.


 * NAV PACK : a PACK that contains PCI and DSI packets—navigation information. There must be one of these at the start of every VOBU.


 * NTSC : National Television System Committee. The original analog broadcast colour TV specification developed in the USA, and one of the two most widely used in the world, the other being PAL. Defines a frame rate of 29.97 fps, and digital video frame dimensions of 720x480.


 * Overscan : a limitation of early TV sets, due to their inability to accurately match up the boundaries of the visible picture with the boundaries of the set’s picture display area, meant that a few percent of the image was lost all round the edges. Even with today’s pixel-accurate LCD and plasma TVs, the convention has carried over of throwing away a few percent of the video image around the edges, because broadcasters often allow weird rubbish to creep into this area, like flashing timecode dots. For this reason, when composing a video image, or laying out a DVD-Video menu, or such, remember not to put anything important too close to the edges of the picture, or it might get chopped off when viewed on a TV.


 * PACK/Pack : (I’m not sure whether it’s capitalized or not) a part of an MPEG file, beginning with a PACK header and including the following packets up to but not including the next PACK header. In DVD-Video, each PACK must be 2048 bytes in size (the DVD sector size).


 * Packet : a unit of information in an MPEG file. Begins with 2 bytes of 0, then one byte containing 1, then another byte containing an ID code which indicates the packet type, followed by two bytes containing the length of the following packet data.


 * PAL : Phase Alternation Line. An analog broadcast colour TV specification developed in Germany, and one of the two most widely used in the world, the other being NTSC. Defines a frame rate of 25 fps, and digital video frame dimensions of 720x576.


 * Pan &amp; Scan : a way of displaying widescreen footage on a narrowscreen TV, by chopping off the sides of the picture to fit. Instead of always chopping equal amounts on both sides, it is possible to define a “picture origin” in the MPEG video stream, which moves to track the area of greatest importance in the image. Compare Letterbox.


 * PCI : Presentation Control Information. A packet that contains information about buttons.


 * P-Frame : a video frame which is encoded as the difference from a preceding I-Frame. This reduces the size of the video data greatly, the downside being that if you randomly skip around in the video stream, playback can only start properly from an I-Frame.


 * PGC : Program Chain. A unit consisting of one or more programs. A menu or title is made up of one or more PGCs. A title PGC may have PTTs defined in it.


 * Pixel : a sample of an image, defining intensity and colour at a single point. Contrary to common misconception, a pixel is not a little square; its ideal shape is a dimensionless point.


 * Pixel Density : the number of pixels per unit distance in a displayed image. For our purposes, the important issue is whether the pixel density is uniform (equal both horizontally and vertically) or not. Computer displays (all the ones worth using, anyway) have uniform pixel densities, and computer image-manipulation software commonly assumes it is dealing with images with uniform pixel densities. Digital video in NTSC and PAL formats, however, does not have uniform pixel densities.


 * Private Stream : the MPEG format reserves two stream IDs for purposes not further defined in MPEG. DVD-Video uses these as follows:
 * * Private Stream 1: for additional non-MPEG audio formats, as well as subpicture streams
 * * Private Stream 2: for PCI and DSI packets.


 * Program : A unit consisting of one or more cells. The smallest unit that can be directly addressed for navigation purposes. A program may also be marked as a PTT (chapter point).


 * PTS : Presentation Timestamp. In order to ensure that playback is properly synchronized, a minimum proportion of the MPEG packets must have PTS values attached to them. These are a monotonically-increasing 33-bit counter in units of a 90kHz clock. (90000 just happens to be easily divisible by the common frame rates like 24, 25, 30.)


 * PTT : Part of Title. A program that can be directly addressed via a jump VM instruction from outside the containing PGC. The DVD-Video term for what ordinary people refer to as a chapter.


 * Reference Frame : an I-Frame or P-Frame.


 * Resolution : an ambiguous term, which can refer to either the total number of pixels or to the pixel density. In this book, I will use it to refer to the total numbers of pixels across and down in the following form: widthxheight, e.g. 720x480 for full-resolution NTSC images, 720x576 for full-resolution PAL.


 * SCR : System Clock Reference. An additional timestamp found in PACK headers, with additional resolution so its units are a 27MHz clock rather than 90kHz.


 * SECAM : Séquentiel Couleur A Mémoire. An analog broadcast colour TV specification developed in France. In terms of recorded media compatibility (DVD-Video as well as the older VHS tapes), it can be lumped in with PAL.


 * Stream : one of the concurrent information channels in an MPEG movie file. Video is a stream, and audio is another stream. DVD-Video also defines other streams used for various purposes (navigation, subtitles).


 * Subpicture : a separate image stream defined by DVD-Video which can be overlaid on top of the video. Can represent arbitrary graphics, except it is limited to two bits per pixel (unlike the video, which can be in full colour). Used to display subtitles and menu button highlights; because it is graphics, and not text, it can represent arbitrary languages and writing systems, use any fonts, etc. Can have transparent areas where the video shows through.


 * Uops : User Operations. Classification of the operations the user can perform (e.g. pause, skip) into different categories that can be selectively prohibited in various places. You see these prohibitions in action every time a disc shows you an ad or copyright warning that can’t be skipped.


 * VM : Virtual Machine. A term with many uses in computing, but in DVD-Video it specifically refers to the machine language defined for implementing button operations and other interactive functions.


 * VMG : Video Manager. The domain in the DVD-Video disc structure that gives access to all the others. There must be exactly one of these. Can only contain menus, not titles.


 * VMGM : Video Manager Menu. A menu in the VMG.


 * VOBU : Video Object Unit. A part of a DVD-Video MPEG file beginning with a NAV PACK and followed by PACKs containing data for the other streams (video, audio, subpicture) as appropriate. The video packets must make up one or more complete GOPs.


 * VTS : Video Titleset. A domain in the DVD-Video disc structure. There can be one or more of these. Can contain menus, titles, or both.


 * VTSM : Video Titleset Menu. A menu in a VTS.


 * Widescreen : in DVD-Video and current-generation digital TVs, an image aspect ratio of 16:9. Most modern cinema movies are filmed in a much wider ratio than this, but 16:9 was presumably considered not to be too jarring a step up from the older 4:3.