[3.4] What are the video details?

DVD-Video is an application of DVD-ROM, according to the specification created by the DVD Forum (see 6.1 ). DVD-Video is also an application of MPEG-1, MPEG-2, Dolby Digital, DTS, and other formats. This means the DVD-Video format defines subsets of these standards and formats to be applied in practice to make discs intended for DVD-Video players. DVD-ROM can contain any desired digital information, but DVD-Video is limited to certain data types designed for television reproduction.

A disc has one track (stream) of MPEG-2 constant bit rate (CBR) or variable bit rate (VBR) compressed digital video. A restricted version of MPEG-2 Main Profile at Main Level (MP@ML) is used. SP@ML is also supported. MPEG-1 CBR and VBR video is also allowed. 525/60 (NTSC, 29.97 interlaced frames/sec) and 625/50 (PAL/SECAM, 25 interlaced frames/sec) video display systems are expressly supported. Coded frame rates of 24 fps progressive from film, 25 fps interlaced from PAL video, and 29.97 fps interlaced from NTSC video are typical. MPEG-2 progressive_sequence is not allowed, but interlaced sequences can contain progressive pictures and progressive macroblocks. In the case of 24 fps source, the encoder embeds MPEG-2 repeat_first_field flags into the video stream to make the decoder either perform 2-3 pulldown for 60Hz NTSC displays (actually 59.94Hz) or 2-2 pulldown (with resulting 4% speedup) for 50Hz PAL/SECAM displays. In other words, the player doesn't "know" what the encoded rate is, it simply follows the MPEG-2 encoder's instructions to produce the predetermined display rate of 25 fps or 29.97 fps. This is one of the main reasons there are two kinds of discs, one for NTSC and one for PAL. (Very few players convert from PAL to NTSC or NTSC to PAL. See 1.19 .)

Because film transfers for NTSC and PAL usually use the same coded picture rate (24 fps) but PAL resolution is higher, the PAL version takes more space on the disc. The raw increase before encoding is 20% (480 to 576), but the final result is closer to 15%, depending on encoder efficiency. This translates to an increase of 600 to 700 megabytes on PAL discs compared to NTSC discs.

It's interesting to note that even interlaced source video can be rendered as progressive-structured MPEG pictures by a good encoder, with interlaced field-encoded macroblocks used only when needed for motion. Most film sources are encoded at 24 frames per second (the inverse telecine process during encoding removes duplicate 2-3 pulldown fields from the videotape source, and the remaining field pairs, although technically in interlaced form, can be re-interleaved by a progressive player). Most video sources are encoded at 25 or 30 interlaced frames per second. These may be mixed on the same disc, such as an interlaced-source logo followed by a progressive-source movie.

See 3.8 for an explanation of progressive and interlaced scanning. See 1.40 for progressive-scan players. See the MPEG page < http://www.mpeg.org/ > for more information on MPEG-2 video.

Picture dimensions are at maximum 720x480 (for 525/60 NTSC display) or 720x576 (for 625/50 PAL/SECAM display). Pictures are subsampled from 4:2:2 ITU-R BT.601 down to 4:2:0 before encoding, allocating an average of 12 bits/pixel in Y'CbCr format. (Color depth is 24 bits, since color samples are shared across 4 pixels.) DVD pixels are not square (see 3.5 ). The uncompressed source is 124.416 Mbps for video source (720x480x12x30 or 720x576x12x25), or 99.533 or 119.439 Mbps for film source (720x480x12x24 or 720x576x12x24). In analog output terms, lines of horizontal resolution is usually around 500, but can go up to 540 (see 3.4.1 ). Typical luma frequency response maintains full amplitude to between 5.0 and 5.5 MHz. This is below the 6.75 MHz native frequency of the MPEG-2 digital signal (in other words, most players fall short of reproducing the full quality of DVD). Chroma frequency response is half that of luma.

Allowable picture resolutions are:
MPEG-2, 525/60 (NTSC): 720x480, 704x480, 352x480, 352x240
MPEG-2, 625/50 (PAL): 720x576, 704x576, 352x576, 352x288
MPEG-1, 525/60 (NTSC): 352x240
MPEG-1, 625/50 (PAL): 352x288

Different players use different numbers of bits for the video digital-to-analog converter, wit the best-quality players using 10 or 12 bits. This has nothing to do with the MPEG decoding process, since each original component signal is limited to 8 bits per sample. More bits in the player provide more "headroom" and more signal levels during digital-to-analog conversion, which can help produce a better picture.

Maximum video bit rate is 9.8 Mbps. The "average" video bit rate is around 4 Mbps but depends entirely on the length, quality, amount of audio, etc. This is a 31:1 reduction from uncompressed 124 Mbps video source (or a 25:1 reduction from 100 Mbps film source). Raw channel data is read off the disc at a constant 26.16 Mbps. After 8/16 demodulation it's down to 13.08 Mbps. After error correction the user data stream goes into the track buffer at a constant 11.08 Mbps. The track buffer feeds system stream data out at a variable rate of up to 10.08 Mbps. After system overhead, the maximum rate of combined elementary streams (audio + video + subpicture) is 10.08. MPEG-1 video rate is limited to 1.856 Mbps with a typical rate of 1.15 Mbps.

Still frames (encoded as MPEG I-frames) are supported and can be displayed for a specific amount of time or indefinitely. These are used for menus or slideshows. Still frames can be accompanied by audio.

A disc also can have up to 32 subpicture streams that overlay the video for subtitles, captions for the hard of hearing, captions for children, karaoke, menus, simple animation, etc. These are full-screen, run-length-encoded bitmaps with two bits per pixel, giving four color values and four transparency values. For each group of subpictures, four colors are selected from a palette of 16 (from the YCbCr gamut), and four contrast values are selected out of 16 levels from transparent to opaque. Since one of the four values is usually 100% transparency (to let the video show through), only three combinations of colors and transparencies are left, making overlay graphics rather crude. Subpicture display command sequences can be used to create effects such as scroll, move, color/highlight, and fade. The maximum subpicture data rate is 3.36 Mbps, with a maximum size per frame of 53220 bytes.

In addition to subtitles in subpicture streams, DVD also supports NTSC Closed Captions. Closed Caption text is stored in the video stream as MPEG-2 user data (in packet headers) and is regenerated by the player as a line-21 analog waveform in the video signal, which then must be decoded by a Closed Caption decoder in the television. Although the DVD-Video spec mentions NTSC only, there is no technical reason PAL/SECAM DVD players could not be made to output the Closed Caption text in World System Teletext (WST) format; the only trick is to deal with frame rate differences. Unfortunate note: DVD Closed Caption MPEG-2 storage format is slightly different than the ATSC format. See 1.45 for more about Closed Captions.

Everyone gets confused by the term "lines of horizontal resolution," also known as LoHR or TVL. It's a carryover from analog video, it's poorly understood, and it's inconsistently measured and reported by manufacturers, but we're stuck with it until all video is digital and we can simply report resolution in pixels.

Technically, lines of horizontal resolution refers to visually resolvable vertical lines per picture height . In other words, it's measured by counting the number of vertical black and white lines that can be distinguished an area that is as wide as the picture is high. The idea is to make the measurement independent of the aspect ratio. Lines of horizontal resolution applies both to television displays and to signal formats such as that produced by a DVD player. Most TVs have ludicrously high numbers listed for their horizontal resolution.

Since DVD has 720 horizontal pixels (on both NTSC and PAL discs), the horizontal resolution can be calculated by dividing 720 by 1.33 (from the 4:3 aspect ratio) to get 540 lines. On a 1.78 (16:9) display, you get 405 lines. In practice, most DVD players provide about 500 lines instead of 540 because of filtering and low-quality digital-to-analog converters. VHS has about 230 (172 widescreen) lines, broadcast TV has about 330 (248 widescreen), and laserdisc has about 425 (318 widescreen).

Don't confuse lines of horizontal resolution (resolution along the x axis) with scan lines (resolution along the y axis). DVD produces exactly 480 scan lines of active picture for NTSC and 576 for PAL. The NTSC standard has 525 total scan lines, but only 480 to 483 or so are visible. (The extra lines contain sync pulses and other information, such as the Closed Captions that are encoded into line 21). PAL has 625 total scan lines, but only about 576 to 580 are visible. Since all video formats (DVD, VHS, LD, broadcast, and so on) have the same number of scan lines, it's the horizontal resolution that makes the big difference in picture quality.

For more information, see Allan Jayne's TV and Video Resolution Explained .