MPEG-2
MPEG-2 (1994) isdesignation forgroupaudiovideo coding standards agreed upon by MPEG (Motion Pictures Coding Experts Group),published as ISO standard 13818. MPEG-2typically usedencode audiovideobroadcast signals, including digital satelliteCable TV. MPEG-2,some modifications,alsocoding format used by standard commercial DVD movies.
MPEG-2similarMPEG-1, but also provides supportinterlaced video (the format used by broadcast TV systems.) MPEG-2 videonot optimizedlow bit-rates (less than 1 Mbit/s), but outperforms MPEG-1 at 3 Mbit/sabove. MPEG-2 also introducesdefines Transport Streams, whichdesignedcarry digital videoaudio over unreliable media, andusedbroadcast applications. With some enhancements, MPEG-2alsocurrent standardHDTV transmission. A standards-compliant MPEG-2 decoder should be capableplaying back MPEG-1 streams.
MPEG-2 audio, definedPart 3 ofstandard, enhances MPEG-1's audio by allowingcodingaudio programsmore than two channels. Part 3 ofstandard allows thisbe done inbackwards compatible way, allowing MPEG-1 audio decodersdecodetwo main stereo components ofpresentation, or innon backwards compatible way, which allows encodersmake better useavailable bandwidth. MPEG-2 supports various audio formats, including MPEG-2 AAC.
| Tablecontents |
|
2 MPEG-2 video coding (simplified) 3 MPEG-2 audio encoding 4 MPEG-2 on DVD 5 MPEG-2 on DVB 6 MPEG-2 standards |
MPEG-2,Standard
General information about MPEG-2 VideoMPEG-2 Audio excluding modifications when used on DVD / DVB.
A MPEG-2 System Stream typically consiststwo elements:
- video data + time stamps
- audio data + time stamps
MPEG-2 video coding (simplified)
MPEG-2forgeneric codingmoving picturesassociated audiocreatesvideo stream outthree typesframe data (intra frames, forward predictive framesbidirectional predicted frames) arranged inspecified order calledGOP structure (GOP = Group Of Pictures - see below).
Typicallyoriginating material isvideo sequence atpre-set pixel resolution at 25 (CCIR) or 29.97 (FCC) frames/secondsound.
MPEG-2 supports both interlacedprogressive scan video streams. In progressive scan streams,basic unitencoding isframe, whileinterlaced streams,basic unit isfield. Indiscussion below,generic terms "picture""image" refereither fields or frames, depending ontypestream.
The MPEG-2 streammade up ofseriesdata frames encoding pictures. The three waysencodingpicture are: intra-coded (I picture), forward predictive (P picture)bidirectional predictive (B picture).
The video imageseparated into one luminance (Y)two chrominance channels (also called color difference signals UV). Italso divided into "macroblocks", which arebasic unitcoding withinpicture. Each macroblockdivided into four 8x8 luminance blocks. The number8x8 chrominance blocks per macroblock depends onchrominance format ofsource image. For example, incommon 4:2:0 format, thereone chrominance block per macroblockeach ofchannels, makingtotalsix blocks per macroblock.
IncaseI pictures,actual image datathen passed throughencoding process described below. PB picturesfirst subjected toprocess"motion compensation",which theycorrelated withprevious (and incaseB pictures,next) image. Each macroblock inP or B picturethen associatedan area inprevious or next image thatwell-correlatedit. The "motion vector" that mapsmacroblockits correlated areaencoded,thendifference betweentwo areaspassed throughencoding process described below.
Each blocktreatedan 8x8 discrete cosine transform. The resulting DCT coefficientsthen quantized according topre-defined scheme, re-orderedmaximizeprobabilitylong runszeros,run-length coded. Finallyfixed-table huffman encoding schemeapplied.
I pictures encodespacial redundancy, PB picturestemporal redundancy. Because adjacent frames invideo streamoften well-correlated, P pictures may be 10% ofsizeI pictures,B pictures 2%their size.
The sequencedifferent frame typescalledGroupPictures (GOP) structure. Theremany possible structures butcommon one15 frames long,hassequence I_BB_P_BB_P_BB_P_BB_P_BB_P_BB_. A similar 12 frame sequencealso common. The ratioI, PB pictures inGOP structuredetermined bynature ofvideo stream andbandwidth constraints onoutput stream, although encoding time may also be an issue. This particularly truelive transmissionin real-time environmentslimited computing resources, asstream containing many B pictures can take three times longerencode than an I-picture-only file.
The output bit-ratean MPEG-2 encoder can be constant or variable, withmaximum bit rate determined byplayback media -exampleDVD movie maximum10.4 Mbit/s. To achieveconstant bit-ratedegreequantizationiteratively alteredachieveoutput bit-rate requirement. Increasing quantization leadsvisible artefacts whenstreamdecoded, generally inform"mosaicing", wherediscontinuities atedgesmacroblocks become more visible as bit ratereduced. Sony Launched The MPEG IMX Professional Videotape For an Bitrate Of Up To 50 Mbit/s @ 4:2:2P@ML
MPEG-2 audio encoding
MPEG-2 also introduces new audio encoding methods. These are
- low bitrate encodinghalved sampling rate (MPEG-1 Layer 1/2/3 LSF)
- multichannel encodingup5.1 channels
MPEG-2 on DVD
Additional restrictionsmodifications on DVD.
- Restrictedone offollowing resolutions
- 720 x 480 pixel, 59.94 fields/sec (FCC)
- 720 x 480 pixel, 29.97 frames/sec (FCC)
- 720 x 576 pixel, 50 fields/sec (CCIR)
- 720 x 576 pixel, 25 frames/sec (CCIR)
- max. 9,8 Mbit/s
- YUV 4:2:0
- additional subtitles possible
- Audio
- MPEG-2 Layer 2 Audio48 kHz possible
- Digital Dolby up448 kbit/s48 kHz possible
- dts754 or 1510 kbit/s possible
- There must be at least one trackMPEG Audio or DD audio
- Restrictions inGOP structure
MPEG-2 on DVB
Additional restrictionsmodifications on DVB-MPEG.
- Restrictedone offollowing resolutions
- 720 × 480 pixel, 59.94 fields/sec (FCC)
- 720 × 480 pixel, 59.94 fields/sec (FCC)
- 480 × 480 pixel, 29.97 frames/sec (FCC)
- 480 × 480 pixel, 29.97 frames/sec (FCC)
- 720 × 576 pixel, 50 fields/sec (CCIR)
- 720 × 576 pixel, 25 frames/sec (CCIR)
- 480 × 576 pixel, 50 fields/sec (CCIR)
- 480 × 576 pixel, 25 frames/sec (CCIR)
MPEG-2 standards
; ISO/IEC 13818-1 : Systems - describes synchronizationmultiplexingvideoaudio. ; ISO/IEC 13818-2 : Video - compression codecinterlacednon-interlaced video signals. ; ISO/IEC 13818-3 : Audio - compression codecperceptual codingaudio signals. A multichannel-enabled extensionMPEG-1 audio (MP3). ; ISO/IEC 13818-4 : Describes procedurestesting compliance. ; ISO/IEC 13818-5 : Describes systemsSoftware simulation. ; ISO/IEC 13818-6 : Describes extensionsDSM-CC (Digital Storage Media CommandControl.) ; ISO/IEC 13818-7 : Advanced Audio Coding (AAC) ; ISO/IEC 13818-9 : Extensionreal time interfaces. ; ISO/IEC 13818-10 : Conformance extensionsDSM-CC.
