MPEG Overview

1
MPEG Overview

• Bert DeKnuydt
• ESAT/VISICS
• K.U.Leuven
• Flanders, Belgium
• bert.deknuydt_at_esat.kuleuven.ac.be
• http//www.esat.kuleuven.ac.be/psi/visics

2
What Is It?

• MPEG stands for Motion Pictures Expert Group,
  part of ISO.
• Started 1988 and still going on.
• Officially ISO/IEC JTC1/SC29/WG11.
• Voluntary standard, called by a convener
• Leonardo Chiaraglione.

3
MPEG Generations

• MPEG1 Coding of moving pictures and associated
  audio for digital storage media at up to about
  1.5 Mbit/s.
• Video-CD, MP3, Internet mpgs.
• MPEG2 Generic coding of moving pictures and
  associated audio.
• DVD, Digital TV, AAC.

4
MPEG Generations

• MPEG4 Multimedia Mobility.
• Face animation, special effects.
• MPEG7 Multimedia Content Description Interface.
• Search interface.
• MPEG21 Multimedia Framework.
• Scheduling, financial and legal aspects.

5
MPEG1 Overview

• DIS 1991, IS 1992.
• More or less generic audio/video coding, targeted
  for 1.5 Mbit/s, for storage only.
• System combining timing, audio and video.
• Video Hybrid DPCM/DCT coding, similar to H.261
  progressive only.
• Audio stereo coder, with psycho-acoustic model.

6
MPEG1 Video
7
MPEG1 Video

• Temporal redundancy
• Prediction using motion vectors (causal and
  non-causal)
• DPCM

8
MPEG1 Video

• Spatial redundancy
• Discrete Cosine Transform
• Zig-Zag scan

9
MPEG1 Examples

• Intra Only
• 188 compression

• IPPPP
• 1822 compression

10
MPEG1 MP3
11
MPEG2 Overview

• DIS 1994.
• Original target coding TV images at around 10
  Mbit/s.
• Target was extended to encompass higher
  resolutions, as MPEG3 was absorbed.
• Video similar to MPEG1, but modular.

12
MPEG2 Profiles Layers

• Layered design
• Profiles coding complexity

13
MPEG2 Levels Profiles

• Levels Memory/Speed requirements

14
MPEG4 Extras

• Temporal and quality scalability.
• Pan scalability 43 in 169.
• 422 profile Jan 1996.
• Multi-View-Profile Oct 1996.

15
MPEG2 Audio

• IS April 1997.
• AAC Advanced Audio Coding.
• Successor of MP3.
• Up to 96 kHz sampling, 48 channels.
• Includes TNS, prediction.
• Used in DAB.
• Only high-Q audio system for MPEG4.

16
MPEG4 Overview

• Version 1 IS 1999.
• Face animation.
• Version 2 IS 2000.
• Extensions for 3D mesh coding.
• Body animation.
• Versions 3,4,5 In the making
• Fine grain scalability, studio profile,
• Animation Framework Extensions (Oct 2002)

17
MPEG4 Functionality

• Coded representation of media objects
• Composition of media objects
• Description and synchronization of streaming data
  for media objects
• Delivery of streaming data
• Interaction with media objects
• Management and identification of Intellectual
  Property

18
MPEG4 General

• Scene description BIFS like VRML.
• but allows for streaming.
• Panoply of levels and profiles (15 for video
  alone!)

19
MPEG4 Audio

• Speech 2 to 24 Kbit/s, low delay option.
• Synthesized speech (TTS), I18N for text and
  phonemes, prosodics, link to face animation.
• Low and High bandwidth generic audio.
• Synthesized audio (MIDI like).
• Extras speed, pitch change, scalability

20
MPEG4 Video

• VLBV Very Low Bitrate Video core. More or less
  H263x, 5 to 64 Kbit/s.
• Coding of rectangular regions, up to CIF size.
• Trick mode (FF, FR, RA).
• High Bitrate up to 4 Mbit/s, support for
  interlace.
• Content based coding of arbitrary regions.
• Region either binary or alpha.

21
MPEG4 Block Diagram
22
MPEG4 Video News

• Global motion compensation
• With 8 affine parameters.
• Based on static sprite.
• Based on dynamic sprite.
• Still image coding, with zerotree wavelets.
• Timing, resync-markers, error concealment.
• RVLC codes.

23
MPEG4 SNHC

• Synthetic, parametric description of the human
  body and face and animation streams of both.
• Static and dynamic mesh coding with texture
  mapping.
• Texture coding for view dependent applications.

24
MPEG4 Faces

• A default average face is built-in.
• FDP Facial Definition Parameters allow permanent
  modification of the faces shape and texture.
• FAP Facial Animation Parameters allow transient
  modification of the shape.
• Filter functions allow autonomous action.

25
MPEG4 2D

• Polygonal mesh in video.
• Nodes can have motion vectors.
• Mesh based texture mapping, useable for
  piece-wise affine motion mapping.
• Also usable for special effects (e.g. waving
  flags).

26
MPEG4 3D

• Efficient representation of general 3D mesh, with
  provision for color, normals, texture
  coordinates. Texture can be anything, including
  video and external applications.
• LOD scalability for mesh.
• Spacial scalability for mesh.

27
MPEG7

• Multimedia Content Description Interface.
• DIS Jul. 2001, IS Sep. 2001.
• Meta information it describes how to describe.
  See further.
• Open issues IPR.

28
MPEG7 D DS

• A descriptor D is a representation of a feature
  that defines the syntax and semantics of the
  feature representation.
• A description scheme DS specifies the structure
  and semantics of the relationships between its
  components, which may be both descriptors and
  description schemes.

29
MPEG7 Descriptors
30
MPEG7 DDL

• A language that specifies/modifies description
  schemes (and possibly descriptors), the
  Description Definition Language (DDL).
• DDL is XML based, but specific extensions are
  added.
• Textual and binary formats for all this.

31
MPEG21 MM framework

• PDTR Oct. 2000 CD expected Dec. 2001.
• Mission statement To enable transparent and
  augmented use of multimedia resources across a
  wide range of networks and devices.
• Probably not so interesting for us
• QoS, IPR

32
MPEG21 Targets