Victorian Electronic Records Strategy - Forever Digital logo

news     faq     contact us     prov
VERS STORY STANDARD COMPLIANCE PROJECTS DIGITAL ARCHIVE TRAINING TOOLKIT PUBLICATIONS
 


Introduction

Standard
(Version 1)
  Specification 1
  Specification 2
  Specification 3

Standard
(Version 2)
  Specification 1
  Specification 2
  Specification 3
  Specification 4
  Specification 5
  Advice 9
  Advice 10
  Advice 11
  Advice 12
  Advice 13
  Advice 14

Home > Standard > Standard (Version 2) > Advice 13 > MPEG 4
Search
    

7.6 MPEG 4

For video encoding we have chosen MPEG-4 (ISO/IEC 14496) as a long term preservation format.

MPEG-4 was designed to support interactive multimedia on viewers ranging from handheld screens (e.g. iPod) to high definition TV (HDTV). It provides an almost bewildering array of technologies. Apart from multiple methods of compressing conventional video and audio, MPEG-4 includes facilities to represent shapes, animations, text, and so on. Unlike a conventional video stream, MPEG-4 is defined in terms of objects within scenes. These objects may be derived from ‘real sources’ (e.g. conventional video), or ‘synthetic sources’ (e.g. CAD, computer animation). Synthetic sources even include symbolic representations of musical scores.

MPEG-4 was chosen over MPEG-2 (the format used on DVDs) for the following reasons:

  • Superior compression of the video stream.
  • Better support for non-television video. This includes non-standard frame sizes, and in particular, non-standard frame rates. This is important in a government archive where much of the preserved footage may not be captured from conventional television footage, for example, CCTV footage.
  • Potential for future improvements in the standard. Whereas MPEG-2 appears to have completed development, MPEG-4 is still being actively developed.

The MPEG-4 standard currently (May 2006) consists of 22 parts, not all of which have been confirmed as international standards (these parts are listed at the end of this section). Because of the complexity of MPEG-4, VERS will only accept MPEG-4 files derived from real sources, that is, conventional video and audio. We accept the following parts of the standard:

  • Part 1 Systems. This part defines the overall structure of an MPEG-4 object. Note that we do not accept the file format defined in the 2001 edition of this part (in Section 13), as this has been removed from the 2004 edition and replaced by the file format defined in Part 14.
  • Part 2 Visual. This defines the compression of video streams.
  • Part 3 Audio. This defines the compression of audio streams.
  • Part 10 Advanced Video Coding (AVC). This is an improved video compression scheme, which is identical to the ITU-T H.264 compression scheme.
  • Part 14 MP4 File Format. This replaces the file format originally defined in Part 1.
  • Part 15 Advanced Video Coding (AVC) File Format

It should be noted that Part 4 of the standard describes testing for conformance to the standard and Part 5 contains a reference implementation. These can be used to ensure that an implementation is correct. The reference implementation in Part 5 conforms to the standard, but may not be efficient or produce good results. (It was noted in one reference that the vendors developing the standard do not have any incentive to make available a free high quality MPEG-4 implementation.) It appears that both Part 4 and 5 are updated as new parts of ISO 14496 become international standards.

Advanced Video Coding (AVC, also known as ITU-T H.264) is a newer video compression algorithm for MPEG-4 that attempts to achieve flexible good quality video at bit rates 50% or less than previous compression without significant increases in decompression complexity. AVC appears to have greater take-up than the original video coding specified in Part 2. We have chosen to support the basic AVC profiles. The Fidelity Range Extensions allow higher quality, but currently lack support.

Conformance to the MPEG-4 standards is specified in terms of profiles and conformance levels (see [Koenen]). A profile is a collection of tools (which represent bit stream syntaxes) which can be used to render onto a device. A level is essentially a measure of complexity; typically a bit rate that will allow a particular level of performance. A conformance point is the combination of a profile at a specific level (very crudely, a set of bit streams that can be handled at a particular bit rate). The standard defines relatively few conformance points.

  • Visual Profiles. We limit video objects to the Main L4 visual profile. This allows an MPEG-4 file to represent a HDTV video stream with a bit rate of 38.4 Megabits/section. We do not accept the Simple Scaleable, N-bit, Scaleable Texture, Simple FA, Basic Animated Texture, or Hybrid profiles. These other profiles mainly represent video information from synthetic sources. For AVC we accept the Baseline, Extended, and Main Profiles up to L4. This allows encoding of HDTV at 60 or 30 frames per second.
  • Audio Profile. We accept the Main L4 audio profile. This profile handles all audio bitstreams (including synthetic sources) defined in MPEG-4 with a bit rate sufficient to encode a 5.1 channel object. We could have restricted the VERS to the Scalable profile (which excludes the synthetic sources), but this also excludes the AAC Main and AAC SSR tools.
  • Graphics Profile. We do not accept any graphics profile, as these represent synthetic sources.
  • System Profiles. There are two aspects to system profiles: Scene Graphs and Object Descriptors. VERS accepts the Core Object Descriptor profile (this is the only profile defined). VERS accepts the Audio and Simple 2D Scene Graph profiles; the Complete 2D and Complete profiles require the display device to be able to manipulate objects.

There are several alternative file formats for MPEG-4 files. Of these, the original format which was defined in Section 13 ISO/IEC 14496-1:2001 seems to be obsolete as it is no longer included in the 2004 edition. The replacement file format is defined in Part 14 (MP File Format), and this is the format that VERS will accept. There is another format (ISO Base Media File Format – BMFF) defined in Part 12. The BMFF is a joint format and is shared with JPEG 2000, however, it does not seem to be widely used. Finally, there is the Advanced Video Coding (AVC) File Format (Part 15) which is used to encode the AVC compressed video.

Where audio is captured along with video, we recommend that MPEG Audio layer 3 compression is used to encode the audio within MPEG-4. MP3 produces small audio files with acceptable quality.

MPEG-4 currently consists of 22 parts. In the following list, the current (May 2006) international standards are identified, together with any approved corrigendum and amendments. Corrigenda correct minor mistakes in the published standard, while amendments extend the standard.

  • Part 1 Systems (ISO/IEC 14496-1:2004, ISO/IEC 14496-1:2004/Amd 1:2005, ISO/IEC 14496-1:2004/Amd8:2004)
  • Part 2 Visual (ISO/IEC 14496-2:2004, ISO/IEC 14496-2:2004/Cor 1:2004, ISO/IEC 14496-2:2004/Amd 1:2004, ISO/IEC 14496-2:2004/Amd 2:2005)
  • Part 3 Audio (ISO/IEC 14496-3:2005)
  • Part 4 Conformance testing (ISO/IEC 14496-4:2004, ISO/IEC 14496-4:2004/Cor 1:2005, ISO/IEC 14496-4:2004/Amd 1:2005, ISO/IEC 14496-4:2004/Amd 1:2005/Cor 1:2005, ISO/IEC 14496-4:2004/Amd 2:2005, ISO/IEC 14496-4:2004/Amd 3:2005, ISO/IEC 14496-4:2004/Amd 4:2005, ISO/IEC 14496-4:2004/Amd 5:2005, ISO/IEC 14496-4:2004/Amd 6:2005, ISO/IEC 14496-4:2004/Amd 7:2005, ISO/IEC 14496-4:2004/Amd 8:2005, ISO/IEC 14496-4:2004/Amd 9:2006, ISO/IEC 14496-4:2004:Amd 10:2005)
  • Part 5 Reference Software (ISO/IEC 14496-5:2001, ISO/IEC 14496-5:2001/Amd 1:2002, ISO/IEC 14496-5:2001/Amd 2:2002, ISO/IEC 14496-5:2001/Amd 3:2002, ISO/IEC 14496-5:2001/Amd 4:2002, ISO/IEC 14496-5:2001/Amd 5:2002, ISO/IEC 14496-5:2001/Amd 6:2002, ISO/IEC 14496-5:2001/Amd 7:2002, ISO/IEC 14496-5:2001/Amd 8:2002)
  • Part 6 Delivery Multimedia Integration Framework (DMIF) (ISO/IEC 14496-6:2000)
  • Part 7 Optimised software for MPEG-4 tools (ISO/IEC 14496-7:2004)
  • Part 8 Carriage of ISO/IEC 14496 objects over IP framework (ISO/IEC 14496-8:2004)
  • Part 9 Reference Hardware Description (ISO/IEC 14496-9:2004)
  • Part 10 Advanced Video Coding (ISO/IEC 14496-10:2005)
  • Part 11 Scene Description and Application Engine (ISO/IEC 14496-11:2005)
  • Part 12 ISO Base Media File Format (ISO/IEC 14496-12:2005, ISO/IEC 14496-12:2005/Cor 1:2005, ISO/IEC 14496-12:2005/Cor 2:2006)
  • Part 13 Intellectual Property Management and Protection (IPMP) Extensions (ISO/IEC 14496-13:2004)
  • Part 14 MP4 File Format (ISO/IEC 14496-14:2003, ISO/IEC 14496-14:2003/Cor 1:2006)
  • Part 15 Advanced Video Coding (AVC) File Format (ISO/IEC 14496-15:2004, ISO/IEC 14496-15:2004/Cor 1:2006, ISO/IEC 14496-15:2003/Amd 1:2006)
  • Part 16 Animation Framework eXtension (AFX) (ISO/IEC 14496-16:2004, ISO/IEC 14496-16:2004/Cor 1:2005, ISO/IEC 14496-16:2004/Cor 2:2005, ISO/IEC 14496-16:2004/Amd 1:2006)
  • Part 17 Streaming Text Format (ISO/IEC 14496-17:2006)
  • Part 18 Font compression and streaming (ISO/IEC 14496-18:2004)
  • Part 19 Synthesized Texture Stream (ISO/IEC 14496-19:2004)
  • Part 20 Lightweight Application Scene Representation
  • Part 21 MPEG-J Extension for rendering
  • Part 22 Open Font Format

back to top

printer friendly

Department for Victorian Communities logo - Link to DVC home

VERS Story | Standard | Compliance | Projects | Digital Archive | Training | Toolkit | Publications

Site last reviewed: 2 Sep 2004
© State of Victoria 2006 | Disclaimer | Privacy Statement

 Public Record Office Victoria logo - Link to PROV home