How does MPEG work?       

MPEG-1  |  MPEG-2  |  MPEG-3  |  MPEG-4 

MPEG stands for "motion pictures expert group" (www.mpeg.org). From this group a file format is developed for the process of compressing and saving Video- or Multimedia files (Video, Picture and sound) in high quality. The MPEG-Standard is divided into four parts: MPEG-1, MPEG-2, MPEG-3 and MPEG-4 whereby MPEG-3 is integrated implanted in MPEG-2.

For playing the big data size of movies (90 minutes of Video, 25 frames per second, high resolution, many colour result in about 120 GByte) with "normal" Computers the MPEG codec uses besides JPEG a way for saving only the changes from picture to picture (not M-MJPEG): The MPEG-Format saves in regular times of twelve pictures Intra-Frames (I-Frames); I-Frames are with JPEG compressed pictures. The pictures between the Intra-Frames are saved, if possible, not in full pictures. MPEG saves the information what parts of a picture are changing from the first and next frame. For this also "Predicated Frames" and "B-Frames" (Bi-directional frame) are used. Because of the fact that this almost never works perfectly is in addition to that per frame the changed part in JPEG-Coded saved. With this method it is possible to reduce the data size for about 99%. The maximal possible compression is at 200:1.



Sample:

MPEG-1

…is developed for smooth Video-playing and has the biggest compression rate. The MPEG-1 compressing or decompressing is originally used as a hardware-based process. Now it is possible with the faster Computer CPUs to Compress and decompress with Software's. The specifications:

  • 25 Frames, no information between interlaced pictures.
  • A resolution of 352 x 288, 352 x 240 (SIF-Resolution)
  • 1,2 to 3 MBit/s (1,2 MBit/s for a VCD)
MPEG-2

The most difference between MPEG-1 and MPEG-2 is that MPEG-2 better works with the in the TV used interlaced picture. The secret of MPEG-2 is in the compression of highest quality so that the video material can be almost 1 : 1 in source quality edited. Consequently MPEG-2 established to the broadcast standard. The specifications:

  • 50 Fields, 25 Frames
  • High-Resolution 1920x1152 and up to 80 MBITs/s (High Definition TV - HDTV)
  • High-Resolution 1440x1440 and up to 60 MBITs/s (HDTV)
  • Main-Resolution 720x576 and up to 15MBITs/s (digital TV and DVD-Video)
  • Low-Resolution 352x288 and up to 4MBITs/s (S-VHS, SIF)
MPEG-3

The part of the MPEG-3 which should be used for High Definition TV-Quality (HDTV) is meanwhile in the MPEG-2-Standard implemented. MPEG-3 is not MP3!.
 

MPEG-4
… is the next version of MPEG-2 and is since 1996 in developing. Although MPEG-4 was originally set as the coding standard for audiovisual data with a low bit rate, was the using much further than only Streaming of linear media files in Internet- und wireless applications. MPEG-4 consists for example efficient mechanisms for the compressing and distribution interactive media content. In addition to that has MPEG-4 3D-potentials which allows to show "artificial intelligent" for example in videoconferences. This happened in a VRML-similar script language.

The compression rate is in MPEG-4 not higher than MPEG-2 but sprites can be compressed much better because for the coding mechanism is available more time. Possibly it is possible to change to Wavelets. The script language allows in a few bytes operations like "moving" faster to finish than the digital compressed form of the same operation would do.

Microsoft first implantation of MPEG-4 is since Windows 98 on every Windows-PC installed and has a limitation 256 Kilobit per second. For a lag-free video in acceptable quality it is not enough). Since the DivX-Hack it was possible to set the bit rate up to 6000 KBit/sec.

Michael