What the difference in Mpeg 1/2/4?

Question

What the difference in Mpeg 1/2/4?

Answer 1

MPEG = Motion Picture Experts Group. This is a standards organization that was put together to help create standards for digital video used in a variety of places. A similar group, the Joint Photographic Experts Group, does the same thing for still images.

The initial goal of MPEG-1 was to put videos on a CD. Since CDs are really too small for video, it was originally designed for 352x240 or so resolution, roughly equivalent to VHS tape. While MPEG-1 can be used at higher resolutions, it usually isn't, and it didn't really catch on. Some digital cameras still use it, and it was briefly used for satellite TV, but it's kind of passed on. You will occasionally see MPEG-1 used on DVDs for low resolution stuff, like ads before the movie, but you probably don't even notice that's it's not MPEG-2.

MPEG-2 was designed for full "D1" class video... full resolution standard definition, such as used in satellite TV and DVD. Keep in mind that the MPEG group didn't define either DVD or digital TV, they just set some standards, and they license them to anyone who wants them.

The main improvements in MPEG-2 were support for interlacing and a few options on how the video was blocked off. All MPEG formats to date for video work in terms of different kinds of "frames" of video... single pictures. An I-Frame is an "independent" frame, which is quite similar to a JPEG image. An uncompressed image is cut up into chunks, usually something like 16x16 pixels, calles a macrocell. Then, a reversible transform is run on this, which is called a Discrete Cosine Transform. This basically converts spatial information into frequency information. That doesn't do any compression, but after that, the higher frequency information (the sharpest bits) are removed, simplifying each cell. Finally, they're huffmann encoded, which is a lossless compression scheme, a little like a ZIP file. It's that simplification, low-pass filtering that results in compression in an I-Frame. If you only chop out a little, no big deal... if you chop out too much, it's pretty visible.

That gets things a little compressed. If you did I-Frame only compression, you would get something like the DV encoding, about 5:1 compression. But there are other frames, P-Frames and B-Frames. These use motion vectors to encode only changes between frames, not the whole next frame. P-Frames (predictive) work only forward, while B-Frames work in both directions.

MPEG-4 actually includes two different video formats: Simpler Profile/Advanced Simple Profile... this is basically what's used in DivX. And Advanced Video Coding, or AVC, which is pretty much the accepted successor to MPEG-2 in the market. These both add better compression with more cleverness and complexity, including variable sized macrocells. MPEG-4 was originally designed for online uses, such as videoconferencing, but it's has grown into many other uses. This added complexity of course makes for far more CPU needed to encode it and decode it, but the win is better quality in the same space...very important for high-definition video on disc, flash-memory based recording, satellite transmissions, as well as internet use.

What happened to MPEG-3, one might ask? That was originally supposed to be the spec for high definition video, but MPEG-2 worked so well, they skipped it.

And MPEG isn't just working on video formats, either. They have a number of specifications for audio (MPEG audio layers 1, 2, and 3, AAC audio as used on Apple's iPod), and they also define a structure for sending around MPEG as a file, kind of what AVI or Quicktime does. In fact, the version of that structure for MPEG-4 was largely influenced by Quicktime's similar structure.

There's also work being done on MPEG-7, which is an XML-based mechanism for tagging multimedia content, not a video standard at all.

Of course, there are other video CODECs out there not from the MPEG folks. One is Macromedia's flash/shockwave video, which often uses a CODEC from On2, Inc. that's designed for very low bitrate connections. Another is Microsoft's own Windows Media 9 video, which has been standardized by the ISO as the VC-1 format, one of the mandatory video formats on Blu-Ray discs (they also support AVC and MPEG-2).

Answer 2

Each one is progressively more efficient, so in other words, if you used Mpeg1/2/4 to compress video so that each clip looked the same quality-wise, Mpeg-4 would have the smallest file size. Otherwise, you might know Mpeg-1 from VCDs, Mpeg-2 from DVDs, and Mpeg-4 from iPod and PSP video. Hope this helps!

Answer 3

mpeg stands for motion picture experts group. this was a committee that established standards for video and audio compression needed for computers. the original standard, mpeg 1 allowed quarter size video (240x320) to run at full speed (30 fps) on a  windows 3 class computer.  it is still used as the video file format for some digital still cameras. mpeg2 is a major improvement and is used as the basis for modern digital television and DVD movies. it can deliver even HD quality video.  mpeg4 is a super compression format intended to allow continuous streaming on an internet connection.

Answer 4

http://en.wikipedia.org/wiki/MPEG-1

http://en.wikipedia.org/wiki/Mpeg2

http://en.wikipedia.org/wiki/Mpeg4