HP ProLiant 1600 Video Streaming Technology - Page 10

ECG068/0798 WHITE PAPER (cont.) MPEG Encoding Scheme 1...0 MJPEG stands for "motion JPEG" and is simply a sequence of JPEG compressed still images to represent a moving picture. Video capture boards sometimes use MJPEG since it is an easily editable format, unlike MPEG, described next. A disadvantage of MJPEG is that it does not handle audio. MPEG The International Standards Organization (ISO) has adopted a series of video codec standards known as MPEG. MPEG stands for Moving Pictures Experts Group, a workgroup of the International Standards Organization (ISO). This group defined several levels of video compression standards known as MPEG-1, MPEG-2, and MPEG-4. MPEG-1 The MPEG-1 standard was defined in January 1992 and was aimed primarily at video conferencing, videophones, computer games, and first generation CD-ROMs. MPEG-1 is also the basis of VideoCD and CD-i players. It was designed to provide consumer-quality video and CDROM quality audio at a data rate of approximately 1.5 Mbps and a frame rate of 30 fps. Originally, MPEG-1 was designed for playing video from 1x CD-ROMs and to be compatible with data rates of T1 lines. MPEG-1 uses Interframe compression to eliminate redundant information between frames, and Intraframe compression within an individual frame using a lossy compression technique similar to JPEG image compression. Three types of frames result from the compression algorithm: Iframes, P-frames, and B-frames. These are illustrated in the following diagram. 15KB Group of Pictures (GOP) Entry Point 8KB Entry Point 3KB I BBP BBP BBP BBP BBI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 I-frames do not reference other previous or future frames. They are stand-alone as "Independent" frames and are compressed only with intraframe encoding. They are thus larger than the other frames. When video is played or indexed into with rewind and fast forward, they are also the entry points into the video since only they represent a complete picture. P-frames, on the other hand contain "Predictive" information with respect to previous I or P frames, containing only the pixels that have changed since the I or P-frame, including taking motion into account. P-frames are therefore smaller than I-frames. I frames are sent at a regular interval, e.g., about every 400ms, and P-frames are sent at some time interval after the I-frame that varies based on the implementation.