Image analysis – Image compression or coding
Reexamination Certificate
1999-04-01
2002-10-01
Couso, Yon J. (Department: 2623)
Image analysis
Image compression or coding
C375S272000, C455S427000
Reexamination Certificate
active
06459811
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to data transmission, and, more particularly, to transmission of compressed video data.
2. Description of the Related Art
Video data consists of video data signals which represent individual video pictures or frames. Each frame may be a still image, but is usually part of a plurality of successive frames of video signal data that represent a motion video sequence.
In various video processing and broadcasting facilities, such as a TV station or studio, there is a need to transmit video signals from one part of the facility to another, to route video signals from selected video sources to selected video destinations or sinks. Video signals nay also be routed from one studio to another. There may be hundreds of video sources (such as the output of a video camera or satellite feed), and hundreds of video sinks (such as a video processor, monitor, video cassette recorder (VCR), or broadcasting unit) that receive video signals from one or more of the video sources. Such facilities also distribute other data signals such as audio signals.
Such systems often operate at a system frame rate, which is typically approximately 30 frames per second (fps). The NTSC standard, for example, operates at (30*1000/1001)≈29.97 fps (referred to subsequently herein for simplicity as 30 fps). Each frame is typically composed of an even field interlaced or interleaved with an odd field. Accordingly, NTSC cameras output 60 fields of analog video signals per second, which includes 30 even fields interlaced with 30 odd fields, to provide video at 30 fps. In such a system, the time required to carry one frame across a given communication path or channel of the system is constant and is the reciprocal of the frame rate: {fraction (1/30)} sec.
There is often a need to switch video signal sources when engaged in operations such as commercial insertion, promo insertion, studio routing, camera switching, tape editing, and the like. In an NTSC system as described above, which does not employ compressed video data, it is relatively simple to switch the input of a given video sink from a first video signal source to a second video signal source. Typically, this is done by switching from one source to another and have the resulting image make a clean cut from one stream to the next. Such switching typically takes place in the vertical interval of the video signal. The associated audio signal is usually also switched simultaneously.
In some newer video facilities (e.g., TV, stations), digital video signals are used to represent video images. The digital video data consists of a compressed or encoded digital bitstream, various segments of which represent a given frame or field. Each frame segment or portion of the compressed video bitstream thus contains compressed data bits that represent the frame. The compressed video bitstream itself represents a sequence of frames (images). In the International Standards Organization (ISO) ISO/IEC 11172 Moving Pictures Experts Group-1 standard (MPEG-1), for example, the display rate and average transmission rate is 30 pictures/second. For the ISO/IEC 13818 (MPEG-2) standard, the display rate and average transmission rate can be 30 frames/second. The MPEG standards support other frame rates as well, including 29.97 and the PAL standard of Europe. In the MPEG standards, the term “picture” refers to a bitstream of data which can represent either a frame of data (i.e., both fields), or a single field of data. In the present application, the general term “frame” will be used for simplicity of explanation.
In such systems, there may be different picture or frame types in the compressed digital stream, such as I frames, P frames, and B frames. I frames, or intra-frames, are self-contained, that is, they are not based on information from previously transmitted and decoded frames. Video frames which are encoded with motion compensation techniques are referred to as predicted frames, or P frames, since their content is predicted from the content of previous I or P frames. P frames may also be utilized as a base for a subsequent P frame. I and P frames are both “anchor” frames, since they may be used as a basis for other frames, such as B or P frames which are predicted based on anchor frames. A “bidirectional” or B frame is predicted from the two anchor frames transmitted most recently relative to the transmission of the B frame. However, because the B frames are typically sent out of order (late), one of the two anchor frames used by a B frame is after the B frame in display order, although it must of course be received by the decoder before the B frame is reconstructed.
I frames typically are the largest in terms of number of encoded bits per frame, while B frames are the smallest, and P frames are somewhere in between. I frames may take many frame times to send (at standard bit rates), while P and B frames often take only a fraction of a frame time. I, P, and B frames are utilized in coding standards such as MPEG-1, while other standards, such as H.261 (P×64), developed by the International Telegraph Union (ITU), utilize only I and P frames.
An encoder at the source end receives unencoded video frame data and compresses the data to provide the compressed digital bitstream. A decoder at the receiving end receives and decompresses (decodes) the data, so that it is put into more useful form, for example for display on a monitor. Referring now to
FIG. 1
, there is shown a prior art compressed video data transmission system, which includes encoder
110
, transmission channel
120
, and decoder
130
. Encoder
110
receives unencoded video frames from a video frame source and is itself a video data source since it provides a compressed video bitstream. Encoder
110
includes a video buffering verifier (VBV)
111
. Decoder
130
, a video sink, comprises buffer
131
and is coupled to a display device such as monitor
132
. The capacity or bandwidth of channel
120
is sufficient to transmit 30 fps on average.
Such a system may be referred to as a constant bit rate (CBR) system. A group of pictures (GOP), which includes at least one I frame, and optionally, a number of B and P frames, is typically transmitted as part of a compressed bitstream by encoder
110
across channel
120
as illustrated in FIG.
2
. As can be seen in video data transmission sequence
200
of
FIG. 2
, in a CBR system such as system
100
, although the size of the bitstream for each consecutive frame may vary, the average size of the frames is equal to the maximum average frame size achievable via channel
120
, so that the entire channel capacity is used to maximize picture quality and to avoid over- or underflowing buffer
131
. Overflowing or underflowing buffer
131
is referred to as a “buffer exception.” Overflow is usually worse than underflow, because underflow requires the decoder to wait for more data to arrive, which usually means a momentary freeze frame and thus some slight temporal distortion, while overflow causes data to be lost, which may result in the loss of all subsequent pictures in the GOP until the next I frame arrives.
Unlike systems in which the time required to transmit a frame from a given source to a given destination is constant (e.g. {fraction (1/30)}s), in a compressed data system such as system
100
the time required to carry the frame or bitstream segment for each frame across channel
120
can vary from great to small from one frame to the next, because the size of the encoded bitstream can vary with the type of frame and also depending on the complexity of the scene information in that frame, how successful the motion compensation block matching was for a given predicted frame, and so on. Unencoded video frame data in is provided on input line
115
to encoder
110
at a constant time per frame (e.g., one frame each {fraction (1/30)} second), and decoded video frames are provided as video out on output line
135
to monitor
132
at the same constant frame rate. However, the time per frame f
Burke William J.
Couso Yon J.
Sarnoff Corporation
LandOfFree
Bursty data transmission of compressed video data does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Bursty data transmission of compressed video data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Bursty data transmission of compressed video data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2981736