Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-04-16
2001-05-15
Le, Vu (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240020
Reexamination Certificate
active
06233282
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the fields of digital video data compression and storage. More specifically, the present invention teaches methods and apparatus for providing reduced bit rate digital video formats.
BACKGROUND OF THE INVENTION
The quality of a digital video image is a function of its resolution, which may be measured in pixels per image. As the number of pixels increases, the resolution of the image increases. As the resolution of an image increases, the amount of digital video data required to store or transmit the image correspondingly increases.
To reduce the number of digital bits required to store and/or transmit an image, compression is often performed. By way of background, current video compression techniques include two major categories: entropy processes, or information preserving processes, and information losing, or so called “lossy” processes. Entropy processes such as run length encoding (RLE) introduce no errors in the encoding/decoding process so that the original signal may be reconstructed exactly. Entropy processes tend to have a small compression ratio, i.e., a small reduction in video bit rate. In contrast, “lossy” compression processes tend to introduce errors in the encoding/decoding process but achieve much higher compression ratios. To strike an acceptable compromise between quality and compression ratio, digital video compression processes typically combine both entropy coding processes and “lossy” coding processes.
A variety of standards have emerged in the video industry for digital video compression. Digital Video (DV) is one such compression standard. The DV compression standard is commonly employed for compressing digital video data generated by, for example, a DV camcorder for storage on digital video tape (DV tape). As will be appreciated, DV is similar to the MPEG encoding technique in that it contains both audio and video data. However, while MPEG has intra-frame compression thereby reducing the redundancy from frame-to-frame, DV, like motion JPEG, has only interframe compression. A further important characteristic of the DV format is that DV maintains a fixed bit rate of about 3.5 Mbytes per second.
According to the DV format, each image frame is converted into a plurality of blocks of a suitable size, i.e., blocks of 8 by 8 pixels. In the DV format, two basic forms of a discrete cosine transform (DCT) are used to transform these blocks into frequency domain components, which may then be encoded into a compressed format. These include a still DCT process used for a still type sample block and a motion DCT process used for a motion type sample block. For simplicity's sake, the following description is limited to the still type sample block.
FIG. 1
represents diagramatically a DCT based video encoding process for a still type 8×8 video sample block
108
. The still type 8×8 sample block
108
includes sixty-four (64) samples
110
. In the DV format, the still type 8×8 sample block
108
is transformed, using a DCT function, into an 8×8 DCT block
112
. The 8×8 DCT block
112
has 64 spatial frequency patterns including a DC spatial frequency pattern
114
and 63 AC spatial frequency patterns
116
.
The DC spatial frequency pattern
114
is located in row zero, column zero of the 8×8 DCT block
112
. The DC spatial frequency pattern
114
has a DC coefficient value and each of the 63 AC spatial frequency patterns
116
has an AC coefficient value. The DC coefficient value of the DCT block
112
is equal to the average of each of the AC coefficient values of the DCT block
112
.
In order to achieve the required bit rate yet optimize picture quality and integrity, the DCT blocks of each image frame in the DV format are encoded into a plurality of fixed length video segments, the video segments each comprising five compressed macroblocks and certain header information. As will be appreciated, each compressed macroblock (CMB) represents a discrete portion of the frame.
FIG. 2
illustrates a typical CMB
120
. The CMB
120
includes header information
122
, four luminance DCT blocks Y
0
, Y
1
, Y
2
, and Y
3
and two chrominance DCT blocks Cr and Cb. The luminance and chrominance blocks have been transformed by a DCT function as described above with reference to FIG.
1
. Further, to achieve the fixed data bit rate of 3.5 Mbytes described above, the luminance DCT blocks Y
0
, Y
1
, Y
2
, and Y
3
each have a predefined storage capacity of 14 bytes, 100 bits of which is dedicated to AC frequency values, and the chrominance DCT blocks Cr and Cb each have a predefined storage capacity of 10 bytes, 68 bits of which is dedicated to AC frequency values.
To create the video segment, a shuffling process first selects five CMBs such as CMB
120
. As will be appreciated by those skilled in the art, shuffling means that five discrete CMB from roving positions on the frame are selected, one from the center portion of the frame, and the other four from the frame's four comers, respectively. Next, the values of the DCT blocks are “quantized,” that is, each value of the DCT matrix is divided by a predetermined number. Because certain frequencies may not appear in the block, or may have small values, the quantized value is often zero. It should be noted that this is a lossy compression method, since every time you quantize you are likely to lose information.
After quantization, the DV format encoding process implements a run length encoding (RLE) compression technique. In the DV format, specific RLE codes are used to compress a digital video bit stream by taking advantage of repetitive patterns of zeros and ones. After implementing the RLE compression, the RLE data is distributed throughout each video segment according to a protocol that takes advantage of any additional space available within the various CMBs. Specifically, the RLE data is distributed throughout the video segment using three passes. In the first pass, RLE data is stored in the allotted areas for each DCT block in a video segment. The second pass finds unused areas in each CMB
120
and stores further RLE bits into those areas. The third pass finds any free space in the video segment and stores any remaining RLE bits in that space until the space runs out or until there are no more bits left.
The DV compression factor is thus a function of the quantization value. DV utilizes this by adjusting the quantization factor for each video frame in order to obtain the fixed bit rate of 3.5 Mbytes per second.
Although the DV format is an immensely popular digital video format, three problems arise under the current DV format. First, the 3.5 Mbyte fixed data rate pushes or exceeds the limit for many computer systems. This results primarily because the storage medium (e.g., the hard disk drive) cannot read or write at the fixed data rate. Second, the DV format is space intensive, requiring significant storage capabilities for working with the data. Third, decompression and compression of the DV data is prohibitively slow, making it difficult for video editors to work with the data unless they have a powerful computing system.
What is needed is a technique for converting, real time, back and forth between DV format and a lower, reduced bit rate (less resolution) digital video compression format. Such a conversion technique would enable the use of video equipment that adheres to the DV formnat together with computer systems that cannot handle the DV data rate.
SUMMARY OF THE INVENTION
In order to achieve the foregoing objectives, the present invention teaches a variety of methods and apparatus for providing reduced bit rate digital video formats. A first method of the present invention teaches encoding digital video data into a second fixed bit rate digital video format that is slower than a first fixed bit rate digital video format in which the digital video data was initially encoded. More specifically, the method retrieves at least one data frame encoded according to the first fixed bit rate digital video format. The at l
Adaptec, Inc.
Le Vu
Martine Penilla & Kim LLP
LandOfFree
Methods and apparatus for providing reduced bit rate digital... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods and apparatus for providing reduced bit rate digital..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods and apparatus for providing reduced bit rate digital... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2543778