Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-11-30
2001-03-06
Lee, Young (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
Reexamination Certificate
active
06198769
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for determining a target amount of code which may be used in a digital video disc (DVD) or the like or in a system for transmitting digital video broadcasting (DVB) digital data, and to a method and apparatus for compressing and coding noncompressed video data.
Techniques for reducing the amount of data during recording by converting video and audio signals to digital data and applying conversion and coding processing to the digital data have been standardized by MPEG (Moving Pictures Expert Group). Such processing may involve inter- or intra-coding of the video data for every macro block. In MPEG, a group of pictures (GOP) may include one frame or picture subjected to intra-coding (an I-picture), and a plurality of frames or pictures subjected to inter-coding which includes predictive coding (P-pictures) and bidirectional predictive coding (B-pictures).
In inter-coding processing, coding processing is applied to the difference between the video data of a current frame and the video data of a past frame. As is to be appreciated, the video data of the current frame may be easily obtained or restored if the video data of the past frame has been already sent.
In determining the above-described difference between frames of video data, motion detection and motion compensation processing may be performed.
Motion detection processing involves detecting or determining the position or macroblock in a past reference frame in which the sum of absolute values of differences of pixel values between such macroblock and the respective macroblock of the current frame is the smallest. Such detection may be performed by determining the number of pixels the macroblock needs to be moved from the position of the respective current frame macroblock in an X- and Y-direction such that the sum of absolute values of differences of the pixel values between the respective current frame macroblock and the reference frame macroblock is the smallest. Data representative of such amount of movement in units of pixels in the X- and Y-directions may be referred to as motion vector data.
Motion compensation processing involves determining a position based upon the motion vector data and the position of the past reference frame macroblock and extracting macroblock data pertaining to such position. Thereafter, the difference between the extracted past reference macroblock and the current respective macroblock, or the motion predictive error, is determined. The past reference frame may be obtained from a local decoder.
The motion predictive error signal may be subjected to orthogonal transformation, such as discrete cosine transformation (DCT) in units of predetermined blocks so as to remove the correlation in the space direction. From such DCT processing, transformed DCT coefficient(s) may be obtained which are subjected to quantization processing (that is, division by a quantization step) to eliminate fine signals whereafter the whole number value of such quotient and a quantization index are outputted. The quantized DCT coefficient(s), the quantized index, and the motion vector data may be variable length coded (VLC) and outputted.
In intra-coding processing, on the other hand, the above-described motion detection and motion compensation processing are not performed. Instead, pixel values are orthogonal transformed in block units, quantized, variable length coded, and outputted.
The above-described processing removes the redundancy in the time and space directions of the video data, thereby compressing the information data. Such compression enables image and audio data to be recorded on a single optical disk (such as a digital video disc-DVD), or a moving picture and audio data to be transmitted within the transmission line capacity of a telecommunications line or a satellite line.
Consider a situation wherein a DVD, a ROM disk, or the like is used. In such situation, the compressed and coded video data should be contained within the capacity of the disc. To accomplish such requirement, so-called variable rate coding may be utilized wherein the amounts of codes are allocated within the range of the amount of code so as to obtain the highest image quality. Such coding method utilizes or accounts for the fact that difficulty in image compression may fluctuate with time due to constant changes of the correlation strength of the video data in the time and space directions.
An example of the above-described fluctuation or change of correlation strength in the time direction will now be presented. A pattern of movement of a moving body in a moving picture may include not only simple parallel movement, but also complex movement patterns such as a change of movement speed, movement direction of the moving body, and shape of the moving body. In such a situation, the motion predictive error data may be increased by motion compensation in a manner as previously described. Alternatively, in moving picture data having little or no correlation in a time direction (such as random noise), the effect of compression by motion compensation may be negligible and motion predictive error data not much different in amount and/or data from that of the current frame may be generated. Such fluctuation in the difficulty of the image compression may exist not only with frame units, but also with other types of units such as GOP units, macroblock units, and so forth.
Accordingly, when compressing and coding data which fluctuates in difficulty (such as video data having a low correlation in the time and space directions) by a uniform bit rate (hereinafter, a “target code rate”) per unit time, the DCT coefficient(s) may be quantized by a relatively large quantization step so that the generated code rate approaches the target code rate. As such, a relatively large quantization error may result and noticeable or conspicuous coding distortion may occur when expanding and decoding the data. On the other hand, when compressing and coding data having a low fluctuation in difficulty (such as video data having a high correlation in the time and space directions), the DCT coefficient (s) may be concentrated at the low frequency component side so as to finely quantize the same by a relatively small quantization step so that the generated code rate approaches the target code rate or target code amount. As a result, distortion after expansion and decoding of the data may be reduced.
Therefore, in recording on DVDs or the like, to avoid coding distortion due to image fluctuation over time and the deterioration of the image quality, variable length coding using a so-called 2 path or 2-pass encoding technique may be utilized. In such 2-pass encoding technique, video data is compressed and coded by a fixed quantization step, and the amount of codes generated is defined as the difficulty data and stored in a first pass or path, and, during a second path or pass, compression and coding are performed based on the difficulty data such that a relatively large amount of code may be allocated to difficult video data and a relatively small amount of code may be allocated to non-difficult video data. (The term “difficult data” refers to the amount of data necessary to obtain a certain constant image quality.)
In the above-described 2-pass encoding technique, the actual encoding (second pass) commences after encoding all of the video data in the first pass. As a result, it is very difficult, if not impossible, to perform such technique on supplied video and audio data with no interruption(s). Due to such limitation, such 2-pass encoding technique may not be suitable for encoding at the time of communication and broadcasts in which relatively long data streams should be encoded without interruption and in real time.
Alternatively, a 2-path or 2-pass technique may be utilized which uses two encoders, that is, a first encoder for encoding data passing through a memory having an FIFO (first-in first-out) configuration and a second encoder for encoding the data not passing through such mem
Frommer William S.
Frommer Lawrence & Haug LLP.
Lee Young
Smid Dennis M.
Sony Corporation
LandOfFree
Method and apparatus for controlling a target amount of code... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for controlling a target amount of code..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for controlling a target amount of code... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2447796