Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2000-03-07
2003-11-25
Philippe, Gims S. (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C382S239000, C370S465000
Reexamination Certificate
active
06654416
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a device and a process for regulating bit rate in a system for the statistical multiplexing of streams of images coded according to the MPEG 2 coding standard. This process and this device are integrated, for example, on each coder of a statistical multiplexing system.
BACKGROUND OF THE INVENTION
A system for coding according to the MPEG 2 video standard uses the properties of the signal to reduce the bit rate thereof.
The coding algorithm implemented describes the images blockwise by exploiting the spatial and temporal redundancies of the images to be coded.
The spatial redundancy is processed, mainly, by virtue of the succession of three operations: an operation commonly called the discrete cosine transform and denoted DCT, an operation of quantizing the coefficients emanating from the DCT and a variable-length coding operation VLC for describing the quantized coefficients emanating from the DCT.
The DCT transformation makes it possible to transform a two-dimensional spatial representation consisting of the digitized video image into a spectral representation in which one is unaware of the phase. This DCT transformation leads to the production of a matrix of the dimension of a block where each coefficient corresponds to a combination of two frequencies, horizontal and vertical, given that the coefficients for the low frequencies are located at the top left of the matrix.
The quantization of the coefficients emanating from the matrix of the DCT transformation consists in expressing these coefficients according to values which are multiples of the quantization interval adopted. In this regard, the coefficients which are less than the quantization interval or level are coded with the value zero. Finally, the variable-length coding consists in coding each coefficient of the matrix once quantized according to a coding of entropy type which associates the smallest binary code with the commonest values of quantized coefficients. The coefficients of the matrix are traversed in a zigzag, starting from the element located at the top left of the matrix (continuous component) and terminating at the one located at the bottom right of the matrix (maximum horizontal and vertical frequencies).
The coding of the coefficients of the matrix terminates when all the coefficients have been traversed or else when the last non-zero coefficient is reached (a specific escape character is then used). Thus, the more zero coefficients there are in the matrix on completion of the quantization, the smaller is the length of the binary train for coding an image, that is to say the volume of data transmitted. On the other hand, the smaller this volume of data, the smaller the quantity of information about the image to be coded, thereby incurring the risk of inducing the appearance of quantization artifacts visible during image reconstruction by the decoder.
The temporal redundancy is analysed via a motion compensation operation which consists in searching, via a translation operation in respect of each block of the current image, for the most similar block located in a reference image. Analysis of the temporal redundancy leads to the determination of a field of translation vectors, commonly called motion vectors, as well as of a prediction error which is the difference between the signal of the current image and the signal of the image predicted by motion compensation. The prediction error is then analysed according to the principle of spatial redundancy.
MPEG 2 coding is a coding of predictive type. It follows that the decoding procedure associated therewith must be regularly reinitialized so as to protect the signal against any transmission error or any signal break due to the toggling of the decoder from one program to another.
To this end, the MPEG 2 standard provides that, periodically, the images must be coded in spatial mode, that is to say according to a mode exploiting spatial redundancy only. The images coded in spatial mode are commonly called INTRA images or I images.
The images coded by exploiting temporal redundancy are of two types: there are, on the one hand the images constructed by reference to a temporally earlier image and on the other hand, the images constructed by reference to a temporally earlier image and to a temporally later image.
The coded images constructed by reference to a temporally earlier image are commonly called predicted images or P images and the coded images constructed by reference to a temporally earlier image and to a temporally later image are commonly called bidirectional images or B images.
An I image is decoded without reference being made to images other than itself. A P image is decoded by referring to the P or I image which precedes it. A B image is decoded by calling on the I or P image which precedes it and on the I or P image which follows it.
The periodicity of the I images defines a group of images widely denoted GOP (the acronym standing for “Group Of Pictures”).
As is known to the person skilled in the art, within a single GOP, the quantity of data contained in an I image is generally greater than the quantity of data contained in a P image and the quantity of data contained in a P image is generally greater than the quantity of data contained in a B image.
In order to manage this disparity in the quantities of data according to the type of image, an MPEG 2 coder comprises a device for slaving the data bit rate.
Such a slaving device makes it possible to control the stream of coded data. It comprises a buffer memory, for storing the coded data, and models the state of the dual buffer memory of a so-called reference decoder. The slaving device smoothes the bit rate of the data exiting the buffer memory in such a way that the sum of the data contained in the coder and in the reference decoder is constant.
Thus, as a function of the type of image (I, P or B) this involves managing the fact that the I images produce a bit rate greater than the mean bit rate (typically equal to 3 to 6 times the mean bit rate), that the P images produce a bit rate of close to the mean bit rate and that the B images produce a bit rate less than the mean bit rate (typically equal to 0.1 to 0.5 times the mean bit rate).
Furthermore, when the coded video signals of a program are intended to be multiplexed with, other signals of the same type from other program, it is known practice to share the global bit rate of the multiplex between the various signals. This configuration is found, for example, during broadcasting of video program by satellite. In this case, the bit rate of the multiplex may reach 40 MB/s, thus permitting the transport of several programs simultaneously (10 programs at 4 MB/s each for example).
A video program emanating from an MPEG 2 type coding at fixed bit rate exhibits, after decoding, a variation in the quality of the image restored. This stems from the variability of the entropy of the video signal over time, this variability being evidenced by a fluctuation in the quantization level of the DCT coefficients.
An appropriate allocation of the bit rates associated with the video program then makes it possible to globally increase the quality of the video programs as a whole and/or to raise the number of programs broadcast. According to the prior art, the result of coding the GOP of order k is then used as prediction of the expected coding difficulty for the GOP of order k+1.
The cost of coding each image is variable. It is the role of the regulating loop to correct the stream variations due to the variations in the complexity of the image (defined hereinbelow) and to transmit as output a stream of data corresponding to a stipulated bit rate. This regulation generally acts on the quantization interval which is dependent on the level of fill of the buffer memory at the output of the coder. This buffer memory makes it possible to absorb the bit rate deviations due to the variations in volume of the coded images depending on their type (I, P or B) as well as to the bit rate deviations due to the resp
Alexandre Patrice
Diascorn Jean-Louis
Lavenan Thierry
Kurdyla Ronald H.
Laks Joseph J.
Philippe Gims S.
Thomson Licensing S.A.
Tripoli Joseph S.
LandOfFree
Device and process for regulating bit rate in system for the... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Device and process for regulating bit rate in system for the..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Device and process for regulating bit rate in system for the... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3167201