Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2001-03-23
2004-10-05
Diep, Nhon (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C348S419100
Reexamination Certificate
active
06801572
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to a method and an apparatus for image signal encoding, and particularly relates to a method and an apparatus for image signal encoding, which encodes image signals at variable bit rates.
2. Description of the Related Art
Recently, such an international standard as MPEG (Moving Picture Experts Group) is employed in an image signal encoding apparatus. In such an apparatus, a VBV (Video Buffering Verifier) is virtually connected to a bitstream output point of the image signal encoding apparatus so that the bitstream shall not collapse a receiving buffer of the image signal decoding apparatus by preventing a collapse at the VBV.
FIG. 1
shows an example of VBV buffer occupancy index that varies with time. At a time A, a bitstream starts to flow into the VBV, and the occupancy increases as time elapses. At a time B, the occupancy decreases by data volume D which is the volume of the data of an image frame then decoded. From the time B to a time C, the buffer occupancy index increases. At the time C, the occupancy decreases by data volume E that is the volume of the data of an image frame then decoded. In
FIG. 1
, a slope of the buffer occupancy while the buffer occupancy index increases represents a transmission rate of the bitstream supplied to the VBV.
In the image signal encoding apparatus, it is necessary to control the encoding so that the buffer occupancy is between 0 and a given buffer size. If the control is not appropriate, then the VBV will collapse as shown in
FIGS. 2 and 3
.
FIG. 2
shows an example of an underflow that occurs at the VBV. At a time F, a bitstream starts flowing into the VBV, increasing the buffer occupancy index. At a time G, the occupancy decreases by data volume I which is the volume of the data of an image frame then decoded. The buffer occupancy index increases from the time G to a time H.
At the time H, not all of the bitstream for an image frame to be decoded has arrived. The image frame cannot be decoded. This situation is called a buffer underflow.
Conversely,
FIG. 3
shows an example of an overflow. The bit rate control method in a constant rate encoding mode is different from that in a variable bit rate encoding mode. The overflow of the VBV occurs in the constant bit rate encoding mode.
At a time K, a bitstream starts flowing into the VBV to increase the occupancy. At a time L, an image frame with data volume N is decoded, and decreases the buffer occupancy index by N. From the time L to a time M, the buffer occupancy index increases.
At the time M, an image frame is to be decoded. However, volume of data that was decoded at the time L was small, which causes the buffer occupancy index of the VBV to reach the given buffer size before the time M. For this reason, at the time M, not all of necessary bitstream has been received and a normal decoding is impossible. This situation is called an overflow.
FIG. 4
shows an example of the buffer occupancy index behavior with time in the variable bit rate encoding mode. In the variable encoding mode, there is no buffer overflow. When the buffer occupancy index reaches at the given buffer size, then the bitstream stops flowing. As the buffer occupancy index decreases, the bitstream resumes flowing.
At a time P, a bitstream starts flowing into the VBV, increasing the buffer occupancy index. At a time Q, an image frame of which data size is T is decoded, thereby decreasing the occupancy index by T. From the time Q to a time R, the buffer occupancy index increases.
At the time R, the buffer occupancy index reaches at the given buffer size, thereby stopping the bitstream flowing to the VBV. At a time S, an image frame of which data volume is U is decoded, which makes the buffer occupancy index decrease by U. Then, the bitstream to the VBV resumes flowing.
The overflow and underflow of the buffer occur where the originated data volume control at encoding is inappropriate. A proper control of the originated data volume at encoding is necessary. In order to control the originated data volume at encoding, an appropriate target data volume is set to an image frame. The originated data volume is controlled so as to meet the target volume.
Conventionally, such a control as follows has been practiced in order to make the originated data volume to approximate the target volume. In the following description of an example, T represents the target data volume of an image frame, d
0
represents an initial occupancy of a quantization control buffer, d
j
represents an occupancy of the quantization control buffer immediately before encoding an encoding block (macroblock) j, B
j
represents an originated data volume from a first encoding block (macroblock) of the image frame to the j-th encoding block (macroblock), MB
cnt
represents a number of encoding blocks (macroblocks) in an image frame, Q represents a quantization scale code by which the j-th encoding block (macroblock) is quantized and r represents an reaction parameter.
The reaction parameter r is given by a formula (1) presented below by a frame rate of a moving image “picture_rate” and an encoding bit rate “bitrate”.
r=
2×(bitrate)/(picture_rate) (1)
An encoding block (macroblock) is made of a plurality of pixels. As shown in
FIG. 5
, the occupancy of the quantization control buffer when encoding the j-th encoding block (macroblock) is given by a formula (2) below. In
FIG. 5
, the hatched blocks represent encoded macroblocks and others represent macroblocks that have not been encoded.
d
j
=d
0
+B
j−1
−(
T
×(
j
−1))/
MB
cnt
(2)
Thus calculated quantization control buffer occupancy index is substituted to a formula (3) below, to obtain a quantization scale code by which the j-th macroblock is to be quantized.
Q=d
j
×31
/r
(3)
The scale code Q thus calculated by the formula (3) is substituted to a formula (4) below to obtain the quantization scale QS.
QS=Q×
2 (4)
By performing the process described above for all the macroblocks in the image frame, the originated data volume approximates to the target data volume. If the originated data volume is greater than the target data volume, then the quantization scale QS is enlarged to reduce the originated data volume, and if the originated data volume is smaller than the target data volume, then the quantization scale QS is made smaller to increase the originated data volume. In this manner, the receiving buffer collapse has been suppressed.
In a conventional image signal encoding apparatus, the constant bit rate encoding mode has been employed, which transfers encoded image signal at a constant encoding bit rate due to requirements from networks or the like. In the constant bit rate encoding mode, however, there is a tendency to build picture quality differences between images that are difficult to compress because of complexity and abrupt motions in the image and relatively still images that are easy to compress. On the other hand, in the variable bit rate encoding mode, a higher encoding bit rate is used for images that are difficult to compress and a lower encoding bit rate is used for images that are easy to compress so that there is no quality difference between the hard-to-compress images and the easy-to-compress images.
With recent availability of wideband asynchronous networks such as IP (Internet Protocol) networks and developments of storage media compliant with the variable bit rate encoding mode, the variable bit rate encoding mode for compressed image signals has been developed and put into practice.
A variable bit rate encoding mode used in DVD (Digital Versatile Disk) is called an n-path variable bit rate encoding method and requires two or more encoding steps to compress-encode moving pictures. For example, 2-path variable bit rate encoding method first determines a degree of difficulty in compression of each compressed image frame by encoding the frame at a constant compression ratio. T
Sakai Kiyoshi
Yamada Kohji
Diep Nhon
Fujitsu Limited
Staas & Halsey , LLP
LandOfFree
Method and apparatus for image signal encoding 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 image signal encoding, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for image signal encoding will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3264308