Variable bit-rate encoding device

Details Variable bit-rate encoding device Variable bit-rate encoding device

2000-09-14

2002-12-31

Diep, Nhon (Department: 2613)

Pulse or digital communications

Bandwidth reduction or expansion

Television or motion video signal

C375S240160, C348S419100

Reexamination Certificate

active

06501800

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates to a variable bit-rate encoding device. More particularly, the present invention relates to a two-pass-system encoding device that encodes moving picture (image) data at a variable bit-rate.
The moving image data encoding technique has a constant bit-rate encoding method and a variable bit-rate encoding method.
FIG. 3
is a diagram illustrating the relationship between time and encoded bit-rate in the case where a certain set of moving image data is subjected to the constant bit-rate encoding and the variable bit-rate encoding.
In the constant bit rate encoding shown in
FIG. 3
, the bit rate is always constant. On the other hand, in the variable bit-rate encoding, the bit rate varies due to the complexity of image data. Since the duration (a) is for image data relatively difficult to be handled, encoding is performed at a high bit rate. Since the duration (b
1
) and the duration (b
2
) are for image data relatively easy to be handled, encoding is performed at a low bit rate.
In the variable bit-rate encoding, a large quantity of codes is allocated to complicated images while a small quantity of codes is allocated to simple images. Consequently, moving image data can be encoded with the image quality maintained uniformly.
In the constant bit-rate encoding, information may not be sufficient in the duration (a) in which much information are produced as shown in the area A shown in FIG.
3
. Information may be encoded redundantly in the duration (b
1
) or (b
2
) in which less information are produced, as shown in the area B shown in FIG.
3
.
Accordingly, it may be generally said that the variable bit-rate encoding method is effective, in regard to the image quality.
The so-called two-pass system variable bit rate encoding scheme is known as one of variable bit-rate encoding schemes.
In the two-pass system variable bit-rate encoding method, an information amount generated per unit section of all sequences of moving images is analyzed in the first pass and a target code amount for each unit section is allocated based on the analyzed result. Thereafter, while the generated code amount for each unit section in the second pass is controlled to bring close to a target code amount, an actual encoding is performed.
As described above, the two-path system variable bit-rate encoding method is characterized in that an ideal variable bit rate control is performed because a target code amount for each unit section is allocated by analyzing the information generation amount of all sequences for a moving image, prior to an actual encoding.
FIG. 11
is a block diagram illustrating a conventional two-pass variable bit-rate encoding device.
As shown in
FIG. 11
, the conventional variable bit rate encoding device consists of an input buffer
202
with an input terminal
201
, a subtracter
203
, a DCT (Discrete Cosine Transform) unit
204
, a quantizer
205
, a variable length encoder
206
, an inverse quantizer
207
, an inverse DCT unit
208
, an adder
209
, a frame memory
210
, a motion compensator
211
, a motion vector searcher
212
, a code amount allocator
213
, a code amount controller
214
, and an output terminal
215
.
Image data input from the input terminal
201
is supplied in line units to the input buffer
202
.
As shown in
FIG. 4
, input image data to encoded as one of an I (Intra) frame, a P (Prediction) frame, or a B (Bi-directional) frame.
I frame is a frame encoded in an inter frame process only and does not undergo an inter motion compensation frame predictive encoding. I frame also is used as a reference frame for the successive P and B frames.
P frame is a reference frame an I or P frame in the forward direction on the time axis and is used to perform an inter motion compensation frame predictive encoding in one direction. P frame is used as a reference frame for the other successive P and B frames.
B frame is a reference frame for the I or P frame in the forward direction and backward direction on the time axis and is used to perform an inter motion compensation predictive encoding in two directions. The B frame is not used in the reference frame of other frame.
The input buffer
202
rearranges input image data in an encoding order and outputs the output signal to the subtracter
203
and the motion vector searcher
212
in encoded block units or in macroblock units.
In
FIG. 4
as previously described, the third P frame, for example, is encoded earlier than the first and second B frames to use the third P frame as a reference frame. Hence, the input buffer
202
rearranges images.
FIG. 5
shows the rearrangement operation. The image data input shown with input order in
FIG. 5
are supplied in the order shown with encoding order in FIG.
5
.
The macroblock is a minimum unit for a motion compensation prediction and corresponds to image data of 16×16 pixels. The macroblock data is configured of the brightness block Y of 8×3 pixels and the color-difference blocks Cr and Cb. The total block number depends on a color-difference data thinned-out type.
When the thinned-out type is 4:2:0, as shown in
FIG. 6A
, there are 6 blocks including 4 brightness blocks Y, one color-difference block Cr, and one color-difference block Cb. When the thinned-out type is 4:2:2. as shown in
FIG. 6B
, there are 8 blocks including 4 brightness blocks Y, 2 color-difference blocks Cr, and 2 color-difference blocks Cb. When the thinned-out type is 4:4:4. as shown in FIG.
6
C. there are 12 blocks including 4 brightness blocks Y, 4′ color-difference blocks Cr and 4 color-difference blocks Cb.
The subtracter
203
receives compressed block data output from the input buffer
202
and reference macroblock data motion-compensated with the motion compensator
211
and then outputs these interframe predictive error signals to the DCT unit
204
.
The DCT unit
204
subjects each interframe predictive error signal to a DCT process to separate it into horizontal and vertical frequency components. Thus, the DCT unit
204
outputs the horizontal and vertical frequency components to the quantizer
205
.
The quantizer
205
quantizes the output from the DCT unit
204
to reduce the code amount. Thus, the quantizer
205
outputs the result to the variable length encoder
206
and the inverse quantizer
207
. The quantizer
205
performs a fixed quantization with the quantization step size of a fixed value in the first pass tentative encoding. The quantizer
205
also performs a fixed quantization with the quantization step size set by the code amount controller
214
in an actual second pass encoding.
The inverse quantizer
207
inverse-quantizes the output of the quantizer
205
and then outputs the result to the inverse DCT unit
208
.
The inverse DCT unit
208
subjects the output of the inverse quantizer
207
to an inverse DCT process and then outputs the result to the adder
209
.
The adder
209
adds the reference macroblock data motion-compensated by the motion compensator
211
and the output of the inverse DCT unit
208
. Then, the adder
209
outputs the resultant sum to the frame memory
210
.
The frame memory
210
outputs macroblock data acting as new reference frames data to the motion compensator
211
and the motion vector searcher
212
.
The motion vector searcher
212
receives the macroblock data of a compressed frame from the input buffer
202
as well as the macroblock data of a reference frame from the frame memory
210
. The motion vector seacher
212
implements the block matching between the input macroblock data sets to obtain a motion vector between the reference frame and the compressed frames and then outputs it to the motion compensator
211
.
The frame memory
210
provides reference frame data to the motion compensator
211
. The motion compensator
211
motion-compensates the reference frame data with the motion vector from the motion vector searcher
212
to create reference macro block data. Thus, the motion compensator
211
supplies it to the subtracter
203
.
The variable length encoder
206
su

Affiliated with

Also associated with

Rating

Variable bit-rate encoding device does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Variable bit-rate encoding device, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Variable bit-rate encoding device will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2987720