U.S. patent application number 15/940393 was filed with the patent office on 2018-10-04 for zero block detection using adaptive rate model.
The applicant listed for this patent is QUALCOMM Incorporated. Invention is credited to In Suk Chong, Marta Karczewicz, Xianglin Wang, Yang Yu.
Application Number | 20180288420 15/940393 |
Document ID | / |
Family ID | 63670176 |
Filed Date | 2018-10-04 |
United States Patent
Application |
20180288420 |
Kind Code |
A1 |
Yu; Yang ; et al. |
October 4, 2018 |
ZERO BLOCK DETECTION USING ADAPTIVE RATE MODEL
Abstract
A video coding device may be configured to estimate, based on a
combination of a first parameter and a number of non-zero
coefficients in a frame, a number of bits for non-zero coefficients
of the frame, to encode the frame based on the estimated number of
bits for the non-zero coefficients, to collect an actual number of
bits used to encode the non-zero coefficients of the frame and an
actual number of the non-zero coefficients in the frame, to update,
based on the actual number of bits used to encode the non-zero
coefficients of the frame and the actual number of non-zero
coefficients in the frame, only the first parameter to form an
updated first parameter, to form a rate estimation model using the
updated first parameter and a second parameter, and to select,
based on the rate estimation model, a coding mode for each block in
the frame.
Inventors: |
Yu; Yang; (San Diego,
CA) ; Chong; In Suk; (San Diego, CA) ; Wang;
Xianglin; (San Diego, CA) ; Karczewicz; Marta;
(San Diego, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
QUALCOMM Incorporated |
San Diego |
CA |
US |
|
|
Family ID: |
63670176 |
Appl. No.: |
15/940393 |
Filed: |
March 29, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62478924 |
Mar 30, 2017 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04N 19/147 20141101;
H04N 19/176 20141101; H04N 19/647 20141101; H04N 19/124 20141101;
H04N 19/149 20141101; H04N 19/14 20141101; H04N 19/194 20141101;
H04N 19/132 20141101; H04N 19/172 20141101 |
International
Class: |
H04N 19/147 20060101
H04N019/147; H04N 19/124 20060101 H04N019/124; H04N 19/176 20060101
H04N019/176; H04N 19/64 20060101 H04N019/64 |
Claims
1. A method of encoding video data, the method comprising:
determining an amount of movement in a frame included in a sequence
of frames; selecting a parameter set for a rate estimation model
based on the determined amount of movement in the frame; and
encoding the frame based on the rate estimation model.
2. The method of claim 1, wherein the frame is a current frame of
the sequence of frames.
3. The method of claim 1, wherein the frame is a previously-encoded
frame of the sequence of frames.
4. The method of claim 1, further comprising identifying, from a
plurality of transform units of the frame, one or more zero
transform units to be zeroed out.
5. The method of claim 4, wherein identifying the one or more zero
transform units to be zeroed out comprises comparing a first rate
distortion cost associated with zeroing out each respective
transform unit of the plurality against a second rate distortion
cost associated with not zeroing out each respective transform unit
of the plurality.
6. The method of claim 5, further comprising estimating the second
rate distortion cost based on a sum of a distortion when the
coefficients of the respective transform unit are encoded and a
Lagrangian multiplier multiplied by a number of bits required to
encode coefficients of the respective transform unit.
7. The method of claim 4, further comprising applying the rate
estimation model to estimate a number of bits to encode a number of
non-zero coefficients in each respective transform unit of the
plurality of transform units.
8. The method of claim 1, wherein selecting the parameter set for
the rate estimation model comprises selecting the parameter set
further based on a quantization parameter (QP) value.
9. The method of claim 1, further comprising: applying the rate
estimation model to estimate a number of bits to encode a number of
non-zero coefficients in a transform unit based on a product of a
first parameter of the parameter set multiplied by a sum of a
number of non-zero coefficients in the frame added to a second
parameter of the parameter set; and selecting, based on the
application of the rate estimation model, from a plurality of
available coding modes, an optimal coding mode with respect each
respective block of the frame.
10. The method of claim 9, wherein selecting the parameter set
comprises: collecting a number of non-zero coefficients of a
previously-encoded frame of the sequence of frames; collecting a
number of bits used to encode the non-zero coefficients of the
previously-encoded frame; and updating only a first parameter of
the parameter set based on the number of bits used to encode the
non-zero coefficients of the previously-encoded frame and the
number of non-zero coefficients of the previously-encoded frame to
form the adaptively selected parameter set.
11. The method of claim 10, wherein updating only the first
parameter of the parameter set comprises updating the first
parameter based on a weighted average of a previous value of the
first parameter and a newly determined value for the first
parameter.
12. The method of claim 1, wherein determining the amount of
movement in the frame comprises collecting: a number of
inter-encoded blocks used to encode a previously-encoded frame of
the sequence, and a degree of motion of the frame of the
sequence.
13. The method of claim 12, wherein determining the amount of
movement in the frame comprises determining that both of a first
condition and a second condition are true, wherein the first
condition comprises the number of inter-encoded blocks used to
encode the previously-encoded frame being greater than a first
threshold value, and wherein the second condition comprises the
degree of motion of the frame being less than a product of the
number of inter-encoded blocks used to encode the frame multiplied
by a second threshold value.
14. The method of claim 13, wherein selecting the parameter set for
the rate model based on an amount of movement in the frame or the
previous frame comprises performing one of: selecting a first
parameter set associated with a static sequence based on both of
the first condition and the second condition being true; or
selecting a second parameter set associated with a non-static
sequence based on at least one of the first condition or the second
condition being false.
15. The method of claim 1, wherein the frame is a current frame,
and wherein determining the amount of movement in the current frame
comprises estimating horizontal and vertical motion vectors and a
distance metric determining the distance between the current frame
and a previously-encoded frame.
16. The method of claim 15, wherein determining the amount of
movement in the current frame comprises determining whether at
least one of a first condition or a second condition is true,
wherein the first condition comprises a sum of horizontal motion
vectors and a sum of vertical motion vectors both being less than a
first threshold value, and wherein the second condition comprises
the number of blocks with zero motion being greater than a second
threshold value and the distance metric being less than a third
threshold value.
17. The method of claim 16, wherein selecting the parameter set
comprises performing one of: selecting a first parameter set
associated with a static sequence based on at least one of the
first condition or the second condition being true; or selecting a
second parameter set associated with a non-static sequence based on
neither the first condition nor the second condition being
true.
18. A video encoding device comprising: a video data memory
configured to store a sequence of frames of video data; and
processing circuitry in communication with the video data memory,
the processing circuitry being configured to: determine an amount
of movement in a frame included in the sequence of frames stored to
the video data memory; select a parameter set for a rate estimation
model based on the determined amount of movement in the frame; and
encode the frame based on the rate estimation model.
19. The video encoding device of claim 18, further comprising an
output interface in communication with the processing circuitry,
wherein the processing circuitry is configured to signal the
encoded frame via the output circuitry.
20. The video encoding device of claim 18, wherein the processing
circuitry is further configured to identify, from a plurality of
transform units of the frame, one or more zero transform units to
be zeroed out.
21. The video encoding device of claim 20, wherein to identify the
one or more zero transform units to be zeroed out, the processing
circuitry is configured to compare a first rate distortion cost
associated with zeroing out each respective transform unit of the
plurality against a second rate distortion cost associated with not
zeroing out each respective transform unit of the plurality.
22. The video encoding device of claim 21, wherein the processing
circuitry is further configured to estimate the second rate
distortion cost based on a sum of a distortion when the
coefficients of the respective transform unit are encoded and a
Lagrangian multiplier multiplied by a number of bits required to
encode coefficients of the respective transform unit.
23. The video encoding device of claim 18, wherein the processing
circuitry is further configured to: apply the rate estimation model
to estimate a number of bits to encode a number of non-zero
coefficients in a transform unit based on a product of a first
parameter of the parameter set multiplied by a sum of a number of
non-zero coefficients in the frame added to a second parameter of
the parameter set; and select, based on the application of the rate
estimation model, from a plurality of available coding modes, an
optimal coding mode with respect each respective block of the
frame.
24. The video encoding device of claim 18, wherein to determine the
amount of movement in the frame, the processing circuitry is
configured to collect: a number of inter-encoded blocks used to
encode a previously-encoded frame of the sequence, and a degree of
motion of the frame of the sequence,
25. The video encoding device of claim 24, wherein to determine the
amount of movement in the frame, the processing circuitry is
configured to determine that both of a first condition and a second
condition are true, wherein the first condition comprises the
number of inter-encoded blocks used to encode the
previously-encoded frame being greater than a first threshold
value, and wherein the second condition comprises the degree of
motion of the frame being less than a product of the number of
inter-encoded blocks used to encode the frame multiplied by a
second threshold value.
26. The video encoding device of claim 25, wherein to select the
parameter set for the rate model based on an amount of movement in
the frame or the previous frame, the processing circuitry is
configured to one of: select a first parameter set associated with
a static sequence based on both of the first condition and the
second condition being true; or select a second parameter set
associated with a non-static sequence based on at least one of the
first condition or the second condition being false.
27. The video encoding device of claim 18, wherein the frame is a
current frame, and wherein to determine the amount of movement in
the current frame, the processing circuitry is configured to
estimate horizontal and vertical motion vectors and a distance
metric determining the distance between the current frame and a
previously-encoded frame.
28. The video encoding device of claim 27, wherein to determine the
amount of movement in the current frame, the processing circuitry
is configured to determine whether at least one of a first
condition or a second condition is true, wherein the first
condition comprises a sum of horizontal motion vectors and a sum of
vertical motion vectors both being less than a first threshold
value, and wherein the second condition comprises the number of
blocks with zero motion being greater than a second threshold value
and the distance metric being less than a third threshold
value.
29. An apparatus for encoding video data, the apparatus comprising:
means for determining an amount of movement in a frame included in
a sequence of frames; means for selecting a parameter set for a
rate estimation model based on the determined amount of movement in
the frame; and means for encoding the frame based on the rate
estimation model.
30. A non-transitory computer-readable storage medium encoded with
instructions that, when executed, cause one or more processors of a
video encoding device to: determine an amount of movement in a
frame included in a sequence of frames; select a parameter set for
a rate estimation model based on the determined amount of movement
in the frame; and encode the frame based on the rate estimation
model.
Description
[0001] This application claims the benefit of U.S. Provisional
Patent Application 62/478,924, filed 30 Mar. 2017, the entire
content of which is incorporated by reference.
TECHNICAL FIELD
[0002] This disclosure relates to video encoding.
BACKGROUND
[0003] Digital video capabilities can be incorporated into a wide
range of devices, including digital televisions, digital direct
broadcast systems, wireless broadcast systems, personal digital
assistants (PDAs), laptop or desktop computers, tablet computers,
e-book readers, digital cameras, digital recording devices, digital
media players, video gaming devices, video game consoles, cellular
or satellite radio telephones, so-called "smart phones," video
teleconferencing devices, video streaming devices, and the like.
Digital video devices implement video compression techniques, such
as those described in the standards defined by MPEG-2, MPEG-4,
ITU-T H.263, ITU-T H.264/MPEG-4, Part 10, Advanced Video Coding
(AVC), the High Efficiency Video Coding (HEVC) standard presently
under development, and extensions of such standards. The video
devices may transmit, receive, encode, decode, and/or store digital
video information more efficiently by implementing such video
compression techniques.
[0004] Video compression techniques perform spatial (intra-picture)
prediction and/or temporal (inter-picture) prediction to reduce or
remove redundancy inherent in video sequences. For block-based
video coding, a video slice (i.e., a video frame or a portion of a
video frame) may be partitioned into video blocks, which may also
be referred to as CTUs, coding units (CUs) and/or coding nodes.
Video blocks in an intra-coded (I) slice of a picture are encoded
using spatial prediction with respect to reference samples in
neighboring blocks in the same picture. Video blocks in an
inter-coded (P or B) slice of a picture may use spatial prediction
with respect to reference samples in neighboring blocks in the same
picture or temporal prediction with respect to reference samples in
other reference pictures. Pictures may be referred to as frames,
and reference pictures may be referred to a reference frames.
[0005] Spatial or temporal prediction results in a predictive block
for a block to be coded. Residual data represents pixel differences
between the original block to be coded and the predictive block. An
inter-coded block is encoded according to a motion vector that
points to a block of reference samples forming the predictive
block, and the residual data indicating the difference between the
coded block and the predictive block. An intra-coded block is
encoded according to an intra-coding mode and the residual data.
For further compression, the residual data may be transformed from
the pixel domain to a transform domain, resulting in residual
transform coefficients, which then may be quantized. The quantized
transform coefficients, initially arranged in a two-dimensional
array, may be scanned in order to produce a one-dimensional vector
of transform coefficients, and entropy coding may be applied to
achieve even more compression.
SUMMARY
[0006] This disclosure describes techniques for zero block
detection using an adaptive rate model. Often, video frames are
highly correlated temporally, especially in sceneries such as video
telephony (VT) and video surveillance. This means the prediction
error (residual) is usually very small. Encoding techniques may
employ a model to decide whether to zero out a transform unit (TU)
or not based on a rate distortion (RD) cost. Techniques of the
current disclosure adaptively select parameters a and b of a rate
estimation model so that when a video sequence has a lot of noise,
a better coding efficiency may be achieved. Techniques of the
current disclosure may also improve the rate estimation model for
R.sub.coef to include consideration of coefficient positions so
that the estimated number of bits is a function of rate estimation
parameters (a and b), a number of non-zero coefficients in a block
(nnz), and coefficient positions (pos). Here, R.sub.coef represents
the number of bits used to code transform coefficients.
[0007] In one example of the disclosure, a method for encoding
video data includes estimating, based on a combination of a first
parameter and a number of non-zero coefficients in a frame, a
number of bits used by non-zero coefficients of a frame, encoding
the frame based on the estimated number of bits for the non-zero
coefficients of the frame, collecting an actual number of bits used
to encode the non-zero coefficients of the frame, collecting an
actual number of the non-zero coefficients in the frame, updating,
based on the actual number of bits used to encode the non-zero
coefficients of the frame and the actual number of non-zero
coefficients in the frame, only the first parameter to form an
updated first parameter, forming a rate estimation model using the
updated first parameter and a second parameter, and selecting,
based on the rate estimation model, a coding mode for each block in
the frame.
[0008] In another example of this disclosure, a video encoding
device includes a video data memory storing video data and
processing circuitry in communication with the video data memory.
The processing circuitry is configured to estimate, based on a
combination of a first parameter and a number of non-zero
coefficients in a frame, a number of bits used by non-zero
coefficients of a frame, encode the frame based on the estimated
number of bits for the non-zero coefficients of the frame, collect
an actual number of bits used to encode the non-zero coefficients
of the frame, collect an actual number of the non-zero coefficients
in the frame, update, based on the actual number of bits used to
encode the non-zero coefficients of the frame and the actual number
of non-zero coefficients in the frame, only the first parameter to
form an updated first parameter, form a rate estimation model using
the updated first parameter and a second parameter, and select,
based on the rate estimation model, a coding mode for each block in
the frame.
[0009] In another example of this disclosure, an apparatus for
encoding video data includes means for estimating, based on a
combination of a first parameter and a number of non-zero
coefficients in a frame, a number of bits used by non-zero
coefficients of a frame, means for encoding the frame based on the
estimated number of bits for the non-zero coefficients of the
frame, means for collecting an actual number of bits used to encode
the non-zero coefficients of the frame, means for collecting an
actual number of the non-zero coefficients in the frame, means for
updating, based on the actual number of bits used to encode the
non-zero coefficients of the frame and the actual number of
non-zero coefficients in the frame, only the first parameter to
form an updated first parameter, means for forming a rate
estimation model using the updated first parameter and a second
parameter, and means for selecting, based on the rate estimation
model, a coding mode for each block in the frame.
[0010] In another example of this disclosure, a non-transitory
computer-readable storage medium is encoded with instructions that,
when executed, cause one or more processors of a video encoding
device to estimate, based on a combination of a first parameter and
a number of non-zero coefficients in a frame of video data, a
number of bits for non-zero coefficients of the frame, to encode
the frame based on the estimated number of bits for the non-zero
coefficients of the frame, to collect an actual number of bits used
to encode the non-zero coefficients of the frame, to collect an
actual number of the non-zero coefficients in the frame, to update,
based on the actual number of bits used to encode the non-zero
coefficients of the frame and the actual number of non-zero
coefficients in the frame, only the first parameter to form an
updated first parameter, to form a rate estimation model using the
updated first parameter and a second parameter, and to select,
based on the rate estimation model, a coding mode for each block in
the frame.
[0011] In one example of this disclosure, a method for encoding
video data includes determining an amount of movement in a frame
included in a sequence of frames, selecting a parameter set for a
rate estimation model based on the determined amount of movement in
the frame, and encoding the frame based on the rate estimation
model.
[0012] In another example of this disclosure, a video encoding
device includes a video data memory configured to store video data
and processing circuitry in communication with the video data
memory. The processing circuitry is configured to determine an
amount of movement in a frame included in a sequence of frames, to
select a parameter set for a rate estimation model based on the
determined amount of movement in the frame, and to encode the frame
based on the rate estimation model.
[0013] In another example of this disclosure, an apparatus for
encoding video data includes means for determining an amount of
movement in a frame included in a sequence of frames, means for
selecting a parameter set for a rate estimation model based on the
determined amount of movement in the frame, and means for encoding
the frame based on the rate estimation model.
[0014] In another example of this disclosure, a non-transitory
computer-readable storage medium is encoded with instructions that,
when executed, cause one or more processors of a video encoding
device to determine an amount of movement in a frame included in a
sequence of frames, to select a parameter set for a rate estimation
model based on the determined amount of movement in the frame, and
to encode the frame based on the rate estimation model.
[0015] The details of one or more examples of the disclosure are
set forth in the accompanying drawings and the description below.
Other features, objects, and advantages of the disclosure will be
apparent from the description and drawings, and from the
claims.
BRIEF DESCRIPTION OF DRAWINGS
[0016] FIG. 1 is a block diagram illustrating an example video
encoding and decoding system that may implement the techniques
described in this disclosure.
[0017] FIG. 2 is a block diagram illustrating an example video
encoder that may implement the techniques described in this
disclosure.
[0018] FIG. 3 is a block diagram illustrating an example video
decoder.
[0019] FIG. 4 is a flowchart illustrating an example method of
adaptively selecting parameter set (a,b) according to the
techniques described in this disclosure.
[0020] FIG. 5 is a flowchart illustrating an example method of
selection of different parameter sets for static and non-static
sequences based on the previous frame's coding statistics.
[0021] FIG. 6 is a flowchart illustrating an example method of
selection of different parameter sets for static and non-static
sequences based on preprocessing of the current frame.
[0022] FIG. 7 is a flowchart illustrating an example method of the
disclosure.
[0023] FIG. 8 is a flowchart illustrating an example method of the
disclosure.
[0024] FIG. 9 is a block diagram illustrating an example video
encoder that may implement the techniques described in this
disclosure.
DETAILED DESCRIPTION
[0025] A block-based hybrid video coding framework is widely used
in video coding standards, such as MPEG-2, H.264/Advanced Video
Coding (AVC), and High Efficiency Video Coding (HEVC), and others.
For example, according to HEVC, a video encoder divides a video
frame into blocks of different sizes called coding units (CUs). The
video encoder may further split each CU into smaller blocks as
prediction units (PUs) for prediction and/or transform units (TUs)
for transform. The video encoder may predict each PU either within
the current frame in the case of intra mode, or from reference
frame(s) (e.g., temporally neighboring frame(s)) in the case of
inter mode. The video encoder may transform, quantize, and encode a
residual (also referred to as "prediction error") using entropy
coding techniques, such as Context Adaptive Binary Arithmetic
Coding (CABAC). In accordance with the techniques of this
disclosure, the video encoder may achieve high coding efficiency by
zeroing out blocks using an adaptive rate model. The video encoder
may be a hardware encoder or may encode using various other
techniques, e.g., software. Examples of video encoders configured
according to the techniques of this disclosure are described in
further detail below, with reference to the drawings.
[0026] In some aspects, this disclosure describes techniques for
bit rate control in a video encoding process. To achieve higher
coding efficiency, video coding standards such as H.264/AVC and
HEVC use Lagrangian rate distortion optimization to determine an
optimal mode, on a per-block basis, for encoding each coding block.
Specifically, a video encoder may, for every mode in a selection
set, calculate a rate distortion (RD) cost. In turn, the video
encoder may choose the mode with the minimum RD cost as the final
"optimal" mode with respect to the particular block. To achieve
higher coding efficiency, video encoders (such as an HEVC video
encoder) may selectively zero out some blocks by optimizing rate
distortion cost. That is, a video encoder may change all transform
coefficient values in one or more selected block(s) to zero. The
video encoder may use a linear model to estimate a number of bits
needed to signal coefficients of one block, using non-zero
coefficients in the block. The parameters used in the linear model
are fixed through configuration.
[0027] This disclosure describes techniques and device
configurations for adaptively deciding a rate estimation model. In
some examples, a video encoder of this disclosure may select the
rate estimation model based on whether a current frame includes
more or less than a threshold amount of motion, as compared to a
previous frame. In these and other examples, a video encoder of
this disclosure may update one of the parameters used in the rate
estimation model used for a given frame, based on the actual number
of bits to be used to encode non-zero coefficients of the frame and
the actual number of non-zero coefficients in the frame. In these
examples of this disclosure, such a revised rate estimation model
(also referred to herein as an "adaptive" rate model or adaptive
rate estimation model) may improve the coding efficiency of video
sequences, such as when encoding noisy camera sequences such as
internet protocol (IP) camera sequences. In some examples involving
IP camera sequences, the average Bjontegaard-Delta (BD)-rate is
shown to improve by approximately 7% when the encoding techniques
are enhanced using one or more aspects of this disclosure.
[0028] FIG. 1 is a block diagram illustrating an example video
encoding and decoding system 10 that may utilize the adaptive rate
model for zero block detection techniques described in this
disclosure. As shown in FIG. 1, system 10 includes a source device
12 that generates encoded video data to be decoded at a later time
by a destination device 14. Source device 12 and destination device
14 may comprise any of a wide range of devices, including desktop
computers, notebook (i.e., laptop) computers, tablet computers,
set-top boxes, telephone handsets such as so-called "smart" phones,
so-called "smart" pads, televisions, cameras, display devices,
digital media players, video gaming consoles, video streaming
device, or the like. In some cases, source device 12 and
destination device 14 may be equipped for wireless
communication.
[0029] Destination device 14 may receive the encoded video data to
be decoded via a link 16. Link 16 may comprise any type of medium
or device capable of moving the encoded video data from source
device 12 to destination device 14. In one example, link 16 may
comprise a communication medium to enable source device 12 to
transmit encoded video data directly to destination device 14 in
real-time. The encoded video data may be modulated according to a
communication standard, such as a wireless communication protocol,
and transmitted to destination device 14. The communication medium
may comprise any wireless or wired communication medium, such as a
radio frequency (RF) spectrum or one or more physical transmission
lines. The communication medium may form part of a packet-based
network, such as a local area network, a wide-area network, or a
global network such as the Internet. The communication medium may
include routers, switches, base stations, or any other equipment
that may be useful to facilitate communication from source device
12 to destination device 14.
[0030] Alternatively, encoded data may be output from output
interface 22 to a storage device 32. Similarly, encoded data may be
accessed from storage device 32 by input interface. Storage device
32 may include any of a variety of distributed or locally accessed
data storage media such as a hard drive, Blu-ray discs, DVDs,
CD-ROMs, flash memory, volatile or non-volatile memory, or any
other suitable digital storage media for storing encoded video
data. In a further example, storage device 32 may correspond to a
file server or another intermediate storage device that may hold
the encoded video generated by source device 12. Destination device
14 may access stored video data from storage device 32 via
streaming or download. The file server may be any type of server
capable of storing encoded video data and transmitting that encoded
video data to the destination device 14. Example file servers
include a web server (e.g., for a website), an FTP server, network
attached storage (NAS) devices, or a local disk drive. Destination
device 14 may access the encoded video data through any standard
data connection, including an Internet connection. This may include
a wireless channel (e.g., a Wi-Fi connection), a wired connection
(e.g., DSL, cable modem, etc.), or a combination of both that is
suitable for accessing encoded video data stored on a file server.
The transmission of encoded video data from storage device 32 may
be a streaming transmission, a download transmission, or a
combination of both.
[0031] The techniques of this disclosure to use an adaptive rate
model are not necessarily limited to wireless applications or
settings. The techniques may be applied to video coding in support
of any of a variety of multimedia applications, such as
over-the-air television broadcasts, cable television transmissions,
satellite television transmissions, streaming video transmissions,
e.g., via the Internet, encoding of digital video for storage on a
data storage medium, decoding of digital video stored on a data
storage medium, or other applications. In some examples, system 10
may be configured to support one-way or two-way video transmission
to support applications such as video streaming, video playback,
video broadcasting, and/or video telephony.
[0032] In the example of FIG. 1, source device 12 includes a video
source 18, video encoder 20 and an output interface 22. In some
cases, output interface 22 may include a modulator/demodulator
(modem) and/or a transmitter. In source device 12, video source 18
may include a source such as a video capture device, e.g., a video
camera, a video archive containing previously captured video, a
video feed interface to receive video from a video content
provider, and/or a computer graphics system for generating computer
graphics data as the source video, or a combination of such
sources. As one example, if video source 18 is a video camera,
source device 12 and destination device 14 may form so-called
camera phones or video phones. In another example, video source 18
may include an IP camera. However, the techniques described in this
disclosure may be applicable to video coding in general, and may be
applied to wireless and/or wired applications.
[0033] The captured, pre-captured, or computer-generated video may
be encoded by video encoder 20. The encoded video data may be
transmitted directly to destination device 14 via output interface
22 of source device 12. The encoded video data may also (or
alternatively) be stored onto storage device 32 for later access by
destination device 14 or other devices, for decoding and/or
playback.
[0034] Destination device 14 includes an input interface 28, a
video decoder 30, and a display device 31. In some cases, input
interface 28 may include a receiver and/or a modem. Input interface
28 of destination device 14 receives the encoded video data over
link 16. The encoded video data communicated over link 16, or
provided on storage device 32, may include a variety of syntax
elements generated by video encoder 20 for use by a video decoder,
such as video decoder 30, in decoding the video data. Such syntax
elements may be included with the encoded video data transmitted on
a communication medium, stored on a storage medium, or stored a
file server.
[0035] Display device 31 may be integrated with, or external to,
destination device 14. In some examples, destination device 14 may
include an integrated display device and be configured to interface
with an external display device. In other examples, destination
device 14 may be a display device. In general, display device 31
displays the decoded video data to a user, and may comprise any of
a variety of display devices such as a liquid crystal display
(LCD), a plasma display, an organic light emitting diode (OLED)
display, or another type of display device.
[0036] Video encoder 20 and video decoder 30 may operate according
to a video compression standard, such as the High Efficiency Video
Coding (HEVC) standard, and may conform to the HEVC Test Model
(HM). Alternatively, video encoder 20 and video decoder 30 may
operate according to other proprietary or industry standards, such
as the ITU-T H.264 standard, alternatively referred to as MPEG-4,
Part 10, Advanced Video Coding (AVC), or extensions of such
standards. The techniques of this disclosure, however, are not
limited to any particular coding standard. Other examples of video
compression standards include MPEG-2 and ITU-T H.263.
[0037] Although not shown in FIG. 1, in some aspects, video encoder
20 and video decoder 30 may each be integrated with an audio
encoder and decoder, and may include appropriate MUX-DEMUX units,
or other hardware and software, to handle encoding of both audio
and video in a common data stream or separate data streams. If
applicable, in some examples, MUX-DEMUX units may conform to the
ITU H.223 multiplexer protocol, or other protocols such as the user
datagram protocol (UDP).
[0038] Video encoder 20 and video decoder 30 each may be
implemented as any of a variety of suitable encoder or decoder
circuitry, such as one or more microprocessors, digital signal
processors (DSPs), application specific integrated circuits
(ASICs), field programmable gate arrays (FPGAs), discrete logic,
software, hardware, firmware or any combinations thereof. When the
techniques are implemented partially in software, a device may
store instructions for the software in a suitable, non-transitory
computer-readable medium and execute the instructions in hardware
using one or more processors to perform the techniques of this
disclosure. Each of video encoder 20 and video decoder 30 may be
included in one or more encoders or decoders, either of which may
be integrated as part of a combined encoder/decoder (CODEC) in a
respective device.
[0039] The JCT-VC developed the HEVC standard based on an evolving
model of a video coding device referred to as the HEVC Test Model
(HM). The HM presumes several additional capabilities of video
coding devices relative to existing devices according to, e.g.,
ITU-T H.264/AVC. For example, whereas H.264 provides nine
intra-prediction encoding modes, the HM may provide as many as
thirty-three intra-prediction encoding modes. One draft of the HEVC
standard, referred to as "HEVC Working Draft 10" or "WD10," is
described in document JCTVC-L1003v34, Bross et al., "High
efficiency video coding (HEVC) text specification draft 10," Joint
Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and
ISO/IEC JTC1/SC29/WG11, 12th Meeting: Geneva, CH, 14-23 January
2013, which, as of Apr. 8, 2013, is downloadable from
http://phenix.int-evry.fr/jct/doc_end_user/documents/12
Geneva/wg11/JCTVC-L1003-v34.zip.
[0040] In general, the HM describes that a video frame or picture
may be divided into a sequence of coding tree units (CTUs) or
largest coding units (LCU) that include both luma and chroma
samples. A CTU has a similar purpose as a macroblock of the H.264
standard. A slice includes a number of consecutive CTUs in coding
order. A video frame or picture may be partitioned into one or more
slices. Each CTU may be split into coding units (CUs) according to
a quadtree. For example, a CTU, as a root node of the quadtree, may
be split into four child nodes, and each child node may in turn be
a parent node and be split into another four child nodes. A final,
unsplit child node, as a leaf node of the quadtree, comprises a
coding node, i.e., a coded video block. Syntax data associated with
a coded bitstream may define a maximum number of times a CTU may be
split, and may also define a minimum size of the coding nodes.
[0041] A CU includes a coding node, as well as prediction units
(PUs) and transform units (TUs) associated with the coding node. A
size of the CU corresponds to a size of the coding node and may be
square in shape. The size of the CU may range from 8.times.8 pixels
up to the size of the CTU with a maximum of 64.times.64 pixels or
greater. Each CU may contain one or more PUs and one or more TUs.
Syntax data associated with a CU may describe, for example,
partitioning of the CU into one or more PUs. Partitioning modes may
differ between whether the CU is skip or direct mode encoded,
intra-prediction mode encoded, or inter-prediction mode encoded.
PUs may be partitioned to be non-square in shape. Syntax data
associated with a CU may also describe, for example, partitioning
of the CU into one or more TUs according to a quadtree. A TU can be
square or non-square in shape.
[0042] The HEVC standard allows for transformations according to
TUs, which may be different for different CUs. The TUs are
typically sized based on the size of PUs within a given CU defined
for a partitioned LCU, although this may not always be the case.
The TUs are typically the same size or smaller than the PUs. In
some examples, residual samples corresponding to a CU may be
subdivided into smaller units using a quadtree structure known as
"residual quad tree" (RQT). The leaf nodes of the RQT may be
referred to as TUs. Pixel difference values associated with the TUs
may be transformed to produce transform coefficients, which may be
quantized.
[0043] In general, a PU includes data related to the prediction
process. For example, when the PU is intra-mode encoded, the PU may
include data describing an intra-prediction mode for the PU. As
another example, when the PU is inter-mode encoded, the PU may
include data defining a motion vector for the PU. The data defining
the motion vector for a PU may describe, for example, a horizontal
component of the motion vector, a vertical component of the motion
vector, a resolution for the motion vector (e.g., one-quarter pixel
precision or one-eighth pixel precision), a reference picture to
which the motion vector points, and/or a reference picture list
(e.g., List 0, List 1, or List C) for the motion vector.
[0044] In general, a TU is used for the transform and quantization
processes. A given CU having one or more PUs may also include one
or more TUs. Following prediction, video encoder 20 may calculate
residual values corresponding to the PU. The residual values
comprise pixel difference values that may be transformed into
transform coefficients, quantized, and scanned using the TUs to
produce serialized transform coefficients for entropy coding. This
disclosure typically uses the term "video block" to refer to a
coding node of a CU. In some specific cases, this disclosure may
also use the term "video block" to refer to a CTU, i.e., LCU, or a
CU, which includes a coding node and PUs and TUs.
[0045] A video sequence typically includes a series of video frames
or pictures. A group of pictures (GOP) generally comprises a series
of one or more of the video pictures. A GOP may include syntax data
in a header of the GOP, a header of one or more of the pictures, or
elsewhere, that describes a number of pictures included in the GOP.
Each slice of a picture may include slice syntax data that
describes an encoding mode for the respective slice. Video encoder
20 typically operates on video blocks within individual video
slices in order to encode the video data. A video block may
correspond to a coding node within a CU. The video blocks may have
fixed or varying sizes, and may differ in size according to a
specified coding standard.
[0046] As an example, the HM supports prediction in various PU
sizes. Assuming that the size of a particular CU is 2N.times.2N,
the HM supports intra-prediction in PU sizes of 2N.times.2N or
N.times.N, and inter-prediction in symmetric PU sizes of
2N.times.2N, 2N.times.N, N.times.2N, or N.times.N. The HM also
supports asymmetric partitioning for inter-prediction in PU sizes
of 2N.times.nU, 2N.times.nD, nL.times.2N, and nR.times.2N. In
asymmetric partitioning, one direction of a CU is not partitioned,
while the other direction is partitioned into 25% and 75%. The
portion of the CU corresponding to the 25% partition is indicated
by an "n" followed by an indication of "Up", "Down," "Left," or
"Right." Thus, for example, "2N.times.nU" refers to a 2N.times.2N
CU that is partitioned horizontally with a 2N.times.0.5N PU on top
and a 2N.times.1.5N PU on bottom.
[0047] In this disclosure, "N.times.N" and "N by N" may be used
interchangeably to refer to the pixel dimensions of a video block
in terms of vertical and horizontal dimensions, e.g., 16.times.16
pixels or 16 by 16 pixels. In general, a 16.times.16 block will
have 16 pixels in a vertical direction (y=16) and 16 pixels in a
horizontal direction (x=16). Likewise, an N.times.N block generally
has N pixels in a vertical direction and N pixels in a horizontal
direction, where N represents a nonnegative integer value. The
pixels in a block may be arranged in rows and columns. Moreover,
blocks need not necessarily have the same number of pixels in the
horizontal direction as in the vertical direction. For example,
blocks may comprise N.times.M pixels, where M is not necessarily
equal to N.
[0048] Following intra-predictive or inter-predictive coding using
the PUs of a CU, video encoder 20 may calculate residual data for
the TUs of the CU. The PUs may comprise pixel data in the spatial
domain (also referred to as the pixel domain) and the TUs may
comprise coefficients in the transform domain following application
of a transform, e.g., a discrete cosine transform (DCT), an integer
transform, a wavelet transform, or a conceptually similar transform
to residual video data. The residual data may correspond to pixel
differences between pixels of the unencoded picture and prediction
values corresponding to the PUs. Video encoder 20 may form the TUs
including the residual data for the CU, and then transform the TUs
to produce transform coefficients for the CU.
[0049] Following any transforms to produce transform coefficients,
video encoder 20 may perform quantization of the transform
coefficients. Quantization generally refers to a process in which
transform coefficients are quantized to possibly reduce the amount
of data used to represent the coefficients, providing further
compression. The quantization process may reduce the bit depth
associated with some or all of the coefficients. For example, an
n-bit value may be rounded down to an m-bit value during
quantization, where n is greater than m. In some examples, video
encoder 20 uses a quantization parameter to determine a rate model
parameter.
[0050] In some examples, video encoder 20 may utilize a predefined
scan order to scan the quantized transform coefficients to produce
a serialized vector that can be entropy encoded. In other examples,
video encoder 20 may perform an adaptive scan. After scanning the
quantized transform coefficients to form a one-dimensional vector,
video encoder 20 may entropy encode the one-dimensional vector,
e.g., according to context adaptive variable length coding (CAVLC),
context adaptive binary arithmetic coding (CABAC), syntax-based
context-adaptive binary arithmetic coding (SBAC), Probability
Interval Partitioning Entropy (PIPE) coding or another entropy
encoding methodology. Video encoder 20 may also entropy encode
syntax elements associated with the encoded video data for use by
video decoder 30 in decoding the video data.
[0051] To perform CABAC, video encoder 20 may assign a context
within a context model to a symbol to be transmitted. The context
may relate to, for example, whether neighboring values of the
symbol are non-zero or not.
[0052] Video encoder 20 of FIG. 1 may be configured to decide
whether to zero out a TU based on a rate distortion (RD) cost
(compared to an RD cost of not zeroing out the TU). When video
encoder 20 zeroes out a TU, video encoder 20 may set the values of
all transform coefficients of the TU to zero. The RD cost to zero
out the TU (C.sub.zbd) may be set equal to the distortion incurred
when the TU is zeroed out (D.sub.zbd). The cost of not zeroing out
the TU (C.sub.inter) may be set equal to the distortion when
transform coefficients inter, are encoded (D.sub.inter) and the
number of bits for the transform coefficients (R.sub.coef)
multiplied by a Lagrangian multiplier (.lamda.). R.sub.coef may be
modeled as a parameter multiplied by the number of non-zero
coefficients (nnz) in the TU added to a parameter b. In one
example, solutions use a fixed linear model of R.sub.coef where
parameters a and b are set through a configuration file.
[0053] These solutions, however, may produce inefficient results
when encoding certain video data such as video with a high noise
level (from e.g., an IP camera). For example, noise is typically at
a high frequency part of the transform coefficients in the block.
The non-zero transform coefficients at high frequency results in
more bits that the video encoder is required to encode. For
instance, if parameters a and b are selected based on training
sequences without noise, the resulting R.sub.coef may fail to
correctly reflect the actual number of bits used and the cost of
not zeroing out the block, C.sub.inter, may be underestimated. This
may result in not zeroing out the block resulting in a higher
bit-rate and a lower peak signal-to-noise ratio (PSNR) which may
hurt overall rate-distortion performance. In some examples, video
encoder 20 may be configured to use values for parameters a and b
based on training data from noisy static sequences. In some
examples, parameter a may be quantization parameter (QP) dependent
and b is not QP dependent. In such examples, b is not QP dependent
because b is the number of bits when the number of non-zero
coefficients is 0. Thus, in the foregoing examples, parameter a may
take on a number of values depending on the selected QP value,
whereas parameter b would have a static value. In other examples,
both parameters are QP independent values or both are QP dependent
values.
[0054] Video encoder 20 may also select parameters a and b
adaptively. For example, parameters a and b may be initialized by
video encoder 20. Parameters a and b may be initialized from any
set of values whether it is the values for non-static sequences, or
values based on static noisy sequences. The current frame may be
encoded by video encoder 20 with initial values of parameters a and
b as well as the quantization parameter. Video encoder 20 may
collect the number of bits actually generated for frame i
(R.sub.coef,i) and the number of non-zero coefficients (nnz.sub.i)
for frame i. In an example, video encoder 20 may perform a linear
regression to minimize the error to determine new values for
parameters a and b. Alternatively, video encoder 20 may calculate
updated values for new parameters a and/or b. In some examples, new
parameter b may be a fixed value and new parameter a is adaptively
changed. In some examples, parameter a (and/or b) is then updated
using a weighted average of the new parameter a and the previously
used parameter a. In other examples, video encoder 20 may update
parameter a and/or parameter b to use new parameter a and/or
parameter b.
[0055] As will be explained in more detail below, video encoder 20
of FIG. 1 may be configured to determine an amount of motion or
movement in a frame of a sequence of frames. In some examples,
video encoder 20 may collect the number of inter blocks (inter_cnt)
and calculate the sum of amplitudes of motion vectors in the
horizontal direction (my_sum_hor) and in the vertical direction
(my_sum_ver) of a frame after encoding the frame. If video encoder
20 determines that the number of inter blocks is greater than a
threshold T (i.e., there are enough inter blocks) and the average
amplitude of motion vectors horizontally and vertically are less
than a threshold t, then video encoder 20 may determine that the
frame is static. Video encoder 20 may then estimate the next frame
to be static or non-static, based on the determination of the
previous frame.
[0056] Video encoder 20 may select a parameter set for a rate
estimation model (e.g., a linear rate model) for a current frame,
based on an amount of movement between frames. For instance, video
encoder 20 may select the parameters a and b based on whether the
frame is estimated to be static or non-static. In an alternative
example, video encoder 20 may select the parameter set for the rate
estimation model (e.g., linear rate model) based on an amount of
movement in the present (current) frame. For example, video encoder
20 may estimate motion vectors (my_hor and my_ver) between the
current frame and the previous frame and use the motion vectors as
a metric to determine whether the frame is static or non-static.
Video encoder 20 may also calculate a sum of absolute difference
(SAD) between pixel values of the current frame and pixel value of
the previous frame, based on the motion vectors. If the sum of
horizontal motion vectors (my_sum_hor) and the sum of vertical
motion vectors (my_sum_ver) are less than threshold t, and the SAD
is less than threshold T, then the sequence may be determined to be
static. It will be appreciated that SAD is just one example
discussed herein. Other distortion metrics can also be used in
accordance with this disclosure, e.g., sum of squared error (SSE),
etc. Video encoder 20 may also calculate SAD or SSE values based on
zero motion vectors, in determining whether a frame is static or
non-static.
[0057] Based on the determination of static/non-static status,
video encoder 20 may use a trained parameter set for a noisy and/or
static sequence may be used, or otherwise, may use the trained
parameter set for non-static and/or non-noisy sequences. Video
encoder 20 may encode the frame based on the QP and the selected
parameter set.
[0058] According to various examples of this disclosure, video
encoder 20 (and/or various components thereof) may update one
parameter (namely, parameter `a`) of a rate estimation model, based
on: (1) an actual number of bits used to encode the non-zero
coefficients of the frame, and (2) the actual number of non-zero
coefficients in the frame. Video encoder 20 may also determine an
amount of motion (or movement) in a frame, and may use the
determined amount of motion/movement to update the parameter
(namely, parameter `a`) of the rate estimation model.
[0059] By updating parameter `a` of the rate estimation model
before applying the rate estimation model to the blocks of the
frame, video encoder 20 may implement model-updating techniques of
this disclosure to zero out blocks of the frame in a more targeted
fashion. In this way, video encoder 20 may implement the techniques
of this disclosure to reduce bitrate and signaling overhead by
zeroing out blocks, while maintaining or potentially improving
coding precision by zeroing out blocks in a targeted fashion by
updating the rate estimation model from frame to frame.
[0060] In certain use case scenarios, video frames are highly
correlated temporally, especially in scenarios such as those of
video telephony (VT) and video surveillance. In cases of video
frames of a sequence being highly correlated temporally, the
prediction error is usually very small. For example, it is observed
that with certain VT-type frame sequences and under certain coding
configurations, of all the 32.times.32 TU calculations, about 79%
of those TUs are zero blocks. Existing encoding techniques may
employ a model to decide whether or not to zero out a block (e.g.,
a TU), based on RD cost. That is, according to the existing
encoding techniques, a video encoder may compare the RD cost
(C.sub.zbd) of zeroing out the block against the RD cost
(C.sub.inter) of not zeroing out the block. If the RD cost of
zeroing out the block is the smaller than the RD cost of not
zeroing out the block (i.e., C.sub.zbd<C.sub.inter), the video
encoder zeroes out the block. Otherwise, a video encoder applies
the inter mode to the block, and encodes and signals transform
coefficients of the block. In this example, the video encoder may
obtain C.sub.zbd and C.sub.inter as follows:
C.sub.zbd=D.sub.zbd (1)
C.sub.inter=D.sub.inter.lamda.R.sub.coef (2)
where D.sub.zbd is the distortion incurred when the block is zeroed
out, and D.sub.inter is the distortion when the transform
coefficients of the block are encoded. R.sub.coef is the number of
bits for the transform coefficients of the block, and .lamda. is a
Lagrangian multiplier. The video encoder may estimate R.sub.coef
based on the number of non-zero coefficients (nnz). In some
exemplary encoders, the relationship of R.sub.coef and the number
of non-zero coefficients nnz is modeled using the following linear
model:
R.sub.coef=annz+b (3)
In the existing encoding techniques, parameters `a` and `b` of
equation (3) are fixed values set through a configuration file.
[0061] However, the fixed linear rate model described above may not
enable efficient encoding of some types of sequences, such as video
sequences from an IP camera (as one use case example) with a high
noise level. For such sequences, using the fixed linear rate model
may have unexpected encoding results. For example, Table 1, below,
compares two different LP (low delay P-prediction) configurations
of IP camera sequence 720pBayBridge, which includes a relatively
high level of noise. According to LP configuration, a first frame
is encoded as an I-frame with a number of following frames encoded
as P-frame.
TABLE-US-00001 TABLE 1 Comparison of Results of Two Encoding
Configurations Bit-Rate PSNR_Y PSNR_U BayBridge720p (kb/s) (db)
(db) PSNR_V (db) QP_I=20, QP_P=21 1163.39 43.86 46.43 47.17
QP_I=20, QP_P=22 1264.6 43.69 46.39 47.16
[0062] In one LP configuration, I frames are encoded with QP_I=20,
and P frames are encoded with QP_P=21. In the other LP
configuration, I frames are encoded with QP_I=20, and P frames are
encoded with QP_P=22. Intuitively, it would be expected that the
configuration where the video encoder uses QP_P=22 for P frames
would result in a fewer number of bits than the configuration where
the video encoder uses QP_P=21. However, to the contrary, the
results in Table 1 above show that the configuration with QP_P=22
results in a higher bit rate than the configuration with QP_P=21
and a lower peak signal-to-noise ratio (PSNR), which indicates an
inefficient result.
[0063] The higher bit rate may be attributed to the sequence having
a lot of noise. In some examples of video sequences, such as IP
camera sequences, the noise is at the high-frequency part of the
transform coefficients in the block. The non-zero transform
coefficients at the higher-frequency portion results in a greater
number of bits required to be encoded. If parameters `a` and `b` of
the linear fixed rate model are trained for sequences without
noise, then the resulting Rcoer value may not correctly reflect the
real number of bits to be used in noisy sequences. In these
scenarios, the cost of not zeroing out the block (C.sub.inter) may
be underestimated. Nevertheless, in this example, the video encoder
chooses the option of not zeroing out the block more frequently
according to existing encoding techniques, and the result is a
higher bitrate and lower PSNR encoding. Coding noise typically does
not help temporal prediction, and thus may hurt overall
rate-distortion performance.
[0064] Video encoder 20 may implement various techniques of this
disclosure to train the linear rate model to identify blocks to be
zeroed out in a more targeted fashion than in existing encoding
techniques. In some examples, video encoder 20 may use an amount of
motion/movement in a current frame or a previously-encoded frame to
update parameter `a` of the linear fixed model. In these and other
examples, video encoder 20 may update parameter `a` of the linear
fixed model based on an actual number of bits used to encode the
non-zero coefficients of the frame, and the actual number of
non-zero coefficients in the frame. In this way, video encoder 20
may implement the techniques of this disclosure to reduce bitrate
and signaling overhead by zeroing out blocks, while maintaining or
potentially improving coding precision by zeroing out blocks in a
targeted fashion by updating the rate estimation model from frame
to frame.
[0065] FIG. 2 is a block diagram illustrating an example video
encoder 20 that may implement the adaptive rate model for zero
block detection techniques described in this disclosure. Video
encoder 20 may be configured to perform intra- and inter-coding of
video blocks within video slices. Intra-coding relies on spatial
prediction to reduce or remove spatial redundancy in video within a
given video frame or picture. Inter-coding relies on temporal
prediction to reduce or remove temporal redundancy in video within
adjacent frames or pictures of a video sequence. Intra-mode (I
mode) may refer to any of several spatial based compression modes.
Inter-modes, such as uni-directional prediction (P mode) or
bi-prediction (B mode), may refer to any of several temporal-based
compression modes.
[0066] In the example of FIG. 2, video encoder 20 includes video
data memory 40, prediction processing unit 41, reference picture
memory 64, summer 50, transform processing unit 52, quantization
unit 54, rate control unit 51, and entropy encoding unit 56.
Prediction processing unit 41 includes motion estimation unit 42,
motion compensation unit 44, and intra prediction processing unit
46. For video block reconstruction, video encoder 20 also includes
inverse quantization unit 58, inverse transform processing unit 60,
and summer 62. A deblocking filter (not shown in FIG. 2) may also
be included to filter block boundaries to remove blockiness
artifacts from reconstructed video. If desired, the deblocking
filter may filter the output of summer 62. Additional loop filters
(in loop or post loop) may also be used in addition to the
deblocking filter.
[0067] Video data memory 40 may store video data to be encoded by
the components of video encoder 20. The video data stored in video
data memory 40 may be obtained, for example, from video source 18.
Reference picture memory 64 is one example of a decoded picture
buffer (DPB) that stores reference video data for use in encoding
video data by video encoder 20 (e.g., in intra- or inter-coding
modes, also referred to as intra- or inter-prediction coding
modes). Video data memory 40 and reference picture memory 64 may be
formed by any of a variety of memory devices, such as dynamic
random access memory (DRAM), including synchronous DRAM (SDRAM),
magnetoresistive RAM (MRAM), resistive RAM (RRAM), or other types
of memory devices. Video data memory 40 and reference picture
memory 64 may be provided by the same memory device or separate
memory devices. In various examples, video data memory 40 may be
on-chip with other components of video encoder 20, or off-chip
relative to those components.
[0068] Video encoder 20 receives video data, and a partitioning
unit (not shown) partitions the data into video blocks. This
partitioning may also include partitioning into slices, tiles, or
other larger units, as wells as video block partitioning, e.g.,
according to a quadtree structure of LCUs and CUs. Video encoder 20
generally illustrates the components that encode video blocks
within a video slice to be encoded. The slice may be divided into
multiple video blocks (and possibly into sets of video blocks
referred to as tiles). Prediction processing unit 41 may select one
of a plurality of possible coding modes, such as one of a plurality
of intra coding modes or one of a plurality of inter coding modes,
for the current video block based on error results (e.g., coding
rate and the level of distortion). Prediction processing unit 41
may provide the resulting intra- or inter-coded block to summer 50
to generate residual block data and to summer 62 to reconstruct the
encoded block for use as a reference picture.
[0069] Intra prediction processing unit 46 within prediction
processing unit 41 may perform intra-predictive coding of the
current video block relative to one or more neighboring blocks in
the same frame or slice as the current block to be coded to provide
spatial compression. Motion estimation unit 42 and motion
compensation unit 44 within prediction processing unit 41 perform
inter-predictive coding of the current video block relative to one
or more predictive blocks in one or more reference pictures to
provide temporal compression.
[0070] Motion estimation unit 42 may be configured to determine the
inter-prediction mode for a video slice according to a
predetermined pattern for a video sequence. The predetermined
pattern may designate video slices in the sequence as P slices or B
slices. Motion estimation unit 42 and motion compensation unit 44
may be highly integrated, but are illustrated separately for
conceptual purposes. Motion estimation, performed by motion
estimation unit 42, is the process of generating motion vectors,
which estimate motion for video blocks. A motion vector, for
example, may indicate the displacement of a PU of a video block
within a current video frame or picture relative to a predictive
block within a reference picture.
[0071] A predictive block may be a block that is found to closely
match the PU of the video block to be coded in terms of pixel
difference, which may be determined by sum of absolute difference
(SAD), sum of square difference (SSD), or other difference metrics.
In some examples, video encoder 20 may calculate values for
sub-integer pixel positions of reference pictures stored in
reference picture memory 64. For example, video encoder 20 may
interpolate values of one-quarter pixel positions, one-eighth pixel
positions, or other fractional pixel positions of the reference
picture. Therefore, motion estimation unit 42 may perform a motion
search relative to the full pixel positions and fractional pixel
positions and output a motion vector with fractional pixel
precision.
[0072] Motion estimation unit 42 calculates a motion vector for a
PU of a video block in an inter-coded slice by comparing the
position of the PU to the position of a predictive block of a
reference picture. The reference picture may be selected from a
first reference picture list (List 0) or a second reference picture
list (List 1), each of which identify one or more reference
pictures stored in reference picture memory 64. Motion estimation
unit 42 sends the calculated motion vector to entropy encoding unit
56 and motion compensation unit 44.
[0073] Motion compensation, performed by motion compensation unit
44, may involve fetching or generating the predictive block based
on the motion vector determined by motion estimation, possibly
performing interpolations to sub-pixel precision. Upon receiving
the motion vector for the PU of the current video block, motion
compensation unit 44 may locate the predictive block to which the
motion vector points in one of the reference picture lists. Video
encoder 20 forms a residual video block by subtracting pixel values
of the predictive block from the pixel values of the current video
block being coded, forming pixel difference values. The pixel
difference values form residual data for the block, and may include
both luma and chroma difference components. Summer 50 represents
the component or components that perform this subtraction
operation. Motion compensation unit 44 may also generate syntax
elements associated with the video blocks and the video slice for
use by video decoder 30 in decoding the video blocks of the video
slice.
[0074] Intra prediction processing unit 46 may perform
intra-prediction on a current block, as an alternative to the
inter-prediction performed by motion estimation unit 42 and motion
compensation unit 44, as described above. In particular, intra
prediction processing unit 46 may determine an intra-prediction
mode to use to encode a current block. In some examples, intra
prediction processing unit 46 may encode a current block using
various intra-prediction modes, e.g., during separate encoding
passes, and intra prediction processing unit 46 may select an
appropriate intra-prediction mode to use from the tested modes. For
example, intra prediction processing unit 46 may calculate
rate-distortion values using a rate-distortion analysis for the
various tested intra-prediction modes, and select the
intra-prediction mode having the best rate-distortion
characteristics among the tested modes. Rate-distortion analysis
generally determines an amount of distortion (or error) between an
encoded block and an original, unencoded block that was encoded to
produce the encoded block, as well as a bit rate (that is, a number
of bits) used to produce the encoded block. Intra prediction
processing unit 46 may calculate ratios from the distortions and
rates for the various encoded blocks to determine which
intra-prediction mode exhibits the best rate-distortion value for
the block.
[0075] In any case, after selecting an intra-prediction mode for a
block, intra prediction processing unit 46 may provide information
indicative of the selected intra-prediction mode for the block to
entropy encoding unit 56. Entropy encoding unit 56 may encode the
information indicating the selected intra-prediction mode in
accordance with the techniques of this disclosure. Video encoder 20
may include in the transmitted bitstream configuration data, which
may include a plurality of intra-prediction mode index tables and a
plurality of modified intra-prediction mode index tables (also
referred to as codeword mapping tables), definitions of encoding
contexts for various blocks, and indications of a most probable
intra-prediction mode, an intra-prediction mode index table, and a
modified intra-prediction mode index table to use for each of the
contexts.
[0076] After prediction processing unit 41 generates the predictive
block for the current video block via either inter-prediction or
intra-prediction, video encoder 20 forms a residual video block by
subtracting the predictive block from the current video block. The
residual video data in the residual block may be included in one or
more TUs and applied to transform processing unit 52. Transform
processing unit 52 transforms the residual video data into residual
transform coefficients using a transform, such as a discrete cosine
transform (DCT) or a conceptually similar transform. Transform
processing unit 52 may convert the residual video data from a pixel
domain to a transform domain, such as a frequency domain.
[0077] Transform processing unit 52 may send the resulting
transform coefficients to quantization unit 54. Quantization unit
54 quantizes the transform coefficients to further reduce bit rate.
The quantization process may reduce the bit depth associated with
some or all of the coefficients. The degree of quantization may be
modified by adjusting a quantization parameter. In some examples,
quantization unit 54 may then perform a scan of the matrix
including the quantized transform coefficients. Alternatively,
entropy encoding unit 56 may perform the scan.
[0078] Following quantization, entropy encoding unit 56 entropy
encodes the quantized transform coefficients. For example, entropy
encoding unit 56 may perform context adaptive variable length
coding (CAVLC), context adaptive binary arithmetic coding (CABAC),
syntax-based context-adaptive binary arithmetic coding (SBAC),
probability interval partitioning entropy (PIPE) coding or another
entropy encoding methodology or technique. Following the entropy
encoding by entropy encoding unit 56, the encoded bitstream may be
transmitted to video decoder 30, or archived for later transmission
or retrieval by video decoder 30. Entropy encoding unit 56 may also
entropy encode the motion vectors and the other syntax elements for
the current video slice being coded.
[0079] Inverse quantization unit 58 and inverse transform
processing unit 60 apply inverse quantization and inverse
transformation, respectively, to reconstruct the residual block in
the pixel domain for later use as a reference block of a reference
picture. Motion compensation unit 44 may calculate a reference
block by adding the residual block to a predictive block of one of
the reference pictures within one of the reference picture lists.
Motion compensation unit 44 may also apply one or more
interpolation filters to a reference frame to calculate sub-integer
pixel values for use in motion estimation. Summer 62 adds the
reconstructed residual block to the motion compensated prediction
block produced by motion compensation unit 44 to produce a
reference block for storage in reference picture memory 64. The
reference block may be used by motion estimation unit 42 and motion
compensation unit 44 as a reference block to inter-predict a block
in a subsequent video frame or picture.
[0080] Video encoder 20 may further include a rate control unit 51
which is configured to control the rate of video encoding (i.e.,
the number of bits that are used to represent the encoded video).
Rate control unit 51 may control the rate of video encoding by
determining a rate control parameter. The rate control parameter
generated by rate control unit 51 may be used by both prediction
processing unit 41 and quantization unit 54 to control the rate of
video encoding. As will be explained in detail elsewhere in this
disclosure, prediction processing unit 41 may use the rate control
parameter in a rate-distortion analysis used to select a prediction
mode (e.g., the best inter or intra prediction mode). In addition,
quantization unit 54 may use the rate control parameter to
determine a quantization parameter. Both the determined prediction
mode and the quantization parameter affect the encoding rate of
video encoder 20.
[0081] In accordance with the techniques of this disclosure, and as
described elsewhere in this disclosure, video encoder 20 may be
configured to estimate the number of bits for the transform
coefficients of a block, R.sub.coef In one example, video encoder
20 may estimate R.sub.coef using the linear model
R.sub.coef=annz+b, of equation (3) above. In another example, video
encoder 20 may improve the R.sub.coef determination by including,
for consideration, coefficient position information. Coefficient
position information refers to data indicating the position of one
or more coefficients in a block. By considering coefficient
position information in the determination of R.sub.coef, video
encoder 20 may determine R.sub.coef such that the estimated number
of bits is a function of parameters `a` and `b` and of nnz and
coefficient position information (pos). Equation (4) below
describes the R.sub.coef determination techniques that include pos
in the R.sub.coef determination:
R.sub.coef=f(a, b, nnz, pos) (4)
When using the model for R.sub.coef of equation 4, video encoder 20
may use the adaptively-chosen model parameter `a` and an unchanged
value for parameter `b` to compensate for deficiencies of this
model.
[0082] After each frame is encoded, video encoder 20 may collect
the number of bits actually generated and the number of non-zero
coefficients. According to a linear regression, video encoder 20
may estimate parameters `a` and `b` by minimizing the following
function:
F ( a , b ) = min ( a , b ) ( R coef , i - a nnz i - b ) 2 ( 5 )
##EQU00001##
R.sub.coef,i and nnz.sub.i are the collected number of bits for the
block i and number of the non-zero coefficients in the block i in
equation (5) above. It can be shown that the values of a and b that
minimize the function F (equation 5) are:
a = 1 N R coef , i nnz i - 1 N R coef , i 1 N nnz i 1 N nnz i 2 - (
1 N nnz i ) 2 = cov ( R coef , nnz ) var ( nnz ) ( 6 ) b = 1 N R
coef , i - a 1 N nnz i ( 7 ) ##EQU00002##
In equations (6) and (7), cov() is the covariance function, and
var() is the variance function.
[0083] In cases where the number of non-zero coefficients is zero,
video encoder 20 may be configured to signal header bits (e.g.
header data such as PPS, slice header, etc. as well as motion
vector bits and block prediction mode bits), the number of which is
usually a constant. In an example, parameter `b` may be set to a
constant number (to indicate the number of header bits, directly or
indirectly), and video encoder 20 may adaptively change parameter
`a` according to collected nnz and R.sub.coef values. In this case,
the calculation of parameter `a` is represented by equation (8)
below:
a = ( nnz i R coef , i - b nnz i ) nnz i 2 = S_xy i - b S_x i S_x2
i ( 8 ) S_xy = nnz i R coef , i ( 9 ) S_x = nnz i ( 10 ) S_x2 = nnz
i 2 ( 11 ) ##EQU00003##
[0084] Video encoder 20 may designate initial values to parameters
a and b. In turn, after encoding each frame, video encoder 20 may
collect sum values of the number of non-zero coefficients nnz, and
the number of bits of the coefficients R.sub.coef,i from every
block i. Video encoder 20 may compute values for S_xy using
equation (9), S_x using equation (10), and S_x2 using equation
(11). After calculating the values of S_xy, S_x, and S_x2, video
encoder 20 may obtain a new value of parameter `a` (new_a) using
equation (8). For each of variables S_xy, S_x, and S_x2, x
represents nnz and y represents R.sub.coef. Again, each of
variables S_xy, S_x, and S_x2, is referred to herein as a "sum
value" and as such, video encoder 20 may determine sum values
variables S_xy, S_x, and S_x2 as part of updating parameter `a`
according to the techniques of this disclosure.
[0085] In some examples, video encoder 20 may update the parameter
`a` using weighted averaging of the new value of parameter `a`
(new_a) and a previous value of parameter `a` (prev_a), as shown in
equation (12) below:
a=w*new_a+(1-w)*prev_a (12)
[0086] In further examples, video encoder 20 may use a group of
pictures (GOP) structure with multiple temporal layers, each
temporal layer may have its own rate estimation model and the
current frame may refer to the rate estimation model of the
previous frame in the same temporal layer instead of its immediate
previous frame. In further examples, video encoder 20 may improve
the accuracy of estimation by using more complex models than a
linear model, because rate estimation may not follow linear models
for some types of content. For example, video encoder 20 may use
Nth-order models or a neural network-based model.
[0087] Video encoder 20 may use different sets of parameters `a`
and `b` for different types of content. For example, video encoder
20 may use separate rate estimation model for static (e.g., IP
camera sequences for video surveillance) sequences and non-static
(e.g. higg-motion) sequences respectively, because the percentage
of bits consumed by noise relative to total bits used may be
different between static and non-static sequences. As a result, the
rate estimation models may indicate preferences to different
parameters. In an example technique of this disclosure, video
encoder 20 may use a different set of parameters `a` and `b` for
static sequences than for non-static sequences.
[0088] In some examples, the values of parameters `a` and `b` are
dependent upon the QP used by video encoder 20. In some examples,
parameter a may be found to be QP dependent and parameter b may be
found to not be QP dependent. Parameter b may not be QP dependent
because the value of parameter `b` is based on the number of bits
when the number of non-zero coefficients is zero (0). For different
QP values, video encoder 20 may train the parameters `a` and `b`
that provide the best coding efficiency. In one implementation,
parameters `a` and `b` are 16-bit unsigned integers for noisy
sequences, and parameters a and b are set as follows:
TABLE-US-00002 UInt16 param_a[52] = { //0, 1, 2, 3, 4, 5, 6, 7
41862, 41862, 41862, 41862, 41862, 41862, 41862, 41862, // 8, 9,
10, 11, 12, 13, 14, 15 41862, 41862, 41862, 41862, 41862, 41862,
41862, 41862, //16, 17, 18, 19, 20, 21, 22, 23 41862, 41862, 51862,
61862, 36862, 46862, 56862, 36862, //24, 25, 26, 27, 28, 29, 30, 31
46862, 41862, 51862, 56862, 51862, 61862, 61862, 61862, //32, 33,
34, 35, 36, 37, 38, 39, 61862, 51862, 56862, 56862, 56862, 56862,
56862, 56862, //40, 41, 42, 43, 44, 45, 46, 47 56862, 56862, 56862,
56862, 56862, 56862, 56862, 56862, //48, 49, 50, 51 56862, 56862,
56862, 56862}; UInt16 param_b = 28405;
[0089] In this example, the value of parameter `a`, shown above, is
indexed by QP values. For example, when QP=22, parameter `a` may be
determined as follows: a=param_a[QP]=param_a[22]=56862. Parameter b
is independent of QP and may be a predetermined value, such as
28405.
[0090] Video encoder 20 may determine whether the current frame is
static or non-static. For instance, video encoder 20 may collect
encoding statistics such as the absolute sum of amplitudes of
motion vectors and the number of inter blocks of the previous
frame. In an example, when the average amplitude of motion vectors
in the frame is less than a threshold, video encoder 20 may
determine that the current frame is static, and may use the
parameter set (a.sub.1,b.sub.1) for noisy and/or static sequences.
Otherwise, video encoder 20 may use another parameter set (a.sub.2,
b.sub.2) for non-noisy and/or non-static sequences. In another
example with increased preprocessing, video encoder 20 may detect
static sequences based on an estimate of the motion of the current
frame and a sum of absolute differences (SAD) between the current
frame pixel values and a previous frame pixel values. For examples,
if a sum of motion vectors of the current frame is less than a
threshold (e.g., a predetermined threshold specified by video
encoder 20) and the SAD is less than a threshold (e.g., a
predetermined threshold specified by video encoder 20), video
encoder 20 may determine that the current frame is static, and may
therefore select the parameter set (a.sub.1,b.sub.1) for noisy
sequences. Otherwise, video encoder 20 may select the parameter set
(a.sub.2, b.sub.2) for non-noisy and/or non-static sequences.
[0091] Video encoder 20 may adaptively select parameter set (a,b)
based on detecting whether a sequence is static or not using
collected encoding statistics regarding the number of inter blocks
in frames of the sequence and amplitudes of motion vectors of
frames in the sequence. After encoding one frame of the sequence,
video encoder 20 may collect the number of inter blocks of the
frame (inter_cnt) and the sum of amplitudes of motion vectors of
the inter blocks of the frame in the horizontal direction
(mv_sum_hor) and the sum of amplitudes of motion vectors of the
inter blocks of the frame in the vertical direction (mv_sum_ver).
If the number of inter blocks of the frame is bigger (greater) than
a threshold T and the average amplitude of motion vectors of the
frame horizontally and vertically are less than a threshold t,
video encoder 20 may determine that the frame is static. In this
scenario, video encoder 20 may estimate that the next frame will be
static, and video encoder 20 may therefore use the parameter set
(a.sub.1,b.sub.1) for a static sequence. Otherwise, video encoder
20 may use the parameter set (a.sub.2,b.sub.2) for a non-static
sequence. In some examples, video encoder 20 selects threshold T to
ensure that there are enough inter blocks. For instance, video
encoder 20 may calculate T as 0.95*total_8.times.8_blocks. In other
examples, threshold T is predefined in a video coding standard. In
some examples, video encoder 20 may select threshold t to ensure
that motion in the frame is limited. In one implementation, the
value of t may be 1.5. In other examples, threshold t is predefined
in a video coding standard.
[0092] FIG. 3 is a block diagram illustrating an example video
decoder 30. In the example of FIG. 3, video decoder 30 includes
video data memory 83, entropy decoding unit 80, prediction
processing unit 81, inverse quantization unit 86, inverse
transformation processing unit 88, summer 90, and reference picture
memory 92. Prediction processing unit 81 includes motion
compensation unit 82 and intra prediction processing unit 84. Video
decoder 30 may, in some examples, perform a decoding pass generally
reciprocal to the encoding pass described with respect to video
encoder 20 from FIG. 2.
[0093] Video data memory 83 may store video data to be decoded by
the components of video decoder 30. The video data stored in video
data memory 83 may be obtained, for example, from video encoder 20.
Reference picture memory 92 is one example of a decoded picture
buffer (DPB) that stores reference video data for use in decoding
video data by video decoder 30 (e.g., in intra- or inter-coding
modes, also referred to as intra- or inter-prediction coding
modes). Video data memory 83 and reference picture memory 92 may be
formed by any of a variety of memory devices, such as dynamic
random access memory (DRAM), including synchronous DRAM (SDRAM),
magnetoresistive RAM (MRAM), resistive RAM (RRAM), or other types
of memory devices. Video data memory 83 and reference picture
memory 92 may be provided by the same memory device or separate
memory devices. In various examples, video data memory 83 may be
on-chip with other components of video decoder 30, or off-chip
relative to those components.
[0094] During the decoding process, video decoder 30 receives an
encoded video bitstream that represents video blocks of an encoded
video slice and associated syntax elements from video encoder 20.
Entropy decoding unit 80 of video decoder 30 entropy decodes the
bitstream to generate quantized coefficients, motion vectors, and
other syntax elements. Entropy decoding unit 80 forwards the motion
vectors and other syntax elements to prediction processing unit 81.
Video decoder 30 may receive the syntax elements at the video slice
level and/or the video block level.
[0095] When the video slice is coded as an intra-coded (I) slice,
intra prediction processing unit 84 of prediction processing unit
81 may generate prediction data for a video block of the current
video slice based on a signaled intra prediction mode and data from
previously decoded blocks of the current frame or picture. When the
video frame is coded as an inter-coded (i.e., B or P) slice, motion
compensation unit 82 of prediction processing unit 81 produces
predictive blocks for a video block of the current video slice
based on the motion vectors and other syntax elements received from
entropy decoding unit 80. The predictive blocks may be produced
from one of the reference pictures within one of the reference
picture lists. Video decoder 30 may construct the reference frame
lists, List 0 and List 1, using default construction techniques
based on reference pictures stored in reference picture memory
92.
[0096] Motion compensation unit 82 determines prediction
information for a video block of the current video slice by parsing
the motion vectors and other syntax elements, and uses the
prediction information to produce the predictive blocks for the
current video block being decoded. For example, motion compensation
unit 82 uses some of the received syntax elements to determine a
prediction mode (e.g., intra- or inter-prediction) used to code the
video blocks of the video slice, an inter-prediction slice type
(e.g., B slice, P slice, or GPB slice), construction information
for one or more of the reference picture lists for the slice,
motion vectors for each inter-encoded video block of the slice,
inter-prediction status for each inter-coded video block of the
slice, and other information to decode the video blocks in the
current video slice.
[0097] Motion compensation unit 82 may also perform interpolation
based on interpolation filters. Motion compensation unit 82 may use
interpolation filters as used by video encoder 20 during encoding
of the video blocks to calculate interpolated values for
sub-integer pixels of reference blocks. In this case, motion
compensation unit 82 may determine the interpolation filters used
by video encoder 20 from the received syntax elements and use the
interpolation filters to produce predictive blocks.
[0098] Inverse quantization unit 86 inverse quantizes, i.e.,
de-quantizes, the quantized transform coefficients provided in the
bitstream and decoded by entropy decoding unit 80. The inverse
quantization process may include use of a quantization parameter
calculated by video encoder 20 for each video block in the video
slice to determine a degree of quantization and, likewise, a degree
of inverse quantization that should be applied. Inverse transform
processing unit 88 applies an inverse transform, e.g., an inverse
DCT, an inverse integer transform, or a conceptually similar
inverse transform process, to the transform coefficients in order
to produce residual blocks in the pixel domain.
[0099] After motion compensation unit 82 generates the predictive
block for the current video block based on the motion vectors and
other syntax elements, video decoder 30 forms a decoded video block
by summing the residual blocks from inverse transform processing
unit 88 with the corresponding predictive blocks generated by
motion compensation unit 82. Summer 90 represents the component or
components that perform this summation operation. If desired, a
deblocking filter may also be applied to filter the decoded blocks
in order to remove blockiness artifacts. Other loop filters (either
in the coding loop or after the coding loop) may also be used to
smooth pixel transitions, or otherwise improve the video quality.
The decoded video blocks in a given frame or picture are then
stored in reference picture memory 92, which stores reference
pictures used for subsequent motion compensation. Reference picture
memory 92 also stores decoded video for later presentation on a
display device, such as display device 31 of FIG. 1.
[0100] FIG. 4 is a flowchart illustrating an example method of
adaptively selecting parameter set (a,b). The techniques of FIG. 4
may be implemented by one or more components of video encoder
20.
[0101] In one example of the disclosure, video encoder 20 is
configured to initialize parameters a and b with some predefined
values (400). The initialized parameters a and b may be based on
data in a configuration file for non-static sequences for
parameters a and b or may be derived based on a set of training
data. For example, parameters a and b may be set based on values of
parameters a and b in param_a and param_b as discussed above. In
such an example, video encoder 20 may use the QP value to select an
initial value for parameter a.
[0102] Video encoder 20 may determine whether the current frame is
the last frame in the sequence (402). If the current frame is the
last frame ("YES" branch of 402), video encoder 20 may end encoding
(404). If the current frame is not the last frame ("NO" branch of
402), video encoder 20 may encode the current frame with the QP and
parameter set (a,b) (406). Video encoder 20 may collect the number
of non-zero coefficients (nnz.sub.i) and the number of bits on the
coefficient R.sub.coef,i used in the encoding in step 406 (408).
Video encoder 20 may calculate values for S_xy.sub.i, S_x.sub.i and
S_x2.sub.i (410). S_xy.sub.i may be calculated from equation 9.
S_.sub.i may be calculated from equation 10. S_x2.sub.imay be
calculated from equation 11. Video encoder 20 may calculate a new
value of parameter a (new_a) (412). Video encoder 20 may calculate
the new value of parameter a (new_a) from equation 8. Furthermore,
in the example of FIG. 4, video encoder 20 determine an updated
value of parameter a based on the weighted average of the
previously used value of parameter a and new_a (414). Video encoder
20 may calculate the updated value of parameter a using equation
12. Video encoder 20 may then continue with the next frame and step
402 of the method.
[0103] FIG. 5 is a flowchart illustrating an example method of
selection of parameter set (a.sub.1,b.sub.1) for static and
parameter set (a.sub.2,b.sub.2) for non-static sequences based on
the previous frame's coding statistics. The techniques of FIG. 5
may be implemented by one or more components of video encoder
20.
[0104] In one example of the disclosure, video encoder 20 is
configured to initialize parameters a and b with predefined values
(500). The initialized parameters a and b may be based on data in a
configuration file for non-static sequences for parameters a and b
or may be derived based on a set of training data. For example,
parameters a and b may be set based on values of parameters a and b
in param a and param b as discussed above. In such an example, the
QP value may be used to select an initial value for parameter a.
Video encoder 20 may determine whether the current frame is the
last frame in the sequence (502). If the current frame is the last
frame ("YES" branch of 502), video encoder 20 may end encoding
(504). If the current frame is not the last frame ("NO" branch of
502), video encoder 20 may encode the current frame with the QP and
parameter set (a,b) (506). Video encoder 20 may collect the number
of inter blocks of the current frame (inter_cnt) and the sum of
amplitude of motion vector of the current frame in the horizontal
direction (mv_sum_hor) and in the vertical direction of the current
frame (mv_sum_ver) (508).
[0105] Video encoder 20 may be configured to determine if the
number of inter blocks (inter_cnt) is greater than a first
threshold (T) and average amplitude of motion vector horizontally
(e.g., mv_sum_hor divided by inter_cnt) and vertically (e.g.,
mv_sum_ver divided by inter_cnt) are less than a second threshold
(t), video encoder 20 may determine the frame is static ("YES"
branch of 510). Video encoder 20 may estimate that the next frame
will be static and video encoder 20 may use the parameter set (a,b)
for the static sequence (512). Otherwise video encoder 20 may
determine the frame is non-static ("NO" branch of 510). Video
encoder 20 may estimate that the next frame will be non-static and
video encoder 20 may use the parameter set (a,b) for the non-static
sequence (514). Video encoder 20 may then continue with the next
frame and step 502 of the method. The first threshold (T) may be
calculated using the formula k*total_8.times.8 blocks, where k is a
constant, e.g. 0.9, 0.95, or others. The value for first threshold
T may be based on there being a great enough number of inter
encoded blocks. In some examples, the second threshold (t) may be a
constant e.g. 1.5. In other examples, the value of T and t may be
made adaptive based on other factors such as video frame spatial
resolution, video frame temporal resolution (i.e. frames per
second), etc. The second threshold `t` is set to ensure there is
limited motion in a frame using the parameter set (a,b) for the
static sequence. Both thresholds (T and t) are empirical values and
together they make sure the motion in the frame is small and there
are enough inter-encoded blocks.
[0106] The process illustrated in FIG. 5 has a one frame delay and
relies on the assumption that if the previous frame is static,
quite likely the current frame is static. This assumption holds
most of the time and may save time for pre-processing, but there
are times it may fail. When the assumption fails, the selection of
(a,b) will be improper for the frame and coding efficiency may be
penalized. When intensive preprocessing is allowed by video encoder
20, motion may be estimated before encoding, which involves some
delay and coding efficiency may not be penalized, as shown in FIG.
6.
[0107] FIG. 6 is a flowchart illustrating an example method of
selection of parameter set (a.sub.1,b.sub.1) for static and
parameter set (a.sub.2,b.sub.2) for non-static sequences based on
preprocessing of the current frame. The techniques of FIG. 6 may be
implemented by one or more components of video encoder 20.
[0108] In one example of the disclosure, video encoder 20 may be
configured to initialize parameters a and b with some predefined
values (600). The initialized parameters a and b may be based on
data in a configuration file for non-static sequences for
parameters a and b or may be derived based on a set of training
data. For example, parameters a and b may be set based on values of
parameters a and b in param_a and param_b as discussed above. In
such an example, the QP value may be used to select an initial
value for parameter a. Video encoder 20 may determine whether the
current frame is the last frame in the sequence (602). If the
current frame is the last frame ("YES" branch of 602), video
encoder 20 may end encoding (604). If the current frame is not the
last frame ("NO" branch of 602), before encoding the current frame,
video encoder 20 may estimate the motion vectors between the
current frame and the previous frame and obtain a sum of absolute
difference (SAD) between the current frame and the previous
frame(606). Video encoder 20 may be configured to determine if the
sum of horizontal motion vectors (mv_sum_hor) and the sum of
vertical motion vectors (mv_sum_ver) are both less than a first
threshold (t) multiplied by the total number of 8.times.8 blocks
(total_8.times.8 blocks), or the number of blocks with zero motion
(zero_motion_blocks) is greater than a second threshold (s)
multiplied by the total number of 8.times.8 blocks (total_8.times.8
blocks), and the SAD is less than a third threshold (T) multiplied
by the total number of pixels (total_pixels), then video encoder 20
may determine that the current sequence is static ("YES" branch of
608) and use the trained parameter set (a,b) for the static (and/or
noisy) sequence (610). Otherwise ("NO" branch of 608), video
encoder 20 may use the trained parameter (a,b) for non-static
(and/or non-noisy) sequence (612). Then based on the QP and
selected parameter set (a,b), video encoder 20 encodes the frame
(614). Video encoder 20 may then continue with the next frame and
step 602 of the method. Threshold tin FIG. 6 may be the same as in
FIG. 5 and may be set to ensure there is limited motion in a frame
using the parameter set (a,b) for the static sequence. In an
example, threshold t may be set to 1.5. Threshold T can be set as a
small value such that the SAD between frames is small. In an
example, threshold T may be set to 2.5. In an example, threshold S
may be set to be 0.4. Thresholds S, T and t are empirical values,
and thresholds S, T, and t together ensure the motion in the frame
is small and the scene is relatively static.
[0109] FIG. 7 is a flowchart illustrating an example process
according to techniques of the present disclosure. The techniques
of FIG. 7 may be implemented by video encoder 20, one or more
components thereof, or a variety of other devices, in accordance
with the aspects of this disclosure. For instance, processing
circuitry of video encoder 20 (and/or processing circuitry of one
or more components thereof) may perform aspects of the process
illustrated in FIG. 7, as well as the other examples described
below.
[0110] As shown in FIG. 7, video encoder 20 may determine an amount
of movement in a frame of a sequence of frames (700). The frame may
be a current frame (e.g., a frame currently being encoded) or a
previously-encoded frame of the sequence. In one example, video
encoder 20 may be configured to determine the movement in the frame
if the number of inter blocks in the frame (inter_cnt) is greater
than a first threshold (T) and average amplitudes of motion vector
horizontally (mv_sum_hor divided by inter_cnt) and vertically
(my_sum_ver divided by inter_cnt) are less than a second threshold
(t), video encoder 20 may determine the frame is static. Video
encoder 20 may estimate that the next frame will be static and
video encoder 20 may use the parameter set (a,b) for the static
sequence. In another example, video encoder 20 may determine that
if the sum of horizontal motion vectors of either the current frame
or a previously-encoded frame of the sequence (mv_sum_hor) and the
sum of vertical motion vectors of either the current frame or a
previously-encoded frame of the sequence (mv_sum_ver) are both less
than a first threshold (t) multiplied by the total number of
8.times.8 blocks (total_8.times.8 blocks), or the number of blocks
with zero motion (zero_motion_blocks) is greater than a second
threshold (s) multiplied by the total number of 8.times.8 blocks
(total_8.times.8 blocks), and the SAD is less than a third
threshold (T) multiplied by the total number of pixels
(total_pixels), then video encoder 20 may determine that the
current sequence is static and use the trained parameter set (a,b)
for the static (and/or noisy) sequence. Otherwise, video encoder 20
may use the trained parameter (a,b) for non-static sequences.
[0111] Video encoder 20 may be configured to select a parameter set
for a rate estimation model (e.g., a linear rate model) based on
the determined amount of movement in the frame (702). The rate
model may be a linear rate model to estimate a number of non-zero
coefficients in a transform unit. The rate model may also be based
on a position of non-zero coefficients in the frame. Selection of
parameters of the rate model may follow the algorithms illustrated
in FIGS. 5 and 6.
[0112] Video encoder 20 may be configured to encode the frame based
on the rate estimation model (e.g., a linear rate model) (704). In
various examples, video encoder 20 may identify, from a plurality
of transform units of the frame, one or more transform units to be
zeroed out. For instance, video encoder 20 may be configured to
implement the rate estimation model (e.g., a linear rate model) to
each block of the frame, to estimate rate distortion costs
associated with decisions to either zero out each block, or to not
zero out each block. Video encoder 20 may compare the first rate
distortion cost of zeroing out a block of the transform unit
against the second rate distortion cost of not zeroing out the
block. The second rate distortion cost is estimated as being the
sum of a distortion when the coefficients of the block are encoded
and the Lagrangian multiplier multiplied by the number of bits
required to encode coefficients of the block.
[0113] In some examples, video encoder 20 may apply the rate
estimation model to estimate a number of bits to encode a number of
non-zero coefficients in each respective transform unit of the
plurality of transform units. In some examples, video encoder 20
may select the parameter set for the rate estimation model
comprises selecting the parameter set further based on a
quantization parameter (QP) value. In some examples, video encoder
20 may apply the rate estimation model to estimate a number of bits
to encode a number of non-zero coefficients in a transform unit
based on a product of a first parameter of the parameter set
multiplied by a sum of a number of non-zero coefficients in the
frame added to a second parameter of the parameter set. In some
examples, video encoder 20 may select, based on the application of
the rate estimation model, from a plurality of available coding
modes, an optimal coding mode with respect each respective block of
the frame.
[0114] In some examples, video encoder 20 may collect a number of
non-zero coefficients of a previously-encoded frame of the sequence
of frames, collect a number of bits used to encode the non-zero
coefficients of the previously-encoded frame, and update only a
first parameter of the parameter set based on the number of bits
used to encode the non-zero coefficients of the previously-encoded
frame and the number of non-zero coefficients of the
previously-encoded frame to form the adaptively selected parameter
set. In some examples, video encoder 20 may update only the first
parameter of the parameter set comprises updating the first
parameter based on a weighted average of a previous value of the
first parameter and a newly determined value for the first
parameter.
[0115] In some examples, video encoder 20 may determine the amount
of movement in the frame by collecting: (i) a number of
inter-encoded blocks used to encode a previously-encoded frame of
the sequence, and (ii) a degree of motion of the frame of the
sequence. In some examples, video encoder 20 may determine the
amount of movement in the frame based on a determination that both
of a first condition and a second condition are true, where the
first condition comprises the number of inter-encoded blocks used
to encode the previously-encoded frame being greater than a first
threshold value, and the second condition comprises the degree of
motion of the frame being less than a product of the number of
inter-encoded blocks used to encode the frame multiplied by a
second threshold value.
[0116] In some examples, video encoder 20 may select the parameter
set for the rate model based on an amount of movement in the frame
or the previous frame by either selecting a first parameter set
associated with a static sequence based on both of the first
condition and the second condition being true, or by selecting a
second parameter set associated with a non-static sequence based on
at least one of the first condition or the second condition being
false. In some examples, the frame is a current frame, and video
encoder 20 may determine the amount of movement in the current
frame by estimating horizontal and vertical motion vectors and a
distance metric determining the distance between the current frame
and a previously-encoded frame.
[0117] In some examples, video encoder 20 may determine the amount
of movement in the current frame by determining whether at least
one of a first condition or a second condition is true, where the
first condition comprises a sum of horizontal motion vectors and a
sum of vertical motion vectors both being less than a first
threshold value, and the second condition comprises the number of
blocks with zero motion being greater than a second threshold value
and the distance metric being less than a third threshold value. In
some examples, video encoder 20 may select the parameter set by
either selecting a first parameter set associated with a static
sequence based on at least one of the first condition or the second
condition being true, or by selecting a second parameter set
associated with a non-static sequence based on neither the first
condition nor the second condition being true.
[0118] FIG. 8 is a flowchart illustrating an example process
according to techniques of the present disclosure. The techniques
of FIG. 8 may be implemented by video encoder 20, one or more
components thereof, or a variety of other devices, in accordance
with aspects of this disclosure. For instance, processing circuitry
of video encoder 20 (and/or processing circuitry of one or more
components thereof) may perform aspects of the process illustrated
in FIG. 7, as well as the other examples described below.
[0119] Video encoder 20 may estimate, based on a combination of a
first parameter and a number of non-zero coefficients (nnz) in the
frame, a number of bits for non-zero coefficients of a frame
(R.sub.coef) (800). Video encoder 20 may be configured to encode
the frame based on the estimated number of bits for the non-zero
coefficients of the frame (802). Video encoder 20 may be configured
to collect an actual number of bits used to encode the non-zero
coefficients and an actual number of non-zero coefficients in the
frame (804). Video encoder 20 may be configured to update, based on
the actual number of bits used by non-zero coefficients used to
encode the frame and the actual number of non-zero coefficients in
the frame, only the first parameter (806). As used herein, "only
the first parameter" refers to a selection of the first parameter
from the pool of the first parameter and second parameter. As such,
it will be appreciated that values outside of the first parameter
and second parameter may also be updated based on the listed
criteria, but that according to certain techniques of this
disclosure, video encoder 20 may update the first parameter based
on the listed criteria, while leaving the second parameter value
unchanged.
[0120] Video encoder 20 may update the first parameter based on a
weighted average of a previous value of the parameter and a newly
determined value for the at least one parameter. Updating the at
least one parameter may include determining values for S_xy.sub.i,
S_x.sub.i, and S_x2.sub.i, as shown in equations 8-11, discussed
above. The number of bits used by non-zero coefficients of the
frame may be based on a linear rate modeled after the formula:
parameter a multiplied by the number of non-zero coefficients in
the frame added to parameter b. The number of bits used by non-zero
coefficients of the frame may be further estimated based on a
position of non-zero coefficients in the frame.
[0121] Video encoder 20 may form a rate estimation model using the
updated first parameter (new_a) and a second parameter (b) (808).
For instance, the value of the second parameter `b` may be
unchanged, while `new_a` represents the updated value of the first
parameter `a`. In turn, video encoder 20 may select, based on the
rate estimation model, a coding mode for each block in the frame
(810). When the techniques in the present disclosure are tested,
the results show that for IP camera sequences where there is a lot
of noise, on average the method can improve the BD-rate by about 7%
from 12.6% (where parameters a and b are fixed) to 20.1% (where
parameter a is selected adaptively). When the technique of
selection of (a.sub.1,b.sub.1) for static sequences and
(a.sub.2,b.sub.2) for non-static sequences based on previous
frame's coding statistics (illustrated in FIG. 5) is applied, the
BD-rate can also improve by about or over 7%, e.g. from 12.6% to
21.1%. For JCTVC sequences where there is little camera noise, the
BD-rate of the adaptive method is almost same as using the fixed
parameters.
TABLE-US-00003 TABLE 2 Performance Gain of the Proposed Methods ZBD
on (Fixed a, b) Adaptive ZBD Nest sequence (VBR) -12.6% -20.1%
JCTVC sequence (VBR) -4.0% -5.3%
[0122] Table 2 shows the performance gain compared with when zero
block detection (ZBD) is off. When ZBD is on with fixed a, b
parameters, the gain is 12.6% for NEST sequences (e.g., IP camera
sequences with noise), and the gain is 4.0% for JCTVC sequences
(e.g., clean sequences). When parameters a and b are adaptively
selected (e.g., in the manner illustrated in FIG. 5), the gain is
20.1% for NEST sequences and 5.3% for JCTVC sequences.
[0123] In some examples, video encoder 20 may determine the actual
number of bits used to encode the non-zero coefficients of the
frame based on a multiplication product of the updated first
parameter and the number of non-zero coefficients in the frame. In
some examples, video encoder 20 may determine the actual number of
bits used to encode the non-zero coefficients of the frame further
comprises adding the second parameter to the multiplication
product. In some examples, to estimate the number of bits used for
the non-zero coefficients of the frame, video encoder 20 may
estimate the number of bits used for the non-zero coefficients of
the frame further based on respective positions of one or more of
the non-zero coefficients in the frame.
[0124] In some examples, to update only the first parameter, video
encoder 20 may update only the first parameter based on a weighted
average of a previous value of the first parameter and a newly
determined value for the first parameter. In some examples, the
selected coding mode is a first coding mode included in a plurality
of coding modes that are available for encoding the frame, and to
select the coding mode, video encoder 20 may perform rate
estimation for each block in the frame with respect to each
respective coding mode of the plurality, based on respective
applications of the rate estimation model, to obtain a respective
rate distortion cost corresponding to each respective coding mode
of the plurality, and identify, as an optimal coding mode for the
frame, the selected coding mode associated with the lowest
respective rate distortion cost of the respective rate distortion
costs obtained by performing the rate estimations. In some
examples, video encoder 20 may encode the frame according to the
identified optimal coding mode, the method further comprising
zeroing out one or more blocks of the frame to form one or more
zeroed-out blocks. In some examples, to zero out the one or more
blocks of the frame, video encoder 20 may set all transform
coefficients of each of the one or more zeroed-out blocks to a
respective zero value.
[0125] FIG. 9 is a block diagram illustrating an example video
encoder that may implement the techniques described in this
disclosure. For example, processing circuitry of the video encoder
illustrated in FIG. 9 (and/or processing circuitry of one or more
components thereof) may implement various techniques described
above.
[0126] In one or more examples, the functions described may be
implemented in hardware, software, firmware, or any combination
thereof. If implemented in software, the functions may be stored on
or transmitted over, as one or more instructions or code, a
computer-readable medium and executed by a hardware-based
processing unit. Computer-readable media may include
computer-readable storage media, which corresponds to a tangible
medium such as data storage media, or communication media including
any medium that facilitates transfer of a computer program from one
place to another, e.g., according to a communication protocol. In
this manner, computer-readable media generally may correspond to
(1) tangible computer-readable storage media which is
non-transitory or (2) a communication medium such as a signal or
carrier wave. Data storage media may be any available media that
can be accessed by one or more computers or one or more processors
to retrieve instructions, code and/or data structures for
implementation of the techniques described in this disclosure. A
computer program product may include a computer-readable
medium.
[0127] By way of example, and not limitation, such
computer-readable storage media can comprise RAM, ROM, EEPROM,
CD-ROM or other optical disk storage, magnetic disk storage, or
other magnetic storage devices, flash memory, or any other medium
that can be used to store desired program code in the form of
instructions or data structures and that can be accessed by a
computer. Also, any connection is properly termed a
computer-readable medium. For example, if instructions are
transmitted from a website, server, or other remote source using a
coaxial cable, fiber optic cable, twisted pair, digital subscriber
line (DSL), or wireless technologies such as infrared, radio, and
microwave, then the coaxial cable, fiber optic cable, twisted pair,
DSL, or wireless technologies such as infrared, radio, and
microwave are included in the definition of medium. It should be
understood, however, that computer-readable storage media and data
storage media do not include connections, carrier waves, signals,
or other transient media, but are instead directed to
non-transient, tangible storage media. Disk and disc, as used
herein, includes compact disc (CD), laser disc, optical disc,
digital versatile disc (DVD), floppy disk and Blu-ray disc, where
disks usually reproduce data magnetically, while discs reproduce
data optically with lasers. Combinations of the above should also
be included within the scope of computer-readable media.
[0128] Instructions may be executed by one or more processors, such
as one or more digital signal processors (DSPs), general purpose
microprocessors, application specific integrated circuits (ASICs),
field programmable logic arrays (FPGAs), or other equivalent
integrated or discrete logic circuitry. Accordingly, the term
"processor," as used herein may refer to any of the foregoing
structure or any other structure suitable for implementation of the
techniques described herein. In addition, in some aspects, the
functionality described herein may be provided within dedicated
hardware and/or software modules configured for encoding and
decoding, or incorporated in a combined codec. Also, the techniques
could be fully implemented in one or more circuits or logic
elements.
[0129] The techniques of this disclosure may be implemented in a
wide variety of devices or apparatuses, including a wireless
handset, an integrated circuit (IC) or a set of ICs (e.g., a chip
set). Various components, modules, or units are described in this
disclosure to emphasize functional aspects of devices configured to
perform the disclosed techniques, but do not necessarily require
realization by different hardware units. Rather, as described
above, various units may be combined in a codec hardware unit or
provided by a collection of interoperative hardware units,
including one or more processors as described above, in conjunction
with suitable software and/or firmware.
[0130] Various examples have been described. These and other
examples are within the scope of the following claims.
* * * * *
References