Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02329426 2000-10-18
WO 99/55013 1 PCT/US99/08673
METHOD AND APPARATUS OF SUPPORTING
A VIDEO PROTOCOL IN A NETWORK ENVIRONMENT
BACKGROUND OF THE INVENTION
1. FIELD OF THE INVENTION
This invention relates to the field of digital video, and, more
specifically, to digital video applications in a network environment.
2. BACKGROUND ART
Computers and computer networks are used to exchange information
in many fields such as media, commerce, and telecommunications, for
example. One form of information that is commonly exchanged is video data
(or image data), i.e., data representing a digitized image or sequence of
images.
A video conferencing feed is an example of telecommunication information
which includes video data. Other examples of video data include video
streams or files associated with scanned images, digitized television
performances, and animation sequences, or portions thereof, as well as other
forms of visual information that are displayed on a display device. It is also
possible to synthesize video information by artificially rendering video data
from two or three-dimensional computer models.
For the purposes of this discussion, the exchange of information
between computers on a network occurs between a "transmitter" and a
"receiver." In video applications, the information contains video data, and
the services provided by the transmitter are associated with the processing
CA 02329426 2000-10-18
WO 99/55013 2 PCT/US99/08673
and transmission of the video data. A problem with current network systems
is that multiple services provided by one or more transmitters may provide
video data using different video protocols. The complexity of the receiver is
necessarily increased by the need to accommodate each of the different video
protocols. Also, the amount of data associated with video applications is very
large. The transmission of such large amounts of data over a network can
result in bandwidth utilization concerns. The following description of video
technology and an example network scheme are given below to provide a
better understanding of the problems involved in transmitting video data
over a network.
General Video Techn,QhT
In digital video technology, a display is comprised of a two
dimensional array of picture elements, or "pixels," which form a viewing
plane. Each pixel has associated visual characteristics that determine how a
pixel appears to a viewer. These visual characteristics may be limited to the
perceived brightness, or "luminance," for monochrome displays, or the
visual characteristics may include color, or "chrominance," information.
Video data is commonly provided as a set of data values mapped to an array
of pixels. The set of data values specify the visual characteristics for those
pixels that result in the display of a desired image. A variety of color
models
exist for representing the visual characteristics of a pixel as one or more
data
values.
RGB color is a commonly used color model for display systems. RGB
color is based on a "color model" system. A color model allows convenient
specification of colors within a color range, such as the RGB (red, green,
blue)
CA 02329426 2000-10-18
WO 99/55013 3 PCTNS99/08673
primary colors. A color model is a specification of a three-dimensional color
coordinate system and a three-dimensional subspace or "color space" in the
coordinate system within which each displayable color is represented by a
point in space. Typically, computer and graphic display systems are three-
s phosphor systems with a red, green and blue phosphor at each pixel location.
The intensities of the red, green and blue phosphors are varied so that the
combination of the three primary colors results in a desired output color.
The RGB color model uses a Cartesian coordinate system. The
subspace of interest in this coordinate system is known as the "RGB color
cube" and is illustrated in Figure 1. Each corner of the cube represents a
color
that is theoretically one-hundred percent pure - that is, the color at that
location contains only the color specified, and contains no amount of other
colors. In the RGB color cube, the corners are defined to be black, white,
red,
green, blue, magenta, cyan, and yellow. Red, green and blue are the primary
colors, black is the absence of color, white is the combination of all colors,
and
cyan, magenta and yellow are the complements of red, green and blue.
Still referring to Figure 1, the origin of the coordinate system
corresponds to the black corner of the color cube. The cube is a unit cube so
that the distance between the origin and adjacent corners is 1. The red corner
is thus at location, {1,0,0). The axis between the origin (black) and the red
corner is referred to as the red axis 110.
The green corner is at location (0,1,0) and the axis 120 between the black
origin and the green corner is referred to as the green axis. The blue corner
is
at location (0,0,1) and the blue axis 130 is the axis between the blue corner
and
the origin.
CA 02329426 2000-10-18
WO 99/55013 4 PCT/US99/08673
Cyan is at corner (0,1,1), magenta is at corner (1,0,1) and yellow is at
corner (1,1,0). The corner opposite the origin on the cube's diagonal at
location (1,1,1) is the white corner.
A color is defined in the color cube by a vector having red, green and
blue components. For example, vector 180 is the resultant of vectors 1808,
{along the red axis), vector 1806 (along the green axis) and vector 180B
(along
the blue axis). The end point of vector 180 can be described mathematically by
0.258 + 0.506 + 0.75B. The end of this vector defines a point in color space
represented mathematically by the sum of its red, green and blue
components.
An example of a system for displaying RGB color is illustrated in
Figure 2. A refresh buffer 140, also known as a video RAM, or VRAM, is
used to store color information for each pixel on a video display, such as CRT
display 160. A DRAM can also be used as buffer 140. The VRAM 140 contains
one memory location for each pixel location on the display 160. For example,
pixel 190 at screen location XpYp corresponds to memory location 150 in the
VRAM 140. The number of bits stored at each memory location for each
display pixel varies depending on the amount of color resolution required.
For example, for word processing applications or display of text, two
intensity
values are acceptable so that only a single bit need be stored at each memory
location (since the screen pixel is either "on" or "off"). For color images,
however, a plurality of intensities must be definable. For certain high end
color graphics applications, it has been found that twenty-four bits per pixel
produces acceptable images.
CA 02329426 2000-10-18
WO 99/55013 5 PCT/US99/08b73
Consider, for example, that in the system of Figure 2, twenty-four bits
are stored for each display pixel. At memory location 150, there are then
eight
bits each for the red, green and blue components of the display pixel. The
eight most significant bits of the VRAM memory location could be used to
represent the red value, the next eight bits represent the green value and the
eight least significant bits represent the blue value. Thus, 256 shades each
of
red, green and blue can be defined in a twenty-four bit per pixel system.
When displaying the pixel at X0, Y0, the bit values at memory location 150 are
provided to video driver 170. The bits corresponding to the R component are
provided to the R driver, the bits representing the green component are
provided to the G driver, and the bits representing the blue component are
provided to the blue driver. These drivers activate the red, green and blue
phosphors at the pixel location 190. The bit values for each color, red, green
and blue, determine the intensity of that color in the display pixel. By
varying the intensities of the red, green and blue components, different
colors
may be produced at that pixel location.
Color information may be represented by color models other than
RGB. One such color space is known as the YUV (or Y'CbCr as specified in
ITU.BT-601) color space which is used in the commercial color TV
broadcasting system. The YUV color space is a recoding of the RGB color
space, and can be mapped into the RGB color cube. The RGB to YUV
conversion that performs the mapping is defined by the following matrix
equation:
CA 02329426 2000-10-18
WO 99/55013 6 PCT/US99/08673
Y' Y' 0.299 0.587 0.114 R'
U' - Cb - -0.169 -0.331 0.500 G'
V' Cr 0.500 -0.419 -0.081 B'
The inverse of the matrix is used for the reverse conversion. The Y
axis of the YUV color model represents the luminance of the display pixel,
and matches the luminosity response curve for the human eye. U and V are
chrominance values. In a monochrome receiver, only the Y value is used.
In a color receiver, all three axes are used to provide display information.
In operation, an image may be recorded with a color camera, which is
an RGB system, and converted to YUV for transmission. At the receiver, the
YUV information is then retransformed into RGB information to drive the
color display.
Many other color models are also used to represent video data. For
example, CMY (cyan, magenta, yellow) is a color model based on the
complements of the RGB components. There are also a variety of color
models, similar to YUV, which specify a luminance value and multiple
chrominance values, such as the YIQ color model. Each color model has its
own color transformation for converting to a common displayable video
format such as RGB. Most transformations may be defined with a transform
matrix similar to that of the YIQ color space.
There are many color formats used in the prior art for transmitting
image and video data over networks. Some examples of existing color
CA 02329426 2000-10-18
WO 99/55013 7 PCT/US99/08673
formats axe H.261 and H.263, which are used in digital telephony, and MPEG1,
MPEG2 and MJPEG. These color formats use compression schemes to reduce
the amount of data being transmitted. For example, many color formats use a
variation of DCT (discrete cosine transform) compression to perform
compression in the frequency domain. A form of variable length Huffman
encoding may also be implemented to reduce bandwidth requirements.
Specialized compression/decompression hardware or software is often used
to perform the non-trivial conversion of data in these color formats to data
for display.
Network Transmission Of Video Data
As has been described, there exist a variety of color formats for video
data. These variations allow for a large number of different possible video
protocols. It becomes problematic for a receiver on a network to handle all
possible video protocols that might be used by different transmitters and
services acting as video data sources on the network. The problems
associated with multiple video protocols are described below with reference
to the sample network system illustrated in Figure 3. Figure 3 illustrates a
sample network system comprising multiple transmitters 300A-300C for
sourcing video data and a single receiver 303. Receiver 303 is equipped with
one or more display devices for providing video output associated with
received video data.
In the example of Figure 3, transmitters 300A, 300B and 300C, and
receiver 303 are coupled together via network 302, which may be, for
example, a local area network (LAN). Transmitter 300A transmits video data
along network connection 301A to network 302 using video protocol A.
CA 02329426 2000-10-18
WO 99/55013 8 PCT/US99/08673
Transmitter 300B transmits video data along network connection 301B to
network 302 using video protocol B. Transmitter 300C transmits video data
along network connection 301C to network 302 using video protocol C. Thus,
receiver 303 may receive video data over network connection 305 from
network 302 under any of video protocols A, B or C, as well as any other
protocols used by other transmitters connected to network 302, or used by
multiple services embodied within one of transmitters 300A-300C.
Receiver 303 may be equipped with different video cards (i.e.,
specialized hardware for video processing) or software plug-ins to support
each video protocol, but this increases the complexity of the receiver, and
necessitates hardware or software upgrades when new video protocols are
developed. For systems wherein it is a goal to minimize processing and
hardware requirements ,for a receiver, the added complexity of supporting
I5 multiple protocols is undesirable.
An issue in all network applications is utilization of network
bandwidth. For video applications, bandwidth is an even greater concern due
to the large amounts of data involved in transmitting one or more frames of
video data. For example, consider a raw workstation video signal of twenty-
four bit RGB data, sent in frames of 1280 x 1024 pixels at thirty frames per
second. The raw workstation video represents 240 MBps (megabytes per
second) of continuous data. Even for smaller frame sizes, video data can
represent a significant load on a network, resulting in poor video
performance if the required bandwidth cannot be provided. Further, other
applications on the network may suffer as bandwidth allocations are reduced
to support the video transmission.
CA 02329426 2000-10-18
WO 99/55013 9 PCT/US99/08673
SUMMARY OF THE INVENTION
A method and apparatus of supporting a video protocol in a network
environment is described. In an embodiment of the invention, video
processing and hardware requirements associated with a receiver are
minimized by specifying a single video protocol for transmission of video
data between transmitters and receivers an a network. The protocol specifies
a color format that allows for high video quality and minimizes the
complexity of the receiver. Transmitters are equipped with transformation
mechanisms that provide for conversion of video data into the designated
protocol as needed. Compression of the components of the color format is
provided to reduce transmission bandwidth requirements.
In one embodiment of the invention, aspects of the designated
protocol compensate for problems associated with transmitting video data
over a network. The designated protocol specifies a color format including a
luminance value and two chrominance values. Quantized differential
coding is applied to the luminance value and subsampling is performed on
the chrominance values to reduce transmission bandwidth requirements. In
one embodiment of the invention, upscaiing of video data is performed at
the receiver, whereas downscaling is performed at the transmitter. Various
display sizes can thus be accommodated with efficient use of network
bandwidth.
CA 02329426 2000-10-18
WO 99/55013 1~ PCT/US99/08673
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is a diagram of an RGB color space.
Figure 2 is a block diagram of a video display apparatus.
Figure 3 is a block diagram of a network system having a single
receiver and multiple transmitters.
Figure 4A is a flow diagram illustrating a luma compression scheme in
accordance with an embodiment of the invention.
Figure 4B is a flow diagram illustrating a luma decompression scheme
in accordance with an embodiment of the invention.
Figure 5 is a diagram illustrating a subsampling and upsampling
process in accordance with an embodiment of the invention.
Figure 6A is a flow diagram illustrating a video data transmission
process in accordance with an embodiment of the invention.
Figure 6B is a flow diagram illustrating a video data reception process
in accordance with an embodiment of the invention.
Figure 7 is a block diagram of a computer execution environment.
Figure 8 is a block diagram of a human interface device computer
system.
Figure 9 is a block diagram of an embodiment of a human interface
device.
CA 02329426 2000-10-18
WO 99/55013 11 PCT/US99/08673
DETAILED DESCRIPTION OF THE INVENTION
The invention is a method and apparatus of supporting a video
protocol in a network environment. In the following description, numerous
specific details are set forth to provide a more thorough description of
embodiments of the invention. It will be apparent, however, to one skilled
in the art, that the invention may be practiced without these specific
details.
in other instances, well known features have not been described in detail so
as not to obscure the invention.
Single Video Protocnl For Networked Transmissions
In an embodiment of the invention, a single video protocol is used for
transmission of video data between a transmitter and a receiver. The
transmitter of the video data is responsible for supplying video data in
accordance with the designated protocol. For example, a transmitter and its
internal video services are configured to perform any necessary protocol
transformations to bring video data into conformance with the designated
protocol before transmission to a receiver. Hardware and processing
requirements of the receiver are minimized as only one video protocol need
be supported at the receiver.
Though discussed in this specification primarily as being applied to
video data transmitted one-way from a transmitter to a receiver, video data
may also be transmitted from the receiver to the transmitter using the
designated protocol. The transmitter may then process the video data in the
form of the designated protocol, or the transmitter may convert the video
data into another video protocol for further processing.
CA 02329426 2000-10-18
WO 99155013 12 PCTNS99/08673
The designated protocol is chosen to give high video quality to
satisfactorily encompass all other video protocols while permitting strategic
compression based on knowledge of human perception of luminance and
chrominance. The high video quality of the designated protocol ensures that
any necessary protocol transformations by a transmitter do not result in a
significant loss of video quality from the original video data. An example of
a protocol that provides high video quality with compression is a protocol
specifying a color format with quantized differential coding of the luminance
value and subsampling of chrominance values.
A transmitter may support video applications using the designated
protocol, and the transmitter may be configured with mechanisms, such as
hardware cards or software plug-ins or drivers, to convert between other
video protocols and the designated protocol, for example, using color model
matrix transformations.
In an embodiment of the invention, data packets are used to transmit
variably sized blocks of video data between a transmitter and a receiver using
a connectionless datagram scheme. A connectionless scheme means that
each packet of video data, i.e., each video block, is processed as an
independent unit, and the loss of a data packet does not affect the processing
of other data packets. This independence provides for robust video
processing even on unreliable networks where packet loss may be
commonplace.
Some networks are prone to periodic packet loss, i.e., packet loss at
regular intervals. This periodic behavior can result in the stagnation of
CA 02329426 2000-10-18
WO 99/55013 13 PCT/US99/08673
portions of the video display as the same video blocks are repeatedly lost. To
prevent video block stagnation, the spatial order in which video blocks are
sent to the receiver for display may be pseudo-randomly determined to
disrupt any periodicity in packet performance.
In one embodiment, the data packets containing video data are
provided with a sequence number. By tracking the sequence numbers, the
receiver can note when a sequence number is skipped, indicating that the
packet was lost during transmission. The receiver can then return to the
transmitter a list or range of sequence numbers identifying the lost packet or
packets. When the transmitter receives the list or range of sequences, the
transmitter can decide whether to ignore the missed packets, resend the
missed packets {such as for still images), or send updated packets (such as
for
streaming video that may have changed since the packet was lost).
In one embodiment of the invention, the video data packet comprises
the following information:
Sequence number - A video stream is processed as a series of blocks of
video data. The sequence number provides a mechanism for
the receiver to tell the transmitter what sequence numbers have
been missed (e.g., due to packet loss), so that the transmitter may
determine whether to resend, update or ignore the associated
video block.
X field - The X field designates the x-coordinate of the receiver's display
device wherein the first pixel of the video block is to be
displayed.
Y field - The Y field designates the y-coordinate of the receiver's display
device wherein the first pixel of the video block is to be
displayed.
CA 02329426 2000-10-18
WO 99/55013 14 PCT/US99/08673
Width - The width field specifies the width of the destination rectangle
on the receiver's display device wherein the video block is to be
displayed.
Height - The height field specifies the height of the destination
rectangle on the receiver's display device wherein the video
block is to be displayed.
Source_w - The source width specifies the width of the video block in
pixels. Note that the source width may be smaller than the
width of the destination rectangle on the receiver's display
device. If this is so, the receiver will upscale the video block
horizontally to fill the width of the destination rectangle. The
source width should not be larger than the width of the
destination rectangle as this implies downscaling, which should
be performed by the transmitter for efficiency.
Source_h - The source height specifies the height of the video block in
pixels. Note that, as with source_w, the source height may be
smaller than the height of the destination rectangle on the
receiver's display device. As above, the receiver will upscale the
video block vertically to fill the height of the destination
rectangle. The source height should not be larger than the
height of the destination rectangle as this implies downscaling,
which should be performed by the transmitter for efficiency.
Luma encoding - The luma encoding field allows the transmitter to
designate a particular luma encoding scheme from a set of
specified luma encoding schemes.
Chroma_sub_X - This field allows the transmitter to designate the
degree of horizontal subsampling performed on the video data
chroma values.
Chroma_sub Y - This field allows the transmitter to designate the
degree of vertical subsampling performed on the video data
chroma values.
Video data - The video data includes (source_w * source_h) pixel luma
values (Y), and ((source_w/chroma_sub_x)
(source_h/chroma_sub_y)) signed chroma values (U, V or Cb,
Cr).
CA 02329426 2000-10-18
WO 99/55013 15 PC'TNS99/08673
YUV ~Y'CbCr) Color
In an embodiment of the invention, the color model of the chosen
protocol is specified by the International Telecommunications Union in
ITU.BT-601 referring to an international standard for digital coding of
television pictures using video data components Y'CbCr, where Y' is a
luminance or "iuma" value, Cb (or U') is a first chromaticity or "chroma"
value represented as a blue color difference proportional to (B'-Y') and Cr
(or
V') is a second chroma value represented as a red color difference
proportional to (R'-Y') (Note that primed values such as Y' indicate a gamma
corrected value). This ITU specification is independent of any scanning
standard and makes no assumptions regarding the "white" point or CRT
gamma. For 0 <_ (R,G,B) <= 1, the range for Y' is 0 <= Y' <= 1 and the range
for Cb and Cr is -0.5 <_ (Cb, Cr) <= 0.5.
The R'G'B' <--> Y'CbCr color transforms are as follows:
Y' Y' 0.299 0.587 0.114 R'
U' - Cb - -0.169 -0.331 0.500 G'
V' Cr 0.500 -0.419 -0.081 B'
R' 1 0 1.403 Y'
G' - 1 -0.344 -0.714 . U'
B' 1 1.773 0 V'
CA 02329426 2000-10-18
WO 99/55013 16 PCT/US99/08673
Under the specified protocol, the transmitter performs any
transformations required to convert the video data into the YUV format.
This may include performing the RGB to YUV matrix conversion shown
above to convert RGB data. Transformations may also include
decompression from other color formats (e.g., H.261, MPEG1, etc.). The
receiver can drive an RGB display device by performing the above matrix
operation to convert incoming YUV (Y'CbCr) data received from a
transmitter into RGB data for display at the display rectangle identified in
the
data packet. No other color transformations are necessary at the receiver.
The receiver is also able to accept RGB data in the same video block format
because RGB data is directly supported in the receiver. For transmission
efficiency, however, any sizable video data transfers between a transmitter
and receiver should be performed in the YUV color format to take advantage
of the compression schemes described below.
puma Compression
In each data packet containing a video block, there are (source_w
source_h) luma values -- one for each pixel. If the luma encoding field
indicates that no encoding is being performed, the luma values are unsigned
eight-bit values. If, however, luma encoding is indicated in the luma
encoding field, the luma values are encoded to achieve a compression ratio of
2:1. In an embodiment of the invention, the luma value "Y" is compressed
using a quantized differential coding (QDC) scheme described below. in other
embodiments, other compression schemes may be specified in the luma
encoding field.
CA 02329426 2000-10-18
WO 99/55013 17 PCT/US99/08673
The luma compression herein described is based on the premise that
luma values do not tend to vary significantly from one pixel to another. It is
therefore possible to transmit the difference value between luma values for
consecutive pixels rather than the luma values themselves. Further, the
luma difference values can be satisfactorily quantized to one of sixteen
quantization levels, each of which is identified by a four-bit code word. The
quantization is non-linear, with more quantization levels near zero where
luma differences between consecutive pixels are more likely to occur.
In one embodiment, the luma difference quantization is performed
according to the following table:
Difference Range Code (Binary Quantized Difference
Lev,
-255 to -91 0 (0000) -100
-90 to -71 1 (0001 ) -80
-70 to -51 2 (0010) -60
-50 to -31 3 (0011) -40
-30 to -16 4 (0100) -20
-15 to -8 5 (0101) -10
-7 to -3 6 (0110) -4
-2 to 0 7 (0111 ) -1
1 to 2 8 (1000) 1
3to7 9(1001) 4
8to15 10(1010) 10
16 to 30 11 (1011) 20
31 to 50 12 (1100) 40
51 to 70 13 (1101) 60
71 to 90 14 (1110) 80
91 to 255 15 (1111) 100
Figure 4A is a flow diagram describing how the luma compression
process is performed in accordance with an embodiment of the invention.
The scheme is based on a subtraction of a "last_value" from the current pixel
luma value to generate the luma difference. "Last value" is used to model
the luma value of the preceding pixel. To prevent divergence of the
CA 02329426 2000-10-18
WO 99/55013 PCTNS99/08673
18
compression and decompression processes, the "last value" is modeled to
account for the previous quantized luma difference rather than to match the
actual luma value of the last pixel. The modeled "last value" in the
compression process therefore matches the corresponding modeled
"last value" extracted in the decompression process.
Because the compression scheme is based on differences in luma
values in rows of luma data, the first luma value in each row has no luma
value with which to form a difference. To provide a starting point, in step
400, an initial "last value" is assigned from the middle of the luma range. In
step 401, the first luma value in the row of pixels is set as the current luma
value. In step 402, the "last value" is subtracted from the current luma
value to generate a current luma difference value. The current luma
difference is applied to a quantization function, in step 403, that outputs
the
quantized difference code. In step 404, the difference code is placed in the
video block data packet.
In step 405, the quantized difference level corresponding to the
difference code is determined, and, in step 406, the "last value" is updated
by
incrementing by the quantized difference level. In step 407, "last value" is
clamped to prevent overflow. The clamping function is:
lam
0 x<0
255 x > 255
x otherwise
In step 408, if there are more pixel luma values in the row, then
process flows to step 409 wherein the next luma value is set as the current
Iuma value. After step 409, the process returns to step 402. If there is no
CA 02329426 2000-10-18
WO 99/55013 19 PCTNS99/08673
further pixel luma value in the row at step 408, then, in step 410, a
determination is made whether there are further rows to process in the video
block. If there are further rows to compress, the next row is designated in
step
411 and the process returns to step 400. If there are no further rows at step
410, the luma compression is completed for the current video block.
A luma decompression process is illustrated in the flow diagram of
Figure 4B. In step 412, the "last value" is set to the same midrange value as
is done for the beginning of a row in the compression scheme. In step 413,
the first luma difference code is set as the current luma difference code. The
quantized difference value is determined from the current luma difference
code in step 414. In step 415, the "last_value" is incremented by the
quantized
difference value. In step 416, "last value" is clamped to prevent overflow.
In step 417, "last value," now representing the decompressed current luma
value, is written to a buffer.
If, in step 418, there are further luma difference codes in the current
row of the video block, the next difference code is set as the current luma
difference code in step 419, and the process returns to step 414. If, in step
418,
there are no further luma difference codes in the current row, the process
continues to step 420. In step 420, if there are no further rows in the video
block, decompression is complete for the current video block. If, in step 420,
there are further rows of luma difference codes, the next row of luma
difference codes is set as the current row in step 421, and the process
returns
to step 412.
CA 02329426 2000-10-18
WO 99/55013 2o PCTNS99/08673
Chroma Compression
The human eye is less sensitive to chroma information than to luma
information, particularly in a spatial sense. For example, if, in generating
an
image, some of the chroma information is spread beyond the actual edges of
an object in the image, the human eye will typically pick up on the edge
queues provided by the luma information and overlook the inaccuracies in
the spatial location of the chroma information. For this reason, some
latitude can be taken with the manner in which chroma information is
provided. Specifically, subsampling may be performed without significantly
degrading visual quality. Subsampling may consist of sampling a single
chroma value from
In accordance with an embodiment of the invention, the amount of
chroma information, and hence the amount of chroma compression, is
specified by the chroma_sub_X and chroma_sub Y fields in the video block
data packet. If the values for both of those fields are zero, then there is no
chroma information and the video block is monochrome, i.e., luma only.
One possible specification for chroma subsampling is:
0 - No chroma values; monochrome image
1 - Subsample by one (i.e., no subsampling)
2 - Subsample by two
3 - Subsample by four
Further subsample arrangements may be provided by extending the above
specification. Chroma_sub_X and chroma_sub Y independently specify
subsampling along respective axes. Several subsampling arrangements
achieved by different combinations of chroma sub X and chroma sub Y, as
defined above, are:
CA 02329426 2000-10-18
WO 99/55013 21 PCT/US99/08673
~hroma sub X chroma ub Y one chroma value comprP~~,sion
s per
0 0 0, no chroma data --
0 1 not permitted --
1 0 not permitted --
1 1 pixel 1:1
2 1 1 x 2 pixel array 2:1
1 2 2 x 1 pixel array 2:1
3 1 1 x 4 pixel array 4:1
1 3 4 x 1 pixel array 4:1
3 2 2 x 4 pixel array 8:1
2 3 4 x 2 pixel array 8:1
3 3 4 x 4 pixel array 16:1
Subsampling may be performed when packing data into a video block
data packet by taking data only at the specified intervals in the specified
directions. For example, for (chroma_sub X, chroma_sub Y) _ (3, 2),
chroma data would be taken at every fourth pixel along each row, and every
other row would be skipped. Other schemes may be used to select a single
pixel from the subsampling matrix, such as pseudo-random assignments.
Further, the chroma values from each pixel in a subsampling matrix may be
used to calculate a single set of average chroma values (U, V) for each
subsampling matrix .
Subsampling is performed as the video block data packet is being
packed and may occur before or after luma compression as luma and chroma
compression are substantially independent. When the video block data
packet reaches the receiver, the chroma subsarnples are upsampled prior to
being converted to RGB. Upsampling may be accomplished by taking the
subsampled chroma information and duplicating the chroma values for each
pixel in the associated subsampling matrix.
CA 02329426 2000-10-18
WO 99/55013 22 PCTNS99/08673
Figure 5 illustrates subsampling and upsampling processes carried out
on an 8 x 8 array of pixels with subsampling performed in 2 x 4 matrices
(chroma_sub_X = 2, chroma_sub Y = 3). 8 x 8 pixel array 500 represents the
original video data prior to subsampling. 4 x 2 pixel array 501 represents the
video data after subsampling by the transmitter, and includes the data that
would be transmitted to the receiver. 8 x 8 pixel array 502 represents the
video data after upsampling at the receiver.
The subsampling matrices are identified in array 500 as those pixel cells
having the same index number. For example, all of the pixel cells containing
a "1" are in the same subsampling matrix. 4 x 2 array 501 contains the
subsampled data from array 500. The chroma values associated with those
pixels with index "1" are averaged into chroma average value A1 (A1
comprises an averaged U value and an averaged V value) placed into the first
cell of array 501. Similarly, the chroma values for those pixels with index
"2"
are averaged into chroma average value A2 and placed into the second
location in array 501. The other subsampling matrices indexed as "3"-"8" are
averaged similarly. The compression ratio seen between array 500 and array
501 is 8:1.
Array 501 is upsampled into array 502 by placing the averaged chroma
values A1-AS into the positions corresponding to the respective original
subsampling matrices. For example, averaged chroma value A1 is placed into
each of the pixels in the upper left corner of 502 shown as containing "A1."
The insensitivity of the human eye to spatial errors in chroma information
allows the averaged chroma values to provide satisfactory viewing results.
CA 02329426 2000-10-18
WO 99/55013 23 PCTNS99/08673
U~scaling And Downscaling Of Vid~Q Data
In an embodiment of the invention, the pixel array size of the source
video block may differ from the size of the destination rectangle on the
receiver's display. This size variation allows for a receiver with a large
display to "blow up" or upscale a small video scene to make better use of the
display resources. For example, a receiver may wish to upscale a 640 x 480
video stream to fill a 1024 x 1024 area on a large display device. Also, a
receiver may have a smaller display than the size of a video stream. For this
case, the video stream should be scaled down to be fully visible on the small
display.
In accordance with an embodiment of the invention, upscaling is
performed by the receiver, whereas downscaling is performed by the
transmitter. One reason for this segregation of scaling duties is that scaled
down video data requires lower network bandwidth to transmit. By
downscaling video data on its own, the transmitter avoids sending video data
that would be later discarded by the receiver. This also permits some
simplification of the receiver in that resources, such as software code for
downscaling video data, are not needed at the receiver.
Upscaling typically involves duplication of video data. It would be
inefficient to send duplicated video data over a network. Therefore, the
receiver performs all upscaling operations after receipt of the video data in
its
smaller form. Upscaling of video data is supported in the fields associated
with the video data packet. Specifically, the video protocol provides separate
fields for specifying the video source pixel array size and the destination
CA 02329426 2000-10-18
WO 99/55013 24 PCTNS99/08673
display rectangle size. The amount of horizontal scaling is (width/source_w),
and the amount of vertical scaling is (height/source_h).
Upscaling is performed after the video data has been decompressed and
transformed into RGB format, though in certain embodiments upscaling may
precede, or be combined with, the decompression steps. The receiver expands
the video data vertically, horizontally or both as need to make the video data
fill the designated display rectangle. Expanding video data may be performed
as simply as doubling pixel values, but more advanced image filtering
techniques may be used to affect re-sampling of the image for better display
quality.
Video Data Process Implementing Protocol
Figure 6A is a flow diagram illustrating how a transmitter processes
video data in accordance with an embodiment of the invention. In step 600,
the transmitter acquires video data for transmission to a receiver. The video
data may be acquired by any mechanism, such as capture of a video signal
using a hardware capture board, generation of video data by a video service,
or input of video data from a video input device such as a video camera.
In step 601, if necessary, the video data is decompressed or converted
into YUV color format in accordance with the established protocol. In step
602, the transmitter downscales the video data if the transmitter determines
that downscaling is needed. The luma values of the YUV video data are
compressed, in step 603, using the quantized differential coding (QDC)
scheme described herein, and loaded into a data packet. In step 604, the
transmitter subsamples the chroma values of the YUV video data and loads
CA 02329426 2000-10-18
WO 99/55013 25 PCT/US99/08673
the subsampled chroma values into the data packet. The completed data
packet containing the video data for a video block is sent to a receiver in
step
605. After transmitting the data packet, the process returns to step 600.
Figure 6B is a flow diagram illustrating how a receiver processes video
data in accordance with an embodiment of the invention. In block 606, the
receiver receives a compressed/subsampled YUV video block data packet
from the transmitter. The receiver decompresses the luma values of the data
packet in step 607, and upsamples the chroma values in step 608. With full
YUV video data, the receiver performs a color transformation to convert the
YUV video data to RGB data. If the destination display rectangle noted in the
data packet header is larger than the source video data, the receiver performs
any necessary upscaling to fill the designated display rectangle with the
source
video data. In step 611, the video data is loaded into a video buffer for
display
on a the receiver's display device. After loading the video data into the
video
buffer, the process returns to step 606.
Embodiment of Computer Execution Environment~Hardware)
An embodiment of the invention can be implemented as computer
software in the form of computer readable code executed on a general
purpose computers such as computer 700 illustrated in Figure 7, or in the
form of bytecode class files executable within a JavaTM runtime environment
running on such a computer. A keyboard 710 and mouse 711 are coupled to a
bi-directional system bus 718. The keyboard and mouse are for introducing
user input to the computer system and communicating that user input to
processor 713. Other suitable input devices may be used in addition to, or in
place of, the mouse 711 and keyboard 710. I/O (input/output) unit 719
CA 02329426 2000-10-18
WO 99/55013 26 PCT/US99/08673
coupled to bi-directional system bus 718 represents such I/O elements as a
printer, A/V (audio/video) I/O, etc.
Computer 700 includes a video memory 714, main memory 715 and
mass storage 712, all coupled to bi-directional system bus 718 along with
keyboard 710, mouse 711 and processor 713. The mass storage 712 may
include both fixed and removable media, such as magnetic, optical or
magnetic optical storage systems or any other available mass storage
technology. Bus 718 may contain, for example, thirty-two address lines for
addressing video memory 714 or main memory 715. The system bus 718 also
includes, for example, a 32-bit data bus for transferring data between and
among the components, such as processor 713, main memory 715, video
memory 714 and mass storage 712. Alternatively, multiplex data/address
lines may be used instead of separate data and address lines.
In one embodiment of the invention, the processor 713 is a
microprocessor manufactured by Motorola, such as the 680X0 processor or a
microprocessor manufactured by Intel, such as the 80X86, or Pentium
processor, or a SPARCTM microprocessor from Sun MicrosystemsTM, Inc.
However, any other suitable microprocessor or microcomputer may be
utilized. Main memory 715 is comprised of dynamic random access memory
(DRAM). Video memory 714 is a dual-ported video random access memory.
One port of the video memory 714 is coupled to video amplifier 716. The
video amplifier 716 is used to drive the cathode ray tube {CRT) raster monitor
717. Video amplifier 716 is well known in the art and may be implemented
by any suitable apparatus. This circuitry converts pixel data stored in video
memory 714 to a raster signal suitable for use by monitor 717. Monitor 717 is
a type of monitor suitable for displaying graphic images. Alternatively, the
CA 02329426 2000-10-18
WO 99/55013 27 PCT/US99/08673
video memory could be used to drive a flat panel or liquid crystal display
(LCD), or any other suitable data presentation device.
Computer 700 may also include a communication interface 720
coupled to bus 718. Communication interface 720 provides a two-way data
communication coupling via a network link 721 to a local network 722. For
example, if communication interface 720 is an integrated services digital
network (ISDN) card or a modem, communication interface 720 provides a
data communication connection to the corresponding type of telephone line,
which comprises part of network link 721. If communication interface 720 is
a local area network (LAN) card, communication interface 720 provides a data
communication connection via network link 721 to a compatible LAN.
Communication interface 720 could also be a cable modem or wireless
interface. In any such implementation, communication interface 720 sends
and receives electrical, electromagnetic or optical signals which carry
digital
data streams representing various types of information.
Network link 721 typically provides data communication through one
or more networks to other data devices. For example, network link 721 may
provide a connection through local network 722 to local server computer 723
or to data equipment operated by an Internet Service Provider (ISP) 724. ISP
724 in turn provides data communication services through the world wide
packet data communication network now commonly referred to as the
"Internet" 725. Local network 722 and Internet 725 both use electrical,
electromagnetic or optical signals which carry digital data streams. The
signals through the various networks and the signals on network link 721
and through communication interface 720, which carry the digital data to and
CA 02329426 2000-10-18
WO 99/55013 28 PCT/US99/08b73
from computer 700, are exemplary forms of carrier waves transporting the
information.
Computer 700 can send messages and receive data, including program
code, through the networks}, network link 721, and communication
interface 720. In the Internet example, remote server computer 726 might
transmit a requested code for an application program through Internet 725,
ISP 724, local network 722 and communication interface 720.
The received code may be executed by processor 713 as it is received,
and/or stored in mass storage 712, or other non-volatile storage for later
execution. In this manner, computer 700 may obtain application code in the
form of a carrier wave.
Application code may be embodied in any form of computer program
product. A computer program product comprises a medium configured to
store or transport computer readable code or data, or in which computer
readable code or data may be embedded. Some examples of computer
program products are CD-ROM disks, ROM cards, floppy disks, magnetic
tapes, computer hard drives, servers on a network, and carrier waves.
hluman Interface Device Computer ~rstem
The invention has application to computer systems where the data is
provided through a network. The network can be a local area network, a
wide area network, the Internet, world wide web, or any other suitable
network configuration. One embodiment of the invention is used in
CA 02329426 2000-10-18
WO 99/55013 29 PCT/US99/08673
computer system configuration referred to herein as a human interface
device computer system.
In this system the functionality of the system is partitioned between a
display and input device, and data sources or services. The display and input
device is a human interface device (HID). The partitioning of this system is
such that state and computation functions have been removed from the HID
and reside on data sources or services. In one embodiment of the invention,
one or more services communicate with one or more HIDs through some
interconnect fabric, such as a network. An example of such a system is
illustrated in Figure 8. Referring to Figure 8, the system consists of
computational service providers 800 communicating data through
interconnect fabric 801 to HIDs 802.
Computational Service Providers - In the HID system, the
computational power and state maintenance is found in the service
providers, or services. The services are not tied to a specific computer, but
may be distributed over one or more traditional desktop systems such as
described in connection with Figure 7, or with traditional servers. One
computer may have one or more services, or a service may be implemented
by one or more computers. The service provides computation, state, and data
to the HIDs and the service is under the control of a common authority or
manager. In Figure 8, the services are found on computers 810, 811, 812, 813,
and 814. In an embodiment of the invention, any of computers 810-814 could
be implemented as a transmitter.
Examples of services include X11 /Unix services, archived video
services, Windows NT service, JavaTM program execution service, and others.
CA 02329426 2000-10-18
WO 99/55013 3~ PCT/US99/08673
A service herein is a process that provides output data and responds to user
requests and input.
Interconnection Fabric - The interconnection fabric is any of multiple
suitable communication paths for carrying data between the services and the
HIDs. In one embodiment the interconnect fabric is a local area network
implemented as an Ethernet network. Any other local network may also be
utilized. The invention also contemplates the use of wide area networks, the
Internet, the world wide web, and others. The interconnect fabric may be
implemented with a physical medium such as a wire or fiber optic cable, or it
may be implemented in a wireless environment.
HIDs - The HID is the means by which users access the computational
services provided by the services. Figure 8 illustrates HIDs 821, 822, and
823.
A HID consists of a display 826, a keyboard 824, mouse 825, and audio speakers
827. The HID includes the electronics need to interface these devices to the
interconnection fabric and to transmit to and receive data from the services.
In an embodiment of the invention, an HID is implemented as a receiver.
A block diagram of the HID is illustrated in Figure 9. The components
of the HID are coupled internally to a PCI bus 912. A network control block
902 communicates to the interconnect fabric, such as an ethernet, through
line 914. An audio codec 903 receives audio data on interface 916 and is
coupled to block 902. USB data communication is provided on lines 913 to
USB controller 901.
An embedded processor 904 may be, for example, a Sparc2ep with
coupled flash memory 905 and DRAM 906. The USB controller 901, network
CA 02329426 2000-10-18
WO 99/55013 31 PCTNS99/08673
controller 902 and embedded processor 904 are all coupled to the PCI bus 912.
Also coupled to the PCI 912 is the video controller 909. The video controller
909 may be for example, and ATI RagePro+ frame buffer controller that
provides SVGA output on line 915. NTSC data is provided in and out of the
video controller through video decoder 910 and video encoder 911
respectively. A smartcard interface 908 may also be coupled to the video
controller 909.
The computer systems described above are for purposes of example
only. An embodiment of the invention may be implemented in any type of
computer system or programming or processing environment.
Thus, a method and apparatus of supporting a video protocol in a
network environment have been described in conjunction with one or more
specific embodiments. The invention is defined by the claims and their full
scope of equivalents.
