CN101616328B - Device and method thereof for converting coding coefficients of video signals - Google Patents
Device and method thereof for converting coding coefficients of video signals Download PDFInfo
- Publication number
- CN101616328B CN101616328B CN 200910160034 CN200910160034A CN101616328B CN 101616328 B CN101616328 B CN 101616328B CN 200910160034 CN200910160034 CN 200910160034 CN 200910160034 A CN200910160034 A CN 200910160034A CN 101616328 B CN101616328 B CN 101616328B
- Authority
- CN
- China
- Prior art keywords
- coefficient
- code coefficient
- code
- ict
- vision signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention discloses a device and a method for converting coding coefficients of video signals, which are used for converting a first coding coefficient into a second coding coefficient. The device comprises a first storage device, a second storage device, a multiplier, an adder and a shifter, wherein the first storage device is used for storing the first coding coefficient; the second storage device is used for storing a conversion matrix for converting the first coding coefficient into the second coding coefficient; the multiplier is used for executing multiplying operation for the first coding coefficient and corresponding parameters in the conversion matrix; and the adder and the shifter are used for carrying out addition and shift operations for the result value of the multiplying operation executed by adopting the multiplier so as to acquire the second coding coefficient. The device and the method can carry out conversion between various compression signals adopting different transformation cores so as to acquire compatibility among the various compression signals; therefore, more standards are compatible and the signals have versatility.
Description
The application be submitted on December 8th, 2005, title divides an application for No. 200510134530.6 patent application of " conversion equipment of the code coefficient of vision signal and method thereof ".
Technical field
The present invention relates to a kind of conversion equipment and method thereof, relate in particular to a kind of device and conversion method thereof that the code coefficient of digital video signal is changed, promptly by in vision signal is handled, the video compression coding coefficient being changed, thereby change mutually between the signal of realizing adopting the signal of discrete cosine transform (DCT) compressed encoding form and adopting integer cosine transformation (ICT) compressed encoding form, thereby the invention still further relates to by the video compression coding coefficient being changed device and its implementation of mutual conversion between the signal that adopts different IC T video compression coding.
Background technology
The compression of video image and encoding and decoding technique development some comparatively mature technique standards occurred so far.MPEG1 for example, H.264 H.263 H.261 MPEG2, H.262, reach and wait.And in these existing standards, video compression mostly is to adopt the method for the 8th rank DCT.And H.264/MPEG-4 the standard that adopts the 8th rank ICT to carry out video signal compression has the FRExt of (Part 10) AVC.
H.264/MPEG-4 (Part 10) AVC represents H.264/MPEG-4 (Part 10) Advanced Video Coding (hereinafter to be referred as H.264/AVC) of ITU-T.H.264/AVC be a member in the up-to-date series of international standards.This standard is by (the VideoCoding Experts Group of video coding expert group of ITU-T, VCEG) and (the MovingPicture Experts Group of the Motion Picture Experts Group of ISO/IEC, MPEG) (JointVideo Team JVT) researches and develops in the video joint working group that brainstrust is formed.(on average) usefulness of this standard is about the twice of the MPEG-2 that circulates now.In July, 2004, this working group proposes a new annex, i.e. Fidelity RangeExtensions (FRExt).This further promotes its usefulness.Also admitting in this annex utilizes the method for the 8th rank ICT that vision signal is compressed.This mainly is because ICT adopts integer arithmetic, and it is compared more simple in realization and saves resource with DCT.
Microsoft Windows Media 9 (WMV9) is a kind of new technology, is widely used in diverse network and device, and can be used for broadcasting.Its video core is a video codec that is called WindowsMedia 9.It is reasonably averaging out between computing and high-quality image quality.It is one of unified standard of the Society of Motion Picture and Television Engineers (The Society of Motion Picture andTelevision Engineers), adopts ICT to carry out video signal compression.
Along with adopting ICT to carry out the appearance of the standard of video compression such as H.264/AVC waiting, can the more standards of compatibility and have more versatility in order to make H.264/AVC, produced the needs of between the compressed signal of DCT and this dual mode of ICT, changing.
In addition, though H.264/AVC signal, AVS (Audio Video Coding Standard, the audio/video encoding/decoding technical standard) signal and WMV9 signal all adopt ICT to carry out video signal compression, but because various signals adopt different conversion cores, so it is also incompatible between the two, in order to make both can be compatible, and the more standards of compatibility, thereby have more versatility, therefore the needs of changing between the compressed signal of these three kinds of modes of AVS and WMV9 have also just been produced H.264/AVC.
Summary of the invention
The objective of the invention is to satisfy between the DCT of compressed signal and the ICT code coefficient and the needs of changing between the various ICT code coefficient.And then a kind of compression of video data code coefficient conversion equipment and method of being used for proposed.
The present invention also aims to provide a kind of apparatus and method of carrying out the quick conversion of video data compaction coding coefficient in simple mode.
According to an aspect of the present invention, provide a kind of conversion equipment of code coefficient of vision signal, be used for first code coefficient is converted to second code coefficient, described device can comprise: first storage device, store described first code coefficient; Second storage device, storage is used for described first code coefficient is converted to the transition matrix of described second code coefficient; Multiplier is used for the described relevant parameter of stating first code coefficient and described transition matrix is carried out multiply operation; Adder and shift unit carry out addition and shifting function to the result value that adopts described multiplier to carry out multiply operation, obtain described second code coefficient.
According to the first embodiment of the present invention, described first code coefficient is for adopting the video coding coefficient of DCT coded system coding, and described second code coefficient is for adopting the video coding coefficient of ICT coded system coding.Second embodiment according to the present invention, described first code coefficient is for adopting the video coding coefficient of ICT coded system coding, and described second code coefficient is also for adopting the video coding coefficient of ICT coded system coding.Third embodiment according to the present invention, described first code coefficient is for adopting the video coding coefficient of ICT coded system coding, and described second code coefficient is for adopting the video coding coefficient of DCT coded system coding.Wherein, the video coding coefficient of described employing ICT coded system coding is WMV9 video coding coefficient, AVS video coding coefficient or video coding coefficient H.264/AVC.
According to the present invention, described transition matrix is selected by a kind of like this mode, promptly, make second vision signal and the error between first vision signal that obtain by the respective opposed transcoding, coding transform that described second code coefficient execution that obtains is met ieee standard 1180-1990 error requirements meet ieee standard 1180-1990 error requirements, wherein, described first vision signal obtains by described first code coefficient is carried out the respective opposed transcoding, coding transform that meets ieee standard 1180-1990 error requirements.
Preferably, video coding coefficients by using the 8th rank ICT code coefficient of described employing ICT coded system coding.
As a kind of selection, described conversion equipment also comprises the 3rd storage device, is used to store the result of described adder and the operation of described shift unit.Described shift unit can be with 17,13,15,8,0,12,22 or 16 of the data shift rights in described the 3rd storage device.The figure place that described second code coefficient can have is 15.
According to a second aspect of the invention, provide the conversion method of video signal coding coefficient, be used for first code coefficient is converted to second code coefficient, wherein, described method can comprise: load described first code coefficient and transition matrix; Relevant parameter in described first code coefficient and the described transition matrix is carried out multiply operation; The data that obtain by described multiply operation are carried out add operation and shifting function, thereby obtain described second code coefficient.
According to one embodiment of the invention, described first code coefficient is for adopting the video coding coefficient of DCT coded system coding, and described second code coefficient is for adopting the video coding coefficient of ICT coded system coding.Second embodiment according to the present invention, described first code coefficient is for adopting the video coding coefficient of ICT coded system coding, and described second code coefficient is also for adopting the video coding coefficient of ICT coded system coding.Third embodiment according to the present invention, described first code coefficient is for adopting the video coding coefficient of ICT coded system coding, and described second code coefficient is for adopting the video coding coefficient of DCT coded system coding.Wherein, the video coding coefficient of described employing ICT coded system coding is WMV9 video coding coefficient, AVS video coding coefficient or video coding coefficient H.264/AVC.Described transition matrix is selected by a kind of like this mode, promptly, make second vision signal and the error between first vision signal that obtain by the respective opposed transcoding, coding transform that described second code coefficient execution that obtains is met ieee standard 1180-1990 error requirements meet ieee standard 1180-1990 error requirements, wherein, described first vision signal obtains by described first code coefficient is carried out the respective opposed transcoding, coding transform that meets ieee standard 1180-1990 error requirements.
Preferably, video coding coefficients by using the 8th rank ICT code coefficient of described employing ICT coded system coding.
Adopt the present invention, can between the compressed signal of the different conversion cores of various employings, change, have compatibility thereby make between the various compressed signals, thus more standards of compatibility and have more versatility.
Description of drawings
Figure 1 shows that code coefficient conversion equipment according to a preferred embodiment of the invention;
Fig. 2 has represented figure according to the code coefficient fast conversion method of an embodiment of the invention with the signal flow graph mode;
Fig. 3 has represented code coefficient fast conversion method according to another implementation of the invention with the signal flow graph mode;
Fig. 4 with the signal flow graph mode represented according to embodiment of the present invention be applied to AVS signal encoding coefficient is converted to WMV9 signal encoding coefficient fast conversion method;
Fig. 5 with the signal flow graph mode represented according to embodiment of the present invention be applied to will be H.264/AVC the signal encoding coefficient be converted to DCT signal encoding coefficient fast conversion method;
Fig. 6 is applied to the schematic diagram that vision signal is handled with the device of conversion of the present invention when DCT and ICT code coefficient are changed fast;
Fig. 7 is applied to the schematic diagram that vision signal is handled with the device of conversion of the present invention when H.264/AVC signal encoding coefficient and AVS signal encoding coefficient are changed fast;
Fig. 8 is converted to the H.264/AVC parameter list of the transition matrix of code coefficient according to embodiment of the present invention from the DCT code coefficient;
Fig. 9 is converted to the parameter list of the transition matrix of WMV9 code coefficient according to embodiment of the present invention from the DCT code coefficient;
Figure 10 is converted to the parameter list of the transition matrix of DCT code coefficient according to embodiment of the present invention from the AVS code coefficient;
Figure 11 be according to embodiment of the present invention from AVS to the parameter list of transition matrix H.264/AVC;
Figure 12 is converted to the parameter list of the transition matrix of AVS code coefficient according to embodiment of the present invention from code coefficient H.264/AVC;
Figure 13 is converted to the parameter list of the transition matrix of WMV9 code coefficient according to embodiment of the present invention from code coefficient H.264/AVC;
Figure 14 is converted to the parameter list of the transition matrix of DCT code coefficient according to embodiment of the present invention from the WMV9 code coefficient;
Figure 15 is converted to the parameter list of the transition matrix of AVS code coefficient according to embodiment of the present invention from the WMV9 code coefficient;
Figure 16 shows the value according to each transform coding coefficient of the preferred embodiment of the invention.
Embodiment
Below in conjunction with accompanying drawing the specific embodiment of the present invention is described in detail.This specification is narration DCT coding and three kinds of ICT Methods for Coding earlier, explain the conversion of various coding methods again.
The DCT coding method
The setting video signal indication is a vector X, and it is C through the resulting DCT code coefficient of dct transform
DctX to C
DctConversion can be by the dct transform matrix T
DctObtain:
C
dct=T
dctX (1)
X=T
dct -1C
dct=T
dct tC
dct (2)
The ICT coding method
The setting video signal indication is a vector X, utilizes ICT transformation matrix T
iIt can be become the ICT code coefficient, and be expressed as a vector C.ICT transformation matrix T
iCan adopt 3 kinds of different modes.First kind is T
i=KKE, for example AVS.And second kind be T
i=KE, for example H.264/AVC.The third is T
1=E,
Wherein, for the ICT on the 8th rank (a, b, c, d, e, f, g), matrix K and E are expressed as follows respectively:
k
B3(i) numerical value can obtain by following formula, and wherein, the i of E (i) representing matrix E is capable.
||k
b3(i)E(i)||
2=1
According to the definition of ICT, seven numerical value of a-g can have unlimited combination.But consider compressed capability and realize usefulness that H.264/AVC adopt ICT (12,10,6,3,8,4,8), and AVS mode compressed encoding adopts ICT (10,9,6,2,10,4,8), WMV9 mode compressed encoding adopts ICT (16,15,9,4,16,6,12).Each method will be elaborated for example with ICT (16,15,9,4,16,6,12) below.
The setting video signal indication is a vector X, then at T
i(describe) the ICT transformation matrix T that the AVS standard adopts during=KKE with the ICT coding method that AVS was adopted
iBe T
Avs=K
AvsK
AvsE
Avs, the resulting ICT code coefficient of conversion is C
AvsSo:
C
avs=T
avsX=K
avsK
avsE
avsX (3)
X=T
avs -1C
avs=E
avs tC
avs (4)
At T
iDuring=KE (describing) with the ICT coding method of H.264/AVC being adopted, the ICT transformation matrix T that adopts of standard H.264/AVC
iBe T
H264=K
H264E
H264, the resulting ICT code coefficient of conversion is C
H264So:
C
h264=T
h264X=K
h264E
h264X (5)
X=T
h264 -1C
h264=E
h264 tK
h264C
h264(6)
At T
iDuring=E (describing) with the ICT coding method that WMV9 was adopted, the ICT transformation matrix T that WMV9 adopts
iBe T
Wmv9=E
Wmv9, the resulting ICT code coefficient of conversion is C
Wmv9So:
C
wmv9=T
wmv9X=E
wmv9X (7)
X=T
wmv9 -1C
wmv9=E
wmv9 tK
wmv9K
wmv9C
wmv9?(8)
As previously mentioned, in order to realize DCT, H.264/AVC, H.264/AVC the interoperability between AVS and the WMV9 standard need be implemented in, the mutual conversion between the vision signal of AVS and WMV9 signal and other DCT coding.Comprising 12 kinds of different switching, this specification is divided into four classes with it, and every class is illustrated for three kinds.
If input coding coefficient C
InFor adopting coded system wherein a kind of in above four kinds of standards (being called first coding) to input signal vector X at this
InThe signal encoding coefficient that obtains of encoding, it is through transition matrix T
OutinBe transformed to and adopt the another kind of coded system signal encoding coefficient C of (being called second coding) at this
Out, that is, and this code coefficient C
OutBe output signal vector X
OutCode coefficient.Transition matrix T
OutinImplementation method have a variety ofly, consider transition matrix T
OutinThe usefulness that realizes and the signal coefficient C of conversion
OutAccuracy, in the described embodiment of this specification, select T with following two criterions
Outin:
1) transition matrix T
OutinStructure the simplest.
2) input coding coefficient C
InBy a phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements, can be transformed to signal phasor
Output signal coefficient C
OutBy a phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements, can be transformed to signal phasor X
OutWherein, signal phasor
And X
OutBetween error, must meet the error requirements of ieee standard 1180-1990.
1. be converted to other code signal by the DCT code signal
1.1. when the DCT code coefficient is converted to the AVS code coefficient, can get by equation (2) and (3):
C
avs=T
avsdctC
dct=K
avsK
avsE
avsT
dct tC
dct (9)
Be the transition matrix T of DCT code coefficient to the AVS code coefficient
AvsdctFor:
T
avsdct=K
avsK
avsE
avsT
dct t (10)
And can be approximately
Wherein, (T
a, T
b, T
c, T
d, T
e, T
f, T
g) and b all be integer.B is the accuracy of transformation matrix, and in a kind of conversion, when the numerical value of b is big more, the code coefficient of being exported is accurate more, but matrix is also complicated more.When ICT uses in different coded systems, (T
a, T
b, T
c, T
d, T
e, T
f, T
g) and b also different.According to described two selection criterions before, (T
a, T
b, T
c, T
d, T
e, T
f, T
g) and the selection of b must allow C
AvsThe reverse conversion signal meet the error requirements of ieee standard 1180-1990 and T
AvsdctStructure for the simplest, promptly b is a minimum value.
According to selection criterion, when input DCT code coefficient is necessary for 12 or more position and b and is 17, the C of conversion
AvsThe code coefficient value is with 15 bit representations, then (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(5793,6232,6048 ,-130 ,-74,7 ,-130), as shown in figure 16.Its fast conversion method can be with reference to figure 2.
1.2. when H.264/AVC the DCT code coefficient is converted to code coefficient, can get by equation (2) and (5):
C
h264=T
h264dctC
dct=K
h264E
h264T
dct tC
dct (12)
Be DCT transition matrix extremely H.264/AVC, T
H264dctFor:
T
h264dct=K
h264E
h264T
dct t (13)
And T
H264dctClose with unit matrix, can be approximately
I
8It is the unit matrix on eight rank.According to selection criterion, when input DCT code coefficient is necessary for 12 or more position and b and is 13, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 8 wherein one group, and the C of conversion
H264The value of code coefficient with 15 positions or more than represent so that the H.264/AVC signal of conversion meets the error requirements of ieee standard 1180-1990.According to the preferred embodiment of the invention, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(1 ,-13 ,-21,178,580,398,162), as shown in figure 16.Its fast conversion method can be with reference to figure 3.
1.3. when the DCT code coefficient is converted to the WMV9 code coefficient, can get by equation (2) and (7):
C
wmv9=T
wmv9dctC
dct=E
wmv9T
dct tC
dct (15)
Be the transition matrix of DCT to WMV9, T
Wmv9dctFor:
T
wmv9dct=E
wmv9T
dct t (16)
And can be approximately formula (17)
According to selection criterion, when input DCT code coefficient is necessary for 12 or more position and b and is 15, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 9 wherein one group, and the C of conversion
WmvThe value of code coefficient with 15 positions or more than represent so that the WMV9 signal of conversion meets the error requirements of ieee standard 1180-1990.According to the preferred embodiment of the invention, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(8690,8690,8744 ,-190 ,-297,423 ,-172), as shown in figure 16.Its fast conversion method can be with reference to figure 2.
2. be converted to other code coefficient by the AVS code coefficient
2.1. when the AVS code coefficient is converted to the DCT code coefficient, can get by equation (4) and (1):
C
dct=T
dctavsC
avs=T
dctE
avs tC
avs (18)
Be the transition matrix T of AVS to DCT
DctavsFor:
T
dctavs=T
dctE
avs t (19)
And can be approximately formula (20)
According to selection criterion, input AVS code coefficient is necessary for 12 or more position, and b is when being 8, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 10 wherein one group, and the C of conversion
DctThe value of code coefficient with 15 positions or more than represent so that the DCT signal of conversion meets the error requirements of ieee standard 1180-1990.According to the preferred embodiment of the invention, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(5793,5380,5514,112,67 ,-6,112), as shown in figure 16.
2.2. when H.264/AVC the AVS code coefficient is converted to code coefficient, can get by equation (4) and (5):
C
h264=T
h264avsC
avs=K
h264E
h264E
avs tC
avs (21)
Be AVS transition matrix T extremely H.264/AVC
H264avsFor:
T
h264avs=K
h264E
h264E
avs t (22)
And can be approximately formula (22)
According to selection criterion, when input AVS code coefficient is necessary for 12 or more position and b and is 8, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 11 wherein one group, and the C of conversion
H264The value of code coefficient with 15 positions or more than represent so that the DCT signal of conversion meets the error requirements of ieee standard 1180-1990.According to the preferred embodiment of the invention, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(5793,5367,5495,234,458,256,213), as shown in figure 16.Its fast conversion method can be with reference to figure 2.
2.3. when the AVS code coefficient is converted to the WMV9 code coefficient, can get by equation (4) and (7):
C
wmv9=T
wmv9avsC
avs=E
wmv9E
avs tC
avs (23)
Be the transition matrix T of AVS to WMV9
Wmv9avsFor:
T
wmvavs=E
wmv9E
avs t (24)
And can be approximately formula (25)
According to selection criterion, input AVS code coefficient is necessary for 12 or more position and b is 0 o'clock, the C of conversion
WmvThe code coefficient value is with 15 bit representations, then according to the preferred embodiment of the invention, and (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(768,714,736,0 ,-16,34,0), as shown in figure 16.Its fast conversion method can be with reference to figure 4.
3. be converted to other code coefficient by code coefficient H.264/AVC
3.1. when H.264/AVC code coefficient is converted to the DCT code coefficient, can get by equation (6) and (1):
C
dct=T
dcth264C
h264=T
dctE
h264 tK
h264C
h264 (26)
Promptly H.264/AVC to the transition matrix T of DCT
Dcth264For:
T
dcth264=T
dctK
h264E
h264 t (27)
And can be approximately following formula (28).
According to selection criterion, import H.264/AVC that code coefficient is necessary for 12 or more position and b is, the C of conversion at 12 o'clock
DctThe code coefficient value is with 15 bit representations, then according to the preferred embodiment of the invention, and (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(0 ,-7 ,-10 ,-81 ,-290 ,-199 ,-89), as shown in figure 16, its fast conversion method can be with reference to figure 5.
3.2. when H.264/AVC code coefficient is converted to the AVS code coefficient, can get by equation (6) and (3):
C
avs=T
avsh264C
h264=K
avsK
avsE
avsEh264 tK
h264C
h264 (29)
Promptly H.264/AVC code coefficient to the transition matrix T of AVS code coefficient
Avsh264For:
T
avsh264=K
avsK
avsE
avsE
h264 tK
h264 (30)
And can be approximately:
According to selection criterion, when importing H.264/AVC code coefficient and being necessary for 12 or more position and b and being 17, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 12 wherein one group, and the C of conversion
AvsThe value of code coefficient with 15 positions or more than represent so that the AVS signal of conversion meets the error requirements of ieee standard 1180-1990.According to the preferred embodiment of the invention, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(5793,6217,6064 ,-247 ,-505 ,-296 ,-271), as shown in figure 16.But its fast conversion method reference diagram.
3.3. when H.264/AVC code coefficient is converted to the WMV9 code coefficient, can get by equation (6) and (7):
C
wmv9=T
wmv9h264C
h264=E
wmv9E
h264 tK
h264C
h264 (32)
Promptly H.264/AVC code coefficient to the transition matrix T of WMV9 code coefficient
Wmv9h264For:
T
wmv9h264=E
wmv9E
h264 tK
h264 (33)
And can be approximately formula (34)
According to selection criterion, when importing H.264/AVC code coefficient and being necessary for 12 or more position and b and being 8, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 13 wherein one group, and the C of conversion
Wmv9The value of code coefficient with 15 positions or more than represent so that the WMV9 signal of conversion meets the error requirements of ieee standard 1180-1990.According to the preferred embodiment of the invention, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(8689,8689,8701 ,-362 ,-916,0 ,-362), as shown in figure 16.Its fast conversion method can be with reference to figure 2.
4. be converted to other code coefficient by the WMV9 code coefficient
4.1. when the WMV9 code coefficient is converted to the DCT code coefficient, can get by equation (8) and (1):
C
dct=T
dctwmv9C
wmv9=T
dctE
wmv tK
wmvK
wmvC
wmv9 (35)
Be the transition matrix of WMV9 to DCT, T
DctwmvFor:
T
dctwmv=T
dctEwmv tK
wmvK
wmv (36)
And can be approximately following formula (37).
According to selection criterion, when input WMV9 code coefficient is necessary for 12 or more position and b and is 17, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 14 wherein one group, and the C of conversion
DctThe value of code coefficient with 15 positions or more than represent so that the DCT signal of conversion meets the error requirements of ieee standard 1180-1990.Preferably, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(3862,3849,3833,76,130 ,-187,84), as shown in figure 16.Its fast conversion method can be with reference to figure 2.
4.2. when the WMV9 code coefficient is converted to the AVS code coefficient, can get by equation (8) and (3):
C
avs=T
avswmv9C
wmv9=K
avsK
avsE
avsE
wmv tK
wmvK
wmvC
wmv9 (38)
Be the transition matrix of WMV9 code coefficient to the AVS code coefficient, T
AvswmvFor:
T
avswmv=K
avsK
avsE
avsE
wmv tK
wmvK
wmv (39)
And can be approximately formula (40)
According to selection criterion, be necessary for 12 positions or more position at the WMV9 code coefficient of input, and b is when being 22, T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue be necessary in the table shown in Figure 15 wherein one group, and the C of conversion
AvsThe value of code coefficient with 15 positions or more than represent so that the AVS signal of conversion meets the error requirements of ieee standard 1180-1990.Preferably, (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(5461,5861,5696,0,124 ,-279,0), as shown in figure 16, its fast conversion method can be with reference to figure 2.
4.3. when H.264/AVC the WMV9 code coefficient is converted to code coefficient, can get by equation (8) and (5):
C
h264=T
h264wmv9C
wmv9=K
h264E
h264E
wmv tK
wmvK
wmvC
wmv9 (41)
Be WMV9 transition matrix extremely H.264/AVC, T
H264wmvFor:
T
h264wmv=K
h264E
h264E
wmv tK
wmvK
wmv (42)
And can be approximately formula (43)
According to selection criterion, input WMV9 code coefficient is necessary for 12 or more position and b is 16 o'clock, the C of conversion
H264The code coefficient value is with 15 bit representations, then preferably, and (T
a, T
b, T
c, T
d, T
e, T
f, T
g)=(1931,1924,1907,80,201,0,80).Its fast conversion method can be with reference to figure 2.
Above from the angle of mathematics described of the present invention with vision signal between DCT code coefficient and ICT code coefficient, and the principle of between various ICT compression coding modes, changing.Based on this, provide the device that is used to carry out the code coefficient conversion of vision signal of the present invention.Specifically describe embodiments of the present invention below with reference to accompanying drawings.
Figure 1 shows that code coefficient conversion equipment according to a preferred embodiment of the invention.As shown in Figure 1, the code coefficient conversion equipment of embodiment of the present invention mainly comprises: first storage device 10, second storage device 12, the 3rd storage device 13, multiplier 30, adder 20 and shift unit 40.
Describe first storage device 10, second storage device 12, the 3rd storage device 13, multiplier 30, adder 20 and shift unit 40 below respectively, and to be converted to the AVS code coefficient with the DCT code coefficient be that example illustrates according to preferred implementation of the present invention.
Second storage device, 12 storage transition matrixes are stored the transition matrix T that is converted to the AVS code coefficient from the DCT code coefficient in this embodiment
AvsdctIn this embodiment, make the parameter (T of transition matrix
a, T
b, T
c, T
d, T
e, T
f, T
g)=(5793,6232,6048 ,-130 ,-74,7 ,-130).For different conversions and precision setting, the concrete parameter value difference in the transition matrix, for example, for the transition matrix that the AVS code coefficient is converted to the DCT code coefficient, its parameter value T
a, T
b, T
c, T
d, T
e, T
fAnd T
gValue can be wherein a group in the table shown in Figure 10.
In one embodiment of the invention, the code coefficient of storage is to load by manual mode in transition matrix described here and above-mentioned first storage device 10.That is, before conversion, at first specify coded system that input signal adopts and output signal, and load different code coefficients and transition matrix by preestablishing of program according to specified coded system with the coded system that adopts.Those of ordinary skill in the art should be realized that, can use multiple mode to store, read code coefficient and transition matrix, for example, they can be stored in the known memory device, read from these memory devices by the program appointment, perhaps the parameter of all code coefficients and transition matrix can be programmed in the program code, and by in program coding, preestablishing the different coding coefficient in the working procedure coding and the parameter of transition matrix.
The 3rd storage device 13 is used to store adder 20 and the intermediate object program of shift unit 40 operations and the final result of conversion.Specifically, the 3rd storage device 13 is cleared before conversion, and the effect of adder 20 is with the storing value addition in the first and the 3rd storage device 13, and the result of addition is temporarily stored in the 3rd storage device 13.
The effect of shift unit 40 is that the numerical value in the 3rd storage device 13 is carried out shifting function, for example, this with embodiment in, finish add operation at adder 20, shift unit 40 is with result's predetermined figure place that moves to right of storage in the 3rd storage device 13, and the result of displacement still is stored in the 3rd storage device 13.The result of storage is exactly the code coefficient after the conversion in the 3rd storage device 13, is AVS code coefficient C in this embodiment
Avs
More than though to be converted to the AVS code coefficient with the DCT code coefficient be to understand the example according to each functions of components of code coefficient conversion equipment of the present invention and realize principle, but, those skilled in the art should be realized that, for the conversion between the different coding coefficient, more than the calculating operation order of each parts may be different, for example, in one an of the present invention embodiment who the DCT code coefficient is converted to AVC code coefficient H.264/, carry out shifting function at shift unit 40, adder 20 also will be carried out add operation.For example at the code coefficient C that calculates H.264/AVC
H2642The time, multiplier 30 is carried out code coefficient T earlier
cWith C
Dct2, T
eWith C
Dct6Multiply operation, 20 pairs of these multiplied result of adder are carried out addition, then, the result that 40 pairs of adders of shift unit 20 carry out the phase add operation is shifted, and is last, adder 20 is with the result and the code coefficient C of shift unit 40 displacements
Dct2Addition obtains code coefficient C
H2642Its mode with signal flow can be expressed as:
C
h2642=(T
c*C
dct2-T
e*C
dct6)>>b+C
dct2
Concrete steps for realizing by the code coefficient conversion equipment of this preferred implementation changing between each code coefficient will be described in detail below this paper.
How to change between each code coefficient in a preferred embodiment of this invention in order more clearly describing, will to describe the signal flow graph of various code coefficient conversions by the mode that exemplifies below.
Fig. 2 shows figure according to the fast conversion method of the code coefficient of one embodiment of the present invention with the mode of signal flow graph.Fig. 3 shows the DCT of another execution mode according to the present invention and the fast conversion method between the code coefficient H.264/AVC with the mode of signal flow graph.Because of the parameter of partly changing core is arranged is zero, so fast conversion method can more be simplified, for example the AVS of Fig. 4 goes to the conversion method that H.264/AVC goes to DCT of WMV9 and Fig. 5.
With reference to Fig. 2, the conversion between each DCT code coefficient and the AVS code coefficient can be described as down:
C
avs0=C
dct0*T
a>>b
C
avs4=C
dct4*T
a>>b
C
avs2=(T
c*C
dct2-T
e*C
dct6)>>b
C
avs6=(T
c*C
dct6+T
e*C
dct2)>>b
C
avs1=(T
b*C
dct1+T
g*C
dct3+T
d*C
dct5-T
f*C
dct7)>>b
C
avs3=(-T
d*C
dct1+T
b*C
dct3+T
f*C
dct5-T
g*C
dct7)>>b
C
avs5=(-T
g*C
dct1-T
f*C
dct3+T
b*C
dct5+T
d*C
dct7)>>b
C
avs7=(T
f*C
dct1+T
d*C
dct3-T
g*C
dct5+T
b*C
dct7)>>b
With reference to Fig. 3, each DCT code coefficient and H.264/AVC the conversion between the code coefficient can be described as down:
C
h2640=C
dct0*T
a>>b+C
dct0
C
h2644=C
dct4*T
a>>b+C
dct4
C
h2642=(T
c*C
dct2-T
e*C
dct6)>>b+C
dct2
C
h2646=(T
c*C
dct6+T
e*C
dct2)>>b+C
dct6
C
h2641=(T
b*C
dct1+T
g*C
dct3+T
d*C
dct5-T
f*C
dct7)>>b+C
dct1
C
h2643=(-T
d*C
dct1+T
b*C
dct3+T
f*C
dct5-T
g*C
dct7)>>b+C
dct3
C
h2645=(-T
g*C
dct1-T
f*C
dct3+T
b*C
dct5+T
d*C
dct7)>>b+C
dct5
C
h2647=(T
f*C
dct1+T
d*C
dct3-T
g*C
dct5+T
b*C
dct7)>>b+C
dct7
Wherein, symbol ">>b " the expression b position that moves to right.
With reference to Fig. 1 and Fig. 2 forward conversion method step between DCT code coefficient and the AVS code coefficient is described respectively below, and to calculate C
Avs7Be example.
1) with 13 zero clearing of the 3rd storage device.
2) load the DCT code coefficient of the signal that the DCT modes compress at first storage device 10, as C
Dct
3) load the corresponding numerical value of transition matrixes at second storage device 12, as T
Avsdct
4) utilize multiplier 30 that the numerical value in the transition matrix of storing in described each DCT code coefficient and second storage device 12 of storage in first storage device 10 is multiplied each other.For example, carry out T
f* C
Dct1Deng, and multiplied result is stored to first storage device 10.
5) utilize adder 20 with the storing value addition in the first and the 3rd storage device, for example, calculate T
d* C
Dct3,-T
g* C
Dct5And T
b* C
Dct7Addition result, and the result is stored in the 3rd storage device 13.
6) utilize shift unit with the result of (5) move to right predetermined figure place, i.e. b.The result deposits to the 3rd storage device 13 and is C
Avs7
Wherein, in step 2), 3) in, can only load C at every turn
Dct, T
AvsdctIn a parameter, also can be with C
Dct, T
AvsdctIn disposable all being loaded in first storage device 10 and second storage device 12 of all parameters.Should be appreciated that, only load C at every turn in employing
Dct, T
AvsdctIn the method for a parameter time, above-mentioned steps 2) to 5) will be the step of repetitive cycling, for example at first load T
fAnd C
Dct1, so in step 4), they are multiplied each other, and multiplied result is stored in first storage device 10, in step 5), the storing value addition in this result and the 3rd storage device 13 turns back to step 2) the middle operation of carrying out all the other parameters.
Reverse conversion between DCT code coefficient and the AVS code coefficient (promptly, be converted to the DCT code coefficient by the AVS code coefficient) method similar to the method for above-mentioned forward conversion, by top forward conversion description and in conjunction with Fig. 3, should be conspicuous for the person of ordinary skill of the art to the conversion the DCT code coefficient by the AVS code coefficient.
Fig. 6 has shown the schematic diagram that the code coefficient conversion equipment of changing fast of the present invention is applied to the vision signal processing between DCT and ICT code coefficient.As shown in Figure 6, vision signal is carried out the ICT mode in the past handle the flow process handled with the DCT mode respectively shown in heavy line and fine line among Fig. 6.Both code coefficients are non-interchangeable.By adding above-mentioned conversion equipment of the present invention, both code coefficients can be exchanged.Specifically, vision signal for input, when adopting DCT mode compressed encoding to become digital of digital video data (forward dct transform), extract the DCT code coefficient (shown in the fine dotted line among the figure) in this video data, by code coefficient conversion equipment of the present invention, the ICT code coefficient (shown in thick dashed line) that is converted in the decoding of carrying out the ICT mode (oppositely ICT conversion), is reduced into vision signal output with digital of digital video data.Wherein, extract the step of DCT code coefficient in the video data, do not belong to scope of the present invention, therefore, here repeat no more.
Fig. 7 shown with of the present invention H.264/AVC and the code coefficient conversion equipment of changing fast between the AVS code coefficient be applied to the schematic diagram that vision signal is handled.As shown in Figure 7, vision signal was carried out that the AVS mode is handled in the past and H.264/AVC the flow process handled of mode respectively shown in heavy line and fine line among the figure.Both code coefficients are non-interchangeable.By adding above-mentioned conversion equipment of the present invention, both code coefficients can be exchanged.Specifically, vision signal for input, the mode compressed encoding becomes digital of digital video data (forward is conversion H.264/AVC) when adopting H.264/AVC, extract the H.264/AVC code coefficient (shown in the fine dotted line among the figure) in this video data, by code coefficient conversion equipment of the present invention, be converted into AVS code coefficient (shown in thick dashed line),, digital of digital video data be reduced into vision signal output in the decoding of carrying out the AVS mode (oppositely AVS conversion).On the contrary, when the video signal compression coding that adopts the AVS mode with input becomes digital of digital video data (forward is conversion H.264/AVC), extract the AVS code coefficient (shown in thick dashed line among the figure) in this video data, by conversion equipment of the present invention, H.264/AVC code coefficient shown in transferring to it on fine dotted line, carrying out the H.264/AVC decoding of mode (oppositely H.264/AVC conversion), digital of digital video data is being reduced into vision signal output.All the other as H.264 and the also available similarity method of the conversion between MSWM9 and AVS and the MSWM9 carry out.
These explanations are not to be limitation of the present invention, explanation according to these embodiments, those skilled in the art can understand the spirit and scope of the invention defined by the claims better, and the alternative means that can predict various conspicuous modifications, conversion and be equal to.
Claims (25)
1. a vision signal that is used for having first code coefficient is converted to the device of the vision signal with second code coefficient, wherein said first code coefficient is an ICT video coding coefficient, described second code coefficient is the ICT video coding coefficient that is different from described first code coefficient, and described device comprises:
First memory cell is configured to store described first code coefficient;
Second memory cell, be configured to store transition matrix, described transition matrix is set to make the error between the first vision signal vector and the second vision signal vector to satisfy ieee standard 1180-1990 error requirements, the described first vision signal vector is to obtain by described first code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements, and the described second vision signal vector is to obtain by described second code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements;
Multiplication unit is configured to described first code coefficient and described transition matrix are carried out multiply operation;
Adder unit is configured to the result value of described multiply operation is carried out add operation; And
Shift unit is configured to the result value of described adder unit is carried out shifting function, to obtain described second code coefficient.
2. device as claimed in claim 1, wherein, described ICT video coding coefficient constitutes eight rank ICT encoder matrixs.
One of H.264/AVC 3. device as claimed in claim 1, wherein, described ICT video coding coefficient meets international standard WMV9, AVS and.
4. device as claimed in claim 2 further comprises:
The 3rd memory cell, the output that is configured to store described adder unit and described shift unit.
5. device as claimed in claim 4, wherein said shift unit moves to right 17,13,15,8,0,12,22 or 16 with the described result value of described adder unit.
6. device as claimed in claim 5, wherein said second code coefficient is 15.
7. device as claimed in claim 1, wherein said shift unit further is configured to the gt with the described result value of described adder unit.
8. device as claimed in claim 1, wherein said second code coefficient are at least 15 coefficient.
9. a vision signal that is used for having first code coefficient is converted to the device of the vision signal with second code coefficient, wherein said first code coefficient is an ICT video coding coefficient, described second code coefficient is a DCT video coding coefficient, and described device comprises:
First memory cell is configured to store described first code coefficient;
Second memory cell, be configured to store transition matrix, described transition matrix is set to make the error between the first vision signal vector and the second vision signal vector to satisfy ieee standard 1180-1990 error requirements, the described first vision signal vector is to obtain by described first code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements, and the described second vision signal vector is to obtain by described second code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements;
Multiplication unit is configured to described first code coefficient and described transition matrix are carried out multiply operation;
Adder unit is configured to the result value of described multiply operation is carried out add operation; And
Shift unit is configured to the result value of described adder unit is carried out shifting function, to obtain described second code coefficient.
10. device as claimed in claim 9, wherein, described ICT video coding coefficient constitutes eight rank ICT encoder matrixs.
One of H.264/AVC 11. device as claimed in claim 9, wherein, described ICT video coding coefficient meets international standard WMV9, AVS and.
12. device as claimed in claim 10 further comprises:
The 3rd memory cell, the output that is configured to store described adder unit and described shift unit.
13. device as claimed in claim 12, wherein said shift unit moves to right 17,13,15,8,0,12,22 or 16 with the described result value of described adder unit.
14. device as claimed in claim 13, wherein said second code coefficient is 15.
15. a vision signal that is used for having first code coefficient is converted to the method for the vision signal with second code coefficient, wherein said first code coefficient is an ICT video coding coefficient, described second code coefficient is the ICT video coding coefficient that is different from described first code coefficient, and described method comprises:
Load described first code coefficient and transition matrix, described transition matrix is set to make the error between the first vision signal vector and the second vision signal vector to satisfy ieee standard 1180-1990 error requirements, the described first vision signal vector is to obtain by described first code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements, and the described second vision signal vector is to obtain by described second code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements;
Described first code coefficient and described transition matrix are carried out multiply operation;
Result value to described multiply operation is carried out add operation; And
Result value to described add operation is carried out shifting function, to obtain described second code coefficient.
16. method as claimed in claim 15, wherein, described ICT video coding coefficient constitutes eight rank ICT encoder matrixs.
One of H.264/AVC 17. method as claimed in claim 15, wherein, described ICT video coding coefficient meets international standard WMV9, AVS and.
18. method as claimed in claim 16, wherein said shifting function further comprises:
The described result value of described add operation is moved to right 17,13,15,8,0,12,22 or 16, to obtain described second code coefficient.
19. method as claimed in claim 18, wherein said second code coefficient is 15.
20. method as claimed in claim 15, wherein said second code coefficient are at least 15 coefficient.
21. a vision signal that is used for having first code coefficient is converted to the method for the vision signal with second code coefficient, wherein said first code coefficient is an ICT video coding coefficient, described second code coefficient is a DCT video coding coefficient, and described method comprises:
Load described first code coefficient and transition matrix, described transition matrix is set to make the error between the first vision signal vector and the second vision signal vector to satisfy ieee standard 1180-1990 error requirements, the described first vision signal vector is to obtain by described first code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements, and the described second vision signal vector is to obtain by described second code coefficient is carried out the phase-reversal coding conversion that meets ieee standard 1180-1990 error requirements;
Described first code coefficient and described transition matrix are carried out multiply operation;
Result value to described multiply operation is carried out add operation; And
Result value to described add operation is carried out shifting function, to obtain described second code coefficient.
22. method as claimed in claim 21, wherein, described ICT video coding coefficient constitutes eight rank ICT encoder matrixs.
One of H.264/AVC 23. method as claimed in claim 21, wherein, described ICT video coding coefficient meets international standard WMV9, AVS and.
24. method as claimed in claim 22, wherein said shifting function further comprises:
The described result value of described phase add operation is moved to right 17,13,15,8,0,12,22 or 16, to obtain described second code coefficient.
25. method as claimed in claim 24, wherein said second code coefficient is 15.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200910160034 CN101616328B (en) | 2005-12-08 | 2005-12-08 | Device and method thereof for converting coding coefficients of video signals |
HK10103796.8A HK1138971A1 (en) | 2005-12-08 | 2010-04-20 | An apparatus and a method for transforming coding coefficient of video signals |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200910160034 CN101616328B (en) | 2005-12-08 | 2005-12-08 | Device and method thereof for converting coding coefficients of video signals |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB2005101345306A Division CN100539704C (en) | 2005-12-08 | 2005-12-08 | The conversion equipment of the code coefficient of vision signal and method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101616328A CN101616328A (en) | 2009-12-30 |
CN101616328B true CN101616328B (en) | 2011-08-10 |
Family
ID=41495675
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200910160034 Active CN101616328B (en) | 2005-12-08 | 2005-12-08 | Device and method thereof for converting coding coefficients of video signals |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN101616328B (en) |
HK (1) | HK1138971A1 (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8719321B2 (en) * | 2011-12-11 | 2014-05-06 | Hong Kong Applied Science and Technology Research Institute Company Limited | Adaptive block-size transform using LLMICT |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1445999A (en) * | 2003-03-26 | 2003-10-01 | 浙江大学 | Method and device of integer transform applicable to image coding and video frequency coding |
-
2005
- 2005-12-08 CN CN 200910160034 patent/CN101616328B/en active Active
-
2010
- 2010-04-20 HK HK10103796.8A patent/HK1138971A1/en not_active IP Right Cessation
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1445999A (en) * | 2003-03-26 | 2003-10-01 | 浙江大学 | Method and device of integer transform applicable to image coding and video frequency coding |
Non-Patent Citations (1)
Title |
---|
JP特開2005-86351A 2005.03.31 |
Also Published As
Publication number | Publication date |
---|---|
HK1138971A1 (en) | 2010-09-03 |
CN101616328A (en) | 2009-12-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN100539704C (en) | The conversion equipment of the code coefficient of vision signal and method thereof | |
CN100390777C (en) | Color movement vector integerization | |
CN100417218C (en) | Method and apparatus for variable accuracy inter-picture timing specification for digital video encoding with reduced requirements for division operations | |
CN1697328B (en) | Fast video codec transform implementations | |
JP4425561B2 (en) | 2-D conversion for image and video coding | |
KR100728263B1 (en) | Moving picture decoding method and device | |
CN100593955C (en) | Method and device for coding a video picture in inter or intra mode | |
CN102176748B (en) | Resampling and picture resizing operations for multi-resolution video coding and decoding | |
CN100362510C (en) | Round-off control for multi-stage interpolation | |
CN101631245B (en) | Motion vector detecting device, motion vector detecting method, image encoding device, and program | |
CN103650490B (en) | For the method and apparatus of motion compensated prediction | |
JPH0799659A (en) | Nonexhaustive movement estimation method | |
JP2007221662A (en) | Motion vector calculation apparatus | |
CN102369730A (en) | Compressed dynamic image encoding device, compressed dynamic image decoding device, compressed dynamic image encoding method and compressed dynamic image decoding method | |
KR100818921B1 (en) | Motion vector compression method, video encoder and video decoder using the method | |
CN1703094B (en) | Image interpolation apparatus and methods that apply quarter pel interpolation to selected half pel interpolation results | |
JP4704333B2 (en) | Image encoding device, image decoding device, and integrated circuit used in the same | |
CN101188761A (en) | Method for optimizing DCT quick algorithm based on parallel processing in AVS | |
CN101309404A (en) | Resolution descending video transcoding method and transcoding device thereof | |
CN102801982B (en) | Estimation method applied on video compression and based on quick movement of block integration | |
CN101616328B (en) | Device and method thereof for converting coding coefficients of video signals | |
JP2010081498A (en) | Image compression coding method and apparatus | |
US8588305B2 (en) | Two-dimensional interpolation architecture for motion compensation in multiple video standards | |
JP4847460B2 (en) | Image interpolation method and pixel interpolation device | |
WO2010007719A1 (en) | Image encoding apparatus, image encoding method, image decoding apparatus, and image decoding method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1138971 Country of ref document: HK |
|
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
REG | Reference to a national code |
Ref country code: HK Ref legal event code: GR Ref document number: 1138971 Country of ref document: HK |