Improved direct mode (selection of spatial/temporal candidate without signaling, SAD competition at decoder and encoder)
MB-adaptive WP
Block-based pyramid prediction
16x16 DCT
Directional transforms
Enhanced BALF (two Wiener filters adaptively switched)
Modified de-blocking filter
C Software written from scratch, multiple OS
JCTVC-A108 (JVC)
Geometry transform prediction of MVs
Decoder-side block boundary decision
Refinement of MC by decoder-side estimation
AC ramp prediction using DC and boundary
QALF from KTA
JM16.2
JCTVC-A109 (MediaTek)
MB Size extension (up to 64x64)
Advanced adaptive IF
Decoder-side MV derivation with various adaptations
Enhanced spatio-temporal direct mode
Scaled motion vector prediction
Enhanced intra prediction: Overlapped block, 9 modes for 16x16, context-dependent mode
DCT up to 32x32
Multiple model KLT for Intra
Adaptive LF: QALF with multiple filters and more granular switching
Adaptive scan of transform coefficients for intra
KTA2.6
JCTVC-A110 (LG)
MB Size extension (up to 64x64)
Inter-Intra mixed mode, partial skip
Scaled MV predictor
Template-based illumination compensation
Modified motion vector competition
Intra prediction for 32x32
Chroma-from-luma intra prediction
New MDDT kernels
Adaptive de-blocking filter based on Wiener filter, QALF
JM 11
JCTVC-A111 (Huawei & Hisilicon)
Flexible macroblock partitions: H/V/D at arbitrary positions
Template-based motion derivation
Template-based inter-frame DC offset
Second order prediction for residual signal with directional characteristics
Resample-based intra prediction
Line-based intra prediction
RD optimized transform for intra prediction: Separable KLT
Directional transform (non-separable KLT) to the residual signal of MC prediction
Adaptive frequency-weighted quantization
KTA 2.6
JCTVC-A112 (RWTH Aachen)
Decoder-side motion vector derivation, P and B pictures, extended to larger MB sizes, candidate-based fast search
KTA2.6
JCTVC-A113 (SK telecom, Sejong Univ. & Sungkyunkwan Univ.)
32x32 MB and new partitions
Tree coding for partition type
Adaptive motion vector precision (down to 1/8 pel)
Intra prediction of larger blocks, special coding of modes
Up to 16x16 DCT
MDDT
QALF+HPF
Modified JM 15.2
JCTVC-A114 (France Telecom, NTT, NTT DOCOMO, Panasonic, & Technicolor)
Basic structure: superblocks 8x8
1/8 pel precision for motion
Basic motion unit 8x8, boundary can be diplaced 2 or 4 pixels
MV comp. in P
2 interpolation modes: fixed and Wiener
Zero-tree coding of (differential) motion information, displacement offset etc.
(...) Load balancing
v2v codes as entropy coding engine.
v2v codes in serial modes using primary and secondary Huffman codes
KTA 2.6r1
JCTVC-A121 (Qualcomm)
Extended MB
Geometric partitioning
Adaptive motion vector resolution (1/4 and 1/8)
Switched interpolation filters with offset: 6-tap, 8-tap and 4x4 non-separable
P direct mode
Scaled MV prediction
MDDT
16x16 transform (LLM factorization)
In-loop filter (modified, merge of QALF and post filter, up to 16 filters, diamond-shaped filters to reduce overhead and complexity)
KTA optimized (faster)
JCTVC-A122 (NHK & Mitsubishi)
MB Size extension (up to 64x64)
Hierarchical motion partitions
Adaptive direct vector derivation
Spatio-temporal adaptive PMV decision
Non-rectangular motion partition shapes
New 16x16 intra prediction modes: Global plane (surface) prediction, iterative adjustment prediction
16x16 DCT
Discrete Sine Transform (DST)
Enhanced BALF (two Wiener filters adaptively switched)
Modified de-blocking filter
C Software from scratch, multiple OS (same platform as A107)
JCTVC-A123 (NCTU)
Parametric OBMC
KTA 2.4r1
JCTVC-A124 Samsung & BBC
Flexible block structure (coding unit, prediction unit, transform unit), quad-tree based
Asymmetric partitions such as 1/4:3/4
Signaling of motion pred. choice
DCT based interpolation filter
High accuracy motion (1/12 bit refinement)
New intra pred methods: Arbitrary direction intra prediction, color comp. corr. pred. for intra, pixel-based template matching, combined intra prediction (ADI with local mean)
Integer DCT NxN
(up to 64x64 used)
Rotational DCT
“Logical transform” with a wavelet-like tree for larger blocks
IBDI
Loop filtering: de-blocking, extreme correction, band correction, content adaptive dynamic range, ALF re-using CU partitions
Syntax-based CABAC, coding engine similar to JPEG annex D
Adaptive coefficient scanning (hor/vert/diag)
C/C++ built from scratch (not object oriented)
JCTVC-A125 (BBC & Samsung; A124 without some or slightly modified tools)
No asymm. Partitions
No pixel-based template matching
IBDI
De-blocking
Coding Unit - Based ALF
Same as A124
JCTVC-A126 (Renesas)
Intra vector prediction by block matching
Intra repetitive pixel replenishment based on template matching
KTA
JCTVC-A127 (ETRI)
Extended MB size 32x32
Intra prediction mode 32x32, intra prediction 16x16 with directions
Recursive intra prediction, pre-filter at boundaries before prediction, post filter after prediction
Up to 32x32
MDDT
Simplified deblocking filter, only boundary pixels
Enhanced QALF
KTA2.3
Page: 10 Date Saved: 2010-04-23
Language:English
Score: 692444.1
-
https://www.itu.int/wftp3/av-a...10_04_A_Dresden/JCTVC-A203.doc
Data Source: un
Introduction to H.264 Intra prediction
H.264 FRExt uses several types of spatial predictions such as the 16
´
16, 8
´
8, 4
´
4 Green blocks, 16
´
16 Red block and 16
´
16 Blue block predictions. (...) The 16
´
16 block prediction makes use of four different prediction methods, viz. the vertical, horizontal, DC and planar modes. (...) Each prediction direction uses the previously-decoded block boundary pixels to predict the pixels in the 4
´
4 block.
Language:English
Score: 691810.64
-
https://www.itu.int/wftp3/av-a...te/2005_07_Poznan/JVT-P016.doc
Data Source: un
Mode 0 is ‘DC-prediction’ (see below). The other modes represent directions of predictions as indicated below.
3.5.1 Mode 0: DC prediction
Generally all pixels are predicted by (A+B+C+D+E+F+G+H)//8. (...) A block may therefore always be predicted in this mode.
3.5.2 Mode 1:
This mode is used only if all A,B,C,D are inside the picture.
a is predicted by: (A+B)/2
e is predicted by B
b,i are predicted by (B+C)/2
f,m are predicted by C
c,j are predicted by (C+D)/2
d,g,h,k,l,n,o,p are predicted by D
3.5.3 Mode 2: Vertical prediction
If A,B,C,D are inside the picture, a,e,i,m are predicted by A, b,f,j,n by B etc.
3.5.4 Mode 3: Diagonal prediction
This mode is used only if all A,B,C,D,E,F,G,H,I are inside the picture. This is a 'diagonal' prediction.
m is predicted by: (H+2G+F)//4
i,n are predicted by (G+2F+E)//4
e,j,o are predicted by (F+2E+I)//4
a,f,k,p are predicted by (E+2I+A)//4
b,g,l are predicted by (I+2A+B)//4
c,h are predicted by (A+2B+C)//4
d is predicted by (B+2C+D)//4
3.5.5 Mode 4:Horizontal prediction
If E,F,G,H are inside the picture, a,b,c,d are predicted by E, e,f,g,h by F etc.
3.5.6 Mode 5:
This mode is used only if all E,F,G,H are inside the picture.
a is predicted by: (E+F)/2
b is predicted by F
c,e are predicted by (F+G)/2
f,d are predicted by G
i,g are predicted by (G+H)/2
h,j,k,l,m,n,o,p are predicted by H
3.5.7 Prediction of chroma blocks
For chroma prediction there is only one mode.
Language:English
Score: 691549.3
-
https://www.itu.int/wftp3/av-a...o-site/0104_Aus/VCEG-M81d0.doc
Data Source: un
Mode 0 is ‘DC-prediction’ (see below). The other modes represent directions of predictions as indicated below.
3.5.1 Mode 0: DC prediction
Generally all pixels are predicted by (A+B+C+D+E+F+G+H)//8. (...) A block may therefore always be predicted in this mode.
3.5.2 Mode 1:
This mode is used only if all A,B,C,D are inside the picture.
a is predicted by: (A+B)/2
e is predicted by B
b,i are predicted by (B+C)/2
f,m are predicted by C
c,j are predicted by (C+D)/2
d,g,h,k,l,n,o,p are predicted by D
3.5.3 Mode 2: Vertical prediction
If A,B,C,D are inside the picture, a,e,i,m are predicted by A, b,f,j,n by B etc.
3.5.4 Mode 3: Diagonal prediction
This mode is used only if all A,B,C,D,E,F,G,H,I are inside the picture. This is a 'diagonal' prediction.
m is predicted by: (H+2G+F)//4
i,n are predicted by (G+2F+E)//4
e,j,o are predicted by (F+2E+I)//4
a,f,k,p are predicted by (E+2I+A)//4
b,g,l are predicted by (I+2A+B)//4
c,h are predicted by (A+2B+C)//4
d is predicted by (B+2C+D)//4
3.5.5 Mode 4:Horizontal prediction
If E,F,G,H are inside the picture, a,b,c,d are predicted by E, e,f,g,h by F etc.
3.5.6 Mode 5:
This mode is used only if all E,F,G,H are inside the picture.
a is predicted by: (E+F)/2
b is predicted by F
c,e are predicted by (F+G)/2
f,d are predicted by G
i,g are predicted by (G+H)/2
h,j,k,l,m,n,o,p are predicted by H
3.5.7 Prediction of chroma blocks
For chroma prediction there is only one mode.
Language:English
Score: 691549.3
-
https://www.itu.int/wftp3/av-a...site/h26L/older_tml/tml7d0.doc
Data Source: un
(Temporally) scaled MV prediction
2.1.4. Other prediction methods
2.2. Motion partitioning
2.2.1. (...) Combined spatial/temporal prediction
4. Intra prediction
4.1. Directional prediction
4.1.1. (...) Pre and post filtering
4.2. Plane prediction
4.3. Pyramid/interleaved prediction
4.4.
Language:English
Score: 691507.23
-
https://www.itu.int/wftp3/av-a...10_04_A_Dresden/JCTVC-A202.doc
Data Source: un
Mode 0 is ‘DC-prediction’ (see below). The other modes represent directions of predictions as indicated below.
2.4.1 Mode 0: DC prediction
Generally all pixels are predicted by (A+B+C+D+E+F+G+H)//8. (...) A block may therefore always be predicted in this mode.
2.4.2 Mode 1:
This mode is used only if all A,B,C,D are inside the picture.
a is predicted by: (A+B)/2
e is predicted by B
b,i are predicted by (B+C)/2
f,m are predicted by C
c,j are predicted by (C+D)/2
d,g,h,k,l,n,o,p are predicted by D
2.4.3 Mode 2: Vertical prediction
If A,B,C,D are inside the picture, a,e,i,m are predicted by A, b,f,j,n by B etc.
2.4.4 Mode 3: Diagonal prediction
This mode is used only if all A,B,C,D,E,F,G,H,I are inside the picture. This is a 'diagonal' prediction.
m is predicted by: (H+2G+F)//4
i,n are predicted by (G+2F+E)//4
e,j,o are predicted by (F+2E+I)//4
a,f,k,p are predicted by (E+2I+A)//4
b,g,l are predicted by (I+2A+B)//4
c,h are predicted by (A+2B+C)//4
d is predicted by (B+2C+D)//4
2.4.5 Mode 4:Horizontal prediction
If E,F,G,H are inside the picture, a,b,c,d are predicted by E, e,f,g,h by F etc.
2.4.6 Mode 5:
This mode is used only if all E,F,G,H are inside the picture.
a is predicted by: (E+F)/2
b is predicted by F
c,e are predicted by (F+G)/2
f,d are predicted by G
i,g are predicted by (G+H)/2
h,j,k,l,m,n,o,p are predicted by H
2.4.7 Prediction of chroma blocks
For chroma prediction there is only one mode.
Language:English
Score: 691421.88
-
https://www.itu.int/wftp3/av-a...o-site/0101_Eib/VCEG-L45d0.doc
Data Source: un
Partly because of these advances, the needs of the users of weather services have simultaneously diversified to encompass “environmental” prediction products such as air quality and hydrological predictions.
(...) Two closely related initiatives, the WWRP Polar Prediction Project and the WCRP Polar Climate Predictability Initiative (PCPI), aim to contribute to GIPPS.
(...) The Subseasonal to Seasonal Prediction project
Forecasting for the sub-seasonal time range has so far received much less attention than medium range and seasonal prediction as it has long been considered a “predictability desert.”
Language:English
Score: 690506.45
-
https://public.wmo.int/en/reso...rch-programme-10-year-vision-0
Data Source: un
Mode 0 is ‘DC-prediction’ (see below). The other modes represent directions of predictions as indicated below.
3.4.1 Mode 0: DC prediction
Generally all pixels are predicted by (A+B+C+D+E+F+G+H)//8. (...) A block may therefore always be predicted in this mode.
3.4.2 Mode 1:
This mode is used only if all A,B,C,D are inside the picture.
a is predicted by: (A+B)/2
e is predicted by B
b,i are predicted by (B+C)/2
f,m are predicted by C
c,j are predicted by (C+D)/2
d,g,h,k,l,n,o,p are predicted by D
3.4.3 Mode 2: Vertical prediction
If A,B,C,D are inside the picture, a,e,i,m are predicted by A, b,f,j,n by B etc.
3.4.4 Mode 3: Diagonal prediction
This mode is used only if all A,B,C,D,E,F,G,H,I are inside the picture. This is a 'diagonal' prediction.
m is predicted by: (H+2G+F)//4
i,n are predicted by (G+2F+E)//4
e,j,o are predicted by (F+2E+I)//4
a,f,k,p are predicted by (E+2I+A)//4
b,g,l are predicted by (I+2A+B)//4
c,h are predicted by (A+2B+C)//4
d is predicted by (B+2C+D)//4
3.4.5 Mode 4:Horizontal prediction
If E,F,G,H are inside the picture, a,b,c,d are predicted by E, e,f,g,h by F etc.
3.4.6 Mode 5:
This mode is used only if all E,F,G,H are inside the picture.
a is predicted by: (E+F)/2
b is predicted by F
c,e are predicted by (F+G)/2
f,d are predicted by G
i,g are predicted by (G+H)/2
h,j,k,l,m,n,o,p are predicted by H
3.4.7 Prediction of chroma blocks
For chroma prediction there is only one mode.
Language:English
Score: 690495.26
-
https://www.itu.int/wftp3/av-a...deo-site/0008_Por/q15k59d1.doc
Data Source: un
Mode 0 is ‘DC-prediction’ (see below). The other modes represent directions of predictions as indicated below.
3.4.1 Mode 0: DC prediction
Generally all pixels are predicted by (A+B+C+D+E+F+G+H)//8. (...) A block may therefore always be predicted in this mode.
3.4.2 Mode 1:
This mode is used only if all A,B,C,D are inside the picture.
a is predicted by: (A+B)/2
e is predicted by B
b,i are predicted by (B+C)/2
f,m are predicted by C
c,j are predicted by (C+D)/2
d,g,h,k,l,n,o,p are predicted by D
3.4.3 Mode 2: Vertical prediction
If A,B,C,D are inside the picture, a,e,i,m are predicted by A, b,f,j,n by B etc.
3.4.4 Mode 3: Diagonal prediction
This mode is used only if all A,B,C,D,E,F,G,H,I are inside the picture. This is a 'diagonal' prediction.
m is predicted by: (H+2G+F)//4
i,n are predicted by (G+2F+E)//4
e,j,o are predicted by (F+2E+I)//4
a,f,k,p are predicted by (E+2I+A)//4
b,g,l are predicted by (I+2A+B)//4
c,h are predicted by (A+2B+C)//4
d is predicted by (B+2C+D)//4
3.4.5 Mode 4:Horizontal prediction
If E,F,G,H are inside the picture, a,b,c,d are predicted by E, e,f,g,h by F etc.
3.4.6 Mode 5:
This mode is used only if all E,F,G,H are inside the picture.
a is predicted by: (E+F)/2
b is predicted by F
c,e are predicted by (F+G)/2
f,d are predicted by G
i,g are predicted by (G+H)/2
h,j,k,l,m,n,o,p are predicted by H
3.4.7 Prediction of chroma blocks
For chroma prediction there is only one mode.
Language:English
Score: 690495.26
-
https://www.itu.int/wftp3/av-a...site/h26L/older_tml/tml5d1.doc
Data Source: un
The MBTYPE field indicates whether there is direct-mode prediction, forward prediction, backward prediction, or bi-directional prediction. (...) Backward prediction means prediction from the same macroblock in the next lower layer (zero motion vector).
(...) Forward prediction denotes prediction from the corresponding macroblock in the next lower layer (zero motion vector).
Language:English
Score: 690336
-
https://www.itu.int/wftp3/av-a...deo-site/9702_Nic/lbc97018.doc
Data Source: un
