ISO/IEC 23002-4:2014/Amd 1:2014

Information technology — MPEG video technologies — Part 4: Video tool library — Amendment 1: Graphics tool library (GTL) for the reconfigurable multimedia coding (RMC) framework

Technologies de l'information — Technologies vidéo MPEG — Partie 4: Bibliothèque d'outils vidéo — Amendement 1: Bibliothèque d'outils graphiques (GTL) pour le cadre de codage multimédia reconfigurable (RMC)

General Information

• Status: Withdrawn
• Publication Date: 16-Nov-2014
• Withdrawal Date: 16-Nov-2014
• ICS: 35.040 - Information coding
• ICS: 35.040.40 - Coding of audio, video, multimedia and hypermedia information
• Technical Committee: ISO/IEC JTC 1/SC 29 - Coding of audio, picture, multimedia and hypermedia information
• Drafting Committee: ISO/IEC JTC 1/SC 29 - Coding of audio, picture, multimedia and hypermedia information
• Current Stage: 9599 - Withdrawal of International Standard
• Completion Date: 06-Jun-2019

Relations

• Revised: ISO/IEC 23002-4:2018 - Information technology — MPEG video technologies — Part 4: Video tool library
• Effective Date: 23-Apr-2020
• Amends: ISO/IEC 23002-4:2014 - Information technology — MPEG video technologies — Part 4: Video tool library
• Effective Date: 08-Jan-2022

Standards Content (Sample)

INTERNATIONAL ISO/IEC
STANDARD 23002-4
Second edition
2014-04-15
AMENDMENT 1
2014-11-15

Information technology — MPEG video
technologies —
Part 4:
Video tool library
AMENDMENT 1: Graphics tool library (GTL)
for the reconfigurable multimedia coding
(RMC) framework
Technologies de l'information — Technologies vidéo MPEG —
Partie 4: Bibliothèque d'outils vidéo
AMENDEMENT 1: Bibliothèque d'outils graphiques (GTL) pour le cadre
de codage multimédia reconfigurable (RMC)




Reference number
ISO/IEC 23002-4:2014/Amd.1:2014(E)
©
ISO/IEC 2014

---------------------- Page: 1 ----------------------
ISO/IEC 23002-4:2014/Amd.1:2014(E)



This CD-ROM contains:
1) the publication ISO/IEC 23002-4:2014/Amd.1:2014 in portable document format (PDF), which can be
viewed using Adobe® Acrobat® Reader;
2) electronic inserts containing network specifications for two graphics decoders.
Adobe and Acrobat are trademarks of Adobe Systems Incorporated.
COPYRIGHT PROTECTED DOCUMENT


©  ISO/IEC 2014
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form or by any
means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior written permission.
Permission can be requested from either ISO at the address below or ISO’s member body in the country of the requester.
ISO copyright office
Case postale 56  CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail [e-mail]
Web www.iso.org
Published in Switzerland

ii © ISO/IEC 2014 – All rights reserved

---------------------- Page: 2 ----------------------
ISO/IEC 23002-4:2014/Amd.1:2014(E)
Installation
If this publ
...

INTERNATIONAL ISO/IEC
STANDARD 23002-4
Second edition
2014-04-15
AMENDMENT 1
2014-11-15

Information technology — MPEG video
technologies —
Part 4:
Video tool library
AMENDMENT 1: Graphics tool library (GTL)
for the reconfigurable multimedia coding
(RMC) framework
Technologies de l'information — Technologies vidéo MPEG —
Partie 4: Bibliothèque d'outils vidéo
AMENDEMENT 1: Bibliothèque d'outils graphiques (GTL) pour le cadre
de codage multimédia reconfigurable (RMC)




Reference number
ISO/IEC 23002-4:2014/Amd.1:2014(E)
©
ISO/IEC 2014

---------------------- Page: 1 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
COPYRIGHT PROTECTED DOCUMENT


©  ISO/IEC 2014
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form or by any
means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior written permission.
Permission can be requested from either ISO at the address below or ISO’s member body in the country of the requester.
ISO copyright office
Case postale 56  CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail [e-mail]
Web www.iso.org
Published in Switzerland

ii © ISO/IEC 2014 – All rights reserved

---------------------- Page: 2 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Foreword
ISO (the International Organization for Standardization) and IEC (the International Electrotechnical
Commission) form the specialized system for worldwide standardization. National bodies that are members of
ISO or IEC participate in the development of International Standards through technical committees
established by the respective organization to deal with particular fields of technical activity. ISO and IEC
technical committees collaborate in fields of mutual interest. Other international organizations, governmental
and non-governmental, in liaison with ISO and IEC, also take part in the work. In the field of information
technology, ISO and IEC have established a joint technical committee, ISO/IEC JTC 1.
The procedures used to develop this document and those intended for its further maintenance are described
in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the different types of
document should be noted. This document was drafted in accordance with the editorial rules of the
ISO/IEC Directives, Part 2 (see www.iso.org/directives).
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. ISO and IEC shall not be held responsible for identifying any or all such patent rights. Details of any
patent rights identified during the development of the document will be in the Introduction and/or on the ISO
list of patent declarations received (see www.iso.org/patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation on the meaning of ISO specific terms and expressions related to conformity assessment,
as well as information about ISO's adherence to the WTO principles in the Technical Barriers to Trade (TBT)
see the following URL: Foreword - Supplementary information
The committee responsible for this document is ISO/IEC JTC 1, Information technology, Subcommittee SC 29,
Coding of audio, picture, multimedia and hypermedia information.
© ISO/IEC 2014 – All rights reserved iii

---------------------- Page: 3 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Information technology — MPEG video technologies —
Part 4:
Video Tool Library
AMENDMENT 1: Graphics tool library (GTL) for the reconfigurable
multimedia coding (RMC) framework
In 4.1 "FU Interfaces", before Table 1, add the following text:
 In several FU diagrams, the ports are named with a trailing “_i” for the input port type and with a
trailing “_o” for the output port type.
 Some FU diagrams contains as well the Finite State Machine diagram. The following conventions
apply: INPUT - the action of reading a token or a set of tokens from the input port, OUTPUT - the
action of writing the token or a set of tokens to an output port.
 "Parameter" is set at network configuration stage (cannot be changed during the process) and it is
characteristic for each FU
 Token RANGE: describes the mathematical interval for the token value
Examples:
Token RANGE: { 0, 1} – binary value
Token RANGE: [ 0 . N ], value [ 0, N ] real values, closed interval

 All the FUs require the data to be in little-endian format.

In 4.2 "FU IDs", complete Table 2 with the following lines:
Note: update the FU table.
ID FU Name
107
Algo_Parser_SC3DMC
108
Algo_InverseQuantization1D
109
Algo_InverseQuantizationND
110
Algo_InversePrediction1D
111
Algo_InversePredictionND
112
Algo_ED_AD_StaticBit
113
Algo_ED_AD_AdaptiveBit
114
Algo_ED_VLD
115
Algo_ED_BitPrecision
116
Algo_ED_AD
117
Algo_ED_AD_EG
118
Algo_ContextModeling
© ISO/IEC 2014 – All rights reserved 1

---------------------- Page: 4 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
119
Algo_ContextModeling_SVA_nType
120
Algo_ContextModeling_SVA_Indexes
121
Algo_ContextModeling_SVA_Vertex_Attribute
122
Algo_ED_4bitsD
123
Algo_ED_ FixedLength
124
Algo_LookUpTable1D
125
Algo_DecodeConnectivity_SVA
126
Algo_DecodeConnectivity_TFAN
127
Algo_ExtractMask_SC3DMC
128
Algo_ExtractFaceDirection_SVA
129
Algo_simpleMath_2op
130
Algo_Connectivity_InversePrediction_SVA
131
Mgnt_Replicate_1_2
132
Mgnt_Replicate_1_4
133
Mgnt_Replicate_1_8
134
Mgnt_MUX_2_1
135
Mgnt_MUX_4_1
136
Mgnt_MUX_8_1
137
Mgnt_DEMUX_1_2
138
Mgnt_DEMUX_1_4
139
Mgnt_DEMUX_1_8
140
Mgnt_ExtractSegment
141
Mgnt_ProviderValue
142
Mgnt_RepeatSegment
143
Mgnt_ExtractBytes
144
Mgnt_ExtractBits
145
Mgnt_Provider1D
146
Mgnt_Provider2D

4.3 "Token Pool":
Add the following rows at the end of the table.
ID & Name Description
Token which value is 0 or 1.
55 BOOLEAN
Token which value is 0 or 1.
56 SIGN
Token which value is 0 or 1.
57 FLAG
Unsigned integer on 2 bits.
58 UINT_2
Unsigned integer on 4 bits.
59 UINT_4
Unsigned integer on 8 bits.
60 UINT_8
Unsigned integer on 16 bits.
61 UINT_16
Unsigned integer on 32 bits.
62 UINT_32
2 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 5 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Unsigned integer on 64 bits.
63 UINT_64
Integer on 8 bits.
64 INT_8
Integer on 16 bits.
65 INT_16
Integer on 32 bits.
66 INT_32
Integer on 64 bits.
67 INT_64
Float on 32 bits.
68 FLOAT_32

Add 4.4 "Array data order":
4.4 Array data order
 Row based:
The data is processed or sent in a sequential order, row by row
Example:


 Column based
The data is processed or sent in a sequential order, column by column
Example:



Add 4.5 "Input ports":
4.5 Input ports ( reset_i, init_i, start_i)
An FU does not have an outside synchronization signal or synchronization mechanism. These ports are used
for the purpose of changing the values of the local variables to default or initialization values.

Add 4.6 "FU block diagram notations":
4.6 FU block diagram notations
The notation [EMBED] defines a part of the main FSM schematic that is described as a separate schematic
(for complexity reasons). The [EMBED] schematic is an integrated part of the main FSM schematic
© ISO/IEC 2014 – All rights reserved 3

---------------------- Page: 6 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)

The notation [MODULE] defines a part of the main FSM schematic that is defined as a separate FU. The
module schematic is integrated in the main schematic with the entire FU logic, except the “START” FSM state.
The INPUT/OUTPUT states do not read or write values from the ports, they refer to local variables relative to
the FU that embeds the other schematic.

Add 4.7 "Conventions":
4.7 Conventions
The significance of the “sign” port values is:
Value Significance
0 negative
1 positive

The significance of the “flag” values is:
Value Significance
0 false
1 true

Add 5.2 "General Processing FUs":
5.2 General Processing FUs
5.2.1 Algo_InverseQuantization1D
FU Name Algo_InverseQuantization1D
Port Name Direction Token TYPE
(I/O)
dataIn_i I INT8, INT16,
INT32, INT64
qp_i I UINT_32
quantMin_i I FLOAT
quantRange_i I FLOAT
segmentSize_i I UINT8,
UINT16,
UINT32,

UINT64
quantizationMode_i I UINT_2
dataOut_o O FLOAT
Description

Inverse Quantization Process:
START:
INPUT_SEGMENT_PARAM:
INPUT:
quantizationMode
segmentSize
SegmentSizeCounter = 0
IF quantizatioMode = 0
INPUT:
qp
quantMin
quantRange
IF ( quantRange > 0.0 )
delta = ( ( 1 << qp ) - 1 ) / quantRange
4 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 7 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
ELSE
delta = 1.0
IF quantizationMode = 1
INPUT:
qp
INPUT_DATA_IN:
INPUT:
dataIn
IF quantizatioMode = 0
PROCESS:
dataOut = quantMin + (dataIn / delta )
IF quantizatioMode = 1
PROCESS:
dataOut = dataIn / qp
OUTPUT
dataOut
SegmentSizeCounter ++
IF SegmentSizeCounter < segmentSize
GOTO INPUT_DATA_IN
ELSE
GOTO INPUT_SEGMENT_PARAM

The following table contains the quantization types index used in the Inverse Quantization 1D FU:
Name of quantization mode Value
Uniform Quantization 0
Uniform Texture Quantization 1

The "Inverse Quantization" is an algorithm (step by step procedures) that allows a set of
data to be represented with a limited set of values that are associated with its nearest
representative.
For a number of “segmentSize” of input data (dataIn), it uses the same set of quantMin,
quantRange and quantValue to produce a set of output data (dataOut) of size
“segmentSize”
ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported

5.2.2 Algo_InverseQuantizationND
FU Name Algo_InverseQuantizationND
Port Name Direction Token TYPE
(I/O)
dataIn_i I INT8, INT16,
INT32, INT64
qp_i I UINT_32
quantMin_i I FLOAT
quantRange_i I FLOAT
segmentSize_i I UINT8, UINT16,
UINT32, UINT64
quantizationMode_i I UINT_2
Description dataOut_o O FLOAT


© ISO/IEC 2014 – All rights reserved 5

---------------------- Page: 8 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Inverse Quantization Process:
dimD, if homogeneousQ 0

dimQ 

1, if homogeneousQ 1

START:
INPUT_SEGMENT_PARAM
INPUT:
quantizationMode
segmentSize
SegmentSizeCounter = 0
IF quantizatioMode = 0
INPUT:
qp
quantMin [ dimQ ]
quantRange [ dimQ ]
WHILE dimQ_counter < dimQ
IF ( quantRange [ dimQ_counter ] > 0.0 )
delta [ dimQ_counter ] = ( ( 1 << qp ) - 1 ) / quantRange [ dimQ_counter ]
ELSE
delta [ dimQ_counter ] = 1.0;
dimQ_counter++
IF quantizationMode = 1
INPUT:
Qp [ dimQ ]
IF quantizationMode = 2
INPUT:
qp
PROCESS:
Subdivision = (qp - 3) / 2

INPUT_DATA_IN:
INPUT:
dataIn [ dimQ ]
IF quantizatioMode = 0
PROCESS:
dataOut = quantMin [ segmentSizeCounter%dimD ] + (dataIn / delta [segmentSizeCounter%dimD])
IF quantizatioMode = 1
PROCESS: EMBED Code 2 Normal
IF quantizatioMode = 2
PROCESS:
dataOut = dataIn / qp [ segmentSizeCounter % dimD ]
OUTPUT
dataOut
SegmentSizeCounter ++
IF SegmentSizeCounter < segmentSize
GOTO INPUT_DATA_IN
ELSE
GOTO INPUT_SEGMENT_PARAM


EMBED: Code 2 Normal
Mask = ( 1 << ( 2 * subdivision) ) - 1
tricode = data & mask;

// Find y coordinate by solving 2nd degree equation
factor = 1 << subdivision
2
y = factor - sqrt ( (factor ) - tricode )
tricode = tricode + ( y * ( y - (2 * factor) ) )
x = tricode / 2
upsideDown = tricode % 2

// Calculate coordinates for all vertices in triangle
v1x = x + upsideDown
v1y = y + upsideDown
v2x = x + 1
v2y = y
v3x = x
v3y = y + 1

// Calculate coordinates of barycenter
6 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 9 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
invMaxCoord = 1 / factor
normal [ 0 ] = (v1x + v2x + v3x) * invMaxCoord
normal [ 1 ] = (v1y + v2y + v3y) * invMaxCoord
normal [ 2 ] = 3 - normal [ 0 ] - normal [ 1 ]

// Flip component signs if necessary
octantCode = ( data >> 2 * subdivision ) & 0x7
if (octantCode & 0x4)
normal [ 0 ] = (-1) * normal [ 0 ]
if (octantCode & 0x2)
normal [ 1 ] = (-1) * normal [ 1 ]
if (octantCode & 0x1)
normal [ 2 ] = (-1) * normal [ 2 ]

2 2 2
invNorm:= 1 / sqrt ( (normal[ 0 ]) + (normal [ 1 ]) + (normal [ 2 ]) );

//Write the 3 output values
normal [ 0 ] = normal [ 0 ] * invNorm
normal [ 1 ] = normal [ 1 ] * invNorm
normal [ 2 ] = normal [ 2 ] * invNorm

dataOut = normal
The following table contains the quantization types index used in the Inverse Quantization ND FU:
Name of quantization mode Value
Uniform Quantization 0
Normal Quantization 1
Uniform Texture Quantization 2

The "Inverse Quantization" is an algorithm (step by step procedures) that allows a set of
data to be represented with a limited set of values that are associated with its nearest
representative.
For a number of “segmentSize” x “dimD” of input data (dataIn), it uses the same set of
quantMin, quantRange and quantValue of size “dimD” to produce a set of output data
(dataOut) of size segmentSize” x “dimD”.
For each set of size “dimD” of input data (dataIn) it uses the corresponding value of
quantMin, quantRange and quantValue.
ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported
Parameter
Name Description Type / Range
Type: Integer
Describes the number of tokens of type dataIn_i that are consumed at
dimD
each firing. This parameter is set at the network configuration level.
5
Range: [1 . 2 ]
Describes the number of tokens of type quantRange_i, quantMin_i and
quantValue_i that are necessary for the inverse quantization process. Type: Boolean
homogeneousQ This parameter is set at the network configuration level. The number of
tokens is equal to dimD if this parameter is 0 and the number of tokens Range: {0,1}
is equal to 1 if this parameter is 1


© ISO/IEC 2014 – All rights reserved 7

---------------------- Page: 10 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
5.2.3 Algo_InversePrediction1D
FU Name Algo_InversePrediction1D
This FU allows five modes of inverse prediction. Not all the ports are used for
each prediction mode. This FU can handle one input token at a time.

Port Name Direction Token TYPE
(I/O)
dataIn_i I INT8, INT16,
INT32, INT64
difValue_i I INT8, INT16,
INT32, INT64
predMode_i I UINT_4
fan_i I INT8, INT16,
INT32, INT64
dim_i I UINT8, UINT16,
UINT32, UINT64
n_i I INT8, INT16,
INT32, INT64
max_i I UINT8, UINT16,
UINT32, UINT64
dataOut_o O INT8, INT16,
INT32, INT64
Description

Process Parallelogram Inverse Prediction Schematic (FSM):
8 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 11 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)

Inverse Prediction Process:

Switch ( predMode )
{
Case 0: NP – No Prediction
          dataOut[ j]  dataIn [ j],   j {0 .N1 } .
Case 1: Diff – Differential Prediction
 DataIn j ,  j 0
 
dataOut[]j 

DataIn[ j] DataOut[ j1],  j {1 .N1}


  Case 2: XOR – based prediction
 dataIn[]j ,  j 0
dataOut[]j 

dataIn j dataOut[ j1],  j {1 .N1}



  Case 3: Adaptive Prediction
 dataIn j , if difValue j0 or j 0
   

dataOut[]j 

dataIn jdataOut[ j difValue[ j]], otherwise




  Case 4: Circular Differential Prediction
 dataIn j ,  j 0
 

dataOut[ j] d , if dataIn[ j] dataIn[ j1] , where


do, therwise

dataIn[1j ]M dataIn[ j], if dataIn j dataIn[1j ]
 

d
d 

dataIn jM dataIn[1j], otherwise

 d



© ISO/IEC 2014 – All rights reserved 9

---------------------- Page: 12 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
  Case 5: Parallelogram Inverse Prediction

a = fan ( 0 )
if a not visited
d = 0
WHILE d < dim
dataOut [ a x dim + d ] = dataIn [ a x dim + d ]
d = d + 1
Mark a as visited
If number of vertexes > 3
b = fan ( 1 )
if b not visited
d = 0
WHILE d < dim
dataOut [ b x dim + d ] = dataIn [ b x dim + d ] + dataOut [ a x dim + d ]
d = d + 1
Mark b as visited
c = fan ( 2 )
if c not visited
init h, t
a
WHILE t = (I,j,k) Triangles incident to vertex a
a
If ( i=a & j=b & k≠c and k not visited ) h = k, break
If ( i=a & k=b & j≠c and j not visited ) h = k, break
If ( j=a & i=b & k≠c and k not visited ) h = k, break
If ( j=a & k=b & i≠c and i not visited ) h = k, break
If ( k=a & i=b & j≠c and j not visited ) h = k, break
If ( k=a & b=b & i≠c and i not visited ) h = k, break
If h not set
d = 0
WHILE d < dim
dataOut [ c x dim + d ]=dataIn [ c x dim + d ]+ dataOut [ b x dim + d ]
d = d + 1
else
d = 0
WHILE d < dim
dataOut [ c x dim + d ] =
dataIn [ c x dim + d ] + dataOut [ b x dim + d ]
– dataOut [ h x dim + d ] + dataOut [ a x dim + d ]
d = d + 1
Mark c as visited
g = 3, d = 0
WHILE g < fanSize
q = fan ( g )
if q not visited
WHILE d < dim
dataOut [ q x dim + d ] =
dataIn [ q x dim + d ] + dataOut [ c x dim + d ]
- dataOut [ b x dim + d ] + dataOut [ a x dim + d ]
d = d + 1
b = c
c = q

The following table contains the prediction types index used in the Inverse Prediction 1D FU:
Name of prediction mode Value
No Prediction 0
Differential Prediction 1
XOR based prediction 2
10 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 13 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Adaptive Differential Prediction 3
Circular Differential Prediction 4
Parallelogram Inverse Prediction 5

The detailed description of the inverse parallelogram prediction is described in
Annex F.
ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported
Parameter
Name Description Range


5.2.4 Algo_InversePredictionND
Algo_InversePredictionND
FU Name
This FU allows five modes of inverse prediction. Not all the ports are used for
each prediction mode. This FU can handle a number of dimD input tokens at a
time.
Port Name Direction Token RANGE
(I/O)
dataIn_i I INT8, INT16,
INT32, INT64
difValue_i I INT8, INT16,
INT32, INT64
predMode_i I UINT_4
Description
fan_i I INT8, INT16,
INT32, INT64
dim_i I UINT8, UINT16,
UINT32, UINT64
n_i I INT8, INT16,
INT32, INT64
max_i I UINT8, UINT16,
UINT32, UINT64
dataOut_o O INT8, INT16,
INT32, INT64

Process Schematic (FSM):
© ISO/IEC 2014 – All rights reserved 11

---------------------- Page: 14 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)


Process Parallelogram Inverse Prediction Schematic (FSM):

Inverse Prediction ND Process:
Switch( predMode )
12 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 15 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
{
Case 0: NP – No Prediction

           dataOut[ j]  dataIn [ j],   j {0 . N x dimD1 }
 
  Case 1: Diff – Differential Prediction
 dataIn j ,  j 0
 

dataOut[]j 

dataIn[ j] dataOut[ jdimD],  j {1 . N x dimD1}



  Case 2: XOR – based prediction
 dataIn j ,  j 0
 

dataOut[]j 

dataIn[ j] dataOut[ jdimD],  j {1 . N x dimD1}



Case 3: Adaptive Prediction
 dataIn j , if difValue j0 or j 0
   

dataOut[]j 

dataIn[]j []dataOut j difValue j dimD, otherwise



  Case 4: Circular Differential Prediction
 DataIn j ,  j 0
 

dataOut[]j d, if dataIn[]j dataIn[ j dimD] , where


d , otherwise

dataIn[]jdimD M dataIn[ j ], if dataIn j dataIn[]j dimD
 

d
d 

dataIn jM dataIn[ j dimD], otherwise

 d

}

Case 5: Parallelogram Inverse Prediction

a = fan ( 0 )
if a not visited
d = 0
WHILE d < dim
dataOut [ a x dim + d ] = dataIn [ a x dim + d ]
d = d + 1
Mark a as visited
If number of vertexes > 3
b = fan ( 1 )
if b not visited
d = 0
WHILE d < dim
dataOut [ b x dim + d ] = dataIn [ b x dim + d ] + dataOut [ a x dim + d ]
© ISO/IEC 2014 – All rights reserved 13

---------------------- Page: 16 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
d = d + 1
Mark b as visited
c = fan ( 2 )
if c not visited
init h, t
a
WHILE t = (I,j,k) Triangles incident to vertex a
a
If ( i=a & j=b & k≠c and k not visited ) h = k, break
If ( i=a & k=b & j≠c and j not visited ) h = k, break
If ( j=a & i=b & k≠c and k not visited ) h = k, break
If ( j=a & k=b & i≠c and i not visited ) h = k, break
If ( k=a & i=b & j≠c and j not visited ) h = k, break
If ( k=a & b=b & i≠c and i not visited ) h = k, break
If h not set
d = 0
WHILE d < dim
dataOut [ c x dim + d ]=dataIn [ c x dim + d ]+ dataOut [ b x dim + d ]
d = d + 1
else
d = 0
WHILE d < dim
dataOut [ c x dim + d ] =
dataIn [ c x dim + d ] + dataOut [ b x dim + d ]
– dataOut [ h x dim + d ] + dataOut [ a x dim + d ]
d = d + 1
Mark c as visited
g = 3, d = 0
WHILE g < fanSize
q = fan ( g )
if q not visited
WHILE d < dim
dataOut [ q x dim + d ] =
dataIn [ q x dim + d ] + dataOut [ c x dim + d ]
- dataOut [ b x dim + d ] + dataOut [ a x dim + d ]
d = d + 1
b = c
c = q

The following table contains the prediction types index used in the Inverse Prediction ND FU:
Name of prediction mode Value
No Prediction 0
Differential Prediction 1
XOR based prediction 2
Adaptive Differential Prediction 3
Circular Differential Prediction 4
Parallelogram Inverse Prediction 5

The detailed description of the inverse parallelogram prediction is described in
Annex F.
ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported
14 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 17 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Parameter
Name Description Range
Describes the number of tokens of type dataIn_i Type: Integer
dimD that are consumed at each firing. This parameter is
5
set at the network configuration level. Range: [1 . 2 ]


5.2.5 Algo_ED_AD_StaticBit
FU Name Algo_ED_AD_StaticBit
This FU describes the arithmetic decoding process based on a static bit model.
Port Name Direction Token RANGE
(I/O)
Bit_0_prob_i I UINT8, UINT16,
UINT32, UINT64
valueIn_i I UINT8, UINT16,
UINT32, UINT64
length_i I UINT8, UINT16,
UINT32, UINT64
valueOut_o O UINT8, UINT16,
UINT32, UINT64
lengthOut_o O UINT8, UINT16,

UINT32, UINT64
bit_o O UINT8, UINT16,
UINT32, UINT64

Process Schematic (FSM):
Description
ED_AD_StaticBit Process:
START
INPUT:
Bit_0_prob_i
lengthIn_i
valueIn_i
product = bit_0_prob * ( lengthIn >> BM_LengthShift )
bit = ( valueIn >= product )
if ( bit == 0 )
valueOut = valueIn
lengthOut = product
else
valueOut = valueIn – product
lengthOut = lengthIn – product
OUTPUT:
valueOut_o
lengthOut_o
GOTO START
ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported
© ISO/IEC 2014 – All rights reserved 15

---------------------- Page: 18 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
Parameter
Name Description Range
Describes the threshold for renormalization. This Type:Unsigned Integer
AC__MinLength
32
parameter is set at the network configuration level. Range: [1 . 2 ]
Describes the length bits discarded before
Type:Unsigned Integer
BM__LengthShift multiplication. This parameter is set at the network
5
Range: [1 . 2 ]
configuration level.


5.2.6 Alog_ED_AD_AdaptiveBit
FU Name Algo_ED_AD_AdaptiveBit
This FU describes the arithmetic decoding process based on a adaptive bit model as
presented in [3,4].
Port Name Direction Token RANGE
(I/O)
valueIn_i I UINT8, UINT16,
UINT32, UINT64
length_i I UINT8, UINT16,
UINT32, UINT64
reset_i I BOOLEAN
valueOut_o O UINT8, UINT16,
UINT32, UINT64

lengthOut_o O UINT8, UINT16,

UINT32, UINT64
Process Schematic (FSM):
bit_o O UINT8, UINT16,
UINT32, UINT64

ED_AD_AdaptiveBit Process:
START
If reset = 1
Description
bit_0_count = 1;
bit_count  = 2;
bit_0_prob = 1U << (BM__LengthShift - 1);
update_cycle = bits_until_update = 4;

product = bit_0_prob * ( lengthIn >> BM_LengthShift )
bit = ( valueIn >= product )

if ( bit == 0 )
valueOut = valueIn
lengthOut = product
++bit_0_count
else
valueOut = valueIn – product
lengthOut = lengthIn – product

if (bit_count += update_cycle) > BM__MaxCount
bit_count = (bit_count + 1) >> 1
bit_0_count = (bit_0_count + 1) >> 1
if bit_0_count = bit_count
++bit_count
scale = scaleMax / bit_count;
bit_0_prob = (bit_0_count * scale) >> (31 - BM__LengthShift);
update_cycle = (5 * update_cycle) >> 2;
16 © ISO/IEC 2014 – All rights reserved

---------------------- Page: 19 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
if update_cycle > 64
update_cycle = 64;
bits_until_update = update_cycle;
GOTO START

ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported
Parameter
Name Description Range
Describes the threshold for renormalization. This Type:Unsigned Integer
AC__MinLength
32
parameter is set at the network configuration level. Range: [1 . 2 ]
Describes the length bits discarded before
Type:Unsigned Integer
multiplication. This parameter is set at the network
BM__LengthShift 5
Range: [1 . 2 ]
configuration level.
Describes the max value to compute the scaled bit 0
Type:Unsigned Integer
scaleMax probability. This parameter is set at the network
32
Range: [1 . 2 ]
configuration level.


5.2.7 Algo_ED_VLD
FU Name Algo_ED_VLD
This FU describes the Variable Length Decoding process.

Port Name Direction Token RANGE
(I/O)
dataIn_i I UINT8, UINT16,
UINT32, UINT64
table_i I UINT8, UINT16,
UINT32, UINT64
size_i I UINT8, UINT16,
UINT32, UINT64
position_o O UINT8, UINT16,
UINT32, UINT64
Process Schematic (FSM):
Description

ED_VLD Process:
START
SET valid[size] = 1
SET n = 1
DO
DO
If valid[j] = 1
If dataIn=table[j][n]
If table[j][0] = n
dataOut = j
else
valid[j] = 0
j = j + 1
WHILE j < size
© ISO/IEC 2014 – All rights reserved 17

---------------------- Page: 20 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
n = n + 1
WHILE n < nBits
GOTO START
The matrix tokens have to be send column based.
ISO Standards
ISO/IEC 14496-16:2011
using the FU
Profiles@levels

supported
Parameter
Name Description Range

Describes the length of the bits used for the
Type: Integer
nBits search algorithm. This parameter is set at the
5
Range: [1 . 2 ]
network configuration level.

5.2.8 Algo_ED_BitPrecision
FU Name Algo_ED_BitPrecision
This FU describes the Entropy Decoding Bit Precision processes.
Port Name Direction Token RANGE
(I/O)
dataIn_i I UINT_8
numerOfData_i I UINT8, UINT16,
UINT32, UINT64
dim_i I UINT8, UINT16,
UINT32, UINT64
prefixSize_i I UINT_8
hasSign_i I BOOLEAN
hasPrediction_i I BOOLEAN
BPLTable_i I UINT8, UINT16,
UINT32, UINT64
BPLTableSize_i I UINT8, UINT16,
UINT32, UINT64

dataOut_o O INT8, INT16, INT32,
INT64
preds_o O INT8, INT16, INT32,
INT64
postProcessing_o O BOOLEAN
sign_o O BOOLEAN

Description
ED_BitPrecision Schematic (FSM):

ED_BitPrecision process:
START
INPUT: BPLTableSize_i
INPUT: BPLTable_i [ BPLTableSize ]
INPUT: numberOfData_i
INPUT: dim_i
IF STATIC_CONTEXT = 1
INPUT: hasSign_i
INPUT: hasPrediction_i
IF STATIC_PREFIX_SIZE = 1
INPUT: prefixSize_i
j = 0
18 © ISO/IEC 2014 – All rights reserved

ELSE STATIC_PREFIX_SIZE = 1
ELSE STATIC_CONTEXT = 1
else j < numberOfData * dim
ELSE STATIC_PREFIX_SIZE = 0
STATIC_CONTEXT = 0
ELSE
else nBPL > 2
else nBPL > 2
difValue != 0 else
else nBPL > 2

---------------------- Page: 21 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
PROCESS_CHUNK:
WHILE j < numberOfData * dim
IF STATIC_PREFIX_SIZE = 0
INPUT: prefixSize_i
IF STATIC_CONTEXT = 0
INPUT: hasSign_i
INPUT: hasPrediction_i
IF hasSign = 0 & hasPrediction = 0
INPUT: dataIn { prefixSize }
nBPL = dataIn
IF nBPL > 2
INPUT: dataIn { nBPL – 1 }
nPayLoad = dataIn
ELSE
nPayLoad = 0
difValue = BPLTable [ nBPL ] + nPayLoad
dataOut = difValue
postProcessing = 0
sign = 0
OUTPUT: dataOut
OUTPUT: postProcessing_o
OUTPUT: sign_o
ELSE IF hasSign = 1 & hasPrediction = 0
INPUT: dataIn { prefixSize }
nBPL = dataIn
if nBPL > 2
INPUT: dataIn { nBPL – 1 }
nPayLoad = dataIn
ELSE
nPayLoad = 0
difValue = BPLTable [ nBPL ] + nPayLoad
IF difValue != 0
INPUT: dataIn { 1 }
nSign = dataIn
ELSE
nSign = 1
sign = nSign
postProcessing = 1
dataOut = difValue
OUTPUT: dataOut
OUTPUT: postProcessing_o
OUTPUT: sign_o
ELSE IF hasSign = 1 & hasPrediction = 1
signDef[2] = { 1,-1 }
predsOut [ j ] = 0
DO
INPUT: dataIn { 2 }
nPred = dataIn
preds += nPred
WHILE nPred != 3
INPUT: dataIn { prefixSize }
nBPL = dataIn
IF ( nBPL > 2 )
INPUT: dataIn { nBPL – 1 }
nPayLoad = dataIn
ELSE
nPayLoad = 0
difValue = BPLTable [ nBPL ] + nPayLoad
dataOut = difValue
IF preds != 0
postProcessing = 1
ELSE
postProcessing = 0
IF postProcessing != 0
sign = signDef [ difValue % 2 ]
ELSE
sign = 1
OUTPUT: dataOut_o
OUTPUT: preds_o
OUTPUT: postProcessing_o
OUTPUT: sign_o
ELSE IF hasSign = 1 & hasPrediction = 0
IF STATIC_CONTEXT = 0
PROCESS_CHUNK
ELSE
© ISO/IEC 2014 – All rights reserved 19

---------------------- Page: 22 ----------------------
ISO/IEC 23002-4:2014/Amd. 1:2014(E)
GOTO START
GOTO START

An input of size {N} has the meaning of a input of a N-size bit value.
Note: The case when both hasSign=1 and hasPrediction=0 is not considered. The behaviour in this
case is to return to the START stage (if the parameter STATIC_CONTEXT=1) or to process the
next data (if the parameter STATIC_CONTEXT=0).
ISO Standar
...