ANNEX 1
Overview
1 Introduction
2 Applications
3 Versions
4 The subjective domain
5 Resolution and accuracy
6 Requirements and
limitations
APPENDIX 1 TO ANNEX 1
Applications
1 General
2 Main applications
2.1 Assessment of
implementations
2.2 Perceptual quality
line up
2.3 On-line monitoring
2.4 Equipment or
connection status
2.5 Codec
identification
2.6 Codec development
2.7 Network planning
2.8 Aid to subjective
assessment
2.9 Summary of
applications
3 Test signals
3.1 Selection of
natural test signals
3.2 Duration
4 Synchronization
5 Copyright issues
APPENDIX 2 TO ANNEX 1
Output variables
1 Introduction
2 Model Output Variables
3 Basic Audio Quality
4 Coding Margin
5 User requirements
APPENDIX 3 TO ANNEX 1
Model outline
1 Audio processing
1.1 User-defined
settings
1.2 Psycho-acoustic
model
1.3 Cognitive model
APPENDIX 4 TO ANNEX 1
Principles and characteristics of objective perceptual audio quality
measurement methods
1 Introduction and history
2 General structure of
objective perceptual audio quality measurement methods
3 Psycho-acoustical and
cognitive basics
3.1 Outer and middle
ear transfer characteristic
3.2 Perceptual
frequency scales
3.3 Excitation
3.4 Detection
3.5 Masking
3.6 Loudness and
partial masking
3.7 Sharpness
3.8 Cognitive
processing
Example 1: Separation of linear from non-linear distortions
Example 2: Auditory scene analysis
Example 3: Informational masking
Example 4: Spectral-temporal weighting
4 Models incorporated
4.1 DIX
4.2 NMR
4.3 OASE
4.4 Perceptual Audio
Quality Measure (PAQM)
4.5 PERCEVAL
4.6 POM
4.7 The toolbox
approach
ANNEX 2
Description of the model
1 Outline
1.1 Basic Version
1.2 Advanced Version
2 Peripheral ear model
2.1 FFT-based ear model
2.1.1 Overview
2.1.2 Time processing
2.1.3 FFT
2.1.4 Outer and middle ear
2.1.5 Grouping into critical bands
2.1.6 Adding internal noise
2.1.7 Spreading
2.1.8 Time domain spreading
2.1.9 Masking threshold
2.2 Filter bank-based ear model
2.2.1 Overview
2.2.2 Subsampling
2.2.3 Setting of playback level
2.2.4 DC rejection filter
2.2.5 Filter
bank
2.2.6 Outer
and middle ear filtering
2.2.7 Frequency
domain spreading
2.2.8 Rectification
2.2.9 Time
domain smearing (1) – Backward masking
2.2.10 Adding
of internal noise
2.2.11 Time
domain smearing (2) – Forward masking
3 Pre-processing of
excitation patterns
3.1 Level and pattern
adaptation
3.1.1 Level
adaptation
3.1.2 Pattern
adaptation
3.2 Modulation
3.3 Loudness
3.4 Calculation of the
error signal
4 Calculation of Model Output
Variables
4.1 Overview
4.2 Modulation
difference
4.2.1 RmsModDiffA
4.2.2 WinModDiff1B
4.2.3 AvgModDiff1B and AvgModDiff2B
4.3 Noise loudness
4.3.1 RmsNoiseLoudA
4.3.2 RmsMissingComponentsA
4.3.3 RmsNoiseLoudAsymA
4.3.4 AvgLinDistA
4.3.5 RmsNoiseLoudB
4.4 Bandwidth
4.4.1 Pseudocode
4.4.2 BandwidthRefB and BandwidthTestB
4.5 Noise-to-Mask
Ratio
4.5.1 Total
NMRB
4.5.2 Segmental
NMRB
4.6 Relative Disturbed
FramesB
4.7 Detection
probability
4.7.1 Maximum
Filtered Probability of Detection (MFPDB)
4.7.2 Average
distorted block (ADBB)
4.8 Harmonic structure
of error
4.8.1 EHSB
5 Averaging
5.1 Spectral averaging
5.1.1 Linear
average
5.2 Temporal averaging
5.2.1 Linear
average
5.2.2 Squared
average
5.2.3 Windowed
average
5.2.4 Frame
selection
5.3 Averaging over
audio channels
6 Estimation of the perceived
basic audio quality
6.1 Artificial neural
network
6.2 Basic Version
6.3 Advanced Version
7 Conformance of
implementations
7.1 General
7.2 Selection
7.3 Settings for the
conformance test
7.4 Acceptable
tolerance interval
7.5 Test items
APPENDIX 1 TO ANNEX 2
Validation process
1 General
2 Competitive phase
3 Collaborative phase
4 Verification
4.1 Comparison of SDG
and ODG values
4.2 Correlation
4.3 Absolute Error
Score (AES)
4.4 Comparison of ODG
versus the confidence interval
4.5 Comparison of ODG
versus the tolerance-interval
5 Selection of the optimal
model versions
5.1 Pre-selection
criteria based on correlation
5.2 Analysis of number
of outliers
5.3 Analysis of
severeness of outliers
6 Conclusion
APPENDIX 2 TO ANNEX 2
Descriptions of the reference databases
1 Introduction
2 Items per database
3 Experimental conditions
3.1 MPEG90
3.2 MPEG91
3.3 ITU92DI
3.4 ITU92CO
3.5 ITU93
3.6 MPEG95
3.7 EIA95
3.8 DB2
3.9 DB3
3.10 CRC97
4 Items per condition for DB2
and DB3
4.1 DB2
4.2 DB3