CONTENTS

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