Go to CAE-USC AI Modules

Function
Ref. Model
I/O Data
SubAIMs
JSON MData
Profiles
Ref. Software
Conformance
Performance

1 Functions

  1. Receives Spherical Harmonic Decomposition Coefficients, Transform Audio, Audio Scene Geometry, and Audio Scene Reference.
  2. Eliminates background noise and reverberation which reduce the audio quality, acting as a Passthrough AIM if environmental conditions do not substantially add ambient noise to the desired speech.
  3. Produces Denoised Transform Audio

The Noise Cancellation Module (CAE‑NCM) AIM:

Receives Spherical Harmonics Decomposition Coefficients Result of the transformation of Transform Multichannel Audio into the spherical frequency domain.
Transform Audio Audio Object in the Transform Domain.
Audio Scene Geometry Spatial arrangement of Audio Objects.
Audio Scene Reference Discrete-time and discrete-valued simple acoustic source models used in source separation.
Produces Enhanced Transform Audio Transform Audio whose noise level has been reduced.

2 Reference Model

Figure 1 depicts the Reference Model of the Noise Cancellation Module (CAE‑NCM) AIM.

3 I/O Data

Table 1 – I/O Data of the Noise Cancellation Module (CAE‑NCM) AIM

Input Description
Spherical Harmonics Decomposition Coefficients Result of the transformation of Transform Multichannel Audio into the spherical frequency domain.
Transform Audio Audio Object in the Transform Domain.
Audio Scene Geometry Spatial arrangement of Audio Objects.
Audio Scene Reference Discrete-time and discrete-valued simple acoustic source models used in source separation.
Output Description
Enhanced Transform Audio Transform Audio whose noise level has been reduced.

4 SubAIMs

No SubAIMs.

5 JSON Metadata

https://schemas.mpai.community/CAE1/V2.5/AIMs/NoiseCancellationModule.json

6 Profiles

No Profiles

7 Reference Software

The Noise Cancellation Module can be downloaded from the MPAI Git.

8 Conformance Testing

Table 2 – Conformance Testing Method for the Noise Cancellation Module (CAE‑NCM) AIM

Receives Spherical Harmonics Decomposition Coefficients Shall validate against the Audio Object schema.
The Qualifier shall validate against the Audio Qualifier schema.
The values of any Sub-Type, Format, and Attribute of the Qualifier shall correspond with the Sub-Type, Format, and Attributes of the Audio Object Qualifier schema.
Transform Audio Shall validate against the Audio Object schema.
The Qualifier shall validate against the Audio Qualifier schema.
The values of any Sub-Type, Format, and Attribute of the Qualifier shall correspond with the Sub-Type, Format, and Attributes of the Audio Object Qualifier schema.
Audio Scene Geometry Shall validate against Audio Scene Geometry schema.
Audio Scene Reference Discrete-time and discrete-valued simple acoustic source models used in source separation.
Produces Enhanced Transform Audio Shall validate against the Audio Object schema.
The Qualifier shall validate against the Audio Qualifier schema.
The values of any Sub-Type, Format, and Attribute of the Qualifier shall correspond with the Sub-Type, Format, and Attributes of the Audio Object Qualifier schema.

9 Performance Assessment

Table 58 gives the Enhanced Audioconference Experience (CAE-EAE) Noise Cancellation Means and how they are used.

Table 58 – AIM Means and use of Enhanced Audioconference Experience (CAE-EAE) Noise Cancellation

Means Actions
Performance Assess Dataset DS1: n Test files containing SHD.
DS2: n Test files containing Transform Audio.
DS3: n Test files containing Audio Scene Geometry.
DS4: n Expected Denoised Transform Audio.
Procedure 1.     Feed the AIM under test with the Test files (DS1, DS2, DS3).
2.     Analyse the Denoised Transform Audio (DS4).
Evaluation 1.     Compare the number of Audio Blocks in the Expected Denoised Transform Audio with the number of Audio Blocks in the Denoised Transform Audio Files.
2.     Compute Perception Evaluation of Speech Quality (PESQ) between the Expected and Output Denoised Transform Audio Files [6].
3.     Accept the AIM under test if these two conditions are satisfied:
a.     The number of Audio Blocks in the Denoised Transform Audio is the same with the number of Audio Blocks in the Expected Denoised Transform Audio.
b.     Compare each Denoised Transform Audio with the Expected Denoised Transform Audio.
c.     If the room reverb time (T60) is greater than 0.5 seconds.
i.     Each object’s PESQ between the Expected and Output is greater than P=2.0.
d.     If the room reverb time (T60) is smaller than 0.5 seconds.
i.     Each object’s PESQ between the Expected and Output is greater than P=3.0.

Figure 23 – Noise Cancellation Testing Flow

After the Tests, Performance Assessor shall fill out Table 59.Table 59

Table 59 – Performance Assessment form of Enhanced Audioconference Experience (CAE-EAE) Noise Cancellation

Performance Assessor ID Unique Performance Assessor Identifier assigned by MPAI
Standard, Use Case ID and Version Standard ID and Use Case ID, Version and Profile of the standard in the form “CAE:EAE:1:0”.
Name of AIM Noise Cancellation
Implementer ID Unique Implementer Identifier assigned by MPAI Store.
AIM Implementation Version Unique Implementation Identifier assigned by Implementer.
Neural Network Version* Unique Neural Network Identifier assigned by Implementer.
Identifier of Performance Assessment Dataset Unique Dataset Identifier assigned by MPAI Store.
Test ID Unique Test Identifier assigned by Performance Assessor.
Actual output The Performance Assessor will provide the following matrix containing a limited number of input records (n) with the corresponding outputs. If an input record fails, the tester would specify the reason why the test case fails.

Input data (DS1, DS2, DS3) Expected Output Data

(DS4)

 Data Format PESQ Score
SHD ID1

Transform Audio ID1

Audio Scene Geometry ID1

 

 Denoised Transform Audio ID1

 

 T/F > P
SHD ID2

Transform Audio ID2

Audio Scene Geometry ID2

 

 Denoised Transform Audio ID2

 

 T/F > P
SHD ID3

Transform Audio ID3

Audio Scene Geometry ID3

 

 Denoised Transform Audio ID3

 

 T/F > P
… … … …
SHD IDn

Transform Audio IDn

Audio Scene Geometry IDn

 

 Denoised Transform Audio IDn

 

 T/F > P

Final evaluation : T/F

Denoting with i, the record number in DS1, DS2, and DS3, the matrices reflect the results obtained with input records i with the corresponding outputs i.

DS1 DS2 DS3 DS4 Noise Cancellation output value
(obtained through the AIM under test)
DS1[i] DS2[i] DS3[i] DS4[i] NoiseCancellation[i]
Execution time* Duration of test execution.
Test comment* Comments on test results and possible needed actions.
Test Date yyyy/mm/dd.

* Optional field

Go to CAE-USC AI Modules