1. Function 2. Reference Model 3. Input/Output Data
4. SubAIMs 5. JSON Metadata 6. Profiles
7. Reference Software 8. Conformance Testing 9.Performance Assessment

1. Function

  1. At the start, it calculates the offset between Preservation Audio and the Audio of the Preservation Audio-Visual File.
  2. Sends Audio Irregularity File to and receives Video Irregularity Files from Video Analysis for Preservation.
  3. Extracts the Audio Files corresponding to the Irregularities identified in both Irregularity Files.
  4. Sends the Irregularity merged from the Audio and Video Irregularity Files to Tape Irregularity Classification with the corresponding Audio Files.

2. Reference Model

Figure 1 – Audio Analysis for Preservation AIM

3. Input/Output Data

Input data Semantics
Preservation Audio File The input Audio File resulting from the digitisation of an audio open-reel tape to be preserved and, in case, restored.
Preservation Audio-Visual File The input Audio-Visual File resulting from the digitisation of an audio open-reel tape to be preserved and of the output of the video camera pointed at the reading head of the audio playback.
Video Irregularity File A JSON file containing information about the Irregularities of the Preservation Audio-Visual File received from Video Analysis for Recording.
Output data Semantics
Audio File Audio Segments corresponding to Irregularities of the Preservation Audio File.
Audio Irregularity File A JSON file containing information about Irregularities of the Preservation Audio File sent to Video Analysis for Recording.

4. SubAIMs

No SubAIMs.

5. JSON Metadata

https://schemas.mpai.community/CAE1/V2.4/AIMs/AudioAnalysisForPreservation.json

6. Profiles

No Profiles.

7. Reference Software

The CAE-AAP Reference Software can be downloaded from the MPAI Git.

8. Conformance Testing

Receives Preservation Audio File 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.
Preservation Audio-Visual File Shall validate against the Audio-Visual Object schema.
The Qualifier shall validate against the Audio-Visual 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-Visual Object Qualifier schema.
Video Irregularity File Shall validate against the Irregularity File schema.
Produces Audio File 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 Irregularity File Shall validate against the Irregularity File schema.

9. Performance Assessment

Table 23 gives the Audio Recording Preservation (ARP) Audio Analysis for Preservation Means and how they are used.

Table 23 – AIM Means and use of Audio Recording Preservation (ARP) Audio Analysis for Preservation .

Means Actions
Performance Assessment Dataset DS1: n* Preservation Audio Files.

DS2: n Preservation Audio-Visual Files related to DS1.

DS3: n Irregularity Files related to DS2.

DS4: n output Irregularity Files in the format of port IrregularityFileOutput_1 with all Irregularities correctly identified.

DS5: n output Irregularity Files in the format of port IrregularityFileOutput_2 with the real offset and all Irregularities correctly identified and included from DS3.

* A reasonable n for Assessment is 5<n<=10, since each file generates multiple irregularities to classify
Procedure 1.     Feed Audio Analyser under Assessment with DS1, DS2 and DS3.

2.     Compare the computed offsets with the ones contained in DS5.

3.     Analyse the Irregularity Files resulting from port IrregularityFileOutput_1.

4.     Analyse the Irregularity Files resulting from port IrregularityFileOutput_2.
Evaluation 1.     Verify the conditions:

a.     The Irregularity Files are syntactically correct and performing to the JSON schema provided in CAE Technical Specification.

b.     All Irregularities from DS3 are included in the Irregularity Files coming from port IrregularityFileOutput_2.

c.    , where  is the offset computed by the Audio Analyser under Assessment,  is the real offset and FPSDS3 is the number of frames per second at which the DS3 video has been recorded.

d.     All output Audio Files are performing to RF64 file format [7].

e.     For each of the n tuples of input records, the output Audio Files are extracted from the corresponding input Preservation Audio File at the Time Labels indicated in the Irregularity File coming from port IrregularityFileOutput_2.

2.     By inspecting the Irregularity Files resulting from port IrregularityFileOutput_1, for each of the n tuples of input records, compute the values of Recall (R) and Precision (P).

3.    Compute the average value of Recall ( ) and Precision ( ) measures obtained at point 2.

4.     Accept the AIM under Assessment if:

a.   R’>0.9

b.    P’> 0.9

Figure 10 – Audio Analysis for Preservation.

After the Assessment, Performance Assessor shall fill out Table 24.

Table 24 – Performance Assessment form of Audio Recording Preservation (ARP) Audio Analysis for Preservation.

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:ARP:1:0”.
Name of AIM Audio Analyser
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.
Assessment ID Unique Assessment Identifier assigned by Performance Assessor.
Actual output Actual output provided as a matrix of n rows containing  and  values.

Tuple #
1 Measure 1 Measure 1
… … …
n Measure n Measure n
Execution time* Duration of Assessment execution.
Assessment comment* –
Assessment Date yyyy/mm/dd.

* Optional field