# microsoft/aec-challenge architecture

> Architecture diagram of the GitHub repository microsoft/aec-challenge, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-18). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/microsoft/aec-challenge
- Repository: https://github.com/microsoft/aec-challenge

## Overview

AEC Challenge combines audio datasets, baseline echo-cancellation inference, and objective AECMOS scoring. A typical researcher workflow is to use microphone and loudspeaker-loopback recordings as enhancement inputs, produce enhanced audio, then evaluate it against the original recordings. The README also describes an online subjective-test framework, but no implementation for it appears in the supplied tree. The graph covers documented capabilities and sampled baseline/scoring code; unsampled dataset internals and subjective-test wiring are not inferred.

Main entry points: [`baseline/icassp2022/enhance.py`](https://github.com/microsoft/aec-challenge/blob/main/baseline/icassp2022/enhance.py) (ICASSP 2022 enhancer), [`baseline/interspeech2021/enhance.py`](https://github.com/microsoft/aec-challenge/blob/main/baseline/interspeech2021/enhance.py) (Interspeech 2021 enhancer), [`baseline`](https://github.com/microsoft/aec-challenge/tree/main/baseline) (Enhanced audio), [`AECMOS/aecmos.py`](https://github.com/microsoft/aec-challenge/blob/main/AECMOS/aecmos.py) (AECMOS client)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Audio datasets"]
  node_training_sets[("Training recordings")]
  node_test_sets[("Challenge test recordings")]
  node_ic21_data[("ICASSP 2021 samples")]
end

subgraph group_enhance["Echo cancellation"]
  node_ic22_enhance["ICASSP 2022 enhancer<br/>[enhance.py]"]
  node_is21_enhance["Interspeech 2021 enhancer<br/>[enhance.py]"]
  node_dec_model["DEC inference<br/>[enhance.py]"]
  node_enhanced_audio["Enhanced audio"]
end

subgraph group_evaluate["Evaluation"]
  node_aecmos_client["AECMOS client<br/>[aecmos.py]"]
  node_local_estimator["Local AECMOS<br/>[aecmos.py]"]
  node_onnx_models["AECMOS ONNX models"]
  node_objective_service{{"Objective scoring service"}}
  node_subjective_framework{{"Subjective test framework"}}
end

subgraph group_internal["Signal processing"]
  node_audio_utils["Audio utilities<br/>[audio_utils.py]"]
end

node_researcher(("Researcher"))

node_researcher -->|"uses data"| node_training_sets
node_researcher -->|"uses data"| node_test_sets
node_researcher -->|"runs inference"| node_ic22_enhance
node_researcher -->|"runs inference"| node_is21_enhance
node_ic22_enhance -->|"configures"| node_dec_model
node_is21_enhance -->|"calls"| node_audio_utils
node_ic22_enhance -->|"writes"| node_enhanced_audio
node_is21_enhance -->|"writes"| node_enhanced_audio
node_researcher -->|"scores audio"| node_aecmos_client
node_aecmos_client -->|"posts audio"| node_objective_service
node_enhanced_audio -.->|"evaluates"| node_aecmos_client
node_researcher -->|"runs scoring"| node_local_estimator
node_local_estimator -->|"loads model"| node_onnx_models
node_researcher -.->|"submits results"| node_subjective_framework

click node_training_sets "https://github.com/microsoft/aec-challenge/tree/main/datasets"
click node_test_sets "https://github.com/microsoft/aec-challenge/tree/main/datasets"
click node_ic21_data "https://github.com/microsoft/aec-challenge/tree/main/baseline/icassp2021"
click node_ic22_enhance "https://github.com/microsoft/aec-challenge/blob/main/baseline/icassp2022/enhance.py"
click node_is21_enhance "https://github.com/microsoft/aec-challenge/blob/main/baseline/interspeech2021/enhance.py"
click node_dec_model "https://github.com/microsoft/aec-challenge/blob/main/baseline/icassp2022/enhance.py"
click node_audio_utils "https://github.com/microsoft/aec-challenge/blob/main/baseline/interspeech2021/audio_utils.py"
click node_enhanced_audio "https://github.com/microsoft/aec-challenge/tree/main/baseline"
click node_aecmos_client "https://github.com/microsoft/aec-challenge/blob/main/AECMOS/aecmos.py"
click node_local_estimator "https://github.com/microsoft/aec-challenge/blob/main/AECMOS/AECMOS_local/aecmos.py"
click node_onnx_models "https://github.com/microsoft/aec-challenge/tree/main/AECMOS/AECMOS_local"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_training_sets,node_test_sets,node_ic21_data toneBlue
class node_ic22_enhance,node_is21_enhance,node_dec_model,node_enhanced_audio toneAmber
class node_aecmos_client,node_local_estimator,node_onnx_models,node_objective_service,node_subjective_framework toneMint
class node_audio_utils toneRose
class node_researcher toneIndigo
```

## Components

### Audio datasets

- [Training recordings](https://github.com/microsoft/aec-challenge/tree/main/datasets): `datasets`
- [Challenge test recordings](https://github.com/microsoft/aec-challenge/tree/main/datasets): `datasets`
- [ICASSP 2021 samples](https://github.com/microsoft/aec-challenge/tree/main/baseline/icassp2021): `baseline/icassp2021`

### Echo cancellation

- [ICASSP 2022 enhancer](https://github.com/microsoft/aec-challenge/blob/main/baseline/icassp2022/enhance.py): `baseline/icassp2022/enhance.py`
- [Interspeech 2021 enhancer](https://github.com/microsoft/aec-challenge/blob/main/baseline/interspeech2021/enhance.py): `baseline/interspeech2021/enhance.py`
- [DEC inference](https://github.com/microsoft/aec-challenge/blob/main/baseline/icassp2022/enhance.py): `baseline/icassp2022/enhance.py`
- [Enhanced audio](https://github.com/microsoft/aec-challenge/tree/main/baseline): `baseline`

### Evaluation

- [AECMOS client](https://github.com/microsoft/aec-challenge/blob/main/AECMOS/aecmos.py): `AECMOS/aecmos.py`
- [Local AECMOS](https://github.com/microsoft/aec-challenge/blob/main/AECMOS/AECMOS_local/aecmos.py): `AECMOS/AECMOS_local/aecmos.py`
- [AECMOS ONNX models](https://github.com/microsoft/aec-challenge/tree/main/AECMOS/AECMOS_local): `AECMOS/AECMOS_local`
- **Objective scoring service**
- **Subjective test framework**

### Signal processing

- [Audio utilities](https://github.com/microsoft/aec-challenge/blob/main/baseline/interspeech2021/audio_utils.py): `baseline/interspeech2021/audio_utils.py`

### Other components

- **Researcher**

## Connections

- Researcher → Training recordings: uses data (evidence: `README.md`)
- Researcher → Challenge test recordings: uses data (evidence: `README.md`)
- Researcher → ICASSP 2022 enhancer: runs inference (evidence: `baseline/icassp2022/enhance.py`)
- Researcher → Interspeech 2021 enhancer: runs inference (evidence: `baseline/interspeech2021/enhance.py`)
- ICASSP 2022 enhancer → DEC inference: configures (evidence: `baseline/icassp2022/enhance.py`)
- Interspeech 2021 enhancer → Audio utilities: calls (evidence: `baseline/interspeech2021/enhance.py`)
- ICASSP 2022 enhancer → Enhanced audio: writes (evidence: `baseline/icassp2022/enhance.py`)
- Interspeech 2021 enhancer → Enhanced audio: writes (evidence: `baseline/interspeech2021/enhance.py`)
- Researcher → AECMOS client: scores audio (evidence: `AECMOS/aecmos.py`)
- AECMOS client → Objective scoring service: posts audio (evidence: `AECMOS/aecmos.py`)
- Enhanced audio → AECMOS client: evaluates
- Researcher → Local AECMOS: runs scoring (evidence: `AECMOS/AECMOS_local/aecmos.py`)
- Local AECMOS → AECMOS ONNX models: loads model (evidence: `AECMOS/AECMOS_local/aecmos.py`)
- Researcher → Subjective test framework: submits results (evidence: `README.md`)
