metadata

datasets:
  - Cnam-LMSSC/vibravox
language: fr
library_name: transformers
license: mit
tags:
  - audio
  - audio-to-audio
  - speech
model-index:
  - name: EBEN(M=?,P=?,Q=?)
    results:
      - task:
          type: speech-enhancement
          name: Bandwidth Extension
        dataset:
          name: Vibravox["YOUR_MIC"]
          type: Cnam-LMSSC/vibravox
          args: fr
        metrics:
          - type: stoi
            value: '???'
            name: Test STOI, in-domain training
          - type: n-mos
            value: '???'
            name: Test Noresqa-MOS, in-domain training

Model Card

Developed by: Cnam-LMSSC
Model: EBEN(M=?,P=?,Q=?) (see publication in IEEE TASLP - arXiv link)
Language: French
License: MIT
Training dataset: speech_clean subset of Cnam-LMSSC/vibravox
Samplerate for usage: 16kHz

Overview

This bandwidth extension model, trained on Vibravox body conduction sensor data, enhances body-conducted speech audio by denoising and regenerating mid and high frequencies from low-frequency content.

Disclaimer

This model, trained for a specific non-conventional speech sensor, is intended to be used with in-domain data. Using it with other sensor data may lead to suboptimal performance.

Link to BWE models trained on other body conducted sensors :

The entry point to all EBEN models for Bandwidth Extension (BWE) is available at https://huggingface.co./Cnam-LMSSC/vibravox_EBEN_models.

Training procedure

Detailed instructions for reproducing the experiments are available on the jhauret/vibravox Github repository.

Inference script :

import torch, torchaudio
from vibravox.torch_modules.dnn.eben_generator import EBENGenerator
from datasets import load_dataset

model = EBENGenerator.from_pretrained("Cnam-LMSSC/EBEN_YOUR_MIC")
test_dataset = load_dataset("Cnam-LMSSC/vibravox", "speech_clean", split="test", streaming=True)

audio_48kHz = torch.Tensor(next(iter(test_dataset))["audio.YOUR_MIC"]["array"])
audio_16kHz = torchaudio.functional.resample(audio_48kHz, orig_freq=48_000, new_freq=16_000)

cut_audio_16kHz = model.cut_to_valid_length(audio_16kHz[None, None, :])
enhanced_audio_16kHz = model(cut_audio_16kHz)