TTS MOS estimation with WavLM for LJSpeech

This repository provides all the necessary tools to perform TTS quality evaluation using a WavLM finetuning model. The model attempts to predict the Mean Opinion Score, i.e. averaged human ratings on a scale from 1 to 5.

The model was trained using the SOMOS dataset from Samsung: https://paperswithcode.com/dataset/somos

For a better experience, we encourage you to learn more about SpeechBrain. The model performance on SOMOS test set is:

Release	Pearson R
2023-02-29	0.904

Pipeline description

This system is composed of an WavLM model and a simple forward transformer followed by statistical pooling. The model was trained by first pre-conditioning the model on a simple classifier that attempts to determine whether the rating is above a certain threshold and then fine-tuned on the regression task.

Install SpeechBrain

First of all, please install the development version of SpeechBrain with the following command:

pip install git+https://github.com/speechbrain/speechbrain.git@$develop

Please notice that we encourage you to read our tutorials and learn more about SpeechBrain.

Perform MOS estimation

The RegressionModelSpeechEvaluator interface is used as a high-level wrapper for the MOS estimation task

from speechbrain.inference.eval import RegressionModelSpeechEvaluator
source = "flexthink/ttseval-wavlm-transformer"
eval = RegressionModelSpeechEvaluator.from_hparams(source)

file_names = [
    "LJ002-0181_110.wav",
    "booksent_2012_0005_001.wav",
]
prediction = eval.evaluate_files(file_names)

The prediction is a SpeechEvaluationResult named tuple instance where prediction.score and predictions.details["score"] both indicate the predicted Mean Opinion Score.

Inference on GPU

To perform inference on the GPU, add run_opts={"device":"cuda"} when calling the from_hparams method.

Training

The model was trained with SpeechBrain.

To train it from scratch follows these steps:

Clone SpeechBrain:

git clone https://github.com/speechbrain/speechbrain/

Install it:

cd speechbrain
pip install -r requirements.txt
pip install -e .

Run Training:

cd  recipes/SOMOS/ttseval
python train.py hparams/train.yaml --data_folder=your_data_folder

Limitations

The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.

Citing SpeechBrain

Please, cite SpeechBrain if you use it for your research or business.

@misc{speechbrain,
  title={{SpeechBrain}: A General-Purpose Speech Toolkit},
  author={Mirco Ravanelli and Titouan Parcollet and Peter Plantinga and Aku Rouhe and Samuele Cornell and Loren Lugosch and Cem Subakan and Nauman Dawalatabad and Abdelwahab Heba and Jianyuan Zhong and Ju-Chieh Chou and Sung-Lin Yeh and Szu-Wei Fu and Chien-Feng Liao and Elena Rastorgueva and François Grondin and William Aris and Hwidong Na and Yan Gao and Renato De Mori and Yoshua Bengio},
  year={2021},
  eprint={2106.04624},
  archivePrefix={arXiv},
  primaryClass={eess.AS},
  note={arXiv:2106.04624}
}

About SpeechBrain

Website: https://speechbrain.github.io/
Code: https://github.com/speechbrain/speechbrain/
HuggingFace: https://huggingface.co/speechbrain/