---
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: transcript
    dtype: string
  splits:
  - name: train
    num_bytes: 417515940220.32
    num_examples: 4330930
  download_size: 700809728224
  dataset_size: 417515940220.32
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
task_categories:
- text-to-speech
- automatic-speech-recognition
size_categories:
- 1M<n<10M
---
## Cantonese Audio TTS Dataset
This dataset contains [alvanlii/cantonese-radio](https://huggingface.co/datasets/alvanlii/cantonese-radio), [alvanlii/cantonese-youtube](https://huggingface.co/datasets/alvanlii/cantonese-youtube), plus a dataset of equal size. It is catered towards TTS (text-to-speech) use cases, more than the 2 previously published datasets, as there is more extensive filtering and audio enhancement. For speaker labelling, you can use speaker embedding models like Nvidia's [TitaNet](https://huggingface.co/nvidia/speakerverification_en_titanet_large)

Filtered out:
- Overlapped voices, detected using [pyannote/speaker-diarization-3.1](https://huggingface.co/pyannote/speaker-diarization-3.1)
- Music, detected using a custom model
- Languages not Chinese or English
- Transcriptions that are mostly English
- Deduplication using MinHashLSH
- Empty audio, using VAD models


Audios are 
- enhanced/denoised via [sarulab-speech/sidon-v0.1](https://huggingface.co/sarulab-speech/sidon-v0.1) and [ClearVoice](https://github.com/modelscope/ClearerVoice-Studio?tab=readme-ov-file)
- transcribed using [alvanlii/whisper-small-cantonese](https://huggingface.co/alvanlii/whisper-small-cantonese) and [stepfun-ai/Step-Audio-2-mini](https://huggingface.co/stepfun-ai/Step-Audio-2-mini)