---
license: cc-by-4.0
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: name
    dtype: string
configs:
  - config_name: default
    data_files:
      - split: speech_english
        path: data/speech-english-*.parquet
      - split: speech_french
        path: data/speech-french-*.parquet
      - split: speech_german
        path: data/speech-german-*.parquet
      - split: speech_italian
        path: data/speech-italian-*.parquet
      - split: speech_russian
        path: data/speech-russian-*.parquet
      - split: speech_spanish
        path: data/speech-spanish-*.parquet
      - split: noise
        path: data/noise-*.parquet
      - split: rir_slr26
        path: data/rir-slr26-*.parquet
      - split: rir_slr28
        path: data/rir-slr28-*.parquet
---

# DNS5 Challenge data

This is a mirror of the [DNS5 Challenge data](https://github.com/microsoft/DNS-Challenge/).
The original files were converted from WAV to Opus to reduce the size and accelerate streaming.

⚠️ Only the LibriVox, AudioSet, Freesound, OpenSLR26, and OpenSLR28 data is included. The VCTK, VocalSet, CREMA-D, VoxCeleb2, and DEMAND data is excluded. ⚠️

- **Sampling rate**: 48 kHz
- **Channels**: 1
- **Format**: Opus
- **Splits**:
  - **speech_english**: 245 hours, 186743 files
  - **speech_french**: 95 hours, 60454 files
  - **speech_german**: 137 hours, 119175 files
  - **speech_italian**: 70 hours, 59525 files
  - **speech_russian**: 25 hours, 16566 files
  - **speech_spanish**: 86 hours, 78603 files
  - **noise**: 177 hours, 63810 files
  - **rir_slr26**: 17 hours, 60000 files
  - **rir_slr28**: 0.1 hours, 248 files
- **License:**
  - **LibriVox**: [Public domain](https://librivox.org/pages/public-domain/)
  - **AudioSet**: CC BY 4.0
  - **Freesound**: CC0 1.0
  - **OpenSLR26 and OpenSLR28**: Apache 2.0
- **Source:** [https://github.com/microsoft/DNS-Challenge/](https://github.com/microsoft/DNS-Challenge/)
- **Paper:** [ICASSP 2023 Deep Noise Suppression Challenge](https://arxiv.org/abs/2303.11510)

## Usage

```python
import io

import soundfile as sf
from datasets import Features, Value, load_dataset

for item in load_dataset(
    "philgzl/dns5",
    split="speech_english",
    streaming=True,
    features=Features({"audio": Value("binary"), "name": Value("string")}),
):
    print(item["name"])
    buffer = io.BytesIO(item["audio"])
    x, fs = sf.read(buffer)
    # do stuff...
```

## Citation

```bibtex
@article{dubey2024icassp,
  title = {{ICASSP} {2023} {Deep} {Noise} {Suppression} {Challenge}},
  author = {Dubey, Harishchandra and Aazami, Ashkan and Gopal, Vishak and Naderi, Babak and Braun, Sebastian and  Cutler, Ross and Gamper, Hannes and Golestaneh, Mehrsa and Aichner, Robert},
  journal = {IEEE Open J. Signal Process.},
  volume = {5},
  pages = {725--737},
  year = {2024},
}
```