---
datasets:
  - OOPPEENN/Galgame_Dataset
configs:
- config_name: default
  data_files:
  - split: train
    path: "data.tsv"
---

- [**Viewer in 🤗 spaces here!**](https://huggingface.co/spaces/litagin/Galgame_dataset_stats)

- Statistics of the audio (voice) files in the [OOPPEENN/Galgame_Dataset](https://huggingface.co/datasets/OOPPEENN/Galgame_Dataset) (in TSV format):

```
game_name	num_speakers	num_mono_files	num_stereo_files	num_error_files	total_duration_hours	avg_sample_rate_kHz	avg_precision	avg_bitrate_kbps	codec	total_size_GB
Game1	47	15055	1	8	20.08	48.0	16.0	88.49	Vorbis	0.73
Game2	40	15370	0	7	30.10	47.8	16.0	87.113	Vorbis	1.07
...
```

- For each game, the [specs](https://huggingface.co/datasets/litagin/Galgame_Dataset_stats/tree/main/specs) folder includes spectrogram images from 5 randomly selected audio files.

![image](https://huggingface.co/datasets/litagin/Galgame_Dataset_stats/resolve/main/specs/AKABEiSOFT3_CharaBration!_~Otome_wa_Koi_shite_Charabureru~/spec1.png)

[Spreadsheet link](https://docs.google.com/spreadsheets/d/12gI6sKJ3k_0r1c-n-LUZJy5y8T5Je1DQ5OEn0t1Upeg/edit?usp=sharing)

- Spaces in game names are replaced with underscores ("_").
- `num_error_files` refers to the number of files that both Sox and pydub cannot extract information from (likely indicating that these files are corrupted).