mythicinfinity/librispeech-pc-44khz-opus

---
language:
- en
license: cc-by-4.0
task_categories:
- automatic-speech-recognition
- text-to-speech
pretty_name: LibriSpeech-PC 44kHz Opus
dataset_info:
features:
- name: file
dtype: string
- name: audio
dtype:
audio:
sampling_rate: null
- name: text
dtype: string
- name: text_raw
dtype: string
- name: text_normalized
dtype: string
- name: speaker_id
dtype: int64
- name: chapter_id
dtype: int64
- name: id
dtype: string
configs:
- config_name: clean
data_files:
- split: test
path: clean/test/*.parquet
- split: train.100
path: clean/train.100/*.parquet
- split: train.360
path: clean/train.360/*.parquet
- split: validation
path: clean/validation/*.parquet
- config_name: other
data_files:
- split: test
path: other/test/*.parquet
- split: train.500
path: other/train.500/*.parquet
- split: validation
path: other/validation/*.parquet
- config_name: all
default: true
data_files:
- split: test.clean
path: all/test.clean/*.parquet
- split: test.other
path: all/test.other/*.parquet
- split: train.clean.100
path: all/train.clean.100/*.parquet
- split: train.clean.360
path: all/train.clean.360/*.parquet
- split: train.other.500
path: all/train.other.500/*.parquet
- split: validation.clean
path: all/validation.clean/*.parquet
- split: validation.other
path: all/validation.other/*.parquet
---

# LibriSpeech-PC 44kHz Opus

## Dataset Description

- **Homepage:** https://www.openslr.org/145/
- **Source Audio:** https://www.openslr.org/12
- **Repository:** https://huggingface.co/datasets/openslr/librispeech_asr
- **Language:** English
- **License:** CC BY 4.0

## Summary
This dataset is a high-quality audio replacement variant of `Librispeech PC`. It preserves the row identity and text fields while replacing audio content from the source audio with the highest available quality (usually mp3 128kpbs) which is then encoded as Opus (64 kbps). Sampling rate is increased from 16khz up to 48khz depending the on source audio.

LibriSpeech-PC is a merge of `openslr/librispeech_asr` audio metadata with SLR145 punctuation/capitalization manifests. It preserves the original LibriSpeech config/split layout (`clean`, `other`, `all`) and adds punctuation/casing targets.

## Dataset Structure

### Data Fields

- `file`: path to the original LibriSpeech audio file.
- `audio`: `Audio` feature.
- `text`: punctuated + cased transcript from LibriSpeech-PC manifests.
- `text_raw`: raw transcript from LibriSpeech-PC manifests.
- `text_normalized`: original normalized LibriSpeech ASR transcript.
- `speaker_id`: speaker identifier.
- `chapter_id`: chapter identifier.
- `id`: utterance identifier.

### Data Splits

Split names and configs mirror `openslr/librispeech_asr`. Some rows may be absent because the SLR145 manifests drop samples during punctuation/capitalization restoration.

## Additional Information

### Citation Information

```bibtex
@article{meister2023librispeechpc,
title={LibriSpeech-PC: Benchmark for Evaluation of Punctuation and Capitalization Capabilities of end-to-end ASR Models},
author={Meister, A. and Novikov, M. and Karpov, N. and Bakhturina, E. and Lavrukhin, V. and Ginsburg, B.},
journal={arXiv preprint arXiv:2310.02943},
year={2023}
}

@inproceedings{panayotov2015librispeech,
title={LibriSpeech: An ASR corpus based on public domain audio books},
author={Panayotov, V. and Chen, G. and Povey, D. and Khudanpur, S.},
booktitle={ICASSP},
year={2015},
doi={10.1109/ICASSP.2015.7178964}
}
```

### Source Links

- LibriSpeech-PC (SLR145): https://www.openslr.org/145/
- LibriSpeech (SLR12): https://www.openslr.org/12
- Hugging Face LibriSpeech ASR parquet source: https://huggingface.co/datasets/openslr/librispeech_asr