Datasets:
pretty_name: Cantonese Speech Samples
license: other
license_name: poseidon-sample-preview
license_link: >-
https://huggingface.co/datasets/psdn-ai/cantonese-speech-samples/blob/main/TERMS.md
tags:
- audio
- chinese
- cantonese
- speech
- transcripts
- samples
language:
- yue
task_categories:
- automatic-speech-recognition
size_categories:
- n<1K
Cantonese Speech Samples
This sample shows Cantonese speech with native transcript metadata. It is meant to help buyers review dialect fit, recording quality, and sample structure before requesting broader coverage.
What This Shows
- Cantonese speech audio with paired transcript metadata
- Language-specific metadata for review and delivery planning
- A compact preview of the available sample structure
Dataset Specifications
| Field | Value |
|---|---|
| Modality | Audio |
| Language | Chinese (Cantonese) |
| Audio format | Audio clips |
| Sample rate | 16 kHz |
| Channels | Mono |
| Transcript source | Ground truth |
| Sample size | 1 clip |
Metadata Coverage
The value is not just the audio clip; it is the combination of listenable media, transcript status, language labels, speaker format, speaker count, speaker context, and technical audio metadata that lets a buyer evaluate fit before requesting broader coverage.
Sample metadata includes: sample identifiers, sample titles, media previews, language labels, speaker format, speaker count, dialect, duration, codec, sample rate, channels, bit depth, transcript text, transcript source, PII status, tags, feature flags, acoustic scene, emotion.
Sample Scope
The sample currently includes 1 clip. Native ground-truth transcript metadata is included where available; we can add more review coverage when needed. Larger deliveries can be scoped by coverage, format, volume, metadata depth, and licensing requirements.
For broader coverage or custom requirements, contact hello@psdn.ai or visit psdn.ai/contact.