NCHLT English Auxiliary Speech Corpus With Orthographic Transcriptions

Orthographically transcribed broadband speech is provided for each of South Africa's eleven official languages. Transcriptions are available in XML format. The corpus was authored by Laura Martinus and last updated in March 2026.

Use Cases

• Train automatic speech recognition models on orthographic transcriptions paired with audio for South African English.
• Develop multilingual speech systems using transcribed speech data from eleven official South African languages.
• Analyze phonetic or linguistic features across South Africa's official languages using the transcribed speech corpus.

Strengths

• Covers all eleven official languages of South Africa.
• Provides orthographic transcriptions in a structured XML format.
• Released under a permissive CC-BY-4.0 license.

Limitations

• The dataset scale is indicated as tiny (0.0 B), suggesting a very limited sample size.
• Specific details on audio duration, number of speakers, or recording conditions are unavailable.

Provenance

Source

Laura Martinus via figshare.

Collection Method

Collection and orthographic transcription of broadband speech.

Freshness

Last updated March 2026.

Geography

South Africa.