Afrikaans Simulated Call Centre Speech Dataset (www-za-afr-cx)

This repository contains an Afrikaans simulated telephone speech dataset with segmented audio and fine-grained transcripts. Each row in the metadata corresponds to a short segment from a longer call, with speaker information and timing metadata. The data can be used for automatic speech recognition (ASR), speaker-aware ASR, and related research on Afrikaans conversational speech.