Which one Performs Better? Wav2Vec or Whisper? Applying both in Badini Kurdish Speech to Text (BKSTT)

TL;DR

This study compares Wav2Vec2 and Whisper speech-to-text models on Badini Kurdish, demonstrating that Wav2Vec2 significantly outperforms Whisper in accuracy and readability for this low-resource dialect.

Contribution

It introduces a Badini Kurdish speech dataset and evaluates the performance of Wav2Vec2 and Whisper models, highlighting the superior accuracy of Wav2Vec2 for this dialect.

Findings

Wav2Vec2 achieved 82.67% accuracy, Whisper 53.17%.

Wav2Vec2 provided 90.38% readability, Whisper 65.45%.

Wav2Vec2 outperforms Whisper in Badini Kurdish STT.

Abstract

Speech-to-text (STT) systems have a wide range of applications. They are available in many languages, albeit at different quality levels. Although Kurdish is considered a less-resourced language from a processing perspective, SST is available for some of the Kurdish dialects, for instance, Sorani (Central Kurdish). However, that is not applied to other Kurdish dialects, Badini and Hawrami, for example. This research is an attempt to address this gap. Bandin, approximately, has two million speakers, and STT systems can help their community use mobile and computer-based technologies while giving their dialect more global visibility. We aim to create a language model based on Badini's speech and evaluate its performance. To cover a conversational aspect, have a proper confidence level of grammatical accuracy, and ready transcriptions, we chose Badini kids' stories, eight books including 78 stories, as the textual input. Six narrators narrated the books, which resulted in approximately 17 hours of recording. We cleaned, segmented, and tokenized the input. The preprocessing produced nearly 15 hours of speech, including 19193 segments and 25221 words. We used Wav2Vec2-Large-XLSR-53 and Whisper-small to develop the language models. The experiments indicate that the transcriptions process based on the Wav2Vec2-Large-XLSR-53 model provides a significantly more accurate and readable output than the Whisper-small model, with 90.38% and 65.45% readability, and 82.67% and 53.17% accuracy, respectively.