Preprint
Automatic Speech Recognition for Non-Native English: Accuracy and Disfluency Handling
Michael McGuire, arXiv, 2025
Read the originalRead on October 16, 2026
- What it is
- A 2025 evaluation of five speech recognition systems on non-native English from the L2-ARCTIC corpus, with speakers whose first languages were Arabic, Chinese, Hindi, Korean, Spanish and Vietnamese. It tests both read and spontaneous speech and how each system handles filler words, repetitions and revisions.
- How much weight it carries
- Preprint, not peer-reviewed at posting. It reports mean match error rate rather than word error rate, so its figures cannot be compared directly with WER figures from other studies. The read-speech set has 24 speakers and the spontaneous set has 22.
Facts taken from this source
- A 2025 study of non-native English speech found that speech recognition systems vary significantly in how they handle filler words, repetitions and revisions.
“finding significant variation in performance across systems and disfluency types” (Abstract)
- In the same 2025 study of non-native English, the five recognition systems varied significantly in how they handled disfluencies such as filler words, repetitions and revisions.
“The study also examined how each system handled disfluencies such as filler words, repetitions, and revisions, finding significant variation in performance across systems and disfluency types.” (Abstract)
- On read non-native English from the L2-ARCTIC corpus, Whisper reached a mean match error rate of 0.054 and AssemblyAI 0.056, which the author describes as approaching human-level accuracy.
“Results showed that for read speech, Whisper and AssemblyAI achieved the best accuracy with mean Match Error Rates (MER) of 0.054 and 0.056 respectively, approaching human-level accuracy.” (Abstract)
- On spontaneous non-native English from the L2-ARCTIC corpus, RevAI had the best mean match error rate of the five systems tested, at 0.063.
“For spontaneous speech, RevAI performed best with a mean MER of 0.063.” (Abstract)
Pages that rest on this source
- Verbatim vs Clean Verbatim TranscriptionGuide
- How do you calculate your speaking speed in words per minute?Question
- Non-native speechGlossary
- Can dictation help non-native English speakers write work emails?Question
- How can non-native speakers get accurate English voice typing?Question
- Speech-to-Text APIs Compared: Price, Accuracy, SpeedGuide
- Speech Recognition and Accents: Accuracy and FixesGuide