Fact · Languages and accents
In the Svarah test, the recognizers performed poorly for speakers whose first language is Assamese, Bodo or Nepali, which the authors say are low-resource languages from northeastern India.
Checked against its source on October 17, 2026
In the source's own words
“the models perform poorly for speakers who are native speakers of Assamese, Bodo, and Nepali”
Svarah: Evaluating English ASR Systems on Indian Accents, Tahir Javed et al., arXiv, 2023, Section 5, Results and Discussions.
Open the original- What it measures
- The authors report that this pattern held across the models, and that the relative ranking of models was similar across accents.
- What it does not show
- Three language groups with a small number of speakers each. It shows that accent results vary by region, not the size of any one group's error rate.