Fact · Languages and accents
The Svarah authors found that the two commercial systems they tested were outperformed by the open source Whisper and NVIDIA Conformer models on Indian English.
Checked against its source on October 17, 2026
In the source's own words
“the two commercial models are outperformed by the open-source Whisper and NVIDIA Conformer models”
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 commercial systems were Azure and Google Cloud speech recognition, tested in their US English and India English versions.
- What it does not show
- The authors could not inspect Azure's model details, and the comparison reflects the versions available in 2023.