ODA Voice – Speech Recognition Challenges with Non-English Languages
i am facing an issue with ODA Voice recognition. It does not capture English words in Latin characters when they are spoken with a non-English pronunciation or accent. We have tested multiple voices, but the issue still persists.
For example, a user may speak an English sentence using Arabic pronunciation/transliteration:
“hal yumkinuk musaeadati fi murajaeat rusum alta'akhur fi alsadadi”
which is intended to mean:
“Can you help me review the late payment charges?”
When the same text is typed manually, it works correctly. However, ODA Voice does not reliably recognize and transcribe the spoken input into the expected English/Latin characters.
Tagged:
0