Azure AI Speech
An Azure service that integrates speech processing into apps and services.
2,000 questions
Random "Yes." Issue in Arabic ASR (Multiple Dialects)
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(35.2, 16%, 13%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EIA%3C/text%3E%3C/svg%3E)
Imran Ali
0
Reputation points
Dear team,
we are building a conversation IVR via our server, where the customer can make the call and get the conversation IVR treatment. We have integrated our SoftSwitch (FreeSwtich) to MS ASR. We have been facing a random issue where the MS ASR (Egyptian and Moroccan dialects) is sending a random "Yes." to our Rasa bot, at random times or when the Arabic speaker says "yes" in English or even when the speaker is completely silent. We were hoping that an Arabic ASR would not be transcribing in any other language than Arabic at all (as is the case with English, French and Spanish ASRs). Could you please help us fix this issue ASAP?
Azure AI Speech
{count} votes
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(6.4, 61%, 10%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ES%3C/text%3E%3C/svg%3E)
santoshkc 14,500 Reputation points Microsoft External Staff Moderator2025-05-06T14:46:26.69+00:00 Hi @Imran Ali,
To help us investigate the issue further, could you please confirm which Azure Speech transcription setup you're currently using with ASR? Specifically:
- Are you using Real-time Speech-to-Text (STT), a Custom Speech model, or something else?
- If available, could you please share the documentation you're following for the integration?
This will help us narrow down the potential root cause and align with the correct support path.
-
santoshkc 14,500 Reputation points Microsoft External Staff Moderator
2025-05-07T14:16:17.2333333+00:00 Hi @Imran Ali,
Did you got any chance to share the above details to investigate further? Thank you.
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(153.6, 52%, 24%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EJS%3C/text%3E%3C/svg%3E)
JAYA SHANKAR G S 2,525 Reputation points Microsoft External Staff Moderator2025-05-08T10:22:12.4733333+00:00 Hello Imran Ali,
It still processes the audio and gives transcription even if there is English words in conversation.
Below you can check for **Ammunition. زخيرة. ذخيرة
**
Regarding outputs when speaker is silent please share your setup so that we can reproduce the issue from our end and look into the issue.
Thank you
-
JAYA SHANKAR G S 2,525 Reputation points Microsoft External Staff Moderator
2025-05-09T08:07:12.42+00:00
Sign in to comment
Sign in to answer