#hey so i recently checked out the elevenlabs transcription on the company's website and for initial
1 messages · Page 1 of 1 (latest)
The provided content does not mention any maintenance or affecting transcription or diarization accuracy. However, issues with Arabic transcription and pronunciation have been reported in other contexts, such as with the Modal Eleven Multilingual V2 model. If you're experiencing mixed-language transcription errors, it might be worth checking if the model or voice settings used are optimized for Arabic.
For further assistance, you could try providing more details, such as the exact model used, the type of audio file uploaded, and any specific settings applied. If the issue persists, I recommend reaching out to ElevenLabs' support team directly for clarification on whether maintenance or updates are affecting the service.
the model i am using is scribe v1 for speech to text, even the api key that i have used have the same issues and it is completely out of the blue to see the level of inaccuracy the model is giving interms of both transcription and diarization.
I don’t have specific information about whether the Scribe v1 model is currently under maintenance or if updates are being implemented. However, issues like the ones you’re describing—such as inaccurate transcription, mixed languages, and poor diarization—could occur due to various factors, including temporary system changes or unexpected model behavior.
To address this, I recommend the following steps:
-
Verify Input Quality: Ensure the audio files you’re uploading are clear, with minimal background noise, and in the correct format.
-
Check Settings: Double-check the settings you’re using for transcription and diarization to ensure they align with your requirements.
-
Contact Support: If the issues persist, it’s best to reach out to the support team directly. Provide them with specific examples, including the audio files, the text output, and the settings you used. This will help them investigate the problem further.
You can submit a ticket through the Help Center for assistance.
do i have to write a mail?
To address this issue, it would be best contact the ElevenLabs support team directly. You can provide them with detailed information about the problem, including:
- The specific issues you're facing with transcription and diarization accuracy.
- The model you're using (Scribe v1).
- Whether the problem occurs both on the website and via the API.
- Any relevant audio samples or examples of the inaccuracies.
You can submit a ticket through the Help Center or use the ticket form to report the issue. This will allow the support team to investigate further and provide assistance.
but the arabic transcription was working absolutely fine