From 521e6118bb06cc76c9953351b2dfe0e5883f122b Mon Sep 17 00:00:00 2001 From: Naomi Date: Thu, 15 May 2025 06:51:49 -0700 Subject: [PATCH] feat: add article email-phone-mispronunciation-tts --- ...ispronunciation-tts-1372572174220529758.md | 34 +++++++++++++++++++ 1 file changed, 34 insertions(+) create mode 100644 articles/en/community/email-phone-mispronunciation-tts-1372572174220529758.md diff --git a/articles/en/community/email-phone-mispronunciation-tts-1372572174220529758.md b/articles/en/community/email-phone-mispronunciation-tts-1372572174220529758.md new file mode 100644 index 0000000..8d7b01d --- /dev/null +++ b/articles/en/community/email-phone-mispronunciation-tts-1372572174220529758.md @@ -0,0 +1,34 @@ +# Addressing Mispronunciation of Emails and Phone Numbers in Text-to-Speech + +When using text-to-speech (TTS) technology, there may be instances where certain details such as email addresses and phone numbers are not pronounced accurately. Addressing these challenges can significantly improve the user experience when interacting with automated systems. + +### Common Issues + +1. **Email Mispronunciation**: In some TTS engines, email addresses may not be pronounced as expected. For instance, the domain suffix "au" might be spoken as "aww" rather than individual letters. + +2. **Phone Number Numerical Interpretation**: Phone numbers might be interpreted as large numbers rather than a series of digits (e.g., "1800123456" being read as "one billion eight hundred million..."). + +These challenges were identified in certain requests processed via the Voice Agent API, where such content was not pronounced as anticipated. + +### Solution + +To resolve these issues, upgrading the voice model to a more advanced version, such as `aura-2`, has shown to address the mispronunciation challenges effectively. This update may enhance the naturalness and accuracy of speech outputs, particularly for complex or non-standard text inputs like emails and phone numbers. + +### Implementation + +- **Upgrade Voice Model**: Consider upgrading to an improved voice model. This involves updating your configuration settings to utilize the features of the latest models available in the API you are using. + +- **Verify API Configuration**: Ensure your settings in the Voice Agent API are correctly configured to support advanced pronunciation features. It's crucial to verify and test outputs after changes to confirm improvements. + +### Request IDs for Diagnosis +If you experience similar issues and need in-depth diagnosis, logging request IDs can aid in the support process. For example: +- Email pronunciation issue: `Request ID: 11111111-2222-3333-4444-555555555555` +- Phone number misinterpretation: `Request ID: 66666666-7777-8888-9999-000000000000` + +For persistent issues or inconsistent system behavior, contacting your Deepgram support representative or visiting the community for assistance is recommended: [Deepgram Community on Discord](https://discord.gg/deepgram). + +### Conclusion +Optimizing TTS for better interpretation of email addresses and phone numbers enhances user satisfaction. By keeping your voice model up-to-date and ensuring proper API configurations, you can mitigate these issues effectively. + +### References +- [Deepgram Voice Agent API](https://developers.deepgram.com/docs/voice-agent) \ No newline at end of file