Question
Why is the Speech Input not recognized?
Applies To
- Automated Speech Recognition (ASR)
- Voice API
Answer
ASR, detected using the speech parameter of the Input action with Voice API, confidence can be affected by several background and telephony environments of the call. If you are experiencing issues with successful speech recognition, we recommend adding a Record action before the Input action. You can hear what is received by listening to the audio file. We recommend checking the following:
- Is there a lot of background noise?
- Do the words sound clipped? This may indicate echo cancellation is enabled on their device.
- Could non-verbal noises be affecting detection?
- Is the "context" parameter being used in the Input action?
Articles in this section
- How do I verify that the Voice API Webhook request and it's payload has not been tampered during transit?
- How To Check Call Quality With Voice Inspector
- What is a Conversation ID?
- How to configure the custom webhook timeout for Voice API functionality and what are the allowed timeout ranges ?
- How do I change TTL for the Conversations and Call Legs?
- Why are call progress events not being received on the Event URL defined in the conversation action NCCO?
- Why is there an extended silence at the beginning of a call and how can it be reduced?
- Intermittent WebSocket Disconnections
- What Is An NCCO
- How Does Jitter Impact Voice Calls