Using UtteranceEnd and Endpointing

[celestk]

Hi Team,

I'm building a conversational chatbot and using speech_final = true to detect end of speech. It works well in most cases. My configs:
interim_results: true, smart_format: true, endpointing: 800, utterance_end_ms: 2000, filler_words: true

However there are some cases whereby the speech_final doesn't come through, probably due to background noise.

I understand that according the docs (https://developers.deepgram.com/docs/understanding-end-of-speech-detection#using-utteranceend-and-endpointing), we can "trigger if you receive an UtteranceEnd message with no preceding speech_final=true message and send the last-received transcript for further processing".

My question is when do i start tracking this particular Utterance End event?

I notice when speech_final is working fine, i will get an Utterance End event in the next 2nd or 3rd or 4th transcript event, which i should be ignoring. Am i right to say that i need to start tracking the next Utterance End event?

Thank you.

[deepgram-community[bot]]

Hey there! It looks like you haven't connected your GitHub account to your Deepgram account. You can do this at https://community.deepgram.com - being verified through this process will allow our team to help you in a much more streamlined fashion.

[deepgram-community[bot]]

It looks like we're missing some important information to help debug your issue. Would you mind providing us with the following details in a reply?

• The programming language you are working in (e.g. JavaScript, Python).
• A request ID that triggered your error or issue.

[jkroll-deepgram]

Hi @celestk, as you mention, an UtteranceEnd event will come after a speech_final event, if the speech_final event does occur. If it doesn't, you'll make use of the next UtteranceEnd event.

In cases when speech_final fires:

• {transcript with speech_final=false}
• {transcript with speech_final=true} (informs you that end of speech has been detected!)
• {UtteranceEnd} (ignored)

In cases when speech_final does not fire:

• {transcript with speech_final=false}
• {transcript with speech_final=false} (end of speech is missed due to background noise)
• {UtteranceEnd} (informs you that end of speech has been detected!)