I'm testing nova 2 for livestreaming a phonecall via the websocket, and I'm getting a lot of empty transcription responses like this.
[2023-11-24T02:54:18.188Z] Transcript: {
"type": "Results",
"channel_index": [
0,
1
],
"duration": 4.2800007,
"start": 17.15,
"is_final": true,
"speech_final": false,
"channel": {
"alternatives": [
{
"transcript": "",
"confidence": 0,
"words": []
}
]
},
"metadata": {
"request_id": "1a7593b6-c706-4316-a127-cbfc0ca85c3f",
"model_info": {
"name": "2-phonecall-nova",
"version": "2023-10-12.7356",
"arch": "nova-2"
},
"model_uuid": "d3bfc4af-f650-4441-94b6-e0b0c1c0c9c5"
}
}
Some of the empty transcriptions are marked as speech_final=true and some are not. I can filter these out but it seems to be slowing down the actual transcripts coming through. I'm noticing a 3-4 second delay between speaking and recieving the transcription result.
these are my current settings:
const dgConnection = deepgram.listen.live({
smart_format: false,
interim_results: false,
language: "en-AU",
model: "nova-2-phonecall",
encoding: "mulaw",
sample_rate: 8000
//endpointing: 1000,
//utterance_end_ms: 1001,
//punctuate: true // Ensure punctuation is enabled
});
using node and the sdk v3 beta 6
Any help would be much appreciated.
Cheers
Steve
#Empty Response + slow response.
1 messages · Page 1 of 1 (latest)
Thanks for asking your question. Please be sure to reply with as much detail as possible so we can assist you efficiently. Such as:
- Provide the
request_idif you've a question about a transcription response. - The options you used or the api.deepgram.com URL you sent your request to, including parameters.
- Any code snippets you can include.
- Any audio you can include, or if you can't share it here please email it to us at [email protected] and provide a link to this thread.
one of the request_id's is 013e0be8-fdda-43e3-9a2a-895124aa9267 in this case the delay was more like 6 seconds.
@floral crypt Any idea if this is related to the v3 beta sdk?