Sending KeepAlive's but still getting connection closed - (Speech to Text in Browser-side JavaScript)

[mj1618]

Hi,

Here's a screenshot of my logs:

As you can see it shows the times the KeepAlives are sent, when you see "SpeechStart" that is when it is streaming audio data.
It seems that after a while, if I send any audio data then Deepgram sends me a "Metadata" message and immediately closes the stream without any other message or error code as to why.

The following are the last websocket messages I see before closing and the code I am using, and confirmation the KeepAlive messages are getting through:

...

Client version is latest (3.9)
Example request ID: aba8eea6-7f11-46be-ac85-337bd1019cd1

[deepgram-community[bot]]

Thanks for asking your question. Please be sure to reply with as much detail as possible so the community can assist you efficiently.
_{Consider joining our Discord community for more opportunity to engage with your fellow Deepgram users. You can earn points which can be redeemed for cool stuff by being active in our communities!}

[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 deepgram product you are using (e.g Speech to Text, Agent API)
• A request ID that triggered your error or issue.

[deepgram-community[bot]]

I see multiple errors in our logs for this request:

Error: Deepgram received an empty file or url rather than valid audio data
Suggestion: The audio the customer sent was empty.

Error: A non-audio file was sent to Deepgram.
Suggestion: Check to see what audio file/data was sent.

Error: Query parameter encoding was set to linear32. linear32 is not a valid value for encoding; see the https://developers.deepgram.com/docs/encoding|docs for valid values.

Suggestion: Ask the customer to set encoding to <https://developers.deepgram.com/docs/encoding|one of these values> instead.

Error: The customer sent a base64 encoded audio file.
Suggestion: The customer needs to send the audio file as raw, unencoded binary data.

This message was sent by John Vajda from Deepgram, via our community automation.

[mj1618]

Ok that’s helpful to know what’s going on.

I still need some help I think.

So basically the encoding is erroring but hasn’t been an issue because most of the time I’m getting correct transcriptions back. This also means I’m sending the correct audio binary data and not base64.

Is it getting confused between the binary data and the JSON messages?

My use case is I only send audio data when the person is speaking, so in between audio data I need to send keepalives since it could be a minute or two between speaking.

If I have to remove the keepalives and also Finalize json messages that could be a problem for me.

Any suggestions are much appreciated thank you

[mj1618]

I can confirm that the parameter encoding: "linear32" does indeed work, I get correct speech transcriptions, and also it doesn't work without the parameter. So this parameter must be being used correctly even though it's not in the docs.

[mj1618]

In other words: my issue still exists

[mj1618]

Testing again I still get the exact same result with request ID 1c68be7a-dc73-45e1-9f56-a17ed9feba37

[jacobtang]

I also try to test the python demo，still have the issue: it seems to limit the duration to about 300 seconds, which causes the service connection to disconnect.

[mj1618]

Yep that’s the same amount of time for me