Lipsycn constantly stopping while talking

Reply by jdiego8d | 2024-08-13 20:56:58

The lipsycn its just annoying, but
The voice problem is blocking all the user experience…
I am asking for the voice problem

Reply by k3kalinix | 2024-08-13 20:57:23

When is your meeting?

Reply by freezfast | 2024-08-13 20:57:45

You can remove the lipsync component this will fix th voice but will stop lipsync until a fix is made

Replying to k3kalinix’s Message

Reply by k3kalinix | 2024-08-13 20:57:23
When is your meeting?

Reply by jdiego8d | 2024-08-13 21:08:25

In 14h

Reply by k3kalinix | 2024-08-13 23:31:58

Hello <@592163175189970944>,

We tested on our side.
It is working.

Reply by jdiego8d | 2024-08-13 23:37:50

Did you made any changes ?
I will test it in a moment

Reply by jdiego8d | 2024-08-14 00:13:19

I have tested it for a few mins now…
It is more stable, but this error still appear sometimes

OnStreamFinish: Status:Not Ok | Debug Log: | Error message:Unable to generate audio response. Please verify that you have a valid voice set for the character. | Error Details: | Error Code:2 | Character ID:e4814288-03da-11ef-9539-42010a7be00e | Session ID:b9f52296f8fd810e37355e6ac95bf483
ConvaiChatbotComponentLog: Warning: UConvaiChatbotComponent Get Response Failed! | Character ID : e4814288-03da-11ef-9539-42010a7be00e | Session ID : b9f52296f8fd810e37355e6ac95bf483

Reply by jdiego8d | 2024-08-14 00:14:31

Do you think we have something wrong with the language selecion?
we are testing with diferent voices but all of them give us the same error

  • Please verify that you have a valid voice set for the character.

Replying to jdiego8d’s Message

Reply by jdiego8d | 2024-08-13 23:37:50
Did you made any changes ?
I will test it in a moment

Reply by jdiego8d | 2024-08-14 12:53:18

can you please answer?
Our QA team do not found any improvement
I dont know if it related to time, because last night in my own test it was working
but today, again we are only seeing error and mesages about voice set

Reply by jdiego8d | 2024-08-14 13:05:29

Today we are also getting this problem

ConvaiGRPCLog: Warning: OnStreamFinish: Status:Not Ok | Debug Log: | Error message:No status received | Error Details: | Error Code:2 | Character ID:e4814288-03da-11ef-9539-42010a7be00e | Session ID:b1b20cd698b61779c794236197df6f73
ConvaiChatbotComponentLog: Warning: UConvaiChatbotComponent Get Response Failed! | Character ID : e4814288-03da-11ef-9539-42010a7be00e | Session ID : b1b20cd698b61779c794236197df6f73

Reply by jdiego8d | 2024-08-14 13:09:01

Any thoughs on this?
We dont understand the message: Error message:No status received.

Reply by k3kalinix | 2024-08-14 13:59:20

My teammate is checking.

Reply by jdiego8d | 2024-08-14 21:14:35

Sorry to ask again, but did your teammate find anything? we are still on throubles with our clients :sweat:

Reply by jeeves901 | 2024-08-15 03:41:28

Hi <@592163175189970944> Is the API working for you now? I think the issue was that the pronunciation of the word White added in character pronunciations had some invalid characters, which was causing an issue with our Text to Speech service.

We will take some steps to return better errors here + do better validation upfront to not let this happen.

Reply by k3kalinix | 2024-08-15 20:59:58

<@592163175189970944>

Reply by jdiego8d | 2024-08-15 21:11:57

Yes, Thank you; Our team also found that, we already change the pronunciation of that word
It is working fine, we also change the Foundation model from Claude 3-5 sonnet to Gpt 3.5
we do that thiking it will work better but, it constantly changing again to Claude
Do you know why this is happening? why cant we select one, without comming back to claude.

Reply by k3kalinix | 2024-08-25 10:19:12

Could you please test again?

Reply by jdiego8d | 2024-08-26 17:32:58

Hi, yes I will be doing some test in a few hours
Thank you

Reply by jdiego8d | 2024-08-27 14:49:55

<@1023671043287699568>
I have been testing it, it works well but, there is something that is happening even on the web platform.

It takes 5 to 15 secs to have any answer after triggering one narrative design trigger, we check it on the build from unreal engine, but as I told you the extreme delay is also happening on the web, via the chat box in character section.

Do you know what could be happening, is something related to our narrative design flows or a general bug from the AI?

Reply by k3kalinix | 2024-08-27 14:50:42

Hello <@592163175189970944>,