Hi Team, The spoken audio is too low in volume, Can you please advise how we can boost the volume of the spoken ausio of the digital human?
Also the Lip Sync issue is still there. the audio and lip movement don’t wok in sync all the time, or the digital human keeps talking and the lips stop moving. Or he will stop speaking midway through a statement abruptly.
Is there a fix for this issue Please. 
Hello @Jeasy_Sehgal,
Could you please share the Character ID and a screen recording where we can observe the current voice volume issue? This will help us better understand and investigate the problem.
Hi @K3, The character id is: 60e087de-76af-11ef-80ce-42010a7be011, in the playground the volume seems alright, but in Unreal Engine, the audio volume is lower.
We have to turn the device (Computer) volume to the maximum to hear the digital human clearly.
The Lip Sync issue is still there. The audio and lip movement don’t work in sync all the time, each session, while the digital human is talking and the lips stop moving or he will stop speaking midway through a statement abruptly. This is happening abruptly and randomly. There is no way to record this behaviour as it not consistent.
Tomorrow, Friday 18 July, I am presenting a live demonstration virtually to a symposium in Germany. I am not sure why this is happening. Your team is already aware of it because of multiple tickets about this issue already from different users, what is the solution, is there a hotfix possibly being worked on?
Thanks for providing the details.
We couldn’t detect any issues with the audio, everything seems fine on the Convai side. We’d recommend double-checking your Unreal Engine project settings to ensure there are no device-level or engine-specific audio dampening issues.
Also, if you’re running on Unreal Engine 5.6 with Convai SDK version 3.6.x, please note that these are the most recent versions and still stabilizing. If you’re preparing for a high-stakes live demo like tomorrow’s presentation, we strongly recommend using a more stable version of the SDK.
As for the lip sync and audio cut-off issues, yes, we’re aware of this inconsistency, and our team is actively working on a fix. There’s no need to submit multiple tickets; it’s on our radar and will be addressed in an upcoming update.
We appreciate your patience and hope your symposium presentation goes smoothly!
Thanks @K3 I appreciate the update. Waiting patiently.
@K3 I have, after extensive testing for a few weeks now, found that the Charon voice from GCP Voices paired with GOOGLE Gemini Pro work the smoothest, better lip sync, less audio drop, however, he speaks the emotions and inner thoughts. for example:
(Extending a hand with a warm smile) Greetings. I am Light-Arch, a digital human built from the creative legacy of Jeasy Sehgal. It’s a pleasure to meet you. May I ask your name?
He actually says the words in the brackets, is there anything that can be done to stop this from happening? It’s similar to the issue everyone was having with the ** being spoken or " quotations being spoken.
For more information, this is the session ID: aca089ced285cec4267c44a734cf1f9e. It’s 12:25 AM at night here. I am heading to bed, Got the Symposium presentation remotely at 08:30 AM this morning.
I’d so appreciate it if this little issue can be fixed. Which foundational CORE AI Model is the most stable at the moment to use anyway. Your recommendation can help us save face for crucial demonstrations and investor relations for substantial investments and grants.
Thank you!
Hey @Jeasy_Sehgal,
You can try out gpt-4o perhaps it tends to avoid such extra gestures/emotions, and maybe in your backstory try adding things like “dont speak your emotions or anything in brackets () or ** ** during the response, no gestures only verbal speech should be given in the response tags”
This is the best advice i can give you currently with the bound time period we have. Hope you have a great presentation, best of luck !!
Thanks @Kamal, I’ve added the prompt you suggested and going live in 10 minutes. Fingers crossed. Talk soon!
Hi @Kamal @K3 thanks for the help,
The symposium went well, had a number of glitches but in the end, it worked out alright. We’ve been invited to apply for the John Templeton Grant.
I’m trying to figure out the Long Term Memory and used the Manage Speaker ID widget in the Project Settings > Plugins>Convai. Created a user Id, but it created a second one by mistake. I can’t delete them now. How do I delete the User IDs?. I try and get this
Please create a new post for this issue.
Sure @K3 Will do
