Audio2Face long delay on streaming

AI & Data Science Digital Humans (closed) Audio2Face (closed)
1

Hi! I got the streaming player working in a2f using the demo scene and the sample python script. It works… but the audio/face only starts playing after 1 minute and 20 seconds… I thought it’d be near-real-time. I tried with different audio wav files (stereo, mono, 44.1, 22, 16 kHz, including the demo files, as shown on the tutorial video) – it makes no difference. Can anyone help me troubleshoot the problem?

In the python terminal I see:

Channel created
Sending audio data…
(here we wait for about 1 min 20 sec for the audio to play and the face to move)
SUCCESS
Channel closed

My specs:
i7 + 32GB RAM / RTX 3080 + 10GB VRAM / Linux Mint 21.1

2

Welcome to the forum @Asters

Audio2Face currently supports Ubuntu 20 and above and CentOS 7

Release Notes — Omniverse Audio2Face documentation (nvidia.com)

3

Hi, thank you for answering. Mint is based on Ubuntu 22.04, so it should be compatible. Do you have any other ideas what might be causing such a long delay? This is not normal, right? It should go faster?

4

That’s correct. It should be super fast. I’ll bring this up to the engineering team. In the meantime, if possible, it would be helpful if you can try it on a different OS.

5

One more thing, can you please provide a link to the video tutorial you’re following?

6

I’m following this tutorial:

The only difference is that I’m using a2f 2022.2.0, so some things, like the graphs, are in different places, but the demos are still there and that’s what I’m running.

It would be difficult for me to try it on a different OS, since I don’t have access to another computer with an RTX graphics card right now… Also, the product we’re building will run on Linux, so I don’t even want to try on Windows because it won’t be solving the problem for us. Let’s hope the engineering team has some ideas what I can do…

7

I just tested on Windows and there’s no delay at all.

I ran this code after installing all necessary libraries:

python "C:/Users/{USER}/AppData/Local/ov/pkg/prod-audio2face-2022.2.1/exts/omni.audio2face.player/omni/audio2face/player/scripts/streaming_server/test_client.py" "C:/audio/some_audio.wav" "/World/audio2face/PlayerStreaming"

I’ll update you after another colleague will test this on Linux.

8

Thank you for your help, upgrading to 2022.2.1 seems to have fixed the problem. I can now stream in real time from python.

9

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.