Setup for streaming

DeepFaceLive only provides a video window of the replaced face. Since the face must be processed, this takes time, so the video window shows the stream with some delay.

So, what do we need for streaming?

Capture window (or receive mpegts udp stream) and sound with some delay, transmit to streaming service (e.g. twitch, youtube, ...)

Below, one of the solutions.

Run OBS Studio

add Window capture

choose DeepFaceLive output window

add Audio Input Capture

select your microphone device

open Advanced Audio Properties for Audio Input Capture

set Sync Offset the same as Target Delay in Stream Output

DONE !

Now you can stream yourself to a stream service.

instead of Window capture

you can use Media Source

with configuration:

enable mpegts in Stream Output

1.8 KiB Raw Permalink Blame History