Any Audio Input

Speech To Life supports audio fed from a Submix as well as AudioGenerator sources. With a submix you can pass in audio from an Audio Component or Audio Capture component to get recognition results. A SoundWave asset could contain vocals and those could be passed into Speech To Life to generate subtitles, for example. Any audio source that can be pumped through a Submix is supported.

A very simple example is using an Audio Capture component (which captures microphone audio) and setting a submix.

Create a new submix, call it "VoiceRecognition":

Add an Audio Capture component to your actor:

In the audio capture component details set the "Base Submix" to your created submix:

In the speech-to-life component set Submix to your created submix:

When you want to pull audio from the audio capture component call "Start" on it:

And now if your Speech To Life component is active and a locale has been set the audio being pumped into Speech to Life through the submix will be recognized and results will be generated.

If you would like to mute the submix so audio only ends up recognized and not played through the audio component you can add a nullifying submix after your submix like so:

A created null submix with output volume of 0:

The voice recognition submix output to nullsubmix input: