Audio Input
Capture live audio from microphone or system input.
What it does
Audio Input captures live sound from a microphone or another system input and outputs a raw waveform, an FFT spectrum, an overall volume level, and three energy bands: bass, mid and treble. It has no image output.
Buffer Size and FFT Size control how much audio is analysed and how fine the frequency detail is. Smoothing averages the spectrum over time. Window shapes the FFT analysis window: Hann is the general default, Blackman-Harris trades some responsiveness for a cleaner, more precise spectrum.
When to use it
Visuals, motion or parameters reacting to live sound: music, voice, room ambience
A bass/mid/treble split for audio-reactive rigs without hand-built band-splitting logic
Monitoring the captured source, with Play Audio enabled
| Parameter | Type | Default |
|---|---|---|
buffer_size | Number | 2048 |
fft_size | Number | 512 |
smoothing | Number | 0.800 |
window | String | "Hann" |
play_audio | Boolean | false |
monitor_volume | Number | 0.800 |
Gotchas
Raising FFT Size gives finer frequency resolution at a higher cost per frame. Bass/mid/treble energy alone rarely needs more than the default.
Smoothing is a temporal average. Raise it for a steadier read of a jittery spectrum, at the cost of responsiveness.
Play Audio sends the captured signal straight to your speakers. A microphone input with speakers nearby can feed back. Leave it off unless you are actively monitoring, or use headphones.
Worked example
Add an Audio Input node and pick up a microphone or line-in.
Wire Volume into a scale or brightness parameter elsewhere in the graph for an audio-reactive pulse.
Wire Bass, Mid, and Treble into separate parameters for a band-split reactive look.
If the reaction feels noisy, raise Smoothing; if it feels sluggish, lower it.
Enable Play Audio briefly to confirm the source being captured, then turn it back off.
See also
Audio File · Audio Output · MIDI Input · Numbers, signals & audio