LC Audio Volume π
The gain slider that keeps your soundtrack from clipping the mix
- audio
- AUDIO
Every soundtrack needs a volume control, and this is it - a single slider on an AUDIO wire. Generated audio, especially, comes out at whatever level the model felt like, and when you mux it under a video you quickly find out whether that level was "barely audible" or "blowing out the speakers." LC Audio Volume lets you set the level explicitly before it ever reaches the encoder.
The design philosophy here is honest simplicity: it's one volume slider with one job, so you can drop it inline without thinking. That's the right call for a utility node - you want gain on a wire the way you want brightness on an image, a single understood adjustment, not a panel.
How it works
volume is a linear gain multiplier on the waveform: 0 = muted, 1 = unchanged, 2 = double amplitude, which is +6 dB in the way audio people count. After applying the gain, the node soft-clips the samples to the Β±1.0 range Comfy audio lives in, so a boosted signal saturates gently instead of spewing out-of-range garbage that distorts unpredictably at encode time.
Two implementation details worth knowing:
- Setting
volumeto exactly1.0makes the node a pass-through - it returns your audio untouched, no processing. Leave it at 1 while you're building, then dial it at the end. - Because the gain is linear rather than in dB, the slider doesn't feel like a mixing desk. Don't fight it: think of
2as "twice as loud,"0.5as "half," and nudge from there.
The input that matters
Just audio (the AUDIO to adjust) and volume (0β4, default 1, step 0.01). One output, AUDIO. There's genuinely nothing else to configure, which is the point.
Where it shines: after an LC Audio Separate stem you don't want at full blast, or before LC Save Video's audio mux when a generated track is clipping at unity. Want a ducked or sidechained feel? Not this node - that needs an envelope or mixing, which is a much bigger tool. This is the "make it quieter / a bit louder" knob.
How to install it
Ships in lonecatone23/ComfyUI_LC_AV_nodes, the LC Audio_Video pack (MIT, by lonecatone23, sibling to the image-side ComfyUI_LC123_nodes - the two repos stay separate):
- ComfyUI Manager: search "LC Audio_Video" or
ComfyUI_LC_AV_nodes. - Manual:
cd ComfyUI/custom_nodes && git clone https://github.com/lonecatone23/ComfyUI_LC_AV_nodes, restart ComfyUI (expect[LC AV] total 16 nodeson the console).
No downloads or extra packages needed beyond ComfyUI's built-in Python.
The trap
Soft clipping is not a limiter. If your source is already peaking at full scale and you push volume past 1, you're asking the node to amplify a signal that has no headroom, and the soft clip is just polite distortion - the waveform flattens. The fix is to lower the source level first (gain staging) rather than expecting this slider to rescue an already-hot mix. If the track was generated way too loud, pull it down, don't push more gain in somewhere upstream.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| audio | AUDIO | β | |
| volume | FLOAT | 1.000β4 | Linear gain. 0 = mute, 1 = unchanged, 2 = +6 dB. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| AUDIO | AUDIO | β |