Skip to main content
Audio Mix

Audio Mix · screenshot coming soon

​
What this node does

Audio Mix sums up to four COMFYTV_AUDIO sources (A, B, C, D) into a single stereo track. Each channel has independent gain and pan, so you can balance a voiceover against music and sound effects in one node. At least one source must be wired; empty inputs are skipped. It outputs one COMFYTV_AUDIO snapshot and has a ▶ Run (FFmpeg does the mix). To feed native ComfyUI AUDIO in or out, insert a Bridge — see bridges.md.

​
When to use it

  • Combine narration, a music bed, and SFX into a finished mix.
  • Set relative levels between tracks and spread them across the stereo field.
  • Re-assemble stems (from Audio Stem Split) after processing them individually.

​
Parameters

​
pan_law

How panning affects loudness across the stereo field: audacity (default) or constant_power. constant_power keeps perceived loudness steadier as a source moves left/right; audacity matches Audacity’s convention.

​
dither

Dithering applied when reducing bit depth: none (default), tpdf, or shaped. Add dither (tpdf is a safe general choice) when exporting to a lower bit depth to avoid quantization artifacts.

​
gain_a / gain_b / gain_c / gain_d

Level for each source in decibels. Range -60 to +12 dB, default 0 (unchanged). Set unwired channels are ignored.

​
pan_a / pan_b / pan_c / pan_d

Stereo position for each source. Range -1 to +1, default 0 (center). -1 = hard left, +1 = hard right.

​
Outputs

​
Tips

  • Watch the summed level: several tracks at 0 dB can clip. Pull individual gains down a few dB if the mix distorts.
  • Only wired inputs count — you can use just two or three sources without touching the others.
  • Audio Duck — automatically dip one track under another instead of setting fixed levels.
  • Audio Stem Split — separates a mix into stems that you can reprocess and re-mix here.
  • Audio Crossfade — join two clips in sequence rather than layering them.