🎬 Studio42 Audio Mixer

Studio42 Audio Mixer is the pack's answer to "I have several audio files and I want them in one file, with volume, timing, fades, and maybe a limiter - all inside ComfyUI." It takes up to six AUDIO inputs and mixes them into one, with per-track volume, start-time offset, fade-in/out, and stereo pan, plus a proper master bus with normalization, compression, and limiting. It's the rare node in this suite that is genuinely what it claims to be: a mini mixing console as a node.

This is the node to reach for when you're building a video with a music bed and voiceover, or stitching sound effects onto a soundtrack. The full pipeline lives in the graph: load audio (Studio42 Audio Loader), mix it here, feed the result into a video or save step. The first track is the required backbone; tracks 2–6 are optional layers on top.

The inputs that matter

• audio_1 through audio_6 - the AUDIO inputs. Track 1 is required; the rest appear as you connect them.
• output_duration - total mix length in seconds (0.5–600). Tracks get placed/sliced against this.
• output_format - wav, mp3, or flac.
• sample_rate - 8000 to 192000; the tooltip spells it out: 44100 CD, 48000 pro, 96000 hi-res.
• Per-track: volume (1.0 = original, up to 5), start_time (seconds offset), fade_in/fade_out (up to 10s), and pan (−1 left, 0 center, 1 right) for tracks 2+.

The professional bus is where this gets interesting:

• master_volume - overall gain.
• normalization_mode - off, prevent_clipping, full_normalize, smart_normalize, or broadcast_standard. The default smart_normalize plus the limiter is the combo that stops you clipping.
• enable_compression with ratio (2:1 gentle to 10:1 heavy), threshold (dB), attack/release (ms) - a real dynamic-range compressor.
• enable_limiter (on by default) with limiter_threshold at −0.5 dB, which is why your mix won't brickwall.
• enable_crossfade with crossfade_duration for bleeding tracks into each other.
• target_loudness in LUFS if you're aiming at broadcast levels.

What comes out

The main output is mixed_audio (AUDIO), ready for a video mux or save node. It also exposes total_duration, sample_rate, and several STRING outputs (mix_info, level_analysis, track_info) that describe what it did - handy for logging and debugging.

Installing it

Same pack, 24oiduts-ComfyUI:

• ComfyUI Manager: search 24oiduts, install, restart.
• Manual:

  cd ComfyUI/custom_nodes/
  git clone https://github.com/GeekyGhost/24oiduts-ComfyUI
  pip install -r 24oiduts-ComfyUI/requirements.txt

  then restart.

This node needs torchaudio and soundfile (both in the pack's requirements) for actual audio work. The GitHub README is a stale Studio42 template that claims the project is "not recommended for use" with no license - ignore it; the code runs.

Troubleshooting

• Node won't load / no audio output - torchaudio or soundfile missing; install them in ComfyUI's venv and restart.
• Mix is quiet or clipping - leave smart_normalize on, keep the limiter on, and check target_loudness if you set a manual value.
• Tracks don't line up - set each track's start_time; the mix places them against output_duration, not end-to-end automatically.

For graph-native audio finishing without leaving ComfyUI, this is one of the more genuinely capable nodes in the suite.