ComfyUI_Geeky_AudioMixer

ComfyUI_Geeky_AudioMixer
★ 3

音频混合torchaudio内存高效实时处理
为 ComfyUI 提供专业多轨音频混合节点,基于 torchaudio 实现高质量、内存高效且支持实时处理,带音量与淡入淡出及调试日志。
💡 在 ComfyUI 中对多轨音频进行混合、音量与淡入淡出处理
🍴 3 Forks💻 Python🔄 2025-08-05
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/e58c8376a81b
📦 requirements.txt
#
ComfyUI
Geeky
AudioMixer
Node
Requirements
#
Core
audio
processing
soundfile>=0.12.1
numpy>=1.21.0
#
PyTorch
audio
for
high-quality
resampling
and
audio
processing
torchaudio>=0.13.0
#
Optional
but
recommended
for
additional
audio
processing
scipy>=1.9.0
#
JSON
handling
(usually
included
with
Python)
#
json
-
built-in
module
#
File
operations
(built-in
modules)
#
os,
tempfile,
time
-
built-in
modules
#
Note:
torch
is
already
available
in
ComfyUI
environment
📄 README

🎵 ComfyUI Geeky AudioMixer

https://github.com/user-attachments/assets/71881db4-22b9-4545-8647-4a51518bbefe

A professional-grade audio mixing node for ComfyUI that allows you to combine up to 4 audio tracks with precise volume control and full timing flexibility. Perfect for creating polished audio tracks for lip-sync videos, tutorials, podcasts, or any multimedia content.

✨ Features

🎤 Multi-Track Audio Mixing

  • 1 Required Track: Primary audio (voice, narration, main content)
  • 3 Optional Tracks: Background music, sound effects, additional audio
  • Native ComfyUI Integration: Uses ComfyUI’s audio input system directly
  • 🎛️ Professional Controls

  • 🔥 ACCURATE Volume Control: Individual track volumes are now precisely preserved – no more unexpected level changes!
  • Individual Volume Control: 0-500% range for each track with slider + number display
  • Precise Timing: Start time offset control (0-60 seconds) for each track
  • Fade Effects: Customizable fade in/out (0-5 seconds) for smooth transitions
  • Master Volume: Overall output level control (0-500%)
  • Pre-Gain Boost: Extra amplification (0.1x-10x) for very quiet audio sources
  • 🔊 Advanced Audio Processing

  • 🆕 Smart Normalization Modes: Choose how (or if) to normalize your audio
  • “prevent_clipping” *(default)*: Only reduces levels if clipping would occur
  • “off”: No normalization – preserves exact volume relationships
  • “full_normalize”: Traditional normalization to maximum level
  • “smart_normalize”: Only boosts very quiet signals
  • High-Quality Resampling: Automatic sample rate conversion with sinc interpolation
  • Dynamic Range Compression: Optional compression (1.0-10.0 ratio)
  • Soft Limiting: Prevents harsh clipping (-1dB default threshold)
  • Detailed Level Metering: Real-time RMS and peak level monitoring with debug output
  • 📊 Output Options

  • Multiple Formats: WAV, MP3, FLAC support
  • Sample Rates: 8kHz to 96kHz configurable
  • Stereo Processing: Automatic mono-to-stereo conversion
  • Timeline Mixing: Precise audio placement on timeline
  • 🚀 Installation

    Method 1: ComfyUI Manager (Recommended)

  • Open ComfyUI Manager
  • Search for “Geeky AudioMixer”
  • Click Install
  • Restart ComfyUI
  • Method 2: Manual Installation

  • Clone this repository into your ComfyUI custom_nodes folder:
  • “`bash

    cd ComfyUI/custom_nodes/

    git clone https://github.com/GeekyGhost/ComfyUI_Geeky_AudioMixer.git

    “`

  • Install Dependencies:
  • “`bash

    cd ComfyUI_Geeky_AudioMixer

    pip install -r requirements.txt

    “`

  • Restart ComfyUI and the node will appear in the audio/mixing category.
  • Requirements

  • torch (included with ComfyUI)
  • torchaudio>=0.13.0
  • soundfile>=0.12.1
  • numpy>=1.21.0
  • scipy>=1.9.0 (optional, for enhanced audio processing)
  • 📖 Usage Guide

    Basic Workflow Setup

    [TTS/Audio Generator] ──→ [🎵 Geeky AudioMixer] ──→ [Video Combine Node]
    [Music Loader] ──→ [audio_2 input]           │
    [SFX Loader] ──→ [audio_3 input]             │
    [SFX Loader] ──→ [audio_4 input]             ▼
                                        [Mixed Audio Output]

    Node Inputs

    Required Inputs

  • audio_1: Main audio track (voice, narration, primary content)
  • output_duration: Length of final mixed audio in seconds
  • output_format: Export format (WAV, MP3, FLAC)
  • sample_rate: Target sample rate (44100Hz recommended)
  • Optional Audio Inputs

  • audio_2: Secondary audio (background music, ambient sounds)
  • audio_3: Additional audio (sound effects, notifications)
  • audio_4: Extra audio (more sound effects, voice-overs)
  • Control Groups

    🎵 AUDIO 1 CONTROLS (Main Track)

  • Volume: 150% default (0-500% range)
  • Start Time: When track begins (0-60 seconds)
  • Fade In/Out: Smooth transitions (0-5 seconds)
  • 🎵 AUDIO 2 CONTROLS (Background)

  • Volume: 120% default (perfect for background music)
  • Start Time: Sync with main track
  • Fade In/Out: 1 second defaults for smooth music entry
  • 🎵 AUDIO 3/4 CONTROLS (Sound Effects)

  • Volume: 100% default (adjust as needed)
  • Start Time: Precise timing for sound effects
  • Fade In/Out: Quick or no fades for impact sounds
  • 🎛️ MASTER CONTROLS

  • Master Volume: 100% default *(changed from 200% for better control)*
  • 🆕 Normalization Mode: “prevent_clipping” default – Choose your normalization strategy
  • Pre-Gain Boost: 100% default (extra amplification if needed)
  • Compression: 1.0 default (no compression)
  • Limiter Threshold: -1dB default (soft limiting)
  • 🔥 Volume Control Fix

    The Problem (FIXED!)

    Previous versions applied normalization to the entire mixed output, which destroyed your carefully set volume relationships. If you set Audio 1 to 150% and Audio 2 to 15% (1/10th volume), they would end up at similar levels after normalization.

    The Solution

    Volume relationships are now preserved! The mixer processes each track with its individual volume setting, mixes them together additively, then applies master processing that respects your intended balance.

    New Normalization Modes

  • “prevent_clipping” *(recommended)*: Only reduces levels if they would clip – preserves your volume ratios
  • “off”: No normalization whatsoever – exact volume control
  • “full_normalize”: Old behavior – normalizes to maximum level
  • “smart_normalize”: Only boosts very quiet overall mixes
  • 🎯 Real-World Examples

    Example 1: Tutorial Video with Precise Volume Control

    # Voice narration (clear and prominent)
    audio_1_volume: 1.5        # 150% - main voice
    audio_1_start_time: 0.0    # Start immediately
    audio_1_fade_in: 0.2       # Quick fade in
    audio_1_fade_out: 0.5      # Gentle fade out
    
    # Background music (exactly 1/4 the volume of voice)
    audio_2_volume: 0.375      # 37.5% - exactly 1/4 of voice volume
    audio_2_start_time: 0.0    # Start with voice
    audio_2_fade_in: 2.0       # Slow fade in
    audio_2_fade_out: 3.0      # Slow fade out
    
    # Notification sound (1/2 the volume of voice)
    audio_3_volume: 0.75       # 75% - exactly 1/2 of voice volume
    audio_3_start_time: 5.2    # Play at 5.2 seconds
    audio_3_fade_in: 0.0       # Instant
    audio_3_fade_out: 0.1      # Quick fade
    
    # Master settings
    master_volume: 2.0         # 200% overall boost
    normalization_mode: "prevent_clipping"  # Preserve ratios
    output_duration: 30.0      # 30-second video

    Example 2: Podcast with Perfect Voice/Music Balance

    # Intro music (full volume initially)
    audio_1_volume: 2.0        # 200% - prominent intro
    audio_1_start_time: 0.0
    audio_1_fade_out: 2.0      # Fade as voice comes in
    
    # Voice narration (louder than music)
    audio_2_volume: 2.5        # 250% - 25% louder than music
    audio_2_start_time: 3.0    # Start at 3 seconds
    audio_2_fade_in: 1.0
    
    # Transition sound (moderate level)
    audio_3_volume: 1.0        # 100% - audible but not overwhelming
    audio_3_start_time: 10.0
    
    # Master settings - no normalization for exact control
    master_volume: 1.8         # 180% boost
    normalization_mode: "off"  # Preserve exact relationships

    Example 3: Gaming Content with SFX Hierarchy

    # Game audio/commentary
    audio_1_volume: 1.8        # 180% - main content
    audio_1_start_time: 0.0
    
    # Background music (much quieter)
    audio_2_volume: 0.36       # 36% - exactly 1/5 of main audio
    audio_2_start_time: 0.0
    
    # Achievement sound (prominent but not overwhelming)
    audio_3_volume: 1.35       # 135% - 3/4 of main audio level
    audio_3_start_time: 15.5
    
    # Victory sound (loudest effect)
    audio_4_volume: 2.25       # 225% - 25% louder than main
    audio_4_start_time: 28.0
    
    # Master settings
    master_volume: 1.5         # 150% overall
    normalization_mode: "smart_normalize"  # Boost if too quiet
    pre_gain_boost: 1.2        # 120% - slight extra boost

    🔧 Output Information

    mixed_audio

    The final mixed audio in ComfyUI’s native audio format, ready to connect to video nodes or audio export nodes.

    mix_info (JSON)

    Detailed information about the mixing process:

    {
      "tracks_loaded": 3,
      "processing_steps": [
        "Processed Audio 1 (main track)",
        "Processed Audio 2",
        "Mixed 3 tracks with preserved volume relationships",
        "Applied master volume: 1.5",
        "Normalization: prevent_clipping - no scaling needed"
      ],
      "sample_rate": 44100,
      "duration": 10.0,
      "channels": 2,
      "master_volume": 1.5,
      "normalization_mode": "prevent_clipping",
      "final_levels": {
        "rms": 0.234567,
        "peak": 0.890000,
        "rms_db": -12.6,
        "peak_db": -1.0
      }
    }

    total_duration

    Actual duration of the mixed audio in seconds.

    level_meters (JSON)

    Real-time audio level information:

    {
      "rms_left_db": -12.3,
      "rms_right_db": -12.1,
      "peak_left_db": -1.0,
      "peak_right_db": -0.9,
      "stereo_balance": "centered"
    }

    🎵 Pro Tips for Best Results

    Volume Balancing (Now Actually Works!)

  • Voice/Main Content: 120-200% for clarity
  • Background Music: 30-60% of main content volume for proper background level
  • Sound Effects: 80-150% depending on impact needed
  • Master Volume: 100-200% for final output level
  • Use ratios: If voice is 150%, set music to 37.5% for exactly 1/4 volume
  • Normalization Mode Selection

  • “prevent_clipping”: Best for most use cases – preserves your ratios
  • “off”: Use when you need exact volume control for professional mixing
  • “full_normalize”: Use when you want maximum loudness regardless of ratios
  • “smart_normalize”: Good for content with very quiet overall levels
  • Timing and Fades

  • Voice: Short fades (0.1-0.5s) for natural speech
  • Music: Longer fades (1-3s) for smooth transitions
  • Sound Effects: Quick or no fades for maximum impact
  • Staggered Timing: Offset tracks by 0.1-0.5s for natural feel
  • Audio Quality Settings

  • Sample Rate: 44.1kHz for most content, 48kHz for professional video
  • Format: WAV for editing workflows, MP3 for final delivery
  • Normalization: Use “prevent_clipping” for best balance
  • Pre-Gain Boost: Use 1.2-2.0x for slightly quiet sources, 3.0+ for very quiet sources
  • Common Workflow Patterns

  • Voice + Music: Voice at 150%, music at 37.5% (1/4 ratio), master at 150%
  • Podcast: Voice at 200%, intro music at 120%, master at 130%
  • Gaming: Game audio at 180%, background music at 36% (1/5 ratio), effects at 135%
  • Tutorial: Voice at 150%, music at 30%, notification SFX at 75%
  • 🐛 Troubleshooting

    Volume Relationships Not Working

    Check these settings:

  • Use normalization_mode: "prevent_clipping" or "off"
  • Avoid "full_normalize" if you need precise ratios
  • Check console output for actual applied volumes and levels
  • Verify master_volume isn’t set too high causing clipping
  • Audio Too Quiet

    Solutions:

  • Increase master_volume to 1.5-3.0
  • Use pre_gain_boost of 1.5-3.0 for quiet sources
  • Raise individual track volumes proportionally
  • Try normalization_mode: "smart_normalize" for very quiet mixes
  • Audio Distorted/Clipping

    Solutions:

  • Use normalization_mode: "prevent_clipping" (default)
  • Reduce master_volume below 1.5
  • Lower individual track volumes proportionally
  • Increase limiter_threshold to -3dB or lower
  • Enable compression with ratio 2.0-4.0
  • Volume Ratios Still Wrong

    Debug steps:

  • Check ComfyUI console for detailed level information
  • Look for lines like “Applied volume: 1.5x” and “Final RMS: 0.123456”
  • Verify normalization mode is not “full_normalize”
  • Test with normalization_mode: "off" and master_volume: 1.0 for pure testing
  • Tracks Not Audible

    Solutions:

  • Check audio connections in ComfyUI workflow
  • Verify audio file formats are supported
  • Increase track volume above 100%
  • Check start times aren’t beyond output duration
  • Look at console output for “Track outside timeline bounds” warnings
  • Sample Rate Issues

    Solutions:

  • Use 44100Hz for most content
  • Check console output for resampling warnings
  • Ensure all input audio has valid sample rates
  • Look for “Resampling from XHz to YHz” in console
  • LazyAudioMap Errors

    Solutions:

  • Update ComfyUI to latest version
  • Restart ComfyUI after installing node
  • Check that audio input nodes are compatible
  • Try connecting audio through different nodes
  • 📊 Technical Details

    Supported Audio Formats

  • Input: Any format supported by ComfyUI audio nodes
  • Output: WAV (16/24-bit), MP3, FLAC
  • Channels: Automatic mono-to-stereo conversion
  • Sample Rates: 8kHz to 96kHz with high-quality resampling
  • Audio Processing Pipeline

  • Audio Extraction: Handles ComfyUI’s LazyAudioMap format
  • Format Conversion: Ensures consistent tensor format [channels, samples]
  • Resampling: High-quality sinc interpolation when needed
  • Volume Application: Per-track volume adjustment *(preserves relationships!)*
  • Fade Processing: Smooth fade in/out curves
  • Timeline Mixing: Sample-accurate additive mixing *(maintains ratios!)*
  • Master Volume: Applied to entire mix
  • Smart Normalization: Applied based on selected mode
  • Compression/Limiting: Optional final processing
  • Output Formatting: ComfyUI-compatible audio format
  • New Volume Preservation System

  • Individual Processing: Each track processed with its volume setting
  • Additive Mixing: Tracks combined while preserving relative levels
  • Intelligent Normalization: Only applied when beneficial
  • Detailed Logging: Console shows exact levels at each processing stage
  • Master Control: Final volume adjustment maintains relationships
  • Performance Notes

  • Memory Efficient: Processes tracks sequentially
  • High Quality: Uses torchaudio for professional audio processing
  • Real-time Capable: Optimized for workflow performance
  • Debug Information: Detailed console logging shows exact processing steps
  • 🤝 Contributing

    We welcome contributions! Here’s how you can help:

    Reporting Issues

  • Check existing issues
  • Provide detailed information:
  • ComfyUI version
  • Audio node versions you’re using
  • Error messages from console
  • Steps to reproduce
  • Include console output showing volume levels
  • Feature Requests

  • Open an issue with “Feature Request” label
  • Describe the feature and use case
  • Explain how it would benefit users
  • Pull Requests

  • Fork the repository
  • Create a feature branch: git checkout -b feature-name
  • Make your changes
  • Add tests if applicable
  • Submit a pull request with clear description
  • Development Setup

    git clone https://github.com/GeekyGhost/ComfyUI_Geeky_AudioMixer.git
    cd ComfyUI_Geeky_AudioMixer
    pip install -r requirements.txt

    📄 License

    This project is licensed under the MIT License – see the LICENSE file for details.

    🙏 Acknowledgments

  • ComfyUI Team: For the amazing node-based interface
  • ComfyUI Community: For inspiration and feedback on the volume control fix
  • Audio Processing Libraries: PyTorch Audio team for excellent tools
  • 📋 Changelog

    v1.1.0 – Volume Control Fix

  • 🔥 FIXED: Volume relationships now precisely preserved during mixing
  • 🆕 NEW: Smart normalization modes (prevent_clipping, off, full_normalize, smart_normalize)
  • 🔧 IMPROVED: Detailed console logging showing exact levels at each processing stage
  • ⚙️ CHANGED: Master volume default changed from 200% to 100% for better control
  • 🐛 FIXED: Normalization no longer destroys intended volume ratios
  • v1.0.0 – Initial Release

  • Multi-track audio mixing with up to 4 tracks
  • Individual volume, timing, and fade controls
  • Professional audio processing pipeline
  • ComfyUI native integration

  • Made with ❤️ for the ComfyUI community

    If this node helped you create amazing content, consider giving us a ⭐ on GitHub!