ComfyUI > Nodes > Audio Separation (Demix) > Various using MDX

ComfyUI Node: Various using MDX

Class Name

AudioSeparateVarious

Category
audio/separation
Author
set-soft (Account age: 3460days)
Extension
Audio Separation (Demix)
Latest Updated
2026-02-11
Github Stars
0.02K

How to Install Audio Separation (Demix)

Install this extension via the ComfyUI Manager by searching for Audio Separation (Demix)
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter Audio Separation (Demix) in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

Various using MDX Description

Separates "Other" and "Reverb" audio elements using MDX-Net for precise audio processing.

Various using MDX:

The AudioSeparateVarious node is designed to separate miscellaneous audio components, specifically targeting elements like "Other" and "Reverb" using MDX-Net networks. This node is particularly useful for audio processing tasks where you need to isolate or manipulate these specific audio elements from a mix. By leveraging advanced machine learning models, it provides a sophisticated method for audio separation, allowing you to extract these components with high precision. This can be especially beneficial in creative audio projects where you want to enhance or modify certain aspects of a track without affecting the entire mix. The node's ability to focus on these less common audio elements makes it a versatile tool in the audio separation toolkit.

Various using MDX Input Parameters:

input_sound

This parameter accepts the audio input that you wish to process. It is the primary audio file from which the node will attempt to separate the "Other" and "Reverb" components. The quality and characteristics of the input sound can significantly impact the separation results, so using a high-quality audio file is recommended.

model

This parameter allows you to select the model used for audio separation. The available models are pre-trained and optimized for different types of audio separation tasks. Choosing the right model can affect the accuracy and quality of the separation, with options tailored to specific audio characteristics.

shifts

This integer parameter controls the number of random shifts for equivariant stabilization, with a default value of 0, a minimum of 0, and a maximum of 16. Increasing the number of shifts can improve the quality of the separation by providing more robust results, but it will also increase processing time. A value of 0 disables this feature.

overlap

This float parameter determines the amount of overlap between audio chunks, with a default value of 0.25, a minimum of 0.0, and a maximum of 0.99. Higher overlap values can reduce stitching artifacts in the separated audio but will slow down the processing. Adjusting this parameter can help achieve smoother transitions between audio segments.

custom_segment

This boolean parameter allows you to enable or disable the custom segment length feature. When enabled, you can override the model's default segment length, which can be useful for certain models like HDemucs and Demucs. Disabling this feature uses the recommended segment length from the model file, which is generally suitable for most tasks.

Various using MDX Output Parameters:

Main

This output parameter provides the separated "Other" and "Reverb" components from the input audio. It represents the primary result of the separation process, allowing you to work with these isolated elements in your audio projects.

Complement

This output parameter contains the remaining audio components after the "Other" and "Reverb" have been separated. It essentially represents the inverse of the "Main" output, providing a complete picture of the audio separation process by showing what was left behind.

Various using MDX Usage Tips:

  • Experiment with different models to find the one that best suits your audio material, as different models may yield varying results depending on the characteristics of the input sound.
  • Adjust the overlap parameter to find a balance between processing speed and audio quality, especially if you notice artifacts at the boundaries of separated audio chunks.
  • Use the custom segment feature to fine-tune the separation process for specific models, particularly if you are working with complex audio tracks that require precise separation.

Various using MDX Common Errors and Solutions:

Error during separation: <specific_error_message>

  • Explanation: This error may occur if there is an issue with the input audio file or the selected model.
  • Solution: Ensure that the input audio file is in a supported format and that the correct model is selected. Check for any updates or compatibility issues with the model files.

Model not found: <model_name>

  • Explanation: This error indicates that the specified model file could not be located or loaded.
  • Solution: Verify that the model file is present in the expected directory and that the file name is correct. Ensure that the model file is not corrupted and is compatible with the node.

Various using MDX Related Nodes

Go back to the extension to check out more related nodes.
Audio Separation (Demix)
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.