aspect-ratio-resizer

Image Resizing Nodes - ComfyUI Custom Nodes

Two powerful ComfyUI custom nodes for intelligent image resizing while maintaining aspect ratio. Perfect for handling images of any size with precision and control.

Nodes Included

1. Aspect Ratio Resizer

Precise control over image dimensions with flexible width/height constraints.

2. Auto Megapixel Reducer

Automatically reduces high-resolution images to a target megapixel count.

Features

Aspect Ratio Resizer

• Flexible Constraints: Enable/disable width or height constraints independently
• Single Dimension Resizing: Use only width OR only height to calculate the other dimension
• Smart Scaling: Only resizes when necessary based on your specified constraints
• Multiple Resize Modes: Choose between downscale only, upscale only, or both

Auto Megapixel Reducer

• Automatic Detection: Detects high-resolution images (4K, 8K, etc.)
• Megapixel Targeting: Reduces to specific megapixel count (e.g., 2MP, 5MP)
• Safe Mode: Only reduces size, never upscales small images
• Perfect for Large Images: Handles any resolution while maintaining aspect ratio

Both Nodes

• Aspect Ratio Preservation: Always maintains the original aspect ratio
• Various Interpolation Methods: Support for different interpolation algorithms
• Batch Processing: Handles multiple images in a single operation
• Real-time Feedback: Shows detailed resize information in the console

Installation

1. Copy this entire folder to your ComfyUI custom_nodes directory:

   ComfyUI/custom_nodes/aspect-ratio-resizer/
   

2. Restart ComfyUI

3. Both nodes will appear in the "image/resize" category when you right-click to add a new node

Usage

Aspect Ratio Resizer - Inputs

• images: Input image(s) to resize
• enable_width: Enable/disable width constraint (default: True)
• max_width: Maximum allowed width (default: 1024, range: 64-8192)
• enable_height: Enable/disable height constraint (default: True)
• max_height: Maximum allowed height (default: 1024, range: 64-8192)
• resize_mode:
  • downscale_only: Only resize if image is larger than max dimensions
  • upscale_only: Only resize if image is smaller than max dimensions
  • both: Resize to fit max dimensions regardless of original size
• interpolation: Interpolation method for resizing
  • bilinear: Good balance of quality and speed (default)
  • bicubic: Higher quality, slower
  • nearest: Fastest, pixelated look
  • linear, trilinear, area: Other options for specific use cases

Auto Megapixel Reducer - Inputs

• images: Input image(s) to resize
• max_megapixels: Target maximum megapixels (default: 2.1, range: 0.1-50.0)
• only_reduce: Enable/disable upscaling protection (default: True)
  • True: Only reduce size, never upscale (safe mode)
  • False: Allow upscaling if image is below target megapixels
• interpolation: Interpolation method for resizing (same options as above)

Outputs

• resized_images: The resized image(s) maintaining aspect ratio

How It Works

1. Calculates Scale Factor: Determines the scaling needed to fit within max dimensions
2. Preserves Aspect Ratio: Uses the smaller of width_scale or height_scale to maintain proportions
3. Applies Resize Mode: Only resizes based on your selected mode
4. Ensures Even Dimensions: Rounds to even numbers for better compatibility
5. Provides Feedback: Shows original and new dimensions in the console

Examples

Example 1: Width-Only Constraint

• Input: 1920x1080 image
• Enable Width: True, Max Width: 1024
• Enable Height: False
• Mode: downscale_only
• Result: 1024x576 (maintains 16:9 aspect ratio, height calculated automatically)

Example 2: Height-Only Constraint

• Input: 1920x1080 image
• Enable Width: False
• Enable Height: True, Max Height: 1024
• Mode: downscale_only
• Result: 1820x1024 (maintains 16:9 aspect ratio, width calculated automatically)

Example 3: Both Constraints (Traditional Mode)

• Input: 2048x1536 image
• Enable Width: True, Max Width: 1024
• Enable Height: True, Max Height: 1024
• Mode: downscale_only
• Result: 1024x768 (maintains 4:3 aspect ratio, uses most restrictive constraint)

Example 4: No Constraints

• Input: Any image
• Enable Width: False, Enable Height: False
• Result: Unchanged (no resizing performed)

Auto Megapixel Reducer Examples

Example 1: 4K to 2MP

• Input: 3840x2160 image (8.3MP)
• Max Megapixels: 2.0
• Result: 1884x1060 (2.0MP, maintains 16:9 aspect ratio)

Example 2: 8K to 2MP

• Input: 7680x4320 image (33.2MP)
• Max Megapixels: 2.0
• Result: 1884x1060 (2.0MP, maintains 16:9 aspect ratio)

Example 3: Small Image (Safe Mode)

• Input: 1024x768 image (0.8MP)
• Max Megapixels: 2.0, Only Reduce: True
• Result: 1024x768 (unchanged, already below target)

Example 4: Portrait 4K

• Input: 2160x3840 image (8.3MP, portrait)
• Max Megapixels: 2.0
• Result: 1060x1884 (2.0MP, maintains aspect ratio)

Technical Details

• Uses PyTorch's F.interpolate for high-quality resizing
• Handles batch processing efficiently
• Maintains ComfyUI's image tensor format [B, H, W, C]
• Provides detailed logging for debugging

Troubleshooting

• Node doesn't appear: Make sure you've restarted ComfyUI after installation
• Images look distorted: Check that aspect ratio is being maintained (should never happen with this node)
• Performance issues: Try using 'nearest' interpolation for faster processing
• Dimension errors: Ensure max_width and max_height are reasonable values (64-8192)

License

This custom node is provided as-is for educational and practical use in ComfyUI workflows.