Skip to content

Latest commit

 

History

History
139 lines (110 loc) · 3.9 KB

GSL_Filter_V1.md

File metadata and controls

139 lines (110 loc) · 3.9 KB

GSL Filter v1 Node

A GPU-accelerated image processing node that applies various shader-based effects to images. This node utilizes OpenGL (via ModernGL) to perform high-performance image filtering and effects processing using GPU shaders.

Input Parameters

Required Inputs

  1. images (IMAGE)

    • The input batch of images to process

  2. effect_preset (COMBO) Options:

    • custom: No effect (bypass)
    • grayscale: Converts image to grayscale
    • edge_detection: Highlights edges in the image
    • gaussian_blur: Applies Gaussian blur effect
    • pixelate: Creates a pixelated effect
    • wave_distortion: Applies wave-based distortion
    • chromatic_aberration: Simulates lens color separation

  3. intensity (FLOAT)

    • Default: 1.0
    • Range: 0.0 to 5.0
    • Step: 0.1
    • Controls the overall strength of the effect

  4. blur_radius (FLOAT)

    • Default: 2.0
    • Range: 0.1 to 10.0
    • Step: 0.1
    • Controls the radius of the Gaussian blur effect

  5. edge_threshold (FLOAT)

    • Default: 0.1
    • Range: 0.0 to 1.0
    • Step: 0.01
    • Threshold for edge detection sensitivity

  6. pixelate_factor (INT)

    • Default: 4
    • Range: 1 to 64
    • Step: 1
    • Size of pixelation blocks (higher = more pixelated)

  7. wave_amplitude (FLOAT)

    • Default: 0.1
    • Range: 0.0 to 1.0
    • Step: 0.01
    • Controls the strength of wave distortion

  8. wave_frequency (FLOAT)

    • Default: 5.0
    • Range: 0.1 to 50.0
    • Step: 0.1
    • Controls the frequency of wave distortion

  9. chromatic_shift (FLOAT)

    • Default: 0.01
    • Range: 0.0 to 0.1
    • Step: 0.001
    • Amount of RGB channel separation

Effect Descriptions

Grayscale

  • Converts the image to grayscale using standard luminance weights
  • Intensity controls the blend between original and grayscale
  • Uses weights: R(0.299), G(0.587), B(0.114)

Edge Detection

  • Implements a simple edge detection filter
  • Edge_threshold controls which edges are visible
  • Higher threshold values show only strong edges
  • Lower threshold values show more subtle details

Gaussian Blur

  • Applies a 2D Gaussian blur
  • Blur_radius controls the spread of the blur
  • Implemented as a separable convolution for better performance
  • Quality scales with blur_radius parameter

Pixelate

  • Creates a blocky, pixelated effect
  • Pixelate_factor determines block size
  • Larger values create larger blocks
  • Useful for retro-style effects

Wave Distortion

  • Applies sinusoidal distortion to the image
  • Wave_amplitude controls distortion strength
  • Wave_frequency controls number of waves
  • Creates liquid-like or wavy effects

Chromatic Aberration

  • Simulates lens color fringing
  • Shifts RGB channels separately
  • Chromatic_shift controls separation amount
  • Creates a "prismatic" effect

Technical Implementation

  • Uses ModernGL for GPU-accelerated processing
  • Implements GLSL shaders for real-time effects
  • Processes images in RGBA format
  • Automatically handles color space conversion
  • Supports batch processing of multiple images

Performance Considerations

  • GPU-accelerated processing for high performance
  • Efficient shader-based implementations
  • Automatic resource management
  • Batch processing capability
  • Memory-efficient texture handling

Use Cases

  • Photo and video post-processing
  • Creative visual effects
  • Real-time image filtering
  • Artistic style transfer
  • Visual glitch effects
  • Retro/vintage style effects

Tips

  • For subtle effects, use lower intensity values
  • Combine multiple passes for complex effects
  • Adjust parameters gradually for precise control
  • Monitor GPU memory usage with large batches
  • Consider image resolution impact on performance

Requirements

  • Requires OpenGL-capable GPU
  • ModernGL Python package
  • Numpy for array operations
  • OpenCV for color space conversion