Learn Phoduit

9.4 Filter

9.4.1 Demosaic

The Demosaic node takes an image where each pixel only has the tone for a specific color and intelligently processes it into a typical image where each pixel has red, green, and blue tones that additively produce a color.

Only images imported with a Raw Image Open node and have not already been processed by a Demosaic node will be processed. All other images are passed through unchanged. See Chapter 7. Raw Images for a walkthrough on how to use the Demosaic node.

Input

Image
The image to be demosaiced.

Output

Image
The input image as demosaiced if raw. Otherwise, the input image.

9.4.2 Dilate

The Dilate node finds the brightest nearby colors in each pixel around each point in an image and assigns them to the pixel at that point. It’s useful for shrinking or expanding masks. Use the Erode node for the opposite functionality which finds the darkest nearby colors.

Input

Image
The image to be dilated.
Shape

The shape of the filter.

  • Circle: A circular shape. Increase the radius to make the shape more round.
  • Square: A square shape.
Radius
The range in pixels to be searched at each point.
Passes
The number of times to apply the dilation. More passes will ultimately increase the range but tends to make images look like they are made from lots of large flat blotches.
Channels

The channels to work on. You will usually want this set to RGB.

  • RGB: The red, green, and blue channels.
  • RGBA: The red, green, blue, and alpha channels.

Output

Image
The dilated image.

9.4.3 Erode

The Erode node finds the darkest nearby colors in each pixel around each point in an image and assigns them to the pixel at that point. It’s useful for shrinking or expanding masks. Use the Dilate node for the opposite functionality which finds the brightest nearby colors.

Input

Image
The image to be eroded.
Shape

The shape of the filter.

  • Circle: A circular shape. Increase the radius to make the shape more round.
  • Square: A square shape.
Radius
The range in pixels to be searched at each point.
Passes
The number of times to apply the erosion. More passes will ultimately increase the range but tends to make images look like watercolor paintings.
Channels

The channels to work on. You will usually want this set to RGB.

  • RGB: The red, green, and blue channels.
  • RGBA: The red, green, blue, and alpha channels.

Output

Image
The eroded image.

9.4.4 Gaussian Blur

The Gaussian Blur node applies a low pass filter on an image which causes it to appear blurred.

Input

Image
The image to be blurred.
Horizontal Radius
The amount in pixels to blur horizontally. A higher value produces a stronger blur. Click the Link button on the right to lock or unlock the ratio of the Horizontal Radius to the Vertical Radius.
Vertical Radius
The amount in pixels to blur vertically. A higher value produces a stronger blur. Click the Link button on the right to lock or unlock the ratio of the Horizontal Radius to the Vertical Radius.
Edges

How pixels outside of the image should be treated.

  • Blank: Use pixels that are black and completely translucent.
  • Clamp: Use the closest pixel on the edge of the image.
  • Repeat: Use pixels from the image as if it were repeatedly tiled.

Output

Image
The blurred image.

9.4.5 Median

The Median node removes speckled noise from an image. The changes in the output image can be hard to see unless the Canvas is zoomed in to 100% or greater.

Input

Image
The image to have speckled noise removed.
Radius
The range in pixels to be searched. A larger radius is better at removing more speckles but is much slower.
Passes
The number of times the image should be processed. Very noisy images can often be enhanced further with more passes. Too many passes can cause images to appear flat and “cartoon-like”.

Output

Image
The input image with speckled noise removed.

9.4.6 Noise

The Noise node changes random pixels into random colors.

Input

Image
The image to be filled with random noise.
Seed
Arbitrary number used when randomizing. All values produce the same effect but the location and color of the altered pixels will change.
Coverage
The percentage of pixels in the input image to be filled with noise.

Output

Image
The input image filled with random noise.

9.4.7 Sharpen

The Sharpen node applies a traditional sharpen box filter to an image. You will usually want to use the Unsharp Mask node instead for more control when sharpening your photographs.

Input

Image
The image to sharpen.
Sharpness
How much to make edges stand out.

Output

Image
The input image as sharpened.

9.4.8 Unsharp Mask

The Unsharp Mask selectively makes edges stand out.

See Sharpening in Chapter 4. Essentials for a guide on using the Unsharp Mask node.

Input

Image
The image to be sharpened.
Amount
How much edges are made to stand out. Higher values makes edges stand out more.
Radius
How wide the edges are once sharpened.
Threshold
How much a pixel’s tones must change relative to its neighbors in order to be considered an edge.

Output

Image
The input image as sharpened.