IDXTConvolution interface
This interface performs image convolutions for DXSurface objects.
Members
The IDXTConvolution interface inherits from the IUnknown interface. IDXTConvolution also has these types of members:
- Methods
Methods
The IDXTConvolution interface has these methods.
| Method | Description |
|---|---|
| GetBias | Retrieves the bias value used for the convolution. |
| GetConvertToGray | Retrieves a value that indicates whether the filter output color is converted to grayscale. |
| GetExcludeAlpha | Retrieves a value that indicates whether the alpha channel should be filtered along with the color data. |
| GetFilterType | Retrieves the selected filter type for the convolution. |
| SetBias | Sets the bias value used for the convolution. |
| SetConvertToGray | Sets a value that indicates whether the filter output color is converted to grayscale. |
| SetCustomFilter | Sets the convolution to use a custom filter. |
| SetExcludeAlpha | Sets a value that indicates whether the alpha channel should be filtered with the color data. |
| SetFilterType | Sets the selected filter type for the convolution. |
Remarks
Image convolution is more commonly known as filtering. Convolution works by combining the colors of surrounding samples to determine the color of a sample on the destination surface. The way in which surrounding pixels are combined is determined by a filter, which is a rectangular matrix of numbers. Each number in the matrix is a weighting for the pixel at that position. The contributions from each weighting are added to produce the output color on the destination surface. Negative weighting values are subtracted from the total. Each filter matrix has a center pixel that is scanned over each sample of the input image to produce the output image.
The filter matrix has a width and height that determine how many pixels of the source image will contribute to the color of the destination's center pixel. Most matrices are square and have an odd number of rows and columns, though you can define custom filters of arbitrary size. There are also predefined filters described in the DXCONVFILTERTYPE enumeration.
Sometimes convolution can reduce or increase the brightness of an image. To compensate, you can define a bias that is added to the filtered result.
Requirements
Minimum supported client |
Windows XP |
Minimum supported server |
Windows 2000 Server |
Header |
Dxtmsft.h |
IDL |
Dxtmsft.idl |
DLL |
Dxtrans.dll |