Fast low pass filter

The fast low pass filter utilises the frame grabber ALU to perform a simple low pass convolutions. This is achieved by a sequence of ALU addition operations, each using a shifted version of the image. At each stage the image is added to itself, shifted by a specific number of pixels either horizontally or vertically; the result is divided by two to produce a new image. At each stage the least significant bit of the result is rounded to maintain eight intensity bits. Thus, unlike the software convolution filters in the convolution filter submenu, the final image is subject to a degree of truncation error.

The effective convolution template depends on the specified size of the kernal. For example a 3x3 kernal uses

           1 2 1
           2 4 2
           1 2 1

Note that this filter is always applied to the entire screen.

Filter which buffer?

Any valid image buffer may be filtered. The buffer filtered will be made the current output buffer. If the buffer specified is in extended memory, then the buffer is first copied to buffer 0; after the filtering is finished, the result will be copied back to the extended memory buffer.

Size of kernal (rounded to next odd value) ?

This entry specifies the size of the convolution kernal to be used. The smallest kernal is 3x3 and the largest 25x25. The kernals are always square in pixel coordinates, and extend an odd number of pixels.

Parent menu

Main Menu

DigImage User Documentation

Stuart Dalziel, last page update: 19 February 1996