NAME

SYNOPSIS

wavelet [-1] [-2] [-d] [-r] [-# channels] [-t [c | d | f | i | l | s]] [-w width] [-n scanlines] [-s squaresize] [-S square_out_size] [-W width_out] < file > wavelet

DESCRIPTION

-d perform decomposition.

-r perform reconstruction.

-1 perform a 1-dimensional (horizontal) transform.

-2 perform a 1-dimensional (interleaved horizontal and vertical) transform.

-# channels Indicates the number of values in each sample point. For example, to perform a transform on a pix(5) file properly, specify "-# 3". This causes the red, green and blue channels to be transformed individually.

-t {cdfils} specifies the data type of the input, and hence the data type in which the wavelet calculations will be performed. The letters each stand for the first character of the C programming language data type: character, double, float, int, long, short. Note that if lossless decomposition and reconstruction are desired, then data sets should be converted to the next larger data type before being processed.

-w width Specify the number of samples per scanline.

-n scanline Specify the number of scanlines in the dataset

-s squaresize Specify both width and scanlines simultaneously (square image).

-W width_out Specifies the width of the output image. Normally this is 1 for a full decomposition/reconstruction. You may specify a width at which decomposition or reconstruction should stop. This number should be a power of 2. The result will be an "average" image of the specified size, with the remainder of the original data width converted to "detail" coefficients. See also: -R

-R avg_size Resumes a transform that left off with an average image of avg_size.

-S square_out_size This is the same as the -W option. It implies the square size for a 2D decomposition limit.

EXAMPLES

decompose -# 3 < img.pix | pix-fb will display the horizontal decomposition of the file img.pix.

The command

decompose -# 3 -s 1024 -l 64 < img.pix > will decompose a 1024x1024 image. The decomposition will stop when the image has been decomposed to 64x64. The output image will thus have a 64x64 version of the original in the lower left corner of the image, with detail terms in the remainder.

SEE ALSO

AUTHOR

COPYRIGHT

BUG REPORTS