Next: Normalisation Up: RATIO Prev: Input Output
Method
Consider Fig (1) below. This is the setup if a 2D array was being
ratioed along its first dimension. N6 and N7 indicate the sizes
of the array in the 7D system.
Fig 1:
+-----------------------------------+ ^
Row1-> | |abcdef| |ghij| | |
| | | | | | N7
| | | | | | |
+-----------------------------------+ v
/ / / /
LOW1 HIGH1 LOW2 HIGH2
The data values in the two ranges are summed, row by row, eg. in
ROW1 the data are summed X = ( a+...+f ) and Y = ( g+...+j ) to
produce arrays of dimensionality one less than the input (Fig 2).
Fig 2: +-+ +-+ +-+
|X| |Y| Y/X |Z|
| | | | => | |
| | | | | |
+-+ +-+ +-+
Then these arrays are divided to produce the final output. The
output array is also multiplied by RANGE1 / RANGE2 (the widths of
the ranges selected in the units of the axis) - there might be
too much noise in one range so you can increase the width of the
range to average out the variations. This ensures that the value
of the ratioing process does not depend on small changes in the
width of the range.
Next: Normalisation Up: RATIO Prev: Input Output