3.11 Arrays of Signed and Unsigned Chars

Manipulation ofarrays of bytes (signed or unsigned chars) is provided by the classes RWMathVec<T>, RWGenMat<T>, and RWMathArray<T>. Two typedefs are used:

typedef signed char SChar;
typedef unsigned char UChar;

For applications requiring only small integers, these classes may be used for high efficiency and compactness. They are also very useful for storing and manipulating graphical images and 8-bit data streams from A/D converters, digitizers, and so on. This makes it easy to take a 2-dimensional Fourier transform of a graphical image, for example:

#include <rw/math/genmat.h>
#include <rw/cfft2d.h>
#include <iostream.h>

main()
{
  // Define a matrix of unsigned chars:
  RWGenMat<UChar> image;
  cin >> image;          // Read in the image

  // Allocate a 2-D FFT server:
  DComplexFFT2DServer serve;

  // Calculate the Fourier Transform.
  RWGenMat<DComplex> spectrum = 
       serve.fourier(RWConvertGenMat<UChar,DComplex>(image));

  cout << spectrum;     // Print periodogram
}

Notice that we use the integer representations of the elements for input/output of vectors of signed or unsigned chars, not the literal character constants. This can differ from the behavior of your compiler when a single element is printed on an output stream. For example, the code:

char a = '$';
cout << a;

prints $ with most compilers. However, the code:

RWMathVec<SChar> a(3, '$');
 cout << a;

always prints 36 36 36 with Math.h++. This is because the vector classes are primarily a tool for manipulating numbers, not strings. The Tools.h++ class RWCString should be used for manipulating strings.