在matlab中，help fread发现的。
Any of the following strings, either the MATLAB version, or  their C or Fortran equivalent, may be used.  If not specified, the default precision is \'uchar\'.
        MATLAB    C or Fortran     Description
        \'uchar\'   \'unsigned char\'  unsigned character,  8 bits.
        \'schar\'   \'signed char\'    signed character,  8 bits.
        \'int8\'    \'integer*1\'      integer, 8 bits.
        \'int16\'   \'integer*2\'      integer, 16 bits.
        \'int32\'   \'integer*4\'      integer, 32 bits.
        \'int64\'   \'integer*8\'      integer, 64 bits.
        \'uint8\'   \'integer*1\'      unsigned integer, 8 bits.
        \'uint16\'  \'integer*2\'      unsigned integer, 16 bits.
        \'uint32\'  \'integer*4\'      unsigned integer, 32 bits.
        \'uint64\'  \'integer*8\'      unsigned integer, 64 bits.
        \'single\'  \'real*4\'         floating point, 32 bits.
        \'float32\' \'real*4\'         floating point, 32 bits.
        \'double\'  \'real*8\'         floating point, 64 bits.
        \'float64\' \'real*8\'         floating point, 64 bits.
   

The following platform dependent formats are also supported but they are not guaranteed to be the same size on all platforms.
        MATLAB    C or Fortran     Description
        \'char\'    \'char*1\'         character,  8 bits (signed or unsigned).
        \'short\'   \'short\'          integer,  16 bits.
        \'int\'     \'int\'            integer,  32 bits.
        \'long\'    \'long\'           integer,  32 or 64 bits.
        \'ushort\'  \'unsigned short\' unsigned integer,  16 bits.
        \'uint\'    \'unsigned int\'   unsigned integer,  32 bits.
        \'ulong\'   \'unsigned long\'  unsigned integer,  32 bits or 64 bits.
        \'float\'   \'float\'          floating point, 32 bits.