MakeFloat

Converts mantissa and exponent arrays to float arrays.

Syntax

IppStatus ippsMakeFloat_16s32f (Ipp16s* inmant, Ipp16s* inexp, Ipp32s size, Ipp32f* outfloat);

Parameters

inmant	Array of mantissas.
inexp	Array of exponents.
size	Number of array elements.
outfloat	Array of resulting float arrays.

Description

Th function ippsMakeFloat is declared in the ippac.h header file. This function converts the mantissa and exponent arrays decoded from the bitstream to the float array of spectral samples by the formula:

outfloat[i] = inmant[1] x 2^{-inexp[i] -15}, where i = 0...size.

The conversion serves to improve application performance when decoding bitstreams in the AC3 format.

Below see an example of using ippsMakeFloat_16s32f.

Return Values

ippStsNoErr	Indicates no error.
ippStsNullPtrErr	Indicates an error when the inmant, inexp or outfloat pointer is NULL.
ippStsSizeErr	Indicates an error when size is less than or equal to 0.

ippsMakeFloat Usage

IppStatus makefloat(void)

  Ipp16s inmant[5] = {1234, 56, 25907, 3498, 27854};

  Ipp16s inexp[5] = {-2, -1, 0, 1, 2};

  Ipp32f outfloat[5];

  IppStatus st = ippsMakeFloat_16s32f(inmant, inexp, 5, outfloat);

  int i;

  printf("\n outfloat = ");

  for (i = 0; i < 5; i++)

    printf("%f ", outfloat[i]);

  printf("\n");

//Output:

//   outfloat = 0.150635 0.003418 0.790619 0.053375 0.212509