IppFFT Class Reference

A wrapper around the Intel IPP FFT routines. More...

Public Member Functions
void	complexFFT (const float *in, float *out, int size) const
	Complex inplace FFT (input is aligned Complex<float> array, size is power of two.)
void	complexFFTInplace (float *data, int size) const
	Complex inplace FFT (input is aligned Complex<float> array, size is power of two.)
void	complexFFTInplace (double *data, int size) const
	Complex inplace FFT (input is aligned Complex<double> array, size is power of two.)
void	complexFFTInverse (const float *in, float *out, int size) const
	Complex inverse inplace FFT (input is aligned float array, size is power of two.)
void	complexFFTInverseInplace (float *data, int size) const
	Complex inverse inplace FFT (input is aligned float array, size is power of two.)
void	complexFFTInverseInplace (double *data, int size) const
	Complex inverse inplace FFT (input is aligned Complex<double> array, size is power of two.)
	IppFFT (DataType typeToUse=DataType::ComplexFloat, int maxPowerOfTwo=IPP_FFT_MAX_POWER_OF_TWO, const int flagToUse=8)
	Creates a IPP FFT object. More...
void	realFFT (const float *in, float *out, int size) const
	Real inplace FFT (input is aligned float array, size is power of two.) More...
void	realFFTInplace (float *data, int size) const
	Real inplace FFT (input is aligned float array, size is power of two.) More...
void	realFFTInplace (double *data, int size) const
	Real inplace FFT (input is aligned double array, size is power of two.) More...
void	realFFTInverse (const float *in, float *out, int size) const
	Real inplace FFT (input is aligned float array, size is power of two.)
void	realFFTInverseInplace (float *data, int size) const
	Real inplace FFT (input is aligned float array, size is power of two.)
void	realFFTInverseInplace (double *data, int size) const
	Real inplace inverse FFT (input is aligned double array, size is power of two.)

Detailed Description

A wrapper around the Intel IPP FFT routines.

It uses RAII to manage all required buffers & datas. In order to use it, create an instance once and then call the routines:

IppFFT fft(IppFFT:DataType::ComplexFloat);

fft.realFFT(data, 512);

It is not templated for speed, but it throws assertions if you call it on the wrong type.

Constructor & Destructor Documentation

IppFFT	(	DataType	typeToUse = DataType::ComplexFloat,
		int	maxPowerOfTwo = IPP_FFT_MAX_POWER_OF_TWO,
		const int	flagToUse = 8
	)

Creates a IPP FFT object.

The initialisation time is rather slow, but then it will unleash its power because it allocates all necessary buffers once.

Member Function Documentation

void realFFT	(	const float *	in,
		float *	out,
		int	size
	)		const

Real inplace FFT (input is aligned float array, size is power of two.)

Input: d[] = re[0],re[1],..,re[size-1]. Output: d[] = re[0],re[size/2],re[1],im[1],..,re[size/2-1],im[size/2-1].

void realFFTInplace	(	float *	data,
		int	size
	)		const

Real inplace FFT (input is aligned float array, size is power of two.)

Input: d[] = re[0],re[1],..,re[size-1]. Output: d[] = re[0],re[size/2],re[1],im[1],..,re[size/2-1],im[size/2-1].

void realFFTInplace	(	double *	data,
		int	size
	)		const

Real inplace FFT (input is aligned double array, size is power of two.)

Input: data[] = re[0],re[1],..,re[size-1]. Output: data[] = re[0],re[size/2],re[1],im[1],..,re[size/2-1],im[size/2-1].