dspp Namespace Reference

Digital Signal Processing library for C++. More...

Namespaces
	FFT
	Fast Fourier/Cosine/Sine Transforms.
	window
	Window functions (also known as an apodization functions or tapering functions).

Classes
struct	Mixer

Functions
template<typename T = double>
constexpr T	pi ()
	3.141592653589793238462643383279502884e+00
template<typename T = double>
constexpr T	two_pi ()
	6.283185307179586476925286766559005768e+00
template<typename T1 , typename T2 , typename T3 >
typename::std::enable_if < ::std::is_same< double, decltype(T1()+T2()+T3())> ::value, decltype(T1()+T2()+T3())> ::type	fmac (T1 x, T2 y, T3 z)
	Fast fused multiply and accumulate. Returns x * y + z. More...
template<typename T1 , typename T2 , typename T3 >
typename::std::enable_if < ::std::is_same< float, decltype(T1()+T2()+T3())> ::value, decltype(T1()+T2()+T3())> ::type	fmac (T1 x, T2 y, T3 z)
	Fast fused multiply and accumulate. Returns x * y + z.

Detailed Description

Digital Signal Processing library for C++.

Function Documentation

template<typename T1 , typename T2 , typename T3 >

typename ::std::enable_if< ::std::is_same<double, decltype(T1()+T2()+T3())>::value, decltype(T1()+T2()+T3())>::type dspp::fmac ( T1  x, T2  y, T3  z  )

inline

Fast fused multiply and accumulate. Returns x * y + z.

A fused multiply and accumulate avoids the rounding that would normally happen after multiplication. This offers greater precision. Some microprocessors do not have an FMA instruction so ::std::fma() will be significantly slower than (x*y+z). Most DSP cases prefer speed over precision which this function ensures by using (x*y+z) on processors without an FMA instruction. If you require precision over speed, use ::std::fma() instead.