gam::ipl Namespace Reference

Interpolation functions. More...

Classes
struct	Trunc
	Truncating random-access interpolation strategy. More...
struct	Round
	Nearest neighbor random-access interpolation strategy. More...
struct	Linear
	Linear random-access interpolation strategy. More...
struct	Cubic
	Cubic random-access interpolation strategy. More...
struct	AllPass
	Allpass random-access interpolation strategy. More...
struct	Any
	Dynamically switchable random-access interpolation strategy. More...
Enumerations
enum	Type {   TRUNC = 0, ROUND, LINEAR, CUBIC,   ALLPASS }
	Interpolation types. More...
Functions
template<class Tf , class Tv >
Tv	allpass (Tf frac, const Tv &x, const Tv &y, Tv &o1)
	First order allpass interpolation.
template<class T >
void	lagrange (T *h, T delay, int order)
	Computes FIR coefficients for Waring-Lagrange interpolation.
template<class T >
void	lagrange1 (T *h, T delay)
	Optimized lagrange() for first order.
template<class T >
void	lagrange2 (T *h, T delay)
	Optimized lagrange() for second order.
template<class T >
void	lagrange3 (T *h, T delay)
	Optimized lagrange() for third order.
template<class T >
T	parabolic (T xm1, T x, T xp1)
	Simplified parabolic interpolation of 3 points.
template<class Tf , class Tv >
Tv	cubic (Tf frac, const Tv &w, const Tv &x, const Tv &y, const Tv &z)
	Cubic interpolation.
template<class Tf , class Tv >
Tv	linear (Tf frac, const Tv &x, const Tv &y)
	Linear interpolation. Identical to first order Lagrange.
template<class Tf , class Tv >
Tv	linear (Tf frac, const Tv &x, const Tv &y, const Tv &z)
	Linear interpolation between three elements.
template<class Tf , class Tv >
Tv	nearest (Tf frac, const Tv &x, const Tv &y)
	Nearest neighbor interpolation.
template<class Tf , class Tv >
Tv	quadratic (Tf frac, const Tv &x, const Tv &y, const Tv &z)
	Quadratic interpolation.

---

Detailed Description

Interpolation functions.

The naming convention for values is that their alphabetical order is equivalent to their sequence order (oldest to newest). The interpolated value lies 'frac' distance between 'x' and 'y' where 'frac' lies in [0, 1).

---

Enumeration Type Documentation

enum Type

Interpolation types.

Enumerator:

TRUNC	Truncating interpolation
ROUND	Rounding interpolation
LINEAR	Linear interpolation
CUBIC	Cubic interpolation
ALLPASS	Allpass interpolation

---

Function Documentation

Tv allpass	(	Tf	frac,
		const Tv &	x,
		const Tv &	y,
		Tv &	o1
	)

First order allpass interpolation.

Allpass interpolation preserves the spectral magnitude of the interpolated sequence. It should only be used for interpolating at a regular speed across a sequence. It will fail miserably at random access.

Tv cubic	(	Tf	frac,
		const Tv &	w,
		const Tv &	x,
		const Tv &	y,
		const Tv &	z
	)

Cubic interpolation.

This is a Cardinal spline with a tension of 0 (AKA a Catmull-Rom spline). The resulting value will never exceed the range of the interpolation points.

void lagrange	(	T *	h,
		T	delay,
		int	order
	)

Computes FIR coefficients for Waring-Lagrange interpolation.

'h' are the FIR coefficients and should be of size ('order' + 1).
'delay' is a fractional delay in samples.
As order increases, this converges to sinc interpolation.

T parabolic	(	T	xm1,
		T	x,
		T	xp1
	)

Simplified parabolic interpolation of 3 points.

This assumes the points are spaced evenly on the x axis from [-1, 1]. The output is an offset from 0.