@c DO NOT EDIT!  Generated automatically by munge-texi.pl.

@c Copyright (C) 2007-2025 The Octave Project Developers
@c
@c This file is part of Octave.
@c
@c Octave is free software: you can redistribute it and/or modify it
@c under the terms of the GNU General Public License as published by
@c the Free Software Foundation, either version 3 of the License, or
@c (at your option) any later version.
@c
@c Octave is distributed in the hope that it will be useful, but
@c WITHOUT ANY WARRANTY; without even the implied warranty of
@c MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
@c GNU General Public License for more details.
@c
@c You should have received a copy of the GNU General Public License
@c along with Octave; see the file COPYING.  If not, see
@c <https://www.gnu.org/licenses/>.

@node Interpolation
@chapter Interpolation

@menu
* One-dimensional Interpolation::
* Multi-dimensional Interpolation::
@end menu

@node One-dimensional Interpolation
@section One-dimensional Interpolation

Octave supports several methods for one-dimensional interpolation, most
of which are described in this section.  @ref{Polynomial Interpolation}
and @ref{Interpolation on Scattered Data} describe additional methods.

@c interp1 scripts/general/interp1.m
@anchor{XREFinterp1}
@html
<span style="display:block; margin-top:-4.5ex;">&nbsp;</span>
@end html


@deftypefn  {} {@var{yi} =} interp1 (@var{x}, @var{y}, @var{xi})
@deftypefnx {} {@var{yi} =} interp1 (@var{y}, @var{xi})
@deftypefnx {} {@var{yi} =} interp1 (@dots{}, @var{method})
@deftypefnx {} {@var{yi} =} interp1 (@dots{}, @var{extrap})
@deftypefnx {} {@var{yi} =} interp1 (@dots{}, "left")
@deftypefnx {} {@var{yi} =} interp1 (@dots{}, "right")
@deftypefnx {} {@var{pp} =} interp1 (@dots{}, "pp")

One-dimensional interpolation.

Interpolate input data to determine the value of @var{yi} at the points
@var{xi}.  If not specified, @var{x} is taken to be the indices of @var{y}
(@code{1:length (@var{y})}).  If @var{y} is a matrix or an N-dimensional
array, the interpolation is performed on each column of @var{y}.

The interpolation @var{method} is one of:

@table @asis
@item @qcode{"nearest"}
Return the nearest neighbor.

@item @qcode{"previous"}
Return the previous neighbor.

@item @qcode{"next"}
Return the next neighbor.

@item @qcode{"linear"} (default)
Linear interpolation from nearest neighbors.

@item @qcode{"pchip"}
Piecewise cubic Hermite interpolating polynomial---shape-preserving
interpolation with smooth first derivative.

@item @qcode{"cubic"}
Cubic interpolation (same as @qcode{"pchip"}).

@item @qcode{"spline"}
Cubic spline interpolation---smooth first and second derivatives
throughout the curve.
@end table

Adding '*' to the start of any method above forces @code{interp1}
to assume that @var{x} is uniformly spaced, and only @code{@var{x}(1)}
and @code{@var{x}(2)} are referenced.  This is usually faster,
and is never slower.  The default method is @qcode{"linear"}.

If @var{extrap} is the string @qcode{"extrap"}, then extrapolate values
beyond the endpoints using the current @var{method}.  If @var{extrap} is a
number, then replace values beyond the endpoints with that number.  When
unspecified, @var{extrap} defaults to @code{NA}.

If the string argument @qcode{"pp"} is specified, then @var{xi} should not
be supplied and @code{interp1} returns a piecewise polynomial object.  This
object can later be used with @code{ppval} to evaluate the interpolation.
There is an equivalence, such that @code{ppval (interp1 (@var{x},
@var{y}, @var{method}, @qcode{"pp"}), @var{xi}) == interp1 (@var{x},
@var{y}, @var{xi}, @var{method}, @qcode{"extrap"})}.

Duplicate points in @var{x} specify a discontinuous interpolant.  There
may be at most 2 consecutive points with the same value.
If @var{x} is increasing, the default discontinuous interpolant is
right-continuous.  If @var{x} is decreasing, the default discontinuous
interpolant is left-continuous.
The continuity condition of the interpolant may be specified by using
the options @qcode{"left"} or @qcode{"right"} to select a left-continuous
or right-continuous interpolant, respectively.
Discontinuous interpolation is only allowed for @qcode{"nearest"} and
@qcode{"linear"} methods; in all other cases, the @var{x}-values must be
unique.

An example of the use of @code{interp1} is

@example
@group
xf = [0:0.05:10];
yf = sin (2*pi*xf/5);
xp = [0:10];
yp = sin (2*pi*xp/5);
lin = interp1 (xp, yp, xf);
near = interp1 (xp, yp, xf, "nearest");
pch = interp1 (xp, yp, xf, "pchip");
spl = interp1 (xp, yp, xf, "spline");
plot (xf,yf,"r", xf,near,"g", xf,lin,"b", xf,pch,"c", xf,spl,"m",
      xp,yp,"r*");
legend ("original", "nearest", "linear", "pchip", "spline");
@end group
@end example

@xseealso{@ref{XREFpchip,,pchip}, @ref{XREFspline,,spline}, @ref{XREFinterpft,,interpft}, @ref{XREFinterp2,,interp2}, @ref{XREFinterp3,,interp3}, @ref{XREFinterpn,,interpn}}
@end deftypefn


There are some important differences between the various interpolation
methods.  The @qcode{"spline"} method enforces that both the first and second
derivatives of the interpolated values have a continuous derivative,
whereas the other methods do not.  This means that the results of the
@qcode{"spline"} method are generally smoother.  If the function to be
interpolated is in fact smooth, then @qcode{"spline"} will give excellent
results.  However, if the function to be evaluated is in some manner
discontinuous, then @qcode{"pchip"} interpolation might give better results.

This can be demonstrated by the code

@example
@group
t = -2:2;
dt = 1;
ti =-2:0.025:2;
dti = 0.025;
y = sign (t);
ys = interp1 (t,y,ti,"spline");
yp = interp1 (t,y,ti,"pchip");
ddys = diff (diff (ys)./dti) ./ dti;
ddyp = diff (diff (yp)./dti) ./ dti;
figure (1);
plot (ti,ys,"r-", ti,yp,"g-");
legend ("spline", "pchip", 4);
figure (2);
plot (ti,ddys,"r+", ti,ddyp,"g*");
legend ("spline", "pchip");
@end group
@end example

@ifnotinfo
@noindent
The result of which can be seen in @ref{fig:interpderiv1} and
@ref{fig:interpderiv2}.

@float Figure,fig:interpderiv1
@center @image{interpderiv1,4in}
@caption{Comparison of @qcode{"pchip"} and @qcode{"spline"} interpolation methods for a
step function}
@end float

@float Figure,fig:interpderiv2
@center @image{interpderiv2,4in}
@caption{Comparison of the second derivative of the @qcode{"pchip"} and @qcode{"spline"}
interpolation methods for a step function}
@end float
@end ifnotinfo

Fourier interpolation, is a resampling technique where a signal is
converted to the frequency domain, padded with zeros and then
reconverted to the time domain.

@c interpft scripts/general/interpft.m
@anchor{XREFinterpft}
@html
<span style="display:block; margin-top:-4.5ex;">&nbsp;</span>
@end html


@deftypefn  {} {@var{y} =} interpft (@var{x}, @var{n})
@deftypefnx {} {@var{y} =} interpft (@var{x}, @var{n}, @var{dim})

Fourier interpolation.

If @var{x} is a vector then @var{x} is resampled with @var{n} points.  The
data in @var{x} is assumed to be equispaced.  If @var{x} is a matrix or an
N-dimensional array, the interpolation is performed on each column of
@var{x}.

If @var{dim} is specified, then interpolate along the dimension @var{dim}.

@code{interpft} assumes that the interpolated function is periodic, and so
assumptions are made about the endpoints of the interpolation.
@xseealso{@ref{XREFinterp1,,interp1}}
@end deftypefn


There are two significant limitations on Fourier interpolation.  First,
the function signal is assumed to be periodic, and so non-periodic
signals will be poorly represented at the edges.  Second, both the
signal and its interpolation are required to be sampled at equispaced
points.  An example of the use of @code{interpft} is

@example
@group
t = 0 : 0.3 : pi; dt = t(2)-t(1);
n = length (t); k = 100;
ti = t(1) + [0 : k-1]*dt*n/k;
y = sin (4*t + 0.3) .* cos (3*t - 0.1);
yp = sin (4*ti + 0.3) .* cos (3*ti - 0.1);
plot (ti, yp, "g", ti, interp1 (t, y, ti, "spline"), "b", ...
      ti, interpft (y, k), "c", t, y, "r+");
legend ("sin(4t+0.3)cos(3t-0.1)", "spline", "interpft", "data");
@end group
@end example

@noindent
@ifinfo
which demonstrates the poor behavior of Fourier interpolation for non-periodic
functions.
@end ifinfo
@ifnotinfo
which demonstrates the poor behavior of Fourier interpolation for non-periodic
functions, as can be seen in @ref{fig:interpft}.

@float Figure,fig:interpft
@center @image{interpft,4in}
@caption{Comparison of @code{interp1} and @code{interpft} for non-periodic data}
@end float
@end ifnotinfo

In addition, the support functions @code{spline} and @code{lookup} that
underlie the @code{interp1} function can be called directly.

@c spline scripts/polynomial/spline.m
@anchor{XREFspline}
@html
<span style="display:block; margin-top:-4.5ex;">&nbsp;</span>
@end html


@deftypefn  {} {@var{pp} =} spline (@var{x}, @var{y})
@deftypefnx {} {@var{yi} =} spline (@var{x}, @var{y}, @var{xi})
Return the cubic spline interpolant of points @var{x} and @var{y}.

When called with two arguments, return the piecewise polynomial @var{pp}
that may be used with @code{ppval} to evaluate the polynomial at specific
points.

When called with a third input argument, @code{spline} evaluates the spline
at the points @var{xi}.  The third calling form
@code{spline (@var{x}, @var{y}, @var{xi})} is equivalent to
@code{ppval (spline (@var{x}, @var{y}), @var{xi})}.

The variable @var{x} must be a vector of length @var{n}.

@var{y} can be either a vector or array.  If @var{y} is a vector it must
have a length of either @var{n} or @code{@var{n} + 2}.  If the length of
@var{y} is @var{n}, then the @qcode{"not-a-knot"} end condition is used.
If the length of @var{y} is @code{@var{n} + 2}, then the first and last
values of the vector @var{y} are the values of the first derivative of the
cubic spline at the endpoints.

If @var{y} is an array, then the size of @var{y} must have the form
@tex
$$[s_1, s_2, \cdots, s_k, n]$$
@end tex
@ifnottex
@code{[@var{s1}, @var{s2}, @dots{}, @var{sk}, @var{n}]}
@end ifnottex
or
@tex
$$[s_1, s_2, \cdots, s_k, n + 2].$$
@end tex
@ifnottex
@code{[@var{s1}, @var{s2}, @dots{}, @var{sk}, @var{n} + 2]}.
@end ifnottex
The array is reshaped internally to a matrix where the leading
dimension is given by
@tex
$$s_1 s_2 \cdots s_k$$
@end tex
@ifnottex
@code{@var{s1} * @var{s2} * @dots{} * @var{sk}}
@end ifnottex
and each row of this matrix is then treated separately.  Note that this is
exactly the opposite of @code{interp1} but is done for @sc{matlab}
compatibility.

@xseealso{@ref{XREFpchip,,pchip}, @ref{XREFppval,,ppval}, @ref{XREFmkpp,,mkpp}, @ref{XREFunmkpp,,unmkpp}}
@end deftypefn


@node Multi-dimensional Interpolation
@section Multi-dimensional Interpolation

There are three multi-dimensional interpolation functions in Octave, with
similar capabilities.  Methods using Delaunay tessellation are described
in @ref{Interpolation on Scattered Data}.

@c interp2 scripts/general/interp2.m
@anchor{XREFinterp2}
@html
<span style="display:block; margin-top:-4.5ex;">&nbsp;</span>
@end html


@deftypefn  {} {@var{zi} =} interp2 (@var{x}, @var{y}, @var{z}, @var{xi}, @var{yi})
@deftypefnx {} {@var{zi} =} interp2 (@var{z}, @var{xi}, @var{yi})
@deftypefnx {} {@var{zi} =} interp2 (@var{z}, @var{n})
@deftypefnx {} {@var{zi} =} interp2 (@var{z})
@deftypefnx {} {@var{zi} =} interp2 (@dots{}, @var{method})
@deftypefnx {} {@var{zi} =} interp2 (@dots{}, @var{method}, @var{extrap})

Two-dimensional interpolation.

Interpolate numeric reference data @var{x}, @var{y}, @var{z} to determine
@var{zi} at the coordinates @var{xi}, @var{yi}.  The reference data
@var{x}, @var{y} can be matrices, as returned by @code{meshgrid}, in which
case the sizes of @var{x}, @var{y}, and @var{z} must be equal.  If @var{x},
@var{y} are vectors describing a grid then
@w{@code{length (@var{x}) == columns (@var{z})}}@ and
@w{@code{length (@var{y}) == rows (@var{z})}}.  In either case the input
data must be strictly monotonic.

If called without @var{x}, @var{y}, and just a single reference data matrix
@var{z}, the 2-D region
@code{@var{x} = 1:columns (@var{z}), @var{y} = 1:rows (@var{z})} is assumed.
This saves memory if the grid is regular and the distance between points is
not important.

If called with a single reference data matrix @var{z} and a refinement
value @var{n}, then perform interpolation over a grid where each original
interval has been recursively subdivided @var{n} times.  This results in
@code{2^@var{n}-1} additional points for every interval in the original
grid.  If @var{n} is omitted a value of 1 is used.  As an example, the
interval [0,1] with @code{@var{n}==2} results in a refined interval with
points at [0, 1/4, 1/2, 3/4, 1].

The interpolation @var{method} is one of:

@table @asis
@item @qcode{"nearest"}
Return the nearest neighbor.

@item @qcode{"linear"} (default)
Linear interpolation from nearest neighbors.

@item @qcode{"pchip"}
Piecewise cubic Hermite interpolating polynomial---shape-preserving
interpolation with smooth first derivative.

@item @qcode{"cubic"}
Cubic interpolation using a convolution kernel function---third order
method with smooth first derivative.

@item @qcode{"spline"}
Cubic spline interpolation---smooth first and second derivatives
throughout the curve.
@end table

@var{extrap} is a scalar number.  It replaces values beyond the endpoints
with @var{extrap}.  Note that if @var{extrap} is used, @var{method} must
be specified as well.  If @var{extrap} is omitted and the @var{method} is
@qcode{"spline"}, then the extrapolated values of the @qcode{"spline"} are
used.  Otherwise the default @var{extrap} value for any other @var{method}
is @qcode{"NA"}.
@xseealso{@ref{XREFinterp1,,interp1}, @ref{XREFinterp3,,interp3}, @ref{XREFinterpn,,interpn}, @ref{XREFmeshgrid,,meshgrid}}
@end deftypefn


@c interp3 scripts/general/interp3.m
@anchor{XREFinterp3}
@html
<span style="display:block; margin-top:-4.5ex;">&nbsp;</span>
@end html


@deftypefn  {} {@var{vi} =} interp3 (@var{x}, @var{y}, @var{z}, @var{v}, @var{xi}, @var{yi}, @var{zi})
@deftypefnx {} {@var{vi} =} interp3 (@var{v}, @var{xi}, @var{yi}, @var{zi})
@deftypefnx {} {@var{vi} =} interp3 (@var{v}, @var{n})
@deftypefnx {} {@var{vi} =} interp3 (@var{v})
@deftypefnx {} {@var{vi} =} interp3 (@dots{}, @var{method})
@deftypefnx {} {@var{vi} =} interp3 (@dots{}, @var{method}, @var{extrapval})

Three-dimensional interpolation.

Interpolate numeric reference data @var{x}, @var{y}, @var{z}, @var{v} to
determine @var{vi} at the coordinates @var{xi}, @var{yi}, @var{zi}.  The
reference data @var{x}, @var{y}, @var{z} can be matrices, as returned by
@code{meshgrid}, in which case the sizes of @var{x}, @var{y}, @var{z}, and
@var{v} must be equal.  If @var{x}, @var{y}, @var{z} are vectors describing
a cubic grid then @code{length (@var{x}) == columns (@var{v})},
@code{length (@var{y}) == rows (@var{v})}, and
@code{length (@var{z}) == size (@var{v}, 3)}.  In either case the input
data must be strictly monotonic.

If called without @var{x}, @var{y}, @var{z}, and just a single reference
data matrix @var{v}, the 3-D region
@code{@var{x} = 1:columns (@var{v}), @var{y} = 1:rows (@var{v}),
@var{z} = 1:size (@var{v}, 3)} is assumed.
This saves memory if the grid is regular and the distance between points is
not important.

If called with a single reference data matrix @var{v} and a refinement
value @var{n}, then perform interpolation over a 3-D grid where each
original interval has been recursively subdivided @var{n} times.  This
results in @code{2^@var{n}-1} additional points for every interval in the
original grid.  If @var{n} is omitted a value of 1 is used.  As an
example, the interval [0,1] with @code{@var{n}==2} results in a refined
interval with points at [0, 1/4, 1/2, 3/4, 1].

The interpolation @var{method} is one of:

@table @asis
@item @qcode{"nearest"}
Return the nearest neighbor.

@item @qcode{"linear"} (default)
Linear interpolation from nearest neighbors.

@item @qcode{"cubic"}
Piecewise cubic Hermite interpolating polynomial---shape-preserving
interpolation with smooth first derivative (not implemented yet).

@item @qcode{"spline"}
Cubic spline interpolation---smooth first and second derivatives
throughout the curve.
@end table

@var{extrapval} is a scalar number.  It replaces values beyond the endpoints
with @var{extrapval}.  Note that if @var{extrapval} is used, @var{method}
must be specified as well.  If @var{extrapval} is omitted and the
@var{method} is @qcode{"spline"}, then the extrapolated values of the
@qcode{"spline"} are used.  Otherwise the default @var{extrapval} value for
any other @var{method} is @qcode{"NA"}.
@xseealso{@ref{XREFinterp1,,interp1}, @ref{XREFinterp2,,interp2}, @ref{XREFinterpn,,interpn}, @ref{XREFmeshgrid,,meshgrid}}
@end deftypefn


@c interpn scripts/general/interpn.m
@anchor{XREFinterpn}
@html
<span style="display:block; margin-top:-4.5ex;">&nbsp;</span>
@end html


@deftypefn  {} {@var{vi} =} interpn (@var{x1}, @var{x2}, @dots{}, @var{v}, @var{y1}, @var{y2}, @dots{})
@deftypefnx {} {@var{vi} =} interpn (@var{v}, @var{y1}, @var{y2}, @dots{})
@deftypefnx {} {@var{vi} =} interpn (@var{v}, @var{m})
@deftypefnx {} {@var{vi} =} interpn (@var{v})
@deftypefnx {} {@var{vi} =} interpn (@dots{}, @var{method})
@deftypefnx {} {@var{vi} =} interpn (@dots{}, @var{method}, @var{extrapval})

Perform @var{n}-dimensional interpolation, where @var{n} is at least two.

Each element of the @var{n}-dimensional numeric array @var{v} represents a
value at a location given by the parameters @var{x1}, @var{x2}, @dots{},
@var{xn}.  The parameters @var{x1}, @var{x2}, @dots{}, @var{xn} are either
@var{n}-dimensional arrays of the same size as the array @var{v} in
the @qcode{"ndgrid"} format or vectors.

The parameters @var{y1}, @var{y2}, @dots{}, @var{yn} represent the points at
which the array @var{vi} is interpolated.  They can be vectors of the same
length and orientation in which case they are interpreted as coordinates of
scattered points.  If they are vectors of differing orientation or length,
they are used to form a grid in @qcode{"ndgrid"} format.  They can also be
@var{n}-dimensional arrays of equal size.

If @var{x1}, @dots{}, @var{xn} are omitted, they are assumed to be
@code{x1 = 1 : size (@var{v}, 1)}, etc.  If @var{m} is specified, then
the interpolation adds a point half way between each of the interpolation
points.  This process is performed @var{m} times.  If only @var{v} is
specified, then @var{m} is assumed to be @code{1}.

The interpolation @var{method} is one of:

@table @asis
@item @qcode{"nearest"}
Return the nearest neighbor.

@item @qcode{"linear"} (default)
Linear interpolation from nearest neighbors.

@item @qcode{"pchip"}
Piecewise cubic Hermite interpolating polynomial---shape-preserving
interpolation with smooth first derivative (not implemented yet).

@item @qcode{"cubic"}
Cubic interpolation (same as @qcode{"pchip"} [not implemented yet]).

@item @qcode{"spline"}
Cubic spline interpolation---smooth first and second derivatives
throughout the curve.
@end table

The default method is @qcode{"linear"}.

@var{extrapval} is a scalar number.  It replaces values beyond the endpoints
with @var{extrapval}.  Note that if @var{extrapval} is used, @var{method}
must be specified as well.  If @var{extrapval} is omitted and the
@var{method} is @qcode{"spline"}, then the extrapolated values of the
@qcode{"spline"} are used.  Otherwise the default @var{extrapval} value for
any other @var{method} is @code{NA}.
@xseealso{@ref{XREFinterp1,,interp1}, @ref{XREFinterp2,,interp2}, @ref{XREFinterp3,,interp3}, @ref{XREFspline,,spline}, @ref{XREFndgrid,,ndgrid}}
@end deftypefn


A significant difference between @code{interpn} and the other two
multi-dimensional interpolation functions is the fashion in which the
dimensions are treated.  For @code{interp2} and @code{interp3}, the y-axis is
considered to be the columns of the matrix, whereas the x-axis corresponds to
the rows of the array.  As Octave indexes arrays in column major order, the
first dimension of any array is the columns, and so @code{interpn} effectively
reverses the 'x' and 'y' dimensions.  Consider the example,

@example
@group
x = y = z = -1:1;
f = @@(x,y,z) x.^2 - y - z.^2;
[xx, yy, zz] = meshgrid (x, y, z);
v = f (xx,yy,zz);
xi = yi = zi = -1:0.1:1;
[xxi, yyi, zzi] = meshgrid (xi, yi, zi);
vi = interp3 (x, y, z, v, xxi, yyi, zzi, "spline");
[xxi, yyi, zzi] = ndgrid (xi, yi, zi);
vi2 = interpn (x, y, z, v, xxi, yyi, zzi, "spline");
mesh (zi, yi, squeeze (vi2(1,:,:)));
@end group
@end example

@noindent
where @code{vi} and @code{vi2} are identical.  The reversal of the
dimensions is treated in the @code{meshgrid} and @code{ndgrid} functions
respectively.
@ifnotinfo
The result of this code can be seen in @ref{fig:interpn}.

@float Figure,fig:interpn
@center @image{interpn,4in}
@caption{Demonstration of the use of @code{interpn}}
@end float
@end ifnotinfo