NDLINSPACE (v1.1, feb 2008)

NDLINSPACE - Generalized linearly spaced matrix for multiple points

R = NDLINSPACE(M1,M2, N) returns a matrix of size [size(M1) N] holding
N equally spaced points between corresponding point of M1 and M2. If N
is smaller than 2, M2 is returned. R = NDLINSPACE(M1,M2) uses N = 100.

For scalar inputs, NDLINSPACE(S1,S2,N) mimicks LINSPACE, returning
equally spaced points between the S1 and S2. Example:

ndlinspace(0, 2, 4)
% -> 0 0.67 1.33 2.00

For two M-by-1 column vectors, NDLINSPACE(V1,V2,N) returns a M-by-N
matrix in which the k-th row holds the N equally spaced numbers between
V1(k) and V2(k). Similary, for two 1-by-M row vectors, a N-by-M matrix
is returned. Example:

ndlinspace([0 ; 12 ; 10], [3 ;18 ; 11],4)
% -> 0 1.00 2.00 3.00
% 12.00 14.00 16.00 18.00
% 10.00 10.33 10.67 11.00

For two ND arrays of size P-by-Q-by-R-by-.. matrices, NDLINSPACE
(M1,M2,N) returns a P-by-Q-by-R-by-..-by-N matrix, holding N equally
space points between M1(i,j,k,..) and M2(i,j,k,..). Example:

M1 = reshape(1:2*3*4*5,[2 3 4 5]) ;
M2 = 10 * M1 ;
size(M1) ; % ans = 2 3 4 5
R = ndlinspace(M1,M2,5) ;
squeeze(R(2,1,4,3,:)).'
% ans = 68 221 374 527 680
% Compare with LINSPACE for scalars
linspace(M1(2,1,4,3),M2(2,1,4,3),5)
% ans = 68 221 374 527 680

In general, M1 and M2 have to have the same size. Scalar expansion is
applied when one of the inputs is a scalar. Example:
ndlinspace(0,[4 ; 8 ; -12],5) % expand M1
% ans = 0 1 2 3 4
% 0 2 4 6 8
% 0 -3 -6 -9 -12

See also linspace, logspace, repmat, interp1, squeeze