trying to make data set from non-equally spaced to eqally spaced

I have two sets of data which are not equally spaced. What I am trying to do is to use interpolation to get an equally space data.

wave1 --> data in x-axis (1xN) abs1 --> data in y-axis (1xN)

wave2 = 900:0.1:1089; abs3 = interp1(wave1, abs1, wave2, 'spline');

Are two sentences a right way to get an equally space data which are wave2 & abs2?

Answer 1

[Star Strider]

That appears correct to me. I would question the 'spline' method, since it assumes information that may not be in your data. I would use 'linear', at least at first. Also, the linspace function might be more appropriate than the colon operator to create ‘wave2’.

[JK]

Error using griddedInterpolant The grid vectors must contain unique points.

Error in interp1 (line 161) F = griddedInterpolant(X,V,method);

Error in DataInterpolation (line 5) abs2 = interp1(wave1, abs1, wave2, 'linear');

Do you know why I am getting these errors?

[Star Strider]

I suspect that ‘wave1’ had duplicated points.

The way I always deal with this is to create a vector that adds very small increasing numbers to each element in the vector:

wave1 = [1 2 3 3 4 5 6 7 7];
abs1 = sin(2*pi*wave1/max(wave1));
wave2 = linspace(min(wave1), max(wave1), 20);

wave1 = wave1 + (1:length(wave1))*1E-14;
abs2 = interp1(wave1, abs1, wave2, 'linear','extrap');

The 'extrap' option may be necessary. It is here, because the addition of the small increasing vector means that the first element of ‘wave2’ is very slightly outside the limits of wave1, so the interpolation has to extrapolate.

[JK]

clear all;

abs = importdata('abs.mat'); wave = importdata('wave.mat'); abs1 = abs'; wave1 = wave'; wave1 = wave1+(1:length(wave1))*1E-14;

wave2 = linspace(900,1089,(1089-900+1)*10); abs2 = interp1(wave1, abs1, wave2, 'linear','extrap');

sorry but it still doesn't work. there was no duplicate points of wave.mat. i attached two .mat files. would you be able to provide me one more support? thanks.

[Star Strider]

Actually, ‘wave(101 & 102) = 980’ are duplicates:

[Ws,~,ic] = unique(wave);
Tally = accumarray(ic, 1);
Dupl_Idx = find(Tally > 1);
Duplicate = [wave(Dupl_Idx) Dupl_Idx; wave(Dupl_Idx+1) Dupl_Idx+1]

The multiplier amplitude needs to be reduced because of the magnitude of the original values of ‘wave’. The value I used in my example will get ‘swamped’ due to the limitations of floating-point precision.

Try this:

W = load('JK wave.mat');
A = load('JK abs.mat');
wave = W.wave;
abs1 = A.abs;

wave1 = wave + (1:length(wave))'*1E-8;

wave2 = linspace(900,1089,(1089-900+1)*10);
abs2 = interp1(wave1, abs1, wave2, 'linear','extrap');

figure(2)
plot(wave2, abs2, '-b',  wave1, abs1, '+r')
grid
legend('Interpolated', 'Original', 'Location','NE')

SUCCESS!

[JK]

it was successful. 980 was a problem. Thanks.

[Star Strider]

As always, my pleasure.