upsamling with any number of data

I want to increase my data number that is different matlab functions interp and upsampling. they add numbers after every sample but i want to add them for example two and adding sample like data=(1 3 5 7 9); data_up=(1 3 4 5 7 8 9); i wrote a code but it doesn't work for every numbers sometimes it works with odds sometimes evens can anyone give any suggestion for that?

[Jan]

Please post the code and explain "doesn't work" with any details.

What does "I want to add them for example two and adding sample data" mean? What is the algorithm or idea to get from data to data_up?

[Image Analyst]

All I can see is that an interpolated number was inserted in only 2 of the four possible locations: between the 2nd and 3rd location, and between the 4th and 5th location. Why there are not numbers in between the other elements, I have no idea. Regardless, you can still get this strange output by passing in the proper interpolation coordinates to interp1() (not interp()) - see my code below.

[geem]

first thanks i have add code my aim is that increase 27 to 30 so i add after every 9 sapmle 1 sample more and i calculate it by averaging but something is wrong at that line dty_upye(1,i:i+r+1)=dty_up(j,:) matlab says:Subscripted assignment dimension mismatch.

%% upsampling with any sampling rate

for i=1:27

    %create data
    datay(i)=i+(i+1);
end
byt_datay=length(datay);
% calculate datay length
r=fix(byt_datay/10);
% all  the number of upsampled data 
all=10*(r+1);
loop_nmbr=all-byt_datay;
point=fix(byt_datay/loop_nmbr);
dty_up=zeros(loop_nmbr,point+1);
k=1;
   for j=1:loop_nmbr      
           extra=(datay(1,k+r)+datay(1,k+r+1))/2;
           dty_up(j,:)= [datay(1,k:k+point-1) extra];           
           k=k+point;          
   end
   i=1;
   [a b]=size(dty_up);
   dty_upye=zeros(1,a*b);
   for j=1:loop_nmbr
    dty_upye(1,i:i+r+1)=dty_up(j,:)
    i=i+r+2;
   end
   c=all-a*b;
   dty_upye2=[dty_upye datay(1,(end-(c-1)):end)];

[Image Analyst]

Proper indenting and adding comments to the above code would help. I'm not going to take the time to figure out badly aligned code that's uncommented with a cryptic alphabet soup of non-descriptive variable names. Perhaps someone else likes to do that though.

Answer 1

[Wayne King]

something very similar to what Image Analyst suggests is just to use a linear interpolation filter.

     x = 1:2:9;
     xup = upsample(x,2);
     h = [1/2 1 1/2];
     y = filter(h,1,xup);

In this case you do get a 1/2 in the first element of the output.

[geem]

thank you so much it is helped me

Answer 2

[Image Analyst]

Is this crazy, weird thing what you want?

data = [1, 3, 5, 7, 9]
% Want data_up = [1 3 4 5 7 8 9]
xInterpLocations = [1, 2, 2.5, 3, 4, 4.5, 5]
data_up = interp1(data, xInterpLocations)

In the command window:

data =
     1     3     5     7     9
xInterpLocations =
  Columns 1 through 4
          1      2      2.5     3
  Columns 5 through 7
            4         4.5       5
data_up =
     1     3     4     5     7     8     9