What does tmp(k, n) command mean in matlab?

I got this code for dct transform matrix, but it isnt correct. I dont understand the tmp calling:

function [ C ] = dct_mtx( N )
for k = 0:N-1
    for n = 0:N-1
       tmp(k+1,n+1) = cos((0.5+N)*k*pi/N);
    end
end

tmp(1,:) = tmp(1,:) / sqrt(2);
C = tmp;
end

[Jaroslav Venjarski]

Also I need help with code for dct transform matrix, Im not allowed to use dctmtx.

Answer 1

[KSSV]

unction [ C ] = dct_mtx( N )     % this is to call a function 
for k = 0:N-1   % loop for k variable (row)
    for n = 0:N-1   % loop for n variable (column)
       tmp(k+1,n+1) = cos((0.5+N)*k*pi/N);    % calculation stored in tmp(k+1,n+1)
    end   % end loop
end   % end loop

tmp(1,:) = tmp(1,:) / sqrt(2);    % this will replace the first row of tmp with (tmp first row)/sqrt(2)
C = tmp;    % tmp is assigned to variable C
end

[KSSV]

Use this code instead:

function [ C ] = dct_mtx( N )
tmp = zeros(N) ;
for k = 1:N
    for n = 1:N
       tmp(k,n) = cos((0.5+N)*(k-1)*pi/N);
    end
end
tmp(1,:) = tmp(1,:) / sqrt(2);
C = tmp;
end

[Jaroslav Venjarski]

And what exactly tmp means? And why i need to replace first row with first row/sqrt(2)? It is supposed to be a code for dct matrix.

[KSSV]

tmp is some variable matrix....

Answer 2

[Stephen23]

MATLAB does not require variables to be declared before being indexed into: when they are first referred to the variable will be implicitly created. Here is a simple example which works without error, even though X does not exist in the workspace:

>> X(3) = 6
X =
    0    0    6

You can see that MATLAB simply created a variable big enough to include the index used, and filled the unused elements with zeros. So the rather badly written code in your question uses that feature:

function [ C ] = dct_mtx( N )
for k = 0:N-1
    for n = 0:N-1
       tmp(k+1,n+1) = cos((0.5+N)*k*pi/N);
    end
end

On the very first loop iteration tmp does not exist, but the first time it is referred to (with indexing) MATLAB creates it, with whatever class the data has.

However I would NOT recommend doing this: this method is susceptible to bugs caused by an existing variable of the same name. It can be trivially avoided by preallocating before the loops. Your example is also pointlessly slow because it expands the array on each loop iteration, which means the array must be moved in memory:

https://www.mathworks.com/help/matlab/matlab_prog/preallocating-arrays.html

Summary: Do NOT learn from that pointlessly inefficient code.

Answer 3

[HOUSSAM Touhrach]

I write this code but the error message is :

Invalid expression. When calling a function or indexing a variable, use parentheses. Otherwise, check for mismatched delimiters.

function [P,A] = ExtrusionPresure(ys,D0,D,k,n,Om)

ts=0:50;

P = 1;

for i=2:length(ts)

P = 2 .* ys .* log(D0./D) + FonctionA(n,Om).* k .* (ts(i)).^n) * (1-(D./D0).^(3 .* n));

end

end

[Stephen23]

Your parentheses are mismatched:

P = 2 .* ys .* log(D0./D) + FonctionA(n,Om).* k .* (ts(i)).^n) * (1-(D./D0).^(3 .* n));
% 0               1     0            1    0        1  2 10  -1   0  1     0  1      0-1
%                                                        ^ maybe you
%                                                        should delete this
%                                                        one