Paul Matthews

There are a few things I should have added to the description above.

1. Please read at least one of the papers by Gottwald and Melbourne before using the code. Marco, there is a discussion and explanation of the oversampling issue in section 6 of the paper linked above, and a helpful picture (fig 10). Oversampling of continuous data can lead to an incorrect result of zero for chaotic systems.
2. The two checks for oversampling that lead to the warning are very crude and should not be relied upon. The code sometimes gives the warning when it shouldn't.
3. The test is designed for deterministic systems. It cannot be used to distinguish chaos from noise.

It doesnt seem faster than dct to me.
In fact it is slower for row vectors-

tic; for i=1:3000;x=rand(1,512);dct(x);end; toc % 1.5
tic; for i=1:3000;x=rand(1,512);idct(x);end; toc % 1.5

tic; for i=1:3000;x=rand(1,512);mirt_dctn(x);end; toc % 3.0
tic; for i=1:3000;x=rand(1,512);mirt_idctn(x);end; toc % 2.7

But for 2d it seems about twice as fast as dct.

It is a mystery to me why Matlab have not written a fast dct

There are a few things I should have added to the description above.

1. Please read at least one of the papers by Gottwald and Melbourne before using the code. Marco, there is a discussion and explanation of the oversampling issue in section 6 of the paper linked above, and a helpful picture (fig 10). Oversampling of continuous data can lead to an incorrect result of zero for chaotic systems.
2. The two checks for oversampling that lead to the warning are very crude and should not be relied upon. The code sometimes gives the warning when it shouldn't.
3. The test is designed for deterministic systems. It cannot be used to distinguish chaos from noise.