1
$\begingroup$

So, I'm evaluating the Discrete Sine transform Basis functions and I found on the internet the following picture:

DST Basis Functions

So, I may be getting the point wrong, but, if the 2D DCT equation is:

DCT equation

Then the Discrete Sine transform shall be: DCT equation

Then, it is clear that for the lower frequency (u=0,v=0) all cosines will be 1, but in the case of the sines all will be 0 (for frequency=0), therefore, the basis function for the lower frequency, in the sine case, will be zero, right?

Is the upper left image (u=0,v=0) correct? It should be constant, right? (btw: It shall also be zero, right?)

Improve this question
$\endgroup$

2 Answers 2

Reset to default
1
$\begingroup$

I think I got it:

While in the DCT there are definitions of the DCT that make cos(0) when u and v are zero, this does not apply to DST (see: https://en.wikipedia.org/wiki/Discrete_sine_transform).

In the DST definition, there is always a +1 or a +1/2 adding to the frequency.

This explains why it is not zero and also the graph above. It varies (for u=0 and v=0) because the content of the sine is not zero.

Improve this answer
$\endgroup$
0
$\begingroup$

No, an all-zero function cannot be a base function; therefore, the DST can not be constructed the way you recommend.

Improve this answer
$\endgroup$
1
  • $\begingroup$ I was not recommending doing it that way, based on the wikipedia page: en.wikipedia.org/wiki/Discrete_sine_transform It seems like in the case of the DST a +1 or +1/2 is added to the frequency (u or v) in order to avoid making 0 the sine when u=0 or v=0 for all the DST types. Indeed, that also explains why it is not constant for u=0 and v=0 $\endgroup$
    – f.gallardo
    Dec 3, 2017 at 17:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.