Best way for a 3d matrix

First of all, thanks for your time.

What I intend to do is a modified version of the DCT.

About the image input, i really doesn't know yet how it will be done and if it will be stored into a memory. I believe serial. This week I have an appointment with my leader and we will decide. But After I have a project that makes the 8x8 compactation, I can always call it in a different project as a component.

Like your example, 1280x720, can be decomposed in 14400 8x8 blocks. I will have a project that makes that.

As I said this is only for tests, since I didn't worked with arrays before.

I really have to be able to manipulate the bits of the pixel, since I must use my own adder/subtractor.

--- Quote Start ---

There are already some defined arrays of std_logic - std_logic_vector, signed and unsigned.

--- Quote End ---

Where can I find more info about that?

When I get more data about my project I explain to make it a little bit clear.

Thanks!