Frame buffer to DVI Question (PC graphics cards/hardware/data bits)

[faculaganymede]

Graphics/hardware gurus,

I know a little bit about shaders and have done some shader programming using OpenGL Shading Language.

In a fragment shader, the RGB colors are mapped to a range between 0 and 1 (as floating point values). I assume this is how the pixel RGB data are stored in the frame buffer also (?).

Suppose I am using a high-end graphics card (e.g. ATI x1800), how does the pixel RGB data from frame buffer get down sampled to the DVI output (e.g. 10-bit DVI output)? I am particularly interested in finding out the flow/path of the RGB data bits. Please point me to any documentation that discusses this.


(I am probably not using some of the terminologies correctly here; please feel free to correct me.)

[sharangad]

Graphics/hardware gurus,

I know a little bit about shaders and have done some shader programming using OpenGL Shading Language.

In a fragment shader, the RGB colors are mapped to a range between 0 and 1 (as floating point values). I assume this is how the pixel RGB data are stored in the frame buffer also (?).



The frame buffes are 32 bit RGBA (with 8 bits (integer values) for R, G, B and the alpha channel).

Of those the RGB values are directly sent to the DVI output.

John Carmack of Id Software has been campaigning for floating point framebuffers ( where each of the RGBA values are floating point values between 0.0-1.0 ). In this case, if 6 bits were available for the R value and in the framebuffer the values was 0.7, then what you see on screen would be:

round down ( ( (2^6) - 1 ) * 0.7 ) the nearest int not greater than the value.