1

I'm a bit of a noob with Directx but i have been trying to get this custom bitpacking working all day. I'm trying to pack a float4 and another float into a uint. The float4 is a color and the float is a depth value. I want to use 6 bits each for the color, and the remaining 8 bits for depth. Precision is not important. I thought I understood what i was doing, but when unpacking, it just keeps returning all zeros. Is this even possible? The pixel format is R16G16B16A16_FLOAT.

Here is the code

uint PackColorAndDepthToUint(float4 Color, float Depth)
{

    uint4 u = (int4) (Color * float4(255, 255, 255, 1.0f));
    uint packedOutput = (u.a << 26) | (u.b << 20) | (u.g << 14) | (u.r << 8) | asuint(Depth * 255);
    return packedOutput;
}

void UnpackColorAndDepthFromUint(uint packedInput, out float4 Color, out float Depth)
{
    uint d = (packedInput & 255);
    Depth = ((float) d / 255);
    uint r = (((packedInput >> 8)) & 64);
    uint g = (((packedInput >> 14)) & 64);
    uint b = (((packedInput >> 20)) & 64);
    uint a = (((packedInput >> 26)) & 64);
    uint4 co = uint4(r, g, b, a);

    Color = (((float4) co) / float4(255, 255, 255, 1.0f));
}

if anyone can point me in the right direction i would appreciate it!

kostenickj
  • 11
  • 3
  • How did you check that you only get zero? When and where are you packing and unpacking? Where is the value stored in between? Two minor things: A multiplier of 1 for the alpha channel seems wrong (why not 255 as for the other channels?). Furthermore, you should mask the six bits of your color components when packing (so as to not influence the other components). Better also do it for the depth value. – Nico Schertler Nov 17 '18 at 17:44

0 Answers0