5 replies to this topic

[lpcstr]

When I change my frame buffer format from R8G8B8A8_UNORM to R16G16B16A16_FLOAT, the colors I'm getting from my textures appear to be overexposed.

Should I be doing something different when using a FP frame buffer format?

[MJP]

Were you using a UNORM_SRGB format previously? Automatic sRGB conversion isn't supported on FLOAT formats, so you would need to do it manually in a shader.

[lpcstr]

No, I was just using DXGI_FORMAT_R8G8B8A8_UNORM.

[lpcstr]

R8G8B8A8_UNORM gives me the results I expect while both R16G16B16A16_FLOAT and R8G8B8A8_UNORM_SRGB give me the overexposed look.

[lpcstr]

Nobody has any ideas? Because this makes no sense at all to me. It would seem to indicate that it is, in fact, applying an automatic gamma correction when writing to the R16G16B16A16_FLOAT frame buffer, which makes no sense because that isn't a sRGB texture format.

[kauna]

Can you elaborate a bit your configuration? Are you using a floating point buffer for back buffer or is it just an intermediate color buffer? What is the back buffer format?

Are you loading your texture with correct sRGB flags?

Best regards!