Byte vs. Float based RGBA

Newbie & Debugging Questions
#1

It would seem to me 4 bytes vs. 16 would at least be more compact in terms of storage. When would I want to send float based color values? If precision is needed for color interpolation I would think that the card’s internal registers would convert the individual color comonents to a float.

Regards,

Monty Hall

#2

Don’t use floats for colours, basically.

Cas :slight_smile:

#3

It’s basically only valid to use them when you’re sending something else than colors to the GPU, typically in GPGPU applications.

Or when you’re using HDR, and you’re rendering to a FBO backed by a texture.