srgb section correction

[jeremyselan]

The document's description of srgb - specifically the discussion of the srgb curve - is inaccurate.

As first noted by Brad Friedman, the srgb "gamma 2.4 equation" in the VES color white paper document is NOT suitable for modeling the light from a calibrated srgb display. This equation, while appearing in the srgb spec, is instead the camera encoding equation recommended to use when create imagery suitable for the srgb display.

To frame it in more modern term, this curve is roughly a scene-referred -> display-referred tone rendering transform. (Though not a particularly useful one for modern capture devices). This also sheds lights on the original of the linear toe portion. While not making sense in a monitor response, such linear black areas are actually sensible in camera encoding curves where dark noise is a concern.

The doc should be updated accordingly.