tcu::PixelFormat format = m_context.getRenderTarget().getPixelFormat();
if (format == tcu::PixelFormat(8, 8, 8, 8) || format == tcu::PixelFormat(8, 8, 8, 0))
- return format.getColorThreshold().toIVec().asUint() + tcu::UVec4(2); // Default threshold.
+ return format.getColorThreshold().toIVec().asUint() + tcu::UVec4(6); // Default threshold.
else
return format.getColorThreshold().toIVec().asUint()
* tcu::UVec4(5) + tcu::UVec4(2); // \note Non-scientific ad hoc formula. Need big threshold when few color bits; especially multiple blendings bring extra inaccuracy.