Merge branch 'dev'

This commit is contained in:
Branimir Karadžić 2016-01-11 20:56:06 -08:00
commit 08a31297d5

View file

@ -1122,29 +1122,38 @@ namespace bgfx
// https://www.opengl.org/registry/specs/EXT/texture_shared_exponent.txt
const int32_t expMax = (1<<ExpBits) - 1;
const int32_t expBias = (1<<(ExpBits - 1) ) - 1;
const float sharedExpMax = float( (1 << MantissaBits) - 1) / (1 << MantissaBits) * (1 << (expMax-expBias));
const float sharedExpMax = float(expMax) / float(expMax + 1) * float(1 << (expMax - expBias) );
const float rr = bx::fclamp(_src[0], 0.0f, sharedExpMax);
const float gg = bx::fclamp(_src[1], 0.0f, sharedExpMax);
const float bb = bx::fclamp(_src[2], 0.0f, sharedExpMax);
const float max = bx::fmax3(rr, gg, bb);
const float expShared = bx::fmax(-expBias-1, bx::ffloor(bx::flog2(max) ) ) + 1 + expBias;
const float invExp = 1.0f/bx::fpow(2.0f, expShared - expBias - MantissaBits);
union { float ff; uint32_t ui; } cast = { max };
int32_t expShared = int32_t(bx::uint32_imax(uint32_t(-expBias-1), ( ( (cast.ui>>23) & 0xff) - 127) ) ) + 1 + expBias;
float denom = bx::fpow(2.0f, float(expShared - expBias - MantissaBits) );
_dst[0] = rr * invExp;
_dst[1] = gg * invExp;
_dst[2] = bb * invExp;
_dst[3] = expShared;
if ( (1<<MantissaBits) == int32_t(bx::fround(max/denom) ) )
{
denom *= 2.0f;
++expShared;
}
const float invDenom = 1.0f/denom;
_dst[0] = bx::fround(rr * invDenom);
_dst[1] = bx::fround(gg * invDenom);
_dst[2] = bx::fround(bb * invDenom);
_dst[3] = float(expShared);
}
template<int32_t MantissaBits, int32_t ExpBits>
void decodeRgbE(float* _dst, const float* _src)
{
const int32_t expBias = (1<<(ExpBits - 1) ) - 1;
const float exp = bx::fpow(2.0f, _src[3]-float(expBias-MantissaBits) );
_dst[0] = _src[0] * exp;
_dst[1] = _src[1] * exp;
_dst[2] = _src[2] * exp;
const float exponent = _src[3]-float(expBias-MantissaBits);
const float scale = bx::fpow(2.0f, exponent);
_dst[0] = _src[0] * scale;
_dst[1] = _src[1] * scale;
_dst[2] = _src[2] * scale;
}
// RGB9E5F
@ -1154,9 +1163,9 @@ namespace bgfx
encodeRgbE<9, 5>(tmp, _src);
*( (uint32_t*)_dst) = 0
| (toUnorm(tmp[0], 511.0f) )
| (toUnorm(tmp[1], 511.0f)<< 9)
| (toUnorm(tmp[2], 511.0f)<<18)
| (uint32_t(tmp[0]) )
| (uint32_t(tmp[1]) << 9)
| (uint32_t(tmp[2]) <<18)
| (uint32_t(tmp[3]) <<27)
;
}