blob: 1dfd3d8ef16e69e4310b904e32a3d6ac10fb8623 [file] [log] [blame]
/*-------------------------------------------------------------------------
* drawElements Quality Program Tester Core
* ----------------------------------------
*
* Copyright 2014 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*//*!
* \file
* \brief Texture lookup simulator that is capable of verifying generic
* lookup results based on accuracy parameters.
*//*--------------------------------------------------------------------*/
#include "tcuTexLookupVerifier.hpp"
#include "tcuTexVerifierUtil.hpp"
#include "tcuVectorUtil.hpp"
#include "tcuTextureUtil.hpp"
#include "deMath.h"
namespace tcu
{
using namespace TexVerifierUtil;
// Generic utilities
#if defined(DE_DEBUG)
static bool isSamplerSupported (const Sampler& sampler)
{
return sampler.compare == Sampler::COMPAREMODE_NONE &&
isWrapModeSupported(sampler.wrapS) &&
isWrapModeSupported(sampler.wrapT) &&
isWrapModeSupported(sampler.wrapR);
}
#endif // DE_DEBUG
// Color read & compare utilities
static inline bool coordsInBounds (const ConstPixelBufferAccess& access, int x, int y, int z)
{
return de::inBounds(x, 0, access.getWidth()) && de::inBounds(y, 0, access.getHeight()) && de::inBounds(z, 0, access.getDepth());
}
template<typename ScalarType>
inline Vector<ScalarType, 4> lookup (const ConstPixelBufferAccess& access, const Sampler& sampler, int i, int j, int k)
{
if (coordsInBounds(access, i, j, k))
return access.getPixelT<ScalarType>(i, j, k);
else
return sampleTextureBorder<ScalarType>(access.getFormat(), sampler);
}
template<>
inline Vector<float, 4> lookup (const ConstPixelBufferAccess& access, const Sampler& sampler, int i, int j, int k)
{
// Specialization for float lookups: sRGB conversion is performed as specified in format.
if (coordsInBounds(access, i, j, k))
{
const Vec4 p = access.getPixel(i, j, k);
return isSRGB(access.getFormat()) ? sRGBToLinear(p) : p;
}
else
return sampleTextureBorder<float>(access.getFormat(), sampler);
}
static inline bool isColorValid (const LookupPrecision& prec, const Vec4& ref, const Vec4& result)
{
const Vec4 diff = abs(ref - result);
return boolAll(logicalOr(lessThanEqual(diff, prec.colorThreshold), logicalNot(prec.colorMask)));
}
static inline bool isColorValid (const IntLookupPrecision& prec, const IVec4& ref, const IVec4& result)
{
return boolAll(logicalOr(lessThanEqual(absDiff(ref, result).asUint(), prec.colorThreshold), logicalNot(prec.colorMask)));
}
static inline bool isColorValid (const IntLookupPrecision& prec, const UVec4& ref, const UVec4& result)
{
return boolAll(logicalOr(lessThanEqual(absDiff(ref, result), prec.colorThreshold), logicalNot(prec.colorMask)));
}
struct ColorQuad
{
Vec4 p00; //!< (0, 0)
Vec4 p01; //!< (1, 0)
Vec4 p10; //!< (0, 1)
Vec4 p11; //!< (1, 1)
};
static void lookupQuad (ColorQuad& dst, const ConstPixelBufferAccess& level, const Sampler& sampler, int x0, int x1, int y0, int y1, int z)
{
dst.p00 = lookup<float>(level, sampler, x0, y0, z);
dst.p10 = lookup<float>(level, sampler, x1, y0, z);
dst.p01 = lookup<float>(level, sampler, x0, y1, z);
dst.p11 = lookup<float>(level, sampler, x1, y1, z);
}
struct ColorLine
{
Vec4 p0; //!< 0
Vec4 p1; //!< 1
};
static void lookupLine (ColorLine& dst, const ConstPixelBufferAccess& level, const Sampler& sampler, int x0, int x1, int y)
{
dst.p0 = lookup<float>(level, sampler, x0, y, 0);
dst.p1 = lookup<float>(level, sampler, x1, y, 0);
}
template<typename T, int Size>
static T minComp (const Vector<T, Size>& vec)
{
T minVal = vec[0];
for (int ndx = 1; ndx < Size; ndx++)
minVal = de::min(minVal, vec[ndx]);
return minVal;
}
template<typename T, int Size>
static T maxComp (const Vector<T, Size>& vec)
{
T maxVal = vec[0];
for (int ndx = 1; ndx < Size; ndx++)
maxVal = de::max(maxVal, vec[ndx]);
return maxVal;
}
static float computeBilinearSearchStepFromFloatLine (const LookupPrecision& prec,
const ColorLine& line)
{
DE_ASSERT(boolAll(greaterThan(prec.colorThreshold, Vec4(0.0f))));
const int maxSteps = 1<<16;
const Vec4 d = abs(line.p1 - line.p0);
const Vec4 stepCount = d / prec.colorThreshold;
const Vec4 minStep = 1.0f / (stepCount + 1.0f);
const float step = de::max(minComp(minStep), 1.0f / float(maxSteps));
return step;
}
static float computeBilinearSearchStepFromFloatQuad (const LookupPrecision& prec,
const ColorQuad& quad)
{
DE_ASSERT(boolAll(greaterThan(prec.colorThreshold, Vec4(0.0f))));
const int maxSteps = 1<<16;
const Vec4 d0 = abs(quad.p10 - quad.p00);
const Vec4 d1 = abs(quad.p01 - quad.p00);
const Vec4 d2 = abs(quad.p11 - quad.p10);
const Vec4 d3 = abs(quad.p11 - quad.p01);
const Vec4 maxD = max(d0, max(d1, max(d2, d3)));
const Vec4 stepCount = maxD / prec.colorThreshold;
const Vec4 minStep = 1.0f / (stepCount + 1.0f);
const float step = de::max(minComp(minStep), 1.0f / float(maxSteps));
return step;
}
static float computeBilinearSearchStepForUnorm (const LookupPrecision& prec)
{
DE_ASSERT(boolAll(greaterThan(prec.colorThreshold, Vec4(0.0f))));
const Vec4 stepCount = 1.0f / prec.colorThreshold;
const Vec4 minStep = 1.0f / (stepCount + 1.0f);
const float step = minComp(minStep);
return step;
}
static float computeBilinearSearchStepForSnorm (const LookupPrecision& prec)
{
DE_ASSERT(boolAll(greaterThan(prec.colorThreshold, Vec4(0.0f))));
const Vec4 stepCount = 2.0f / prec.colorThreshold;
const Vec4 minStep = 1.0f / (stepCount + 1.0f);
const float step = minComp(minStep);
return step;
}
static inline Vec4 min (const ColorLine& line)
{
return min(line.p0, line.p1);
}
static inline Vec4 max (const ColorLine& line)
{
return max(line.p0, line.p1);
}
static inline Vec4 min (const ColorQuad& quad)
{
return min(quad.p00, min(quad.p10, min(quad.p01, quad.p11)));
}
static inline Vec4 max (const ColorQuad& quad)
{
return max(quad.p00, max(quad.p10, max(quad.p01, quad.p11)));
}
static bool isInColorBounds (const LookupPrecision& prec, const ColorQuad& quad, const Vec4& result)
{
const tcu::Vec4 minVal = min(quad) - prec.colorThreshold;
const tcu::Vec4 maxVal = max(quad) + prec.colorThreshold;
return boolAll(logicalOr(logicalAnd(greaterThanEqual(result, minVal), lessThanEqual(result, maxVal)), logicalNot(prec.colorMask)));
}
static bool isInColorBounds (const LookupPrecision& prec, const ColorQuad& quad0, const ColorQuad& quad1, const Vec4& result)
{
const tcu::Vec4 minVal = min(min(quad0), min(quad1)) - prec.colorThreshold;
const tcu::Vec4 maxVal = max(max(quad0), max(quad1)) + prec.colorThreshold;
return boolAll(logicalOr(logicalAnd(greaterThanEqual(result, minVal), lessThanEqual(result, maxVal)), logicalNot(prec.colorMask)));
}
static bool isInColorBounds (const LookupPrecision& prec, const ColorLine& line0, const ColorLine& line1, const Vec4& result)
{
const tcu::Vec4 minVal = min(min(line0), min(line1)) - prec.colorThreshold;
const tcu::Vec4 maxVal = max(max(line0), max(line1)) + prec.colorThreshold;
return boolAll(logicalOr(logicalAnd(greaterThanEqual(result, minVal), lessThanEqual(result, maxVal)), logicalNot(prec.colorMask)));
}
static bool isInColorBounds (const LookupPrecision& prec,
const ColorQuad& quad00,
const ColorQuad& quad01,
const ColorQuad& quad10,
const ColorQuad& quad11,
const Vec4& result)
{
const tcu::Vec4 minVal = min(min(quad00), min(min(quad01), min(min(quad10), min(quad11)))) - prec.colorThreshold;
const tcu::Vec4 maxVal = max(max(quad00), max(max(quad01), max(max(quad10), max(quad11)))) + prec.colorThreshold;
return boolAll(logicalOr(logicalAnd(greaterThanEqual(result, minVal), lessThanEqual(result, maxVal)), logicalNot(prec.colorMask)));
}
// Range search utilities
static bool isLinearRangeValid (const LookupPrecision& prec,
const Vec4& c0,
const Vec4& c1,
const Vec2& fBounds,
const Vec4& result)
{
// This is basically line segment - AABB test. Valid interpolation line is checked
// against result AABB constructed by applying threshold.
const Vec4 i0 = c0*(1.0f - fBounds[0]) + c1*fBounds[0];
const Vec4 i1 = c0*(1.0f - fBounds[1]) + c1*fBounds[1];
const Vec4 rMin = result - prec.colorThreshold;
const Vec4 rMax = result + prec.colorThreshold;
bool allIntersect = true;
// Algorithm: For each component check whether segment endpoints are inside, or intersect with slab.
// If all intersect or are inside, line segment intersects the whole 4D AABB.
for (int compNdx = 0; compNdx < 4; compNdx++)
{
if (!prec.colorMask[compNdx])
continue;
// Signs for both bounds: false = left, true = right.
const bool sMin0 = i0[compNdx] >= rMin[compNdx];
const bool sMin1 = i1[compNdx] >= rMin[compNdx];
const bool sMax0 = i0[compNdx] > rMax[compNdx];
const bool sMax1 = i1[compNdx] > rMax[compNdx];
// If all signs are equal, line segment is outside bounds.
if (sMin0 == sMin1 && sMin1 == sMax0 && sMax0 == sMax1)
{
allIntersect = false;
break;
}
}
return allIntersect;
}
static bool isBilinearRangeValid (const LookupPrecision& prec,
const ColorQuad& quad,
const Vec2& xBounds,
const Vec2& yBounds,
const float searchStep,
const Vec4& result)
{
DE_ASSERT(xBounds.x() <= xBounds.y());
DE_ASSERT(yBounds.x() <= yBounds.y());
DE_ASSERT(xBounds.x() + searchStep > xBounds.x()); // step is not effectively 0
DE_ASSERT(xBounds.y() + searchStep > xBounds.y());
if (!isInColorBounds(prec, quad, result))
return false;
for (float x = xBounds.x(); x < xBounds.y()+searchStep; x += searchStep)
{
const float a = de::min(x, xBounds.y());
const Vec4 c0 = quad.p00*(1.0f - a) + quad.p10*a;
const Vec4 c1 = quad.p01*(1.0f - a) + quad.p11*a;
if (isLinearRangeValid(prec, c0, c1, yBounds, result))
return true;
}
return false;
}
static bool isTrilinearRangeValid (const LookupPrecision& prec,
const ColorQuad& quad0,
const ColorQuad& quad1,
const Vec2& xBounds,
const Vec2& yBounds,
const Vec2& zBounds,
const float searchStep,
const Vec4& result)
{
DE_ASSERT(xBounds.x() <= xBounds.y());
DE_ASSERT(yBounds.x() <= yBounds.y());
DE_ASSERT(zBounds.x() <= zBounds.y());
DE_ASSERT(xBounds.x() + searchStep > xBounds.x()); // step is not effectively 0
DE_ASSERT(xBounds.y() + searchStep > xBounds.y());
DE_ASSERT(yBounds.x() + searchStep > yBounds.x());
DE_ASSERT(yBounds.y() + searchStep > yBounds.y());
if (!isInColorBounds(prec, quad0, quad1, result))
return false;
for (float x = xBounds.x(); x < xBounds.y()+searchStep; x += searchStep)
{
for (float y = yBounds.x(); y < yBounds.y()+searchStep; y += searchStep)
{
const float a = de::min(x, xBounds.y());
const float b = de::min(y, yBounds.y());
const Vec4 c0 = quad0.p00*(1.0f-a)*(1.0f-b) + quad0.p10*a*(1.0f-b) + quad0.p01*(1.0f-a)*b + quad0.p11*a*b;
const Vec4 c1 = quad1.p00*(1.0f-a)*(1.0f-b) + quad1.p10*a*(1.0f-b) + quad1.p01*(1.0f-a)*b + quad1.p11*a*b;
if (isLinearRangeValid(prec, c0, c1, zBounds, result))
return true;
}
}
return false;
}
static bool isReductionValid (const LookupPrecision& prec,
const Vec4& c0,
const Vec4& c1,
tcu::Sampler::ReductionMode reductionMode,
const Vec4& result)
{
DE_ASSERT(reductionMode == tcu::Sampler::MIN || reductionMode == tcu::Sampler::MAX);
const Vec4 color = (reductionMode == tcu::Sampler::MIN ? tcu::min(c0, c1) : tcu::max(c0, c1));
return isColorValid(prec, color, result);
}
static bool isReductionValid (const LookupPrecision& prec,
const ColorQuad& quad,
tcu::Sampler::ReductionMode reductionMode,
const Vec4& result)
{
DE_ASSERT(reductionMode == tcu::Sampler::MIN || reductionMode == tcu::Sampler::MAX);
const Vec4 c0 = (reductionMode == tcu::Sampler::MIN ? tcu::min(quad.p00, quad.p01) : tcu::max(quad.p00, quad.p01));
const Vec4 c1 = (reductionMode == tcu::Sampler::MIN ? tcu::min(quad.p10, quad.p11) : tcu::max(quad.p10, quad.p11));
return isReductionValid(prec, c0, c1, reductionMode, result);
}
static bool isReductionValid (const LookupPrecision& prec,
const ColorQuad& quad0,
const ColorQuad& quad1,
tcu::Sampler::ReductionMode reductionMode,
const Vec4& result)
{
DE_ASSERT(reductionMode == tcu::Sampler::MIN || reductionMode == tcu::Sampler::MAX);
const ColorQuad quad =
{
reductionMode == tcu::Sampler::MIN ? tcu::min(quad0.p00, quad1.p00) : tcu::max(quad0.p00, quad1.p00), // p00
reductionMode == tcu::Sampler::MIN ? tcu::min(quad0.p01, quad1.p01) : tcu::max(quad0.p01, quad1.p01), // p01
reductionMode == tcu::Sampler::MIN ? tcu::min(quad0.p10, quad1.p10) : tcu::max(quad0.p10, quad1.p10), // p10
reductionMode == tcu::Sampler::MIN ? tcu::min(quad0.p11, quad1.p11) : tcu::max(quad0.p11, quad1.p11), // p11
};
return isReductionValid(prec, quad, reductionMode, result);
}
static bool is1DTrilinearFilterResultValid (const LookupPrecision& prec,
const ColorLine& line0,
const ColorLine& line1,
const Vec2& xBounds0,
const Vec2& xBounds1,
const Vec2& zBounds,
const float searchStep,
const Vec4& result)
{
DE_ASSERT(xBounds0.x() <= xBounds0.y());
DE_ASSERT(xBounds1.x() <= xBounds1.y());
DE_ASSERT(xBounds0.x() + searchStep > xBounds0.x()); // step is not effectively 0
DE_ASSERT(xBounds0.y() + searchStep > xBounds0.y());
DE_ASSERT(xBounds1.x() + searchStep > xBounds1.x());
DE_ASSERT(xBounds1.y() + searchStep > xBounds1.y());
if (!isInColorBounds(prec, line0, line1, result))
return false;
for (float x0 = xBounds0.x(); x0 < xBounds0.y()+searchStep; x0 += searchStep)
{
const float a0 = de::min(x0, xBounds0.y());
const Vec4 c0 = line0.p0*(1.0f-a0) + line0.p1*a0;
for (float x1 = xBounds1.x(); x1 <= xBounds1.y(); x1 += searchStep)
{
const float a1 = de::min(x1, xBounds1.y());
const Vec4 c1 = line1.p0*(1.0f-a1) + line1.p1*a1;
if (isLinearRangeValid(prec, c0, c1, zBounds, result))
return true;
}
}
return false;
}
static bool is2DTrilinearFilterResultValid (const LookupPrecision& prec,
const ColorQuad& quad0,
const ColorQuad& quad1,
const Vec2& xBounds0,
const Vec2& yBounds0,
const Vec2& xBounds1,
const Vec2& yBounds1,
const Vec2& zBounds,
const float searchStep,
const Vec4& result)
{
DE_ASSERT(xBounds0.x() <= xBounds0.y());
DE_ASSERT(yBounds0.x() <= yBounds0.y());
DE_ASSERT(xBounds1.x() <= xBounds1.y());
DE_ASSERT(yBounds1.x() <= yBounds1.y());
DE_ASSERT(xBounds0.x() + searchStep > xBounds0.x()); // step is not effectively 0
DE_ASSERT(xBounds0.y() + searchStep > xBounds0.y());
DE_ASSERT(yBounds0.x() + searchStep > yBounds0.x());
DE_ASSERT(yBounds0.y() + searchStep > yBounds0.y());
DE_ASSERT(xBounds1.x() + searchStep > xBounds1.x());
DE_ASSERT(xBounds1.y() + searchStep > xBounds1.y());
DE_ASSERT(yBounds1.x() + searchStep > yBounds1.x());
DE_ASSERT(yBounds1.y() + searchStep > yBounds1.y());
if (!isInColorBounds(prec, quad0, quad1, result))
return false;
for (float x0 = xBounds0.x(); x0 < xBounds0.y()+searchStep; x0 += searchStep)
{
for (float y0 = yBounds0.x(); y0 < yBounds0.y()+searchStep; y0 += searchStep)
{
const float a0 = de::min(x0, xBounds0.y());
const float b0 = de::min(y0, yBounds0.y());
const Vec4 c0 = quad0.p00*(1.0f-a0)*(1.0f-b0) + quad0.p10*a0*(1.0f-b0) + quad0.p01*(1.0f-a0)*b0 + quad0.p11*a0*b0;
for (float x1 = xBounds1.x(); x1 <= xBounds1.y(); x1 += searchStep)
{
for (float y1 = yBounds1.x(); y1 <= yBounds1.y(); y1 += searchStep)
{
const float a1 = de::min(x1, xBounds1.y());
const float b1 = de::min(y1, yBounds1.y());
const Vec4 c1 = quad1.p00*(1.0f-a1)*(1.0f-b1) + quad1.p10*a1*(1.0f-b1) + quad1.p01*(1.0f-a1)*b1 + quad1.p11*a1*b1;
if (isLinearRangeValid(prec, c0, c1, zBounds, result))
return true;
}
}
}
}
return false;
}
static bool is3DTrilinearFilterResultValid (const LookupPrecision& prec,
const ColorQuad& quad00,
const ColorQuad& quad01,
const ColorQuad& quad10,
const ColorQuad& quad11,
const Vec2& xBounds0,
const Vec2& yBounds0,
const Vec2& zBounds0,
const Vec2& xBounds1,
const Vec2& yBounds1,
const Vec2& zBounds1,
const Vec2& wBounds,
const float searchStep,
const Vec4& result)
{
DE_ASSERT(xBounds0.x() <= xBounds0.y());
DE_ASSERT(yBounds0.x() <= yBounds0.y());
DE_ASSERT(zBounds0.x() <= zBounds0.y());
DE_ASSERT(xBounds1.x() <= xBounds1.y());
DE_ASSERT(yBounds1.x() <= yBounds1.y());
DE_ASSERT(zBounds1.x() <= zBounds1.y());
DE_ASSERT(xBounds0.x() + searchStep > xBounds0.x()); // step is not effectively 0
DE_ASSERT(xBounds0.y() + searchStep > xBounds0.y());
DE_ASSERT(yBounds0.x() + searchStep > yBounds0.x());
DE_ASSERT(yBounds0.y() + searchStep > yBounds0.y());
DE_ASSERT(zBounds0.x() + searchStep > zBounds0.x());
DE_ASSERT(zBounds0.y() + searchStep > zBounds0.y());
DE_ASSERT(xBounds1.x() + searchStep > xBounds1.x());
DE_ASSERT(xBounds1.y() + searchStep > xBounds1.y());
DE_ASSERT(yBounds1.x() + searchStep > yBounds1.x());
DE_ASSERT(yBounds1.y() + searchStep > yBounds1.y());
DE_ASSERT(zBounds1.x() + searchStep > zBounds1.x());
DE_ASSERT(zBounds1.y() + searchStep > zBounds1.y());
if (!isInColorBounds(prec, quad00, quad01, quad10, quad11, result))
return false;
for (float x0 = xBounds0.x(); x0 < xBounds0.y()+searchStep; x0 += searchStep)
{
for (float y0 = yBounds0.x(); y0 < yBounds0.y()+searchStep; y0 += searchStep)
{
const float a0 = de::min(x0, xBounds0.y());
const float b0 = de::min(y0, yBounds0.y());
const Vec4 c00 = quad00.p00*(1.0f-a0)*(1.0f-b0) + quad00.p10*a0*(1.0f-b0) + quad00.p01*(1.0f-a0)*b0 + quad00.p11*a0*b0;
const Vec4 c01 = quad01.p00*(1.0f-a0)*(1.0f-b0) + quad01.p10*a0*(1.0f-b0) + quad01.p01*(1.0f-a0)*b0 + quad01.p11*a0*b0;
for (float z0 = zBounds0.x(); z0 < zBounds0.y()+searchStep; z0 += searchStep)
{
const float c0 = de::min(z0, zBounds0.y());
const Vec4 cz0 = c00*(1.0f-c0) + c01*c0;
for (float x1 = xBounds1.x(); x1 < xBounds1.y()+searchStep; x1 += searchStep)
{
for (float y1 = yBounds1.x(); y1 < yBounds1.y()+searchStep; y1 += searchStep)
{
const float a1 = de::min(x1, xBounds1.y());
const float b1 = de::min(y1, yBounds1.y());
const Vec4 c10 = quad10.p00*(1.0f-a1)*(1.0f-b1) + quad10.p10*a1*(1.0f-b1) + quad10.p01*(1.0f-a1)*b1 + quad10.p11*a1*b1;
const Vec4 c11 = quad11.p00*(1.0f-a1)*(1.0f-b1) + quad11.p10*a1*(1.0f-b1) + quad11.p01*(1.0f-a1)*b1 + quad11.p11*a1*b1;
for (float z1 = zBounds1.x(); z1 < zBounds1.y()+searchStep; z1 += searchStep)
{
const float c1 = de::min(z1, zBounds1.y());
const Vec4 cz1 = c10*(1.0f - c1) + c11*c1;
if (isLinearRangeValid(prec, cz0, cz1, wBounds, result))
return true;
}
}
}
}
}
}
return false;
}
template<typename PrecType, typename ScalarType>
static bool isNearestSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const PrecType& prec,
const float coordX,
const int coordY,
const Vector<ScalarType, 4>& result)
{
DE_ASSERT(level.getDepth() == 1);
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coordX, prec.coordBits.x(), prec.uvwBits.x());
const int minI = deFloorFloatToInt32(uBounds.x());
const int maxI = deFloorFloatToInt32(uBounds.y());
for (int i = minI; i <= maxI; i++)
{
const int x = wrap(sampler.wrapS, i, level.getWidth());
const Vector<ScalarType, 4> color = lookup<ScalarType>(level, sampler, x, coordY, 0);
if (isColorValid(prec, color, result))
return true;
}
return false;
}
template<typename PrecType, typename ScalarType>
static bool isNearestSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const PrecType& prec,
const Vec2& coord,
const int coordZ,
const Vector<ScalarType, 4>& result)
{
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getHeight(), coord.y(), prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinates - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x());
const int maxI = deFloorFloatToInt32(uBounds.y());
const int minJ = deFloorFloatToInt32(vBounds.x());
const int maxJ = deFloorFloatToInt32(vBounds.y());
// \todo [2013-07-03 pyry] This could be optimized by first computing ranges based on wrap mode.
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
const int x = wrap(sampler.wrapS, i, level.getWidth());
const int y = wrap(sampler.wrapT, j, level.getHeight());
const Vector<ScalarType, 4> color = lookup<ScalarType>(level, sampler, x, y, coordZ);
if (isColorValid(prec, color, result))
return true;
}
}
return false;
}
template<typename PrecType, typename ScalarType>
static bool isNearestSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const PrecType& prec,
const Vec3& coord,
const Vector<ScalarType, 4>& result)
{
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getHeight(), coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 wBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getDepth(), coord.z(), prec.coordBits.z(), prec.uvwBits.z());
// Integer coordinates - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x());
const int maxI = deFloorFloatToInt32(uBounds.y());
const int minJ = deFloorFloatToInt32(vBounds.x());
const int maxJ = deFloorFloatToInt32(vBounds.y());
const int minK = deFloorFloatToInt32(wBounds.x());
const int maxK = deFloorFloatToInt32(wBounds.y());
// \todo [2013-07-03 pyry] This could be optimized by first computing ranges based on wrap mode.
for (int k = minK; k <= maxK; k++)
{
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
const int x = wrap(sampler.wrapS, i, level.getWidth());
const int y = wrap(sampler.wrapT, j, level.getHeight());
const int z = wrap(sampler.wrapR, k, level.getDepth());
const Vector<ScalarType, 4> color = lookup<ScalarType>(level, sampler, x, y, z);
if (isColorValid(prec, color, result))
return true;
}
}
}
return false;
}
bool isLinearSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const LookupPrecision& prec,
const float coordX,
const int coordY,
const Vec4& result)
{
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coordX, prec.coordBits.x(), prec.uvwBits.x());
const int minI = deFloorFloatToInt32(uBounds.x()-0.5f);
const int maxI = deFloorFloatToInt32(uBounds.y()-0.5f);
const int w = level.getWidth();
const TextureFormat format = level.getFormat();
const TextureChannelClass texClass = getTextureChannelClass(format.type);
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
DE_UNREF(texClass);
DE_UNREF(format);
for (int i = minI; i <= maxI; i++)
{
// Wrapped coordinates
const int x0 = wrap(sampler.wrapS, i , w);
const int x1 = wrap(sampler.wrapS, i+1, w);
// Bounds for filtering factors
const float minA = de::clamp((uBounds.x()-0.5f)-float(i), 0.0f, 1.0f);
const float maxA = de::clamp((uBounds.y()-0.5f)-float(i), 0.0f, 1.0f);
const Vec4 colorA = lookup<float>(level, sampler, x0, coordY, 0);
const Vec4 colorB = lookup<float>(level, sampler, x1, coordY, 0);
if (sampler.reductionMode == Sampler::WEIGHTED_AVERAGE)
{
if (isLinearRangeValid(prec, colorA, colorB, Vec2(minA, maxA), result))
return true;
}
else
{
if (isReductionValid(prec, colorA, colorB, sampler.reductionMode, result))
return true;
}
}
return false;
}
bool isLinearSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const Vec4& result)
{
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getHeight(), coord.y(), prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinate bounds for (x0,y0) - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x()-0.5f);
const int maxI = deFloorFloatToInt32(uBounds.y()-0.5f);
const int minJ = deFloorFloatToInt32(vBounds.x()-0.5f);
const int maxJ = deFloorFloatToInt32(vBounds.y()-0.5f);
const int w = level.getWidth();
const int h = level.getHeight();
const TextureChannelClass texClass = getTextureChannelClass(level.getFormat().type);
float searchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
// \todo [2013-07-03 pyry] This could be optimized by first computing ranges based on wrap mode.
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
// Wrapped coordinates
const int x0 = wrap(sampler.wrapS, i , w);
const int x1 = wrap(sampler.wrapS, i+1, w);
const int y0 = wrap(sampler.wrapT, j , h);
const int y1 = wrap(sampler.wrapT, j+1, h);
// Bounds for filtering factors
const float minA = de::clamp((uBounds.x()-0.5f)-float(i), 0.0f, 1.0f);
const float maxA = de::clamp((uBounds.y()-0.5f)-float(i), 0.0f, 1.0f);
const float minB = de::clamp((vBounds.x()-0.5f)-float(j), 0.0f, 1.0f);
const float maxB = de::clamp((vBounds.y()-0.5f)-float(j), 0.0f, 1.0f);
ColorQuad quad;
lookupQuad(quad, level, sampler, x0, x1, y0, y1, coordZ);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep = computeBilinearSearchStepFromFloatQuad(prec, quad);
if (sampler.reductionMode == Sampler::WEIGHTED_AVERAGE)
{
if (isBilinearRangeValid(prec, quad, Vec2(minA, maxA), Vec2(minB, maxB), searchStep, result))
return true;
}
else
{
if (isReductionValid(prec, quad, sampler.reductionMode, result))
return true;
}
}
}
return false;
}
static bool isLinearSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec3& coord,
const Vec4& result)
{
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getHeight(), coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 wBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getDepth(), coord.z(), prec.coordBits.z(), prec.uvwBits.z());
// Integer coordinate bounds for (x0,y0) - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x()-0.5f);
const int maxI = deFloorFloatToInt32(uBounds.y()-0.5f);
const int minJ = deFloorFloatToInt32(vBounds.x()-0.5f);
const int maxJ = deFloorFloatToInt32(vBounds.y()-0.5f);
const int minK = deFloorFloatToInt32(wBounds.x()-0.5f);
const int maxK = deFloorFloatToInt32(wBounds.y()-0.5f);
const int w = level.getWidth();
const int h = level.getHeight();
const int d = level.getDepth();
const TextureChannelClass texClass = getTextureChannelClass(level.getFormat().type);
float searchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
// \todo [2013-07-03 pyry] This could be optimized by first computing ranges based on wrap mode.
for (int k = minK; k <= maxK; k++)
{
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
// Wrapped coordinates
const int x0 = wrap(sampler.wrapS, i , w);
const int x1 = wrap(sampler.wrapS, i+1, w);
const int y0 = wrap(sampler.wrapT, j , h);
const int y1 = wrap(sampler.wrapT, j+1, h);
const int z0 = wrap(sampler.wrapR, k , d);
const int z1 = wrap(sampler.wrapR, k+1, d);
// Bounds for filtering factors
const float minA = de::clamp((uBounds.x()-0.5f)-float(i), 0.0f, 1.0f);
const float maxA = de::clamp((uBounds.y()-0.5f)-float(i), 0.0f, 1.0f);
const float minB = de::clamp((vBounds.x()-0.5f)-float(j), 0.0f, 1.0f);
const float maxB = de::clamp((vBounds.y()-0.5f)-float(j), 0.0f, 1.0f);
const float minC = de::clamp((wBounds.x()-0.5f)-float(k), 0.0f, 1.0f);
const float maxC = de::clamp((wBounds.y()-0.5f)-float(k), 0.0f, 1.0f);
ColorQuad quad0, quad1;
lookupQuad(quad0, level, sampler, x0, x1, y0, y1, z0);
lookupQuad(quad1, level, sampler, x0, x1, y0, y1, z1);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep = de::min(computeBilinearSearchStepFromFloatQuad(prec, quad0), computeBilinearSearchStepFromFloatQuad(prec, quad1));
if (sampler.reductionMode == Sampler::WEIGHTED_AVERAGE)
{
if (isTrilinearRangeValid(prec, quad0, quad1, Vec2(minA, maxA), Vec2(minB, maxB), Vec2(minC, maxC), searchStep, result))
return true;
}
else
{
if (isReductionValid(prec, quad0, quad1, sampler.reductionMode, result))
return true;
}
}
}
}
return false;
}
static bool isNearestMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const LookupPrecision& prec,
const float coord,
const int coordY,
const Vec2& fBounds,
const Vec4& result)
{
const int w0 = level0.getWidth();
const int w1 = level1.getWidth();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w0, coord, prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w1, coord, prec.coordBits.x(), prec.uvwBits.x());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x());
const int maxI0 = deFloorFloatToInt32(uBounds0.y());
const int minI1 = deFloorFloatToInt32(uBounds1.x());
const int maxI1 = deFloorFloatToInt32(uBounds1.y());
for (int i0 = minI0; i0 <= maxI0; i0++)
{
for (int i1 = minI1; i1 <= maxI1; i1++)
{
const Vec4 c0 = lookup<float>(level0, sampler, wrap(sampler.wrapS, i0, w0), coordY, 0);
const Vec4 c1 = lookup<float>(level1, sampler, wrap(sampler.wrapS, i1, w1), coordY, 0);
if (isLinearRangeValid(prec, c0, c1, fBounds, result))
return true;
}
}
return false;
}
static bool isNearestMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const Vec2& fBounds,
const Vec4& result)
{
const int w0 = level0.getWidth();
const int w1 = level1.getWidth();
const int h0 = level0.getHeight();
const int h1 = level1.getHeight();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w0, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w1, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h0, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 vBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h1, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x());
const int maxI0 = deFloorFloatToInt32(uBounds0.y());
const int minI1 = deFloorFloatToInt32(uBounds1.x());
const int maxI1 = deFloorFloatToInt32(uBounds1.y());
const int minJ0 = deFloorFloatToInt32(vBounds0.x());
const int maxJ0 = deFloorFloatToInt32(vBounds0.y());
const int minJ1 = deFloorFloatToInt32(vBounds1.x());
const int maxJ1 = deFloorFloatToInt32(vBounds1.y());
for (int j0 = minJ0; j0 <= maxJ0; j0++)
{
for (int i0 = minI0; i0 <= maxI0; i0++)
{
for (int j1 = minJ1; j1 <= maxJ1; j1++)
{
for (int i1 = minI1; i1 <= maxI1; i1++)
{
const Vec4 c0 = lookup<float>(level0, sampler, wrap(sampler.wrapS, i0, w0), wrap(sampler.wrapT, j0, h0), coordZ);
const Vec4 c1 = lookup<float>(level1, sampler, wrap(sampler.wrapS, i1, w1), wrap(sampler.wrapT, j1, h1), coordZ);
if (isLinearRangeValid(prec, c0, c1, fBounds, result))
return true;
}
}
}
}
return false;
}
static bool isNearestMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec3& coord,
const Vec2& fBounds,
const Vec4& result)
{
const int w0 = level0.getWidth();
const int w1 = level1.getWidth();
const int h0 = level0.getHeight();
const int h1 = level1.getHeight();
const int d0 = level0.getDepth();
const int d1 = level1.getDepth();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w0, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w1, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h0, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 vBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h1, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 wBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, d0, coord.z(), prec.coordBits.z(), prec.uvwBits.z());
const Vec2 wBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, d1, coord.z(), prec.coordBits.z(), prec.uvwBits.z());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x());
const int maxI0 = deFloorFloatToInt32(uBounds0.y());
const int minI1 = deFloorFloatToInt32(uBounds1.x());
const int maxI1 = deFloorFloatToInt32(uBounds1.y());
const int minJ0 = deFloorFloatToInt32(vBounds0.x());
const int maxJ0 = deFloorFloatToInt32(vBounds0.y());
const int minJ1 = deFloorFloatToInt32(vBounds1.x());
const int maxJ1 = deFloorFloatToInt32(vBounds1.y());
const int minK0 = deFloorFloatToInt32(wBounds0.x());
const int maxK0 = deFloorFloatToInt32(wBounds0.y());
const int minK1 = deFloorFloatToInt32(wBounds1.x());
const int maxK1 = deFloorFloatToInt32(wBounds1.y());
for (int k0 = minK0; k0 <= maxK0; k0++)
{
for (int j0 = minJ0; j0 <= maxJ0; j0++)
{
for (int i0 = minI0; i0 <= maxI0; i0++)
{
for (int k1 = minK1; k1 <= maxK1; k1++)
{
for (int j1 = minJ1; j1 <= maxJ1; j1++)
{
for (int i1 = minI1; i1 <= maxI1; i1++)
{
const Vec4 c0 = lookup<float>(level0, sampler, wrap(sampler.wrapS, i0, w0), wrap(sampler.wrapT, j0, h0), wrap(sampler.wrapR, k0, d0));
const Vec4 c1 = lookup<float>(level1, sampler, wrap(sampler.wrapS, i1, w1), wrap(sampler.wrapT, j1, h1), wrap(sampler.wrapR, k1, d1));
if (isLinearRangeValid(prec, c0, c1, fBounds, result))
return true;
}
}
}
}
}
}
return false;
}
static bool isLinearMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const LookupPrecision& prec,
const float coordX,
const int coordY,
const Vec2& fBounds,
const Vec4& result)
{
// \todo [2013-07-04 pyry] This is strictly not correct as coordinates between levels should be dependent.
// Right now this allows pairing any two valid bilinear quads.
const int w0 = level0.getWidth();
const int w1 = level1.getWidth();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w0, coordX, prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w1, coordX, prec.coordBits.x(), prec.uvwBits.x());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x()-0.5f);
const int maxI0 = deFloorFloatToInt32(uBounds0.y()-0.5f);
const int minI1 = deFloorFloatToInt32(uBounds1.x()-0.5f);
const int maxI1 = deFloorFloatToInt32(uBounds1.y()-0.5f);
const TextureChannelClass texClass = getTextureChannelClass(level0.getFormat().type);
const float cSearchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
for (int i0 = minI0; i0 <= maxI0; i0++)
{
ColorLine line0;
float searchStep0;
{
const int x0 = wrap(sampler.wrapS, i0 , w0);
const int x1 = wrap(sampler.wrapS, i0+1, w0);
lookupLine(line0, level0, sampler, x0, x1, coordY);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep0 = computeBilinearSearchStepFromFloatLine(prec, line0);
else
searchStep0 = cSearchStep;
}
const float minA0 = de::clamp((uBounds0.x()-0.5f)-float(i0), 0.0f, 1.0f);
const float maxA0 = de::clamp((uBounds0.y()-0.5f)-float(i0), 0.0f, 1.0f);
for (int i1 = minI1; i1 <= maxI1; i1++)
{
ColorLine line1;
float searchStep1;
{
const int x0 = wrap(sampler.wrapS, i1 , w1);
const int x1 = wrap(sampler.wrapS, i1+1, w1);
lookupLine(line1, level1, sampler, x0, x1, coordY);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep1 = computeBilinearSearchStepFromFloatLine(prec, line1);
else
searchStep1 = cSearchStep;
}
const float minA1 = de::clamp((uBounds1.x()-0.5f)-float(i1), 0.0f, 1.0f);
const float maxA1 = de::clamp((uBounds1.y()-0.5f)-float(i1), 0.0f, 1.0f);
if (is1DTrilinearFilterResultValid(prec, line0, line1, Vec2(minA0, maxA0), Vec2(minA1, maxA1), fBounds, de::min(searchStep0, searchStep1), result))
return true;
}
}
return false;
}
static bool isLinearMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const Vec2& fBounds,
const Vec4& result)
{
// \todo [2013-07-04 pyry] This is strictly not correct as coordinates between levels should be dependent.
// Right now this allows pairing any two valid bilinear quads.
const int w0 = level0.getWidth();
const int w1 = level1.getWidth();
const int h0 = level0.getHeight();
const int h1 = level1.getHeight();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w0, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w1, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h0, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 vBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h1, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x()-0.5f);
const int maxI0 = deFloorFloatToInt32(uBounds0.y()-0.5f);
const int minI1 = deFloorFloatToInt32(uBounds1.x()-0.5f);
const int maxI1 = deFloorFloatToInt32(uBounds1.y()-0.5f);
const int minJ0 = deFloorFloatToInt32(vBounds0.x()-0.5f);
const int maxJ0 = deFloorFloatToInt32(vBounds0.y()-0.5f);
const int minJ1 = deFloorFloatToInt32(vBounds1.x()-0.5f);
const int maxJ1 = deFloorFloatToInt32(vBounds1.y()-0.5f);
const TextureChannelClass texClass = getTextureChannelClass(level0.getFormat().type);
const float cSearchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
for (int j0 = minJ0; j0 <= maxJ0; j0++)
{
for (int i0 = minI0; i0 <= maxI0; i0++)
{
ColorQuad quad0;
float searchStep0;
{
const int x0 = wrap(sampler.wrapS, i0 , w0);
const int x1 = wrap(sampler.wrapS, i0+1, w0);
const int y0 = wrap(sampler.wrapT, j0 , h0);
const int y1 = wrap(sampler.wrapT, j0+1, h0);
lookupQuad(quad0, level0, sampler, x0, x1, y0, y1, coordZ);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep0 = computeBilinearSearchStepFromFloatQuad(prec, quad0);
else
searchStep0 = cSearchStep;
}
const float minA0 = de::clamp((uBounds0.x()-0.5f)-float(i0), 0.0f, 1.0f);
const float maxA0 = de::clamp((uBounds0.y()-0.5f)-float(i0), 0.0f, 1.0f);
const float minB0 = de::clamp((vBounds0.x()-0.5f)-float(j0), 0.0f, 1.0f);
const float maxB0 = de::clamp((vBounds0.y()-0.5f)-float(j0), 0.0f, 1.0f);
for (int j1 = minJ1; j1 <= maxJ1; j1++)
{
for (int i1 = minI1; i1 <= maxI1; i1++)
{
ColorQuad quad1;
float searchStep1;
{
const int x0 = wrap(sampler.wrapS, i1 , w1);
const int x1 = wrap(sampler.wrapS, i1+1, w1);
const int y0 = wrap(sampler.wrapT, j1 , h1);
const int y1 = wrap(sampler.wrapT, j1+1, h1);
lookupQuad(quad1, level1, sampler, x0, x1, y0, y1, coordZ);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep1 = computeBilinearSearchStepFromFloatQuad(prec, quad1);
else
searchStep1 = cSearchStep;
}
const float minA1 = de::clamp((uBounds1.x()-0.5f)-float(i1), 0.0f, 1.0f);
const float maxA1 = de::clamp((uBounds1.y()-0.5f)-float(i1), 0.0f, 1.0f);
const float minB1 = de::clamp((vBounds1.x()-0.5f)-float(j1), 0.0f, 1.0f);
const float maxB1 = de::clamp((vBounds1.y()-0.5f)-float(j1), 0.0f, 1.0f);
if (is2DTrilinearFilterResultValid(prec, quad0, quad1, Vec2(minA0, maxA0), Vec2(minB0, maxB0), Vec2(minA1, maxA1), Vec2(minB1, maxB1),
fBounds, de::min(searchStep0, searchStep1), result))
return true;
}
}
}
}
return false;
}
static bool isLinearMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec3& coord,
const Vec2& fBounds,
const Vec4& result)
{
// \todo [2013-07-04 pyry] This is strictly not correct as coordinates between levels should be dependent.
// Right now this allows pairing any two valid bilinear quads.
const int w0 = level0.getWidth();
const int w1 = level1.getWidth();
const int h0 = level0.getHeight();
const int h1 = level1.getHeight();
const int d0 = level0.getDepth();
const int d1 = level1.getDepth();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w0, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, w1, coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h0, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 vBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, h1, coord.y(), prec.coordBits.y(), prec.uvwBits.y());
const Vec2 wBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, d0, coord.z(), prec.coordBits.z(), prec.uvwBits.z());
const Vec2 wBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, d1, coord.z(), prec.coordBits.z(), prec.uvwBits.z());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x()-0.5f);
const int maxI0 = deFloorFloatToInt32(uBounds0.y()-0.5f);
const int minI1 = deFloorFloatToInt32(uBounds1.x()-0.5f);
const int maxI1 = deFloorFloatToInt32(uBounds1.y()-0.5f);
const int minJ0 = deFloorFloatToInt32(vBounds0.x()-0.5f);
const int maxJ0 = deFloorFloatToInt32(vBounds0.y()-0.5f);
const int minJ1 = deFloorFloatToInt32(vBounds1.x()-0.5f);
const int maxJ1 = deFloorFloatToInt32(vBounds1.y()-0.5f);
const int minK0 = deFloorFloatToInt32(wBounds0.x()-0.5f);
const int maxK0 = deFloorFloatToInt32(wBounds0.y()-0.5f);
const int minK1 = deFloorFloatToInt32(wBounds1.x()-0.5f);
const int maxK1 = deFloorFloatToInt32(wBounds1.y()-0.5f);
const TextureChannelClass texClass = getTextureChannelClass(level0.getFormat().type);
const float cSearchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
for (int k0 = minK0; k0 <= maxK0; k0++)
{
for (int j0 = minJ0; j0 <= maxJ0; j0++)
{
for (int i0 = minI0; i0 <= maxI0; i0++)
{
ColorQuad quad00, quad01;
float searchStep0;
{
const int x0 = wrap(sampler.wrapS, i0 , w0);
const int x1 = wrap(sampler.wrapS, i0+1, w0);
const int y0 = wrap(sampler.wrapT, j0 , h0);
const int y1 = wrap(sampler.wrapT, j0+1, h0);
const int z0 = wrap(sampler.wrapR, k0 , d0);
const int z1 = wrap(sampler.wrapR, k0+1, d0);
lookupQuad(quad00, level0, sampler, x0, x1, y0, y1, z0);
lookupQuad(quad01, level0, sampler, x0, x1, y0, y1, z1);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep0 = de::min(computeBilinearSearchStepFromFloatQuad(prec, quad00), computeBilinearSearchStepFromFloatQuad(prec, quad01));
else
searchStep0 = cSearchStep;
}
const float minA0 = de::clamp((uBounds0.x()-0.5f)-float(i0), 0.0f, 1.0f);
const float maxA0 = de::clamp((uBounds0.y()-0.5f)-float(i0), 0.0f, 1.0f);
const float minB0 = de::clamp((vBounds0.x()-0.5f)-float(j0), 0.0f, 1.0f);
const float maxB0 = de::clamp((vBounds0.y()-0.5f)-float(j0), 0.0f, 1.0f);
const float minC0 = de::clamp((wBounds0.x()-0.5f)-float(k0), 0.0f, 1.0f);
const float maxC0 = de::clamp((wBounds0.y()-0.5f)-float(k0), 0.0f, 1.0f);
for (int k1 = minK1; k1 <= maxK1; k1++)
{
for (int j1 = minJ1; j1 <= maxJ1; j1++)
{
for (int i1 = minI1; i1 <= maxI1; i1++)
{
ColorQuad quad10, quad11;
float searchStep1;
{
const int x0 = wrap(sampler.wrapS, i1 , w1);
const int x1 = wrap(sampler.wrapS, i1+1, w1);
const int y0 = wrap(sampler.wrapT, j1 , h1);
const int y1 = wrap(sampler.wrapT, j1+1, h1);
const int z0 = wrap(sampler.wrapR, k1 , d1);
const int z1 = wrap(sampler.wrapR, k1+1, d1);
lookupQuad(quad10, level1, sampler, x0, x1, y0, y1, z0);
lookupQuad(quad11, level1, sampler, x0, x1, y0, y1, z1);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep1 = de::min(computeBilinearSearchStepFromFloatQuad(prec, quad10), computeBilinearSearchStepFromFloatQuad(prec, quad11));
else
searchStep1 = cSearchStep;
}
const float minA1 = de::clamp((uBounds1.x()-0.5f)-float(i1), 0.0f, 1.0f);
const float maxA1 = de::clamp((uBounds1.y()-0.5f)-float(i1), 0.0f, 1.0f);
const float minB1 = de::clamp((vBounds1.x()-0.5f)-float(j1), 0.0f, 1.0f);
const float maxB1 = de::clamp((vBounds1.y()-0.5f)-float(j1), 0.0f, 1.0f);
const float minC1 = de::clamp((wBounds1.x()-0.5f)-float(k1), 0.0f, 1.0f);
const float maxC1 = de::clamp((wBounds1.y()-0.5f)-float(k1), 0.0f, 1.0f);
if (is3DTrilinearFilterResultValid(prec, quad00, quad01, quad10, quad11,
Vec2(minA0, maxA0), Vec2(minB0, maxB0), Vec2(minC0, maxC0),
Vec2(minA1, maxA1), Vec2(minB1, maxB1), Vec2(minC1, maxC1),
fBounds, de::min(searchStep0, searchStep1), result))
return true;
}
}
}
}
}
}
return false;
}
static bool isLevelSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const Sampler::FilterMode filterMode,
const LookupPrecision& prec,
const float coordX,
const int coordY,
const Vec4& result)
{
if (filterMode == Sampler::LINEAR)
return isLinearSampleResultValid(level, sampler, prec, coordX, coordY, result);
else
return isNearestSampleResultValid(level, sampler, prec, coordX, coordY, result);
}
static bool isLevelSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const Sampler::FilterMode filterMode,
const LookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const Vec4& result)
{
if (filterMode == Sampler::LINEAR)
return isLinearSampleResultValid(level, sampler, prec, coord, coordZ, result);
else
return isNearestSampleResultValid(level, sampler, prec, coord, coordZ, result);
}
static bool isMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const Sampler::FilterMode levelFilter,
const LookupPrecision& prec,
const float coordX,
const int coordY,
const Vec2& fBounds,
const Vec4& result)
{
if (levelFilter == Sampler::LINEAR)
return isLinearMipmapLinearSampleResultValid(level0, level1, sampler, prec, coordX, coordY, fBounds, result);
else
return isNearestMipmapLinearSampleResultValid(level0, level1, sampler, prec, coordX, coordY, fBounds, result);
}
static bool isMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const Sampler::FilterMode levelFilter,
const LookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const Vec2& fBounds,
const Vec4& result)
{
if (levelFilter == Sampler::LINEAR)
return isLinearMipmapLinearSampleResultValid(level0, level1, sampler, prec, coord, coordZ, fBounds, result);
else
return isNearestMipmapLinearSampleResultValid(level0, level1, sampler, prec, coord, coordZ, fBounds, result);
}
bool isLookupResultValid (const Texture2DView& texture, const Sampler& sampler, const LookupPrecision& prec, const Vec2& coord, const Vec2& lodBounds, const Vec4& result)
{
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
DE_ASSERT(isSamplerSupported(sampler));
if (canBeMagnified)
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.magFilter, prec, coord, 0, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
const float minF = de::clamp(minLod - float(level), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(level), 0.0f, 1.0f);
if (isMipmapLinearSampleResultValid(texture.getLevel(level), texture.getLevel(level+1), sampler, getLevelFilter(sampler.minFilter), prec, coord, 0, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
if (isLevelSampleResultValid(texture.getLevel(level), sampler, getLevelFilter(sampler.minFilter), prec, coord, 0, result))
return true;
}
}
else
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.minFilter, prec, coord, 0, result))
return true;
}
}
return false;
}
bool isLookupResultValid (const Texture1DView& texture, const Sampler& sampler, const LookupPrecision& prec, const float coord, const Vec2& lodBounds, const Vec4& result)
{
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
DE_ASSERT(isSamplerSupported(sampler));
if (canBeMagnified)
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.magFilter, prec, coord, 0, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
const float minF = de::clamp(minLod - float(level), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(level), 0.0f, 1.0f);
if (isMipmapLinearSampleResultValid(texture.getLevel(level), texture.getLevel(level+1), sampler, getLevelFilter(sampler.minFilter), prec, coord, 0, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
if (isLevelSampleResultValid(texture.getLevel(level), sampler, getLevelFilter(sampler.minFilter), prec, coord, 0, result))
return true;
}
}
else
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.minFilter, prec, coord, 0, result))
return true;
}
}
return false;
}
static bool isSeamlessLinearSampleResultValid (const ConstPixelBufferAccess (&faces)[CUBEFACE_LAST],
const Sampler& sampler,
const LookupPrecision& prec,
const CubeFaceFloatCoords& coords,
const Vec4& result)
{
const int size = faces[coords.face].getWidth();
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size, coords.s, prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size, coords.t, prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinate bounds for (x0,y0) - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x()-0.5f);
const int maxI = deFloorFloatToInt32(uBounds.y()-0.5f);
const int minJ = deFloorFloatToInt32(vBounds.x()-0.5f);
const int maxJ = deFloorFloatToInt32(vBounds.y()-0.5f);
const TextureChannelClass texClass = getTextureChannelClass(faces[coords.face].getFormat().type);
float searchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
const CubeFaceIntCoords c00 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i+0, j+0)), size);
const CubeFaceIntCoords c10 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i+1, j+0)), size);
const CubeFaceIntCoords c01 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i+0, j+1)), size);
const CubeFaceIntCoords c11 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i+1, j+1)), size);
// If any of samples is out of both edges, implementations can do pretty much anything according to spec.
// \todo [2013-07-08 pyry] Test the special case where all corner pixels have exactly the same color.
if (c00.face == CUBEFACE_LAST || c01.face == CUBEFACE_LAST || c10.face == CUBEFACE_LAST || c11.face == CUBEFACE_LAST)
return true;
// Bounds for filtering factors
const float minA = de::clamp((uBounds.x()-0.5f)-float(i), 0.0f, 1.0f);
const float maxA = de::clamp((uBounds.y()-0.5f)-float(i), 0.0f, 1.0f);
const float minB = de::clamp((vBounds.x()-0.5f)-float(j), 0.0f, 1.0f);
const float maxB = de::clamp((vBounds.y()-0.5f)-float(j), 0.0f, 1.0f);
ColorQuad quad;
quad.p00 = lookup<float>(faces[c00.face], sampler, c00.s, c00.t, 0);
quad.p10 = lookup<float>(faces[c10.face], sampler, c10.s, c10.t, 0);
quad.p01 = lookup<float>(faces[c01.face], sampler, c01.s, c01.t, 0);
quad.p11 = lookup<float>(faces[c11.face], sampler, c11.s, c11.t, 0);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep = computeBilinearSearchStepFromFloatQuad(prec, quad);
if (sampler.reductionMode == Sampler::WEIGHTED_AVERAGE)
{
if (isBilinearRangeValid(prec, quad, Vec2(minA, maxA), Vec2(minB, maxB), searchStep, result))
return true;
}
else
{
if (isReductionValid(prec, quad, sampler.reductionMode, result))
return true;
}
}
}
return false;
}
static bool isSeamplessLinearMipmapLinearSampleResultValid (const ConstPixelBufferAccess (&faces0)[CUBEFACE_LAST],
const ConstPixelBufferAccess (&faces1)[CUBEFACE_LAST],
const Sampler& sampler,
const LookupPrecision& prec,
const CubeFaceFloatCoords& coords,
const Vec2& fBounds,
const Vec4& result)
{
// \todo [2013-07-04 pyry] This is strictly not correct as coordinates between levels should be dependent.
// Right now this allows pairing any two valid bilinear quads.
const int size0 = faces0[coords.face].getWidth();
const int size1 = faces1[coords.face].getWidth();
const Vec2 uBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size0, coords.s, prec.coordBits.x(), prec.uvwBits.x());
const Vec2 uBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size1, coords.s, prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds0 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size0, coords.t, prec.coordBits.y(), prec.uvwBits.y());
const Vec2 vBounds1 = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size1, coords.t, prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinates - without wrap mode
const int minI0 = deFloorFloatToInt32(uBounds0.x()-0.5f);
const int maxI0 = deFloorFloatToInt32(uBounds0.y()-0.5f);
const int minI1 = deFloorFloatToInt32(uBounds1.x()-0.5f);
const int maxI1 = deFloorFloatToInt32(uBounds1.y()-0.5f);
const int minJ0 = deFloorFloatToInt32(vBounds0.x()-0.5f);
const int maxJ0 = deFloorFloatToInt32(vBounds0.y()-0.5f);
const int minJ1 = deFloorFloatToInt32(vBounds1.x()-0.5f);
const int maxJ1 = deFloorFloatToInt32(vBounds1.y()-0.5f);
const TextureChannelClass texClass = getTextureChannelClass(faces0[coords.face].getFormat().type);
const float cSearchStep = texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ? computeBilinearSearchStepForUnorm(prec) :
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ? computeBilinearSearchStepForSnorm(prec) :
0.0f; // Step is computed for floating-point quads based on texel values.
DE_ASSERT(texClass == TEXTURECHANNELCLASS_UNSIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_SIGNED_FIXED_POINT ||
texClass == TEXTURECHANNELCLASS_FLOATING_POINT ||
sampler.reductionMode != Sampler::WEIGHTED_AVERAGE);
for (int j0 = minJ0; j0 <= maxJ0; j0++)
{
for (int i0 = minI0; i0 <= maxI0; i0++)
{
ColorQuad quad0;
float searchStep0;
{
const CubeFaceIntCoords c00 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i0+0, j0+0)), size0);
const CubeFaceIntCoords c10 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i0+1, j0+0)), size0);
const CubeFaceIntCoords c01 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i0+0, j0+1)), size0);
const CubeFaceIntCoords c11 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i0+1, j0+1)), size0);
// If any of samples is out of both edges, implementations can do pretty much anything according to spec.
// \todo [2013-07-08 pyry] Test the special case where all corner pixels have exactly the same color.
if (c00.face == CUBEFACE_LAST || c01.face == CUBEFACE_LAST || c10.face == CUBEFACE_LAST || c11.face == CUBEFACE_LAST)
return true;
quad0.p00 = lookup<float>(faces0[c00.face], sampler, c00.s, c00.t, 0);
quad0.p10 = lookup<float>(faces0[c10.face], sampler, c10.s, c10.t, 0);
quad0.p01 = lookup<float>(faces0[c01.face], sampler, c01.s, c01.t, 0);
quad0.p11 = lookup<float>(faces0[c11.face], sampler, c11.s, c11.t, 0);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep0 = computeBilinearSearchStepFromFloatQuad(prec, quad0);
else
searchStep0 = cSearchStep;
}
const float minA0 = de::clamp((uBounds0.x()-0.5f)-float(i0), 0.0f, 1.0f);
const float maxA0 = de::clamp((uBounds0.y()-0.5f)-float(i0), 0.0f, 1.0f);
const float minB0 = de::clamp((vBounds0.x()-0.5f)-float(j0), 0.0f, 1.0f);
const float maxB0 = de::clamp((vBounds0.y()-0.5f)-float(j0), 0.0f, 1.0f);
for (int j1 = minJ1; j1 <= maxJ1; j1++)
{
for (int i1 = minI1; i1 <= maxI1; i1++)
{
ColorQuad quad1;
float searchStep1;
{
const CubeFaceIntCoords c00 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i1+0, j1+0)), size1);
const CubeFaceIntCoords c10 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i1+1, j1+0)), size1);
const CubeFaceIntCoords c01 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i1+0, j1+1)), size1);
const CubeFaceIntCoords c11 = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, IVec2(i1+1, j1+1)), size1);
if (c00.face == CUBEFACE_LAST || c01.face == CUBEFACE_LAST || c10.face == CUBEFACE_LAST || c11.face == CUBEFACE_LAST)
return true;
quad1.p00 = lookup<float>(faces1[c00.face], sampler, c00.s, c00.t, 0);
quad1.p10 = lookup<float>(faces1[c10.face], sampler, c10.s, c10.t, 0);
quad1.p01 = lookup<float>(faces1[c01.face], sampler, c01.s, c01.t, 0);
quad1.p11 = lookup<float>(faces1[c11.face], sampler, c11.s, c11.t, 0);
if (texClass == TEXTURECHANNELCLASS_FLOATING_POINT)
searchStep1 = computeBilinearSearchStepFromFloatQuad(prec, quad1);
else
searchStep1 = cSearchStep;
}
const float minA1 = de::clamp((uBounds1.x()-0.5f)-float(i1), 0.0f, 1.0f);
const float maxA1 = de::clamp((uBounds1.y()-0.5f)-float(i1), 0.0f, 1.0f);
const float minB1 = de::clamp((vBounds1.x()-0.5f)-float(j1), 0.0f, 1.0f);
const float maxB1 = de::clamp((vBounds1.y()-0.5f)-float(j1), 0.0f, 1.0f);
if (is2DTrilinearFilterResultValid(prec, quad0, quad1, Vec2(minA0, maxA0), Vec2(minB0, maxB0), Vec2(minA1, maxA1), Vec2(minB1, maxB1),
fBounds, de::min(searchStep0, searchStep1), result))
return true;
}
}
}
}
return false;
}
static bool isCubeLevelSampleResultValid (const ConstPixelBufferAccess (&level)[CUBEFACE_LAST],
const Sampler& sampler,
const Sampler::FilterMode filterMode,
const LookupPrecision& prec,
const CubeFaceFloatCoords& coords,
const Vec4& result)
{
if (filterMode == Sampler::LINEAR)
{
if (sampler.seamlessCubeMap)
return isSeamlessLinearSampleResultValid(level, sampler, prec, coords, result);
else
return isLinearSampleResultValid(level[coords.face], sampler, prec, Vec2(coords.s, coords.t), 0, result);
}
else
return isNearestSampleResultValid(level[coords.face], sampler, prec, Vec2(coords.s, coords.t), 0, result);
}
static bool isCubeMipmapLinearSampleResultValid (const ConstPixelBufferAccess (&faces0)[CUBEFACE_LAST],
const ConstPixelBufferAccess (&faces1)[CUBEFACE_LAST],
const Sampler& sampler,
const Sampler::FilterMode levelFilter,
const LookupPrecision& prec,
const CubeFaceFloatCoords& coords,
const Vec2& fBounds,
const Vec4& result)
{
if (levelFilter == Sampler::LINEAR)
{
if (sampler.seamlessCubeMap)
return isSeamplessLinearMipmapLinearSampleResultValid(faces0, faces1, sampler, prec, coords, fBounds, result);
else
return isLinearMipmapLinearSampleResultValid(faces0[coords.face], faces1[coords.face], sampler, prec, Vec2(coords.s, coords.t), 0, fBounds, result);
}
else
return isNearestMipmapLinearSampleResultValid(faces0[coords.face], faces1[coords.face], sampler, prec, Vec2(coords.s, coords.t), 0, fBounds, result);
}
static void getCubeLevelFaces (const TextureCubeView& texture, const int levelNdx, ConstPixelBufferAccess (&out)[CUBEFACE_LAST])
{
for (int faceNdx = 0; faceNdx < CUBEFACE_LAST; faceNdx++)
out[faceNdx] = texture.getLevelFace(levelNdx, (CubeFace)faceNdx);
}
bool isLookupResultValid (const TextureCubeView& texture, const Sampler& sampler, const LookupPrecision& prec, const Vec3& coord, const Vec2& lodBounds, const Vec4& result)
{
int numPossibleFaces = 0;
CubeFace possibleFaces[CUBEFACE_LAST];
DE_ASSERT(isSamplerSupported(sampler));
getPossibleCubeFaces(coord, prec.coordBits, &possibleFaces[0], numPossibleFaces);
if (numPossibleFaces == 0)
return true; // Result is undefined.
for (int tryFaceNdx = 0; tryFaceNdx < numPossibleFaces; tryFaceNdx++)
{
const CubeFaceFloatCoords faceCoords (possibleFaces[tryFaceNdx], projectToFace(possibleFaces[tryFaceNdx], coord));
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
if (canBeMagnified)
{
ConstPixelBufferAccess faces[CUBEFACE_LAST];
getCubeLevelFaces(texture, 0, faces);
if (isCubeLevelSampleResultValid(faces, sampler, sampler.magFilter, prec, faceCoords, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int levelNdx = minLevel; levelNdx <= maxLevel; levelNdx++)
{
const float minF = de::clamp(minLod - float(levelNdx), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(levelNdx), 0.0f, 1.0f);
ConstPixelBufferAccess faces0[CUBEFACE_LAST];
ConstPixelBufferAccess faces1[CUBEFACE_LAST];
getCubeLevelFaces(texture, levelNdx, faces0);
getCubeLevelFaces(texture, levelNdx + 1, faces1);
if (isCubeMipmapLinearSampleResultValid(faces0, faces1, sampler, getLevelFilter(sampler.minFilter), prec, faceCoords, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int levelNdx = minLevel; levelNdx <= maxLevel; levelNdx++)
{
ConstPixelBufferAccess faces[CUBEFACE_LAST];
getCubeLevelFaces(texture, levelNdx, faces);
if (isCubeLevelSampleResultValid(faces, sampler, getLevelFilter(sampler.minFilter), prec, faceCoords, result))
return true;
}
}
else
{
ConstPixelBufferAccess faces[CUBEFACE_LAST];
getCubeLevelFaces(texture, 0, faces);
if (isCubeLevelSampleResultValid(faces, sampler, sampler.minFilter, prec, faceCoords, result))
return true;
}
}
}
return false;
}
static inline IVec2 computeLayerRange (int numLayers, int numCoordBits, float layerCoord)
{
const float err = computeFloatingPointError(layerCoord, numCoordBits);
const int minL = (int)deFloatFloor(layerCoord - err + 0.5f); // Round down
const int maxL = (int)deFloatCeil(layerCoord + err + 0.5f) - 1; // Round up
DE_ASSERT(minL <= maxL);
return IVec2(de::clamp(minL, 0, numLayers-1), de::clamp(maxL, 0, numLayers-1));
}
bool isLookupResultValid (const Texture1DArrayView& texture, const Sampler& sampler, const LookupPrecision& prec, const Vec2& coord, const Vec2& lodBounds, const Vec4& result)
{
const IVec2 layerRange = computeLayerRange(texture.getNumLayers(), prec.coordBits.y(), coord.y());
const float coordX = coord.x();
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
DE_ASSERT(isSamplerSupported(sampler));
for (int layer = layerRange.x(); layer <= layerRange.y(); layer++)
{
if (canBeMagnified)
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.magFilter, prec, coordX, layer, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
const float minF = de::clamp(minLod - float(level), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(level), 0.0f, 1.0f);
if (isMipmapLinearSampleResultValid(texture.getLevel(level), texture.getLevel(level+1), sampler, getLevelFilter(sampler.minFilter), prec, coordX, layer, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
if (isLevelSampleResultValid(texture.getLevel(level), sampler, getLevelFilter(sampler.minFilter), prec, coordX, layer, result))
return true;
}
}
else
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.minFilter, prec, coordX, layer, result))
return true;
}
}
}
return false;
}
bool isLookupResultValid (const Texture2DArrayView& texture, const Sampler& sampler, const LookupPrecision& prec, const Vec3& coord, const Vec2& lodBounds, const Vec4& result)
{
const IVec2 layerRange = computeLayerRange(texture.getNumLayers(), prec.coordBits.z(), coord.z());
const Vec2 coordXY = coord.swizzle(0,1);
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
DE_ASSERT(isSamplerSupported(sampler));
for (int layer = layerRange.x(); layer <= layerRange.y(); layer++)
{
if (canBeMagnified)
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.magFilter, prec, coordXY, layer, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
const float minF = de::clamp(minLod - float(level), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(level), 0.0f, 1.0f);
if (isMipmapLinearSampleResultValid(texture.getLevel(level), texture.getLevel(level+1), sampler, getLevelFilter(sampler.minFilter), prec, coordXY, layer, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
if (isLevelSampleResultValid(texture.getLevel(level), sampler, getLevelFilter(sampler.minFilter), prec, coordXY, layer, result))
return true;
}
}
else
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.minFilter, prec, coordXY, layer, result))
return true;
}
}
}
return false;
}
static bool isLevelSampleResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const Sampler::FilterMode filterMode,
const LookupPrecision& prec,
const Vec3& coord,
const Vec4& result)
{
if (filterMode == Sampler::LINEAR)
return isLinearSampleResultValid(level, sampler, prec, coord, result);
else
return isNearestSampleResultValid(level, sampler, prec, coord, result);
}
static bool isMipmapLinearSampleResultValid (const ConstPixelBufferAccess& level0,
const ConstPixelBufferAccess& level1,
const Sampler& sampler,
const Sampler::FilterMode levelFilter,
const LookupPrecision& prec,
const Vec3& coord,
const Vec2& fBounds,
const Vec4& result)
{
if (levelFilter == Sampler::LINEAR)
return isLinearMipmapLinearSampleResultValid(level0, level1, sampler, prec, coord, fBounds, result);
else
return isNearestMipmapLinearSampleResultValid(level0, level1, sampler, prec, coord, fBounds, result);
}
bool isLookupResultValid (const Texture3DView& texture, const Sampler& sampler, const LookupPrecision& prec, const Vec3& coord, const Vec2& lodBounds, const Vec4& result)
{
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
DE_ASSERT(isSamplerSupported(sampler));
if (canBeMagnified)
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.magFilter, prec, coord, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
const float minF = de::clamp(minLod - float(level), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(level), 0.0f, 1.0f);
if (isMipmapLinearSampleResultValid(texture.getLevel(level), texture.getLevel(level+1), sampler, getLevelFilter(sampler.minFilter), prec, coord, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int level = minLevel; level <= maxLevel; level++)
{
if (isLevelSampleResultValid(texture.getLevel(level), sampler, getLevelFilter(sampler.minFilter), prec, coord, result))
return true;
}
}
else
{
if (isLevelSampleResultValid(texture.getLevel(0), sampler, sampler.minFilter, prec, coord, result))
return true;
}
}
return false;
}
static void getCubeArrayLevelFaces (const TextureCubeArrayView& texture, const int levelNdx, const int layerNdx, ConstPixelBufferAccess (&out)[CUBEFACE_LAST])
{
const ConstPixelBufferAccess& level = texture.getLevel(levelNdx);
const int layerDepth = layerNdx * 6;
for (int faceNdx = 0; faceNdx < CUBEFACE_LAST; faceNdx++)
{
const CubeFace face = (CubeFace)faceNdx;
out[faceNdx] = getSubregion(level, 0, 0, layerDepth + getCubeArrayFaceIndex(face), level.getWidth(), level.getHeight(), 1);
}
}
bool isLookupResultValid (const TextureCubeArrayView& texture, const Sampler& sampler, const LookupPrecision& prec, const IVec4& coordBits, const Vec4& coord, const Vec2& lodBounds, const Vec4& result)
{
const IVec2 layerRange = computeLayerRange(texture.getNumLayers(), coordBits.w(), coord.w());
const Vec3 layerCoord = coord.toWidth<3>();
int numPossibleFaces = 0;
CubeFace possibleFaces[CUBEFACE_LAST];
DE_ASSERT(isSamplerSupported(sampler));
getPossibleCubeFaces(layerCoord, prec.coordBits, &possibleFaces[0], numPossibleFaces);
if (numPossibleFaces == 0)
return true; // Result is undefined.
for (int layerNdx = layerRange.x(); layerNdx <= layerRange.y(); layerNdx++)
{
for (int tryFaceNdx = 0; tryFaceNdx < numPossibleFaces; tryFaceNdx++)
{
const CubeFaceFloatCoords faceCoords (possibleFaces[tryFaceNdx], projectToFace(possibleFaces[tryFaceNdx], layerCoord));
const float minLod = lodBounds.x();
const float maxLod = lodBounds.y();
const bool canBeMagnified = minLod <= sampler.lodThreshold;
const bool canBeMinified = maxLod > sampler.lodThreshold;
if (canBeMagnified)
{
ConstPixelBufferAccess faces[CUBEFACE_LAST];
getCubeArrayLevelFaces(texture, 0, layerNdx, faces);
if (isCubeLevelSampleResultValid(faces, sampler, sampler.magFilter, prec, faceCoords, result))
return true;
}
if (canBeMinified)
{
const bool isNearestMipmap = isNearestMipmapFilter(sampler.minFilter);
const bool isLinearMipmap = isLinearMipmapFilter(sampler.minFilter);
const int minTexLevel = 0;
const int maxTexLevel = texture.getNumLevels()-1;
DE_ASSERT(minTexLevel <= maxTexLevel);
if (isLinearMipmap && minTexLevel < maxTexLevel)
{
const int minLevel = de::clamp((int)deFloatFloor(minLod), minTexLevel, maxTexLevel-1);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod), minTexLevel, maxTexLevel-1);
DE_ASSERT(minLevel <= maxLevel);
for (int levelNdx = minLevel; levelNdx <= maxLevel; levelNdx++)
{
const float minF = de::clamp(minLod - float(levelNdx), 0.0f, 1.0f);
const float maxF = de::clamp(maxLod - float(levelNdx), 0.0f, 1.0f);
ConstPixelBufferAccess faces0[CUBEFACE_LAST];
ConstPixelBufferAccess faces1[CUBEFACE_LAST];
getCubeArrayLevelFaces(texture, levelNdx, layerNdx, faces0);
getCubeArrayLevelFaces(texture, levelNdx + 1, layerNdx, faces1);
if (isCubeMipmapLinearSampleResultValid(faces0, faces1, sampler, getLevelFilter(sampler.minFilter), prec, faceCoords, Vec2(minF, maxF), result))
return true;
}
}
else if (isNearestMipmap)
{
// \note The accurate formula for nearest mipmapping is level = ceil(lod + 0.5) - 1 but Khronos has made
// decision to allow floor(lod + 0.5) as well.
const int minLevel = de::clamp((int)deFloatCeil(minLod + 0.5f) - 1, minTexLevel, maxTexLevel);
const int maxLevel = de::clamp((int)deFloatFloor(maxLod + 0.5f), minTexLevel, maxTexLevel);
DE_ASSERT(minLevel <= maxLevel);
for (int levelNdx = minLevel; levelNdx <= maxLevel; levelNdx++)
{
ConstPixelBufferAccess faces[CUBEFACE_LAST];
getCubeArrayLevelFaces(texture, levelNdx, layerNdx, faces);
if (isCubeLevelSampleResultValid(faces, sampler, getLevelFilter(sampler.minFilter), prec, faceCoords, result))
return true;
}
}
else
{
ConstPixelBufferAccess faces[CUBEFACE_LAST];
getCubeArrayLevelFaces(texture, 0, layerNdx, faces);
if (isCubeLevelSampleResultValid(faces, sampler, sampler.minFilter, prec, faceCoords, result))
return true;
}
}
}
}
return false;
}
Vec4 computeFixedPointThreshold (const IVec4& bits)
{
return computeFixedPointError(bits);
}
Vec4 computeFloatingPointThreshold (const IVec4& bits, const Vec4& value)
{
return computeFloatingPointError(value, bits);
}
Vec4 computeColorBitsThreshold(const IVec4& bits, const IVec4& numAccurateBits)
{
return computeColorBitsError(bits, numAccurateBits);
}
Vec2 computeLodBoundsFromDerivates (const float dudx, const float dvdx, const float dwdx, const float dudy, const float dvdy, const float dwdy, const LodPrecision& prec)
{
const float mux = deFloatAbs(dudx);
const float mvx = deFloatAbs(dvdx);
const float mwx = deFloatAbs(dwdx);
const float muy = deFloatAbs(dudy);
const float mvy = deFloatAbs(dvdy);
const float mwy = deFloatAbs(dwdy);
// Ideal:
// px = deFloatSqrt2(mux*mux + mvx*mvx + mwx*mwx);
// py = deFloatSqrt2(muy*muy + mvy*mvy + mwy*mwy);
// fx, fy estimate lower bounds
const float fxMin = de::max(de::max(mux, mvx), mwx);
const float fyMin = de::max(de::max(muy, mvy), mwy);
// fx, fy estimate upper bounds
const float sqrt2 = deFloatSqrt(2.0f);
const float fxMax = sqrt2 * (mux + mvx + mwx);
const float fyMax = sqrt2 * (muy + mvy + mwy);
// p = max(px, py) (isotropic filtering)
const float pMin = de::max(fxMin, fyMin);
const float pMax = de::max(fxMax, fyMax);
// error terms
const float pMinErr = computeFloatingPointError(pMin, prec.derivateBits);
const float pMaxErr = computeFloatingPointError(pMax, prec.derivateBits);
const float minLod = deFloatLog2(pMin-pMinErr);
const float maxLod = deFloatLog2(pMax+pMaxErr);
const float lodErr = computeFixedPointError(prec.lodBits);
DE_ASSERT(minLod <= maxLod);
return Vec2(minLod-lodErr, maxLod+lodErr);
}
Vec2 computeLodBoundsFromDerivates (const float dudx, const float dvdx, const float dudy, const float dvdy, const LodPrecision& prec)
{
return computeLodBoundsFromDerivates(dudx, dvdx, 0.0f, dudy, dvdy, 0.0f, prec);
}
Vec2 computeLodBoundsFromDerivates (const float dudx, const float dudy, const LodPrecision& prec)
{
return computeLodBoundsFromDerivates(dudx, 0.0f, 0.0f, dudy, 0.0f, 0.0f, prec);
}
Vec2 computeCubeLodBoundsFromDerivates (const Vec3& coord, const Vec3& coordDx, const Vec3& coordDy, const int faceSize, const LodPrecision& prec)
{
const bool allowBrokenEdgeDerivate = false;
const CubeFace face = selectCubeFace(coord);
int maNdx = 0;
int sNdx = 0;
int tNdx = 0;
// \note Derivate signs don't matter when computing lod
switch (face)
{
case CUBEFACE_NEGATIVE_X:
case CUBEFACE_POSITIVE_X: maNdx = 0; sNdx = 2; tNdx = 1; break;
case CUBEFACE_NEGATIVE_Y:
case CUBEFACE_POSITIVE_Y: maNdx = 1; sNdx = 0; tNdx = 2; break;
case CUBEFACE_NEGATIVE_Z:
case CUBEFACE_POSITIVE_Z: maNdx = 2; sNdx = 0; tNdx = 1; break;
default:
DE_ASSERT(DE_FALSE);
}
{
const float sc = coord[sNdx];
const float tc = coord[tNdx];
const float ma = de::abs(coord[maNdx]);
const float scdx = coordDx[sNdx];
const float tcdx = coordDx[tNdx];
const float madx = de::abs(coordDx[maNdx]);
const float scdy = coordDy[sNdx];
const float tcdy = coordDy[tNdx];
const float mady = de::abs(coordDy[maNdx]);
const float dudx = float(faceSize) * 0.5f * (scdx*ma - sc*madx) / (ma*ma);
const float dvdx = float(faceSize) * 0.5f * (tcdx*ma - tc*madx) / (ma*ma);
const float dudy = float(faceSize) * 0.5f * (scdy*ma - sc*mady) / (ma*ma);
const float dvdy = float(faceSize) * 0.5f * (tcdy*ma - tc*mady) / (ma*ma);
const Vec2 bounds = computeLodBoundsFromDerivates(dudx, dvdx, dudy, dvdy, prec);
// Implementations may compute derivate from projected (s, t) resulting in incorrect values at edges.
if (allowBrokenEdgeDerivate)
{
const Vec3 dxErr = computeFloatingPointError(coordDx, IVec3(prec.derivateBits));
const Vec3 dyErr = computeFloatingPointError(coordDy, IVec3(prec.derivateBits));
const Vec3 xoffs = abs(coordDx) + dxErr;
const Vec3 yoffs = abs(coordDy) + dyErr;
if (selectCubeFace(coord + xoffs) != face ||
selectCubeFace(coord - xoffs) != face ||
selectCubeFace(coord + yoffs) != face ||
selectCubeFace(coord - yoffs) != face)
{
return Vec2(bounds.x(), 1000.0f);
}
}
return bounds;
}
}
Vec2 clampLodBounds (const Vec2& lodBounds, const Vec2& lodMinMax, const LodPrecision& prec)
{
const float lodErr = computeFixedPointError(prec.lodBits);
const float a = lodMinMax.x();
const float b = lodMinMax.y();
return Vec2(de::clamp(lodBounds.x(), a-lodErr, b-lodErr), de::clamp(lodBounds.y(), a+lodErr, b+lodErr));
}
bool isLevel1DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const LookupPrecision& prec,
const float coordX,
const int coordY,
const Vec4& result)
{
const Sampler::FilterMode filterMode = scaleMode == TEX_LOOKUP_SCALE_MAGNIFY ? sampler.magFilter : sampler.minFilter;
return isLevelSampleResultValid(access, sampler, filterMode, prec, coordX, coordY, result);
}
bool isLevel1DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const IntLookupPrecision& prec,
const float coordX,
const int coordY,
const IVec4& result)
{
DE_ASSERT(sampler.minFilter == Sampler::NEAREST && sampler.magFilter == Sampler::NEAREST);
DE_UNREF(scaleMode);
return isNearestSampleResultValid(access, sampler, prec, coordX, coordY, result);
}
bool isLevel1DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const IntLookupPrecision& prec,
const float coordX,
const int coordY,
const UVec4& result)
{
DE_ASSERT(sampler.minFilter == Sampler::NEAREST && sampler.magFilter == Sampler::NEAREST);
DE_UNREF(scaleMode);
return isNearestSampleResultValid(access, sampler, prec, coordX, coordY, result);
}
bool isLevel2DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const LookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const Vec4& result)
{
const Sampler::FilterMode filterMode = scaleMode == TEX_LOOKUP_SCALE_MAGNIFY ? sampler.magFilter : sampler.minFilter;
return isLevelSampleResultValid(access, sampler, filterMode, prec, coord, coordZ, result);
}
bool isLevel2DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const IntLookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const IVec4& result)
{
DE_ASSERT(sampler.minFilter == Sampler::NEAREST && sampler.magFilter == Sampler::NEAREST);
DE_UNREF(scaleMode);
return isNearestSampleResultValid(access, sampler, prec, coord, coordZ, result);
}
bool isLevel2DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const IntLookupPrecision& prec,
const Vec2& coord,
const int coordZ,
const UVec4& result)
{
DE_ASSERT(sampler.minFilter == Sampler::NEAREST && sampler.magFilter == Sampler::NEAREST);
DE_UNREF(scaleMode);
return isNearestSampleResultValid(access, sampler, prec, coord, coordZ, result);
}
bool isLevel3DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const LookupPrecision& prec,
const Vec3& coord,
const Vec4& result)
{
const tcu::Sampler::FilterMode filterMode = scaleMode == TEX_LOOKUP_SCALE_MAGNIFY ? sampler.magFilter : sampler.minFilter;
return isLevelSampleResultValid(access, sampler, filterMode, prec, coord, result);
}
bool isLevel3DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const IntLookupPrecision& prec,
const Vec3& coord,
const IVec4& result)
{
DE_ASSERT(sampler.minFilter == Sampler::NEAREST && sampler.magFilter == Sampler::NEAREST);
DE_UNREF(scaleMode);
return isNearestSampleResultValid(access, sampler, prec, coord, result);
}
bool isLevel3DLookupResultValid (const ConstPixelBufferAccess& access,
const Sampler& sampler,
TexLookupScaleMode scaleMode,
const IntLookupPrecision& prec,
const Vec3& coord,
const UVec4& result)
{
DE_ASSERT(sampler.minFilter == Sampler::NEAREST && sampler.magFilter == Sampler::NEAREST);
DE_UNREF(scaleMode);
return isNearestSampleResultValid(access, sampler, prec, coord, result);
}
template<typename PrecType, typename ScalarType>
static bool isGatherOffsetsResultValid (const ConstPixelBufferAccess& level,
const Sampler& sampler,
const PrecType& prec,
const Vec2& coord,
int coordZ,
int componentNdx,
const IVec2 (&offsets)[4],
const Vector<ScalarType, 4>& result)
{
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getWidth(), coord.x(), prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, level.getHeight(), coord.y(), prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinate bounds for (x0, y0) - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x()-0.5f);
const int maxI = deFloorFloatToInt32(uBounds.y()-0.5f);
const int minJ = deFloorFloatToInt32(vBounds.x()-0.5f);
const int maxJ = deFloorFloatToInt32(vBounds.y()-0.5f);
const int w = level.getWidth();
const int h = level.getHeight();
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
Vector<ScalarType, 4> color;
for (int offNdx = 0; offNdx < 4; offNdx++)
{
// offNdx-th coordinate offset and then wrapped.
const int x = wrap(sampler.wrapS, i+offsets[offNdx].x(), w);
const int y = wrap(sampler.wrapT, j+offsets[offNdx].y(), h);
color[offNdx] = lookup<ScalarType>(level, sampler, x, y, coordZ)[componentNdx];
}
if (isColorValid(prec, color, result))
return true;
}
}
return false;
}
bool isGatherOffsetsResultValid (const Texture2DView& texture,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec2& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const Vec4& result)
{
return isGatherOffsetsResultValid(texture.getLevel(0), sampler, prec, coord, 0, componentNdx, offsets, result);
}
bool isGatherOffsetsResultValid (const Texture2DView& texture,
const Sampler& sampler,
const IntLookupPrecision& prec,
const Vec2& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const IVec4& result)
{
return isGatherOffsetsResultValid(texture.getLevel(0), sampler, prec, coord, 0, componentNdx, offsets, result);
}
bool isGatherOffsetsResultValid (const Texture2DView& texture,
const Sampler& sampler,
const IntLookupPrecision& prec,
const Vec2& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const UVec4& result)
{
return isGatherOffsetsResultValid(texture.getLevel(0), sampler, prec, coord, 0, componentNdx, offsets, result);
}
template <typename PrecType, typename ScalarType>
static bool is2DArrayGatherOffsetsResultValid (const Texture2DArrayView& texture,
const Sampler& sampler,
const PrecType& prec,
const Vec3& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const Vector<ScalarType, 4>& result)
{
const IVec2 layerRange = computeLayerRange(texture.getNumLayers(), prec.coordBits.z(), coord.z());
for (int layer = layerRange.x(); layer <= layerRange.y(); layer++)
{
if (isGatherOffsetsResultValid(texture.getLevel(0), sampler, prec, coord.swizzle(0,1), layer, componentNdx, offsets, result))
return true;
}
return false;
}
bool isGatherOffsetsResultValid (const Texture2DArrayView& texture,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec3& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const Vec4& result)
{
return is2DArrayGatherOffsetsResultValid(texture, sampler, prec, coord, componentNdx, offsets, result);
}
bool isGatherOffsetsResultValid (const Texture2DArrayView& texture,
const Sampler& sampler,
const IntLookupPrecision& prec,
const Vec3& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const IVec4& result)
{
return is2DArrayGatherOffsetsResultValid(texture, sampler, prec, coord, componentNdx, offsets, result);
}
bool isGatherOffsetsResultValid (const Texture2DArrayView& texture,
const Sampler& sampler,
const IntLookupPrecision& prec,
const Vec3& coord,
int componentNdx,
const IVec2 (&offsets)[4],
const UVec4& result)
{
return is2DArrayGatherOffsetsResultValid(texture, sampler, prec, coord, componentNdx, offsets, result);
}
template<typename PrecType, typename ScalarType>
static bool isGatherResultValid (const TextureCubeView& texture,
const Sampler& sampler,
const PrecType& prec,
const CubeFaceFloatCoords& coords,
int componentNdx,
const Vector<ScalarType, 4>& result)
{
const int size = texture.getLevelFace(0, coords.face).getWidth();
const Vec2 uBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size, coords.s, prec.coordBits.x(), prec.uvwBits.x());
const Vec2 vBounds = computeNonNormalizedCoordBounds(sampler.normalizedCoords, size, coords.t, prec.coordBits.y(), prec.uvwBits.y());
// Integer coordinate bounds for (x0,y0) - without wrap mode
const int minI = deFloorFloatToInt32(uBounds.x()-0.5f);
const int maxI = deFloorFloatToInt32(uBounds.y()-0.5f);
const int minJ = deFloorFloatToInt32(vBounds.x()-0.5f);
const int maxJ = deFloorFloatToInt32(vBounds.y()-0.5f);
// Face accesses
ConstPixelBufferAccess faces[CUBEFACE_LAST];
for (int face = 0; face < CUBEFACE_LAST; face++)
faces[face] = texture.getLevelFace(0, CubeFace(face));
for (int j = minJ; j <= maxJ; j++)
{
for (int i = minI; i <= maxI; i++)
{
static const IVec2 offsets[4] =
{
IVec2(0, 1),
IVec2(1, 1),
IVec2(1, 0),
IVec2(0, 0)
};
Vector<ScalarType, 4> color;
for (int offNdx = 0; offNdx < 4; offNdx++)
{
const CubeFaceIntCoords c = remapCubeEdgeCoords(CubeFaceIntCoords(coords.face, i+offsets[offNdx].x(), j+offsets[offNdx].y()), size);
// If any of samples is out of both edges, implementations can do pretty much anything according to spec.
// \todo [2014-06-05 nuutti] Test the special case where all corner pixels have exactly the same color.
// See also isSeamlessLinearSampleResultValid and similar.
if (c.face == CUBEFACE_LAST)
return true;
color[offNdx] = lookup<ScalarType>(faces[c.face], sampler, c.s, c.t, 0)[componentNdx];
}
if (isColorValid(prec, color, result))
return true;
}
}
return false;
}
template <typename PrecType, typename ScalarType>
static bool isCubeGatherResultValid (const TextureCubeView& texture,
const Sampler& sampler,
const PrecType& prec,
const Vec3& coord,
int componentNdx,
const Vector<ScalarType, 4>& result)
{
int numPossibleFaces = 0;
CubeFace possibleFaces[CUBEFACE_LAST];
getPossibleCubeFaces(coord, prec.coordBits, &possibleFaces[0], numPossibleFaces);
if (numPossibleFaces == 0)
return true; // Result is undefined.
for (int tryFaceNdx = 0; tryFaceNdx < numPossibleFaces; tryFaceNdx++)
{
const CubeFaceFloatCoords faceCoords(possibleFaces[tryFaceNdx], projectToFace(possibleFaces[tryFaceNdx], coord));
if (isGatherResultValid(texture, sampler, prec, faceCoords, componentNdx, result))
return true;
}
return false;
}
bool isGatherResultValid (const TextureCubeView& texture,
const Sampler& sampler,
const LookupPrecision& prec,
const Vec3& coord,
int componentNdx,
const Vec4& result)
{
return isCubeGatherResultValid(texture, sampler, prec, coord, componentNdx, result);
}
bool isGatherResultValid (const TextureCubeView& texture,
const Sampler& sampler,
const IntLookupPrecision& prec,
const Vec3& coord,
int componentNdx,
const IVec4& result)
{
return isCubeGatherResultValid(texture, sampler, prec, coord, componentNdx, result);
}
bool isGatherResultValid (const TextureCubeView& texture,
const Sampler& sampler,
const IntLookupPrecision& prec,
const Vec3& coord,
int componentNdx,
const UVec4& result)
{
return isCubeGatherResultValid(texture, sampler, prec, coord, componentNdx, result);
}
} // tcu