crossxtex/DirectXTex/DirectXTexConvert.cpp

//-------------------------------------------------------------------------------------
// DirectXTexConvert.cpp
//  
// DirectX Texture Library - Image conversion 
//
// THIS CODE AND INFORMATION IS PROVIDED "AS IS" WITHOUT WARRANTY OF
// ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED TO
// THE IMPLIED WARRANTIES OF MERCHANTABILITY AND/OR FITNESS FOR A
// PARTICULAR PURPOSE.
//
// Copyright (c) Microsoft Corporation. All rights reserved.
//
// http://go.microsoft.com/fwlink/?LinkId=248926
//-------------------------------------------------------------------------------------

#include "directxtexp.h"

using namespace DirectX::PackedVector;
using Microsoft::WRL::ComPtr;

namespace
{
#if DIRECTX_MATH_VERSION < 306
    inline float round_to_nearest( float x )
    {
        // Round to nearest (even)
        float i = floorf(x);
        x -= i;
        if(x < 0.5f)
            return i;
        if(x > 0.5f)
            return i + 1.f;

        float int_part;
        modff( i / 2.f, &int_part );
        if ( (2.f*int_part) == i )
        {
            return i;
        }

        return i + 1.f;
    }
#endif

    inline uint32_t FloatTo7e3(float Value)
    {
        uint32_t IValue = reinterpret_cast<uint32_t *>(&Value)[0];

        if ( IValue & 0x80000000U )
        {
            // Positive only
            return 0;
        }
        else if (IValue > 0x41FF73FFU)
        {
            // The number is too large to be represented as a 7e3. Saturate.
            return 0x3FFU;
        }
        else
        {
            if (IValue < 0x3E800000U)
            {
                // The number is too small to be represented as a normalized 7e3.
                // Convert it to a denormalized value.
                uint32_t Shift = 125U - (IValue >> 23U);
                IValue = (0x800000U | (IValue & 0x7FFFFFU)) >> Shift;
            }
            else
            {
                // Rebias the exponent to represent the value as a normalized 7e3.
                IValue += 0xC2000000U;
            }

            return ((IValue + 0x7FFFU + ((IValue >> 16U) & 1U)) >> 16U)&0x3FFU; 
        }
    }

    inline float FloatFrom7e3( uint32_t Value )
    {
        uint32_t Mantissa = (uint32_t)(Value & 0x7F);

        uint32_t Exponent = (Value & 0x380);
        if (Exponent != 0)  // The value is normalized
        {
            Exponent = (uint32_t)((Value >> 7) & 0x7);
        }
        else if (Mantissa != 0)     // The value is denormalized
        {
            // Normalize the value in the resulting float
            Exponent = 1;

            do
            {
                Exponent--;
                Mantissa <<= 1;
            } while ((Mantissa & 0x80) == 0);

            Mantissa &= 0x7F;
        }
        else                        // The value is zero
        {
            Exponent = (uint32_t)-124;
        }

        uint32_t Result = ((Exponent + 124) << 23) | // Exponent
                          (Mantissa << 16);          // Mantissa

        return reinterpret_cast<float*>(&Result)[0];
    }

    inline uint32_t FloatTo6e4(float Value)
    {
        uint32_t IValue = reinterpret_cast<uint32_t *>(&Value)[0];

        if ( IValue & 0x80000000U )
        {
            // Positive only
            return 0;
        }
        else if (IValue > 0x43FEFFFFU)
        {
            // The number is too large to be represented as a 6e4. Saturate.
            return 0x3FFU;
        }
        else
        {
            if (IValue < 0x3C800000U)
            {
                // The number is too small to be represented as a normalized 6e4.
                // Convert it to a denormalized value.
                uint32_t Shift = 121U - (IValue >> 23U);
                IValue = (0x800000U | (IValue & 0x7FFFFFU)) >> Shift;
            }
            else
            {
                // Rebias the exponent to represent the value as a normalized 6e4.
                IValue += 0xC4000000U;
            }

            return ((IValue + 0xFFFFU + ((IValue >> 17U) & 1U)) >> 17U)&0x3FFU; 
        }
    }

    inline float FloatFrom6e4( uint32_t Value )
    {
        uint32_t Mantissa = (uint32_t)(Value & 0x3F);

        uint32_t Exponent = (Value & 0x3C0);
        if (Exponent != 0)  // The value is normalized
        {
            Exponent = (uint32_t)((Value >> 6) & 0xF);
        }
        else if (Mantissa != 0)     // The value is denormalized
        {
            // Normalize the value in the resulting float
            Exponent = 1;

            do
            {
                Exponent--;
                Mantissa <<= 1;
            } while ((Mantissa & 0x40) == 0);

            Mantissa &= 0x3F;
        }
        else                        // The value is zero
        {
            Exponent = (uint32_t)-120;
        }

        uint32_t Result = ((Exponent + 120) << 23) | // Exponent
                          (Mantissa << 17);          // Mantissa

        return reinterpret_cast<float*>(&Result)[0];
    }
};

namespace DirectX
{
static const XMVECTORF32 g_Grayscale = { 0.2125f, 0.7154f, 0.0721f, 0.0f };

//-------------------------------------------------------------------------------------
// Copies an image row with optional clearing of alpha value to 1.0
// (can be used in place as well) otherwise copies the image row unmodified.
//-------------------------------------------------------------------------------------
void _CopyScanline(_When_(pDestination == pSource, _Inout_updates_bytes_(outSize))
                   _When_(pDestination != pSource, _Out_writes_bytes_(outSize))
                   LPVOID pDestination, _In_ size_t outSize,
                   _In_reads_bytes_(inSize) LPCVOID pSource, _In_ size_t inSize,
                   _In_ DXGI_FORMAT format, _In_ DWORD flags)
{
    assert( pDestination && outSize > 0 );
    assert( pSource && inSize > 0 );
    assert( IsValid(format) && !IsPalettized(format) );

    if ( flags & TEXP_SCANLINE_SETALPHA )
    {
        switch( static_cast<int>(format) )
        {
        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_R32G32B32A32_TYPELESS:
        case DXGI_FORMAT_R32G32B32A32_FLOAT:
        case DXGI_FORMAT_R32G32B32A32_UINT:
        case DXGI_FORMAT_R32G32B32A32_SINT:
            if ( inSize >= 16 && outSize >= 16 )
            {
                uint32_t alpha;
                if ( format == DXGI_FORMAT_R32G32B32A32_FLOAT )
                    alpha = 0x3f800000;
                else if ( format == DXGI_FORMAT_R32G32B32A32_SINT )
                    alpha = 0x7fffffff;
                else
                    alpha = 0xffffffff;

                if ( pDestination == pSource )
                {
                    uint32_t *dPtr = reinterpret_cast<uint32_t*> (pDestination);
                    for( size_t count = 0; count < ( outSize - 15 ); count += 16 )
                    {
                        dPtr += 3;
                        *(dPtr++) = alpha;
                    }
                }
                else
                {
                    const uint32_t * __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
                    uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 15 ); count += 16 )
                    {
                        *(dPtr++) = *(sPtr++);
                        *(dPtr++) = *(sPtr++);
                        *(dPtr++) = *(sPtr++);
                        *(dPtr++) = alpha;
                        sPtr++;
                    }
                }
            }
            return;

        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_R16G16B16A16_TYPELESS:
        case DXGI_FORMAT_R16G16B16A16_FLOAT:
        case DXGI_FORMAT_R16G16B16A16_UNORM:
        case DXGI_FORMAT_R16G16B16A16_UINT:
        case DXGI_FORMAT_R16G16B16A16_SNORM:
        case DXGI_FORMAT_R16G16B16A16_SINT:
        case DXGI_FORMAT_Y416:
            if ( inSize >= 8 && outSize >= 8 )
            {
                uint16_t alpha;
                if ( format == DXGI_FORMAT_R16G16B16A16_FLOAT )
                    alpha = 0x3c00;
                else if ( format == DXGI_FORMAT_R16G16B16A16_SNORM || format == DXGI_FORMAT_R16G16B16A16_SINT )
                    alpha = 0x7fff;
                else
                    alpha = 0xffff;

                if ( pDestination == pSource )
                {
                    uint16_t *dPtr = reinterpret_cast<uint16_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 7 ); count += 8 )
                    {
                        dPtr += 3;
                        *(dPtr++) = alpha;
                    }
                }
                else
                {
                    const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
                    uint16_t * __restrict dPtr = reinterpret_cast<uint16_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 7 ); count += 8 )
                    {
                        *(dPtr++) = *(sPtr++);
                        *(dPtr++) = *(sPtr++);
                        *(dPtr++) = *(sPtr++);
                        *(dPtr++) = alpha;
                        sPtr++;
                    }
                }
            }
            return;

        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_R10G10B10A2_TYPELESS:
        case DXGI_FORMAT_R10G10B10A2_UNORM:
        case DXGI_FORMAT_R10G10B10A2_UINT:
        case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
        case DXGI_FORMAT_Y410:
        case 116 /* DXGI_FORMAT_R10G10B10_7E3_A2_FLOAT */:
        case 117 /* DXGI_FORMAT_R10G10B10_6E4_A2_FLOAT */:
            if ( inSize >= 4 && outSize >= 4 )
            {
                if ( pDestination == pSource )
                {
                    uint32_t *dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 3 ); count += 4 )
                    {
                        *dPtr |= 0xC0000000;
                        ++dPtr;
                    }
                }
                else
                {
                    const uint32_t * __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
                    uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 3 ); count += 4 )
                    {
                        *(dPtr++) = *(sPtr++) | 0xC0000000;
                    }
                }
            }
            return;

        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_R8G8B8A8_TYPELESS:
        case DXGI_FORMAT_R8G8B8A8_UNORM:
        case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
        case DXGI_FORMAT_R8G8B8A8_UINT:
        case DXGI_FORMAT_R8G8B8A8_SNORM:
        case DXGI_FORMAT_R8G8B8A8_SINT:
        case DXGI_FORMAT_B8G8R8A8_UNORM:
        case DXGI_FORMAT_B8G8R8A8_TYPELESS:
        case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
        case DXGI_FORMAT_AYUV:
            if ( inSize >= 4 && outSize >= 4 )
            {
                const uint32_t alpha = ( format == DXGI_FORMAT_R8G8B8A8_SNORM || format == DXGI_FORMAT_R8G8B8A8_SINT ) ? 0x7f000000 : 0xff000000;

                if ( pDestination == pSource )
                {
                    uint32_t *dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 3 ); count += 4 )
                    {
                        uint32_t t = *dPtr & 0xFFFFFF;
                        t |= alpha;
                        *(dPtr++) = t;
                    }
                }
                else
                {
                    const uint32_t * __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
                    uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 3 ); count += 4 )
                    {
                        uint32_t t = *(sPtr++) & 0xFFFFFF;
                        t |= alpha;
                        *(dPtr++) = t;
                    }
                }
            }
            return;

        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_B5G5R5A1_UNORM:
            if ( inSize >= 2 && outSize >= 2 )
            {
                if ( pDestination == pSource )
                {
                    uint16_t *dPtr = reinterpret_cast<uint16_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 1 ); count += 2 )
                    {
                        *(dPtr++) |= 0x8000;
                    }
                }
                else
                {
                    const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
                    uint16_t * __restrict dPtr = reinterpret_cast<uint16_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 1 ); count += 2 )
                    {
                        *(dPtr++) = *(sPtr++) | 0x8000;
                    }
                }
            }
            return;

        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_A8_UNORM:
            memset( pDestination, 0xff, outSize );
            return;

        //-----------------------------------------------------------------------------
        case DXGI_FORMAT_B4G4R4A4_UNORM:
            if ( inSize >= 2 && outSize >= 2 )
            {
                if ( pDestination == pSource )
                {
                    uint16_t *dPtr = reinterpret_cast<uint16_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 1 ); count += 2 )
                    {
                        *(dPtr++) |= 0xF000;
                    }
                }
                else
                {
                    const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
                    uint16_t * __restrict dPtr = reinterpret_cast<uint16_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 1 ); count += 2 )
                    {
                        *(dPtr++) = *(sPtr++) | 0xF000;
                    }
                }
            }
            return;
        }
    }

    // Fall-through case is to just use memcpy (assuming this is not an in-place operation)
    if ( pDestination == pSource )
        return;

    size_t size = std::min<size_t>( outSize, inSize );
    memcpy_s( pDestination, outSize, pSource, size );
}


//-------------------------------------------------------------------------------------
// Swizzles (RGB <-> BGR) an image row with optional clearing of alpha value to 1.0
// (can be used in place as well) otherwise copies the image row unmodified.
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
void _SwizzleScanline( LPVOID pDestination, size_t outSize, LPCVOID pSource, size_t inSize, DXGI_FORMAT format, DWORD flags )
{
    assert( pDestination && outSize > 0 );
    assert( pSource && inSize > 0 );
    assert( IsValid(format) && !IsPlanar(format) && !IsPalettized(format) );

    switch( format )
    {
    //---------------------------------------------------------------------------------
    case DXGI_FORMAT_R10G10B10A2_TYPELESS:
    case DXGI_FORMAT_R10G10B10A2_UNORM:
    case DXGI_FORMAT_R10G10B10A2_UINT:
    case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
        if ( inSize >= 4 && outSize >= 4 )
        {
            if ( flags & TEXP_SCANLINE_LEGACY )
            {
                // Swap Red (R) and Blue (B) channel (used for D3DFMT_A2R10G10B10 legacy sources)
                if ( pDestination == pSource )
                {
                    uint32_t *dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 3 ); count += 4 )
                    {
                        uint32_t t = *dPtr;

                        uint32_t t1 = (t & 0x3ff00000) >> 20;
                        uint32_t t2 = (t & 0x000003ff) << 20;
                        uint32_t t3 = (t & 0x000ffc00);
                        uint32_t ta = ( flags & TEXP_SCANLINE_SETALPHA ) ? 0xC0000000 : (t & 0xC0000000);

                        *(dPtr++) = t1 | t2 | t3 | ta;
                    }
                }
                else
                {
                    const uint32_t * __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
                    uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 3 ); count += 4 )
                    {
                        uint32_t t = *(sPtr++);

                        uint32_t t1 = (t & 0x3ff00000) >> 20;
                        uint32_t t2 = (t & 0x000003ff) << 20;
                        uint32_t t3 = (t & 0x000ffc00);
                        uint32_t ta = ( flags & TEXP_SCANLINE_SETALPHA ) ? 0xC0000000 : (t & 0xC0000000);

                        *(dPtr++) = t1 | t2 | t3 | ta;
                    }
                }
                return;
            }
        }
        break;

    //---------------------------------------------------------------------------------
    case DXGI_FORMAT_R8G8B8A8_TYPELESS:
    case DXGI_FORMAT_R8G8B8A8_UNORM:
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8A8_UNORM:
    case DXGI_FORMAT_B8G8R8X8_UNORM:
    case DXGI_FORMAT_B8G8R8A8_TYPELESS:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8X8_TYPELESS:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
        if ( inSize >= 4 && outSize >= 4 )
        {
            // Swap Red (R) and Blue (B) channels (used to convert from DXGI 1.1 BGR formats to DXGI 1.0 RGB)
            if ( pDestination == pSource )
            {
                uint32_t *dPtr = reinterpret_cast<uint32_t*>(pDestination);
                for( size_t count = 0; count < ( outSize - 3 ); count += 4 )
                {
                    uint32_t t = *dPtr;

                    uint32_t t1 = (t & 0x00ff0000) >> 16;
                    uint32_t t2 = (t & 0x000000ff) << 16;
                    uint32_t t3 = (t & 0x0000ff00);
                    uint32_t ta = ( flags & TEXP_SCANLINE_SETALPHA ) ? 0xff000000 : (t & 0xFF000000);

                    *(dPtr++) = t1 | t2 | t3 | ta;
                }
            }
            else
            {
                const uint32_t * __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
                uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
                size_t size = std::min<size_t>( outSize, inSize );
                for( size_t count = 0; count < ( size - 3 ); count += 4 )
                {
                    uint32_t t = *(sPtr++);

                    uint32_t t1 = (t & 0x00ff0000) >> 16;
                    uint32_t t2 = (t & 0x000000ff) << 16;
                    uint32_t t3 = (t & 0x0000ff00);
                    uint32_t ta = ( flags & TEXP_SCANLINE_SETALPHA ) ? 0xff000000 : (t & 0xFF000000);

                    *(dPtr++) = t1 | t2 | t3 | ta;
                }
            }
            return;
        }
        break;

    //---------------------------------------------------------------------------------
    case DXGI_FORMAT_YUY2:
        if ( inSize >= 4 && outSize >= 4 )
        {
            if ( flags & TEXP_SCANLINE_LEGACY )
            {
                // Reorder YUV components (used to convert legacy UYVY -> YUY2)
                if ( pDestination == pSource )
                {
                    uint32_t *dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    for( size_t count = 0; count < ( outSize - 3 ); count += 4 )
                    {
                        uint32_t t = *dPtr;

                        uint32_t t1 = (t & 0x000000ff) << 8;
                        uint32_t t2 = (t & 0x0000ff00) >> 8;
                        uint32_t t3 = (t & 0x00ff0000) << 8;
                        uint32_t t4 = (t & 0xff000000) >> 8;

                        *(dPtr++) = t1 | t2 | t3 | t4;
                    }
                }
                else
                {
                    const uint32_t * __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
                    uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
                    size_t size = std::min<size_t>( outSize, inSize );
                    for( size_t count = 0; count < ( size - 3 ); count += 4 )
                    {
                        uint32_t t = *(sPtr++);

                        uint32_t t1 = (t & 0x000000ff) << 8;
                        uint32_t t2 = (t & 0x0000ff00) >> 8;
                        uint32_t t3 = (t & 0x00ff0000) << 8;
                        uint32_t t4 = (t & 0xff000000) >> 8;

                        *(dPtr++) = t1 | t2 | t3 | t4;
                    }
                }
                return;
            }
        }
        break;
    }

    // Fall-through case is to just use memcpy (assuming this is not an in-place operation)
    if ( pDestination == pSource )
        return;

    size_t size = std::min<size_t>( outSize, inSize );
    memcpy_s( pDestination, outSize, pSource, size );
}


//-------------------------------------------------------------------------------------
// Converts an image row with optional clearing of alpha value to 1.0
// Returns true if supported, false if expansion case not supported
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
bool _ExpandScanline( LPVOID pDestination, size_t outSize, DXGI_FORMAT outFormat,  
                      LPCVOID pSource, size_t inSize, DXGI_FORMAT inFormat, DWORD flags )
{
    assert( pDestination && outSize > 0 );
    assert( pSource && inSize > 0 );
    assert( IsValid(outFormat) && !IsPlanar(outFormat) && !IsPalettized(outFormat) );
    assert( IsValid(inFormat) && !IsPlanar(inFormat) && !IsPalettized(inFormat) );

    switch( inFormat )
    {
    case DXGI_FORMAT_B5G6R5_UNORM:
        if ( outFormat != DXGI_FORMAT_R8G8B8A8_UNORM )
            return false;

        // DXGI_FORMAT_B5G6R5_UNORM -> DXGI_FORMAT_R8G8B8A8_UNORM
        if ( inSize >= 2 && outSize >= 4 )
        {
            const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
            uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);

            for( size_t ocount = 0, icount = 0; ( ( icount < ( inSize - 1 ) ) && ( ocount < ( outSize - 3 ) ) ); icount += 2, ocount += 4 )
            {
                uint16_t t = *(sPtr++);

                uint32_t t1 = ((t & 0xf800) >> 8) | ((t & 0xe000) >> 13);
                uint32_t t2 = ((t & 0x07e0) << 5) | ((t & 0x0600) >> 5);
                uint32_t t3 = ((t & 0x001f) << 19) | ((t & 0x001c) << 14);

                *(dPtr++) = t1 | t2 | t3 | 0xff000000;
            }
            return true;
        }
        return false;
        
    case DXGI_FORMAT_B5G5R5A1_UNORM:
        if ( outFormat != DXGI_FORMAT_R8G8B8A8_UNORM )
            return false;

        // DXGI_FORMAT_B5G5R5A1_UNORM -> DXGI_FORMAT_R8G8B8A8_UNORM
        if ( inSize >= 2 && outSize >= 4 )
        {
            const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
            uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);

            for( size_t ocount = 0, icount = 0; ( ( icount < ( inSize - 1 ) ) && ( ocount < ( outSize - 3 ) ) ); icount += 2, ocount += 4 )
            {
                uint16_t t = *(sPtr++);

                uint32_t t1 = ((t & 0x7c00) >> 7) | ((t & 0x7000) >> 12);
                uint32_t t2 = ((t & 0x03e0) << 6) | ((t & 0x0380) << 1);
                uint32_t t3 = ((t & 0x001f) << 19) | ((t & 0x001c) << 14);
                uint32_t ta = ( flags & TEXP_SCANLINE_SETALPHA ) ? 0xff000000 : ((t & 0x8000) ? 0xff000000 : 0);

                *(dPtr++) = t1 | t2 | t3 | ta;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B4G4R4A4_UNORM:
        if ( outFormat != DXGI_FORMAT_R8G8B8A8_UNORM )
            return false;

        // DXGI_FORMAT_B4G4R4A4_UNORM -> DXGI_FORMAT_R8G8B8A8_UNORM
        if ( inSize >= 2 && outSize >= 4 )
        {
            const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
            uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);

            for( size_t ocount = 0, icount = 0; ( ( icount < ( inSize - 1 ) ) && ( ocount < ( outSize - 3 ) ) ); icount += 2, ocount += 4 )
            {
                uint16_t t = *(sPtr++);

                uint32_t t1 = ((t & 0x0f00) >> 4) | ((t & 0x0f00) >> 8);
                uint32_t t2 = ((t & 0x00f0) << 8) | ((t & 0x00f0) << 4);
                uint32_t t3 = ((t & 0x000f) << 20) | ((t & 0x000f) << 16);
                uint32_t ta = ( flags & TEXP_SCANLINE_SETALPHA ) ? 0xff000000 : (((t & 0xf000) << 16) | ((t & 0xf000) << 12));

                *(dPtr++) = t1 | t2 | t3 | ta;
            }
            return true;
        }
        return false;
    }

    return false;
}


//-------------------------------------------------------------------------------------
// Loads an image row into standard RGBA XMVECTOR (aligned) array
//-------------------------------------------------------------------------------------
#define LOAD_SCANLINE( type, func )\
        if ( size >= sizeof(type) )\
        {\
            const type * __restrict sPtr = reinterpret_cast<const type*>(pSource);\
            for( size_t icount = 0; icount < ( size - sizeof(type) + 1 ); icount += sizeof(type) )\
            {\
                if ( dPtr >= ePtr ) break;\
                *(dPtr++) = func( sPtr++ );\
            }\
            return true;\
        }\
        return false;

#define LOAD_SCANLINE3( type, func, defvec )\
        if ( size >= sizeof(type) )\
        {\
            const type * __restrict sPtr = reinterpret_cast<const type*>(pSource);\
            for( size_t icount = 0; icount < ( size - sizeof(type) + 1 ); icount += sizeof(type) )\
            {\
                XMVECTOR v = func( sPtr++ );\
                if ( dPtr >= ePtr ) break;\
                *(dPtr++) = XMVectorSelect( defvec, v, g_XMSelect1110 );\
            }\
            return true;\
        }\
        return false;

#define LOAD_SCANLINE2( type, func, defvec )\
        if ( size >= sizeof(type) )\
        {\
            const type * __restrict sPtr = reinterpret_cast<const type*>(pSource);\
            for( size_t icount = 0; icount < ( size - sizeof(type) + 1 ); icount += sizeof(type) )\
            {\
                XMVECTOR v = func( sPtr++ );\
                if ( dPtr >= ePtr ) break;\
                *(dPtr++) = XMVectorSelect( defvec, v, g_XMSelect1100 );\
            }\
            return true;\
        }\
        return false;

#pragma warning(suppress: 6101)
_Use_decl_annotations_ bool _LoadScanline( XMVECTOR* pDestination, size_t count,
                    LPCVOID pSource, size_t size, DXGI_FORMAT format )
{
    assert( pDestination && count > 0 && (((uintptr_t)pDestination & 0xF) == 0) );
    assert( pSource && size > 0 );
    assert( IsValid(format) && !IsTypeless(format, false) && !IsCompressed(format) && !IsPlanar(format) && !IsPalettized(format) );

    XMVECTOR* __restrict dPtr = pDestination;
    if ( !dPtr )
        return false;

    const XMVECTOR* ePtr = pDestination + count;

    switch( static_cast<int>(format) )
    {
    case DXGI_FORMAT_R32G32B32A32_FLOAT:
        {
            size_t msize = (size > (sizeof(XMVECTOR)*count)) ? (sizeof(XMVECTOR)*count) : size;
            memcpy_s( dPtr, sizeof(XMVECTOR)*count, pSource, msize );
        }
        return true;

    case DXGI_FORMAT_R32G32B32A32_UINT:
        LOAD_SCANLINE( XMUINT4, XMLoadUInt4 )

    case DXGI_FORMAT_R32G32B32A32_SINT:
        LOAD_SCANLINE( XMINT4, XMLoadSInt4 )

    case DXGI_FORMAT_R32G32B32_FLOAT:
        LOAD_SCANLINE3( XMFLOAT3, XMLoadFloat3, g_XMIdentityR3 )

    case DXGI_FORMAT_R32G32B32_UINT:
        LOAD_SCANLINE3( XMUINT3, XMLoadUInt3, g_XMIdentityR3 )

    case DXGI_FORMAT_R32G32B32_SINT:
        LOAD_SCANLINE3( XMINT3, XMLoadSInt3, g_XMIdentityR3 )

    case DXGI_FORMAT_R16G16B16A16_FLOAT:
        LOAD_SCANLINE( XMHALF4, XMLoadHalf4 )

    case DXGI_FORMAT_R16G16B16A16_UNORM:
        LOAD_SCANLINE( XMUSHORTN4, XMLoadUShortN4 ) 

    case DXGI_FORMAT_R16G16B16A16_UINT:
        LOAD_SCANLINE( XMUSHORT4, XMLoadUShort4 )

    case DXGI_FORMAT_R16G16B16A16_SNORM:
        LOAD_SCANLINE( XMSHORTN4, XMLoadShortN4 )

    case DXGI_FORMAT_R16G16B16A16_SINT:
        LOAD_SCANLINE( XMSHORT4, XMLoadShort4 )

    case DXGI_FORMAT_R32G32_FLOAT:
        LOAD_SCANLINE2( XMFLOAT2, XMLoadFloat2, g_XMIdentityR3 )

    case DXGI_FORMAT_R32G32_UINT:
        LOAD_SCANLINE2( XMUINT2, XMLoadUInt2, g_XMIdentityR3 )

    case DXGI_FORMAT_R32G32_SINT:
        LOAD_SCANLINE2( XMINT2, XMLoadSInt2, g_XMIdentityR3 )

    case DXGI_FORMAT_D32_FLOAT_S8X24_UINT:
        {
            const size_t psize = sizeof(float)+sizeof(uint32_t);
            if ( size >= psize )
            {
                const float * sPtr = reinterpret_cast<const float*>(pSource);
                for( size_t icount = 0; icount < ( size - psize + 1 ); icount += psize )
                {
                    const uint8_t* ps8 = reinterpret_cast<const uint8_t*>( &sPtr[1] );
                    if ( dPtr >= ePtr ) break;
                    *(dPtr++) = XMVectorSet( sPtr[0], static_cast<float>( *ps8 ), 0.f, 1.f );
                    sPtr += 2;
                }
                return true;
            }
        }
        return false;

    case DXGI_FORMAT_R32_FLOAT_X8X24_TYPELESS:
        {
            const size_t psize = sizeof(float)+sizeof(uint32_t);
            if ( size >= psize )
            {
                const float * sPtr = reinterpret_cast<const float*>(pSource);
                for( size_t icount = 0; icount < ( size - psize + 1 ); icount += psize )
                {
                    if ( dPtr >= ePtr ) break;
                    *(dPtr++) = XMVectorSet( sPtr[0], 0.f /* typeless component assumed zero */, 0.f, 1.f );
                    sPtr += 2;
                }
                return true;
            }
        }
        return false;

    case DXGI_FORMAT_X32_TYPELESS_G8X24_UINT:
        {
            const size_t psize = sizeof(float)+sizeof(uint32_t);
            if ( size >= psize )
            {
                const float * sPtr = reinterpret_cast<const float*>(pSource);
                for( size_t icount = 0; icount < ( size - psize + 1 ); icount += psize )
                {
                    const uint8_t* pg8 = reinterpret_cast<const uint8_t*>( &sPtr[1] );
                    if ( dPtr >= ePtr ) break;
                    *(dPtr++) = XMVectorSet( 0.f /* typeless component assumed zero */, static_cast<float>( *pg8 ), 0.f, 1.f );
                    sPtr += 2;
                }
                return true;
            }
        }
        return false;

    case DXGI_FORMAT_R10G10B10A2_UNORM:
        LOAD_SCANLINE( XMUDECN4, XMLoadUDecN4 );

    case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
#if DIRECTX_MATH_VERSION >= 306
        LOAD_SCANLINE( XMUDECN4, XMLoadUDecN4_XR );
#else
        if ( size >= sizeof(XMUDECN4) )
        {
            const XMUDECN4 * __restrict sPtr = reinterpret_cast<const XMUDECN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( dPtr >= ePtr ) break;

                int32_t ElementX = sPtr->v & 0x3FF;
                int32_t ElementY = (sPtr->v >> 10) & 0x3FF;
                int32_t ElementZ = (sPtr->v >> 20) & 0x3FF;

                XMVECTORF32 vResult = {
                    (float)(ElementX - 0x180) / 510.0f,
                    (float)(ElementY - 0x180) / 510.0f,
                    (float)(ElementZ - 0x180) / 510.0f,
                    (float)(sPtr->v >> 30) / 3.0f
                };

                ++sPtr;

                *(dPtr++) = vResult.v;
            }
            return true;
        }
        return false;
#endif

    case DXGI_FORMAT_R10G10B10A2_UINT:
        LOAD_SCANLINE( XMUDEC4, XMLoadUDec4 );

    case DXGI_FORMAT_R11G11B10_FLOAT:
        LOAD_SCANLINE3( XMFLOAT3PK, XMLoadFloat3PK, g_XMIdentityR3 );

    case DXGI_FORMAT_R8G8B8A8_UNORM:
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
        LOAD_SCANLINE( XMUBYTEN4, XMLoadUByteN4 )

    case DXGI_FORMAT_R8G8B8A8_UINT:
        LOAD_SCANLINE( XMUBYTE4, XMLoadUByte4 )

    case DXGI_FORMAT_R8G8B8A8_SNORM:
        LOAD_SCANLINE( XMBYTEN4, XMLoadByteN4 )

    case DXGI_FORMAT_R8G8B8A8_SINT:
        LOAD_SCANLINE( XMBYTE4, XMLoadByte4 )

    case DXGI_FORMAT_R16G16_FLOAT:
        LOAD_SCANLINE2( XMHALF2, XMLoadHalf2, g_XMIdentityR3 )

    case DXGI_FORMAT_R16G16_UNORM:
        LOAD_SCANLINE2( XMUSHORTN2, XMLoadUShortN2, g_XMIdentityR3 )

    case DXGI_FORMAT_R16G16_UINT:
        LOAD_SCANLINE2( XMUSHORT2, XMLoadUShort2, g_XMIdentityR3 )

    case DXGI_FORMAT_R16G16_SNORM:
        LOAD_SCANLINE2( XMSHORTN2, XMLoadShortN2, g_XMIdentityR3 )

    case DXGI_FORMAT_R16G16_SINT:
        LOAD_SCANLINE2( XMSHORT2, XMLoadShort2, g_XMIdentityR3 )

    case DXGI_FORMAT_D32_FLOAT:
    case DXGI_FORMAT_R32_FLOAT:
        if ( size >= sizeof(float) )
        {
            const float* __restrict sPtr = reinterpret_cast<const float*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(float) + 1 ); icount += sizeof(float) )
            {
                XMVECTOR v = XMLoadFloat( sPtr++ );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v, g_XMSelect1000 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R32_UINT:
        if ( size >= sizeof(uint32_t) )
        {
            const uint32_t* __restrict sPtr = reinterpret_cast<const uint32_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(uint32_t) + 1 ); icount += sizeof(uint32_t) )
            {
                XMVECTOR v = XMLoadInt( sPtr++ );
                v = XMConvertVectorUIntToFloat( v, 0 );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v, g_XMSelect1000 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R32_SINT:
        if ( size >= sizeof(int32_t) )
        {
            const int32_t * __restrict sPtr = reinterpret_cast<const int32_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(int32_t) + 1 ); icount += sizeof(int32_t) )
            {
                XMVECTOR v = XMLoadInt( reinterpret_cast<const uint32_t*> (sPtr++) );
                v = XMConvertVectorIntToFloat( v, 0 );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v, g_XMSelect1000 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_D24_UNORM_S8_UINT:
        if ( size >= sizeof(uint32_t) )
        {
            const uint32_t * sPtr = reinterpret_cast<const uint32_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(uint32_t) + 1 ); icount += sizeof(uint32_t) )
            {
                float d = static_cast<float>( *sPtr & 0xFFFFFF ) / 16777215.f;
                float s = static_cast<float>( ( *sPtr & 0xFF000000 ) >> 24 );
                ++sPtr;
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( d, s, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R24_UNORM_X8_TYPELESS:
        if ( size >= sizeof(uint32_t) )
        {
            const uint32_t * sPtr = reinterpret_cast<const uint32_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(uint32_t) + 1 ); icount += sizeof(uint32_t) )
            {
                float r = static_cast<float>( *sPtr & 0xFFFFFF ) / 16777215.f;
                ++sPtr;
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( r, 0.f /* typeless component assumed zero */, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_X24_TYPELESS_G8_UINT:
        if ( size >= sizeof(uint32_t) )
        {
            const uint32_t * sPtr = reinterpret_cast<const uint32_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(uint32_t) + 1 ); icount += sizeof(uint32_t) )
            {
                float g = static_cast<float>( ( *sPtr & 0xFF000000 ) >> 24 );
                ++sPtr;
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( 0.f /* typeless component assumed zero */, g, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8G8_UNORM:
        LOAD_SCANLINE2( XMUBYTEN2, XMLoadUByteN2, g_XMIdentityR3 )

    case DXGI_FORMAT_R8G8_UINT:
        LOAD_SCANLINE2( XMUBYTE2, XMLoadUByte2, g_XMIdentityR3 )

    case DXGI_FORMAT_R8G8_SNORM:
        LOAD_SCANLINE2( XMBYTEN2, XMLoadByteN2, g_XMIdentityR3 )

    case DXGI_FORMAT_R8G8_SINT:
        LOAD_SCANLINE2( XMBYTE2, XMLoadByte2, g_XMIdentityR3 )

    case DXGI_FORMAT_R16_FLOAT:
        if ( size >= sizeof(HALF) )
        {
            const HALF * __restrict sPtr = reinterpret_cast<const HALF*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(HALF) + 1 ); icount += sizeof(HALF) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( XMConvertHalfToFloat(*sPtr++), 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_D16_UNORM:
    case DXGI_FORMAT_R16_UNORM:
        if ( size >= sizeof(uint16_t) )
        {
            const uint16_t* __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(uint16_t) + 1 ); icount += sizeof(uint16_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++) / 65535.f, 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R16_UINT:
        if ( size >= sizeof(uint16_t) )
        {
            const uint16_t * __restrict sPtr = reinterpret_cast<const uint16_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(uint16_t) + 1 ); icount += sizeof(uint16_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++), 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R16_SNORM:
        if ( size >= sizeof(int16_t) )
        {
            const int16_t * __restrict sPtr = reinterpret_cast<const int16_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(int16_t) + 1 ); icount += sizeof(int16_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++) / 32767.f, 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R16_SINT:
        if ( size >= sizeof(int16_t) )
        {
            const int16_t * __restrict sPtr = reinterpret_cast<const int16_t*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(int16_t) + 1 ); icount += sizeof(int16_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++), 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_UNORM:
        if ( size >= sizeof(uint8_t) )
        {
            const uint8_t * __restrict sPtr = reinterpret_cast<const uint8_t*>(pSource);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++) / 255.f, 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_UINT:
        if ( size >= sizeof(uint8_t) )
        {
            const uint8_t * __restrict sPtr = reinterpret_cast<const uint8_t*>(pSource);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++), 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_SNORM:
        if ( size >= sizeof(int8_t) )
        {
            const int8_t * __restrict sPtr = reinterpret_cast<const int8_t*>(pSource);
            for( size_t icount = 0; icount < size; icount += sizeof(int8_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++) / 127.f, 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_SINT:
        if ( size >= sizeof(int8_t) )
        {
            const int8_t * __restrict sPtr = reinterpret_cast<const int8_t*>(pSource);
            for( size_t icount = 0; icount < size; icount += sizeof(int8_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( static_cast<float>(*sPtr++), 0.f, 0.f, 1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_A8_UNORM:
        if ( size >= sizeof(uint8_t) )
        {
            const uint8_t * __restrict sPtr = reinterpret_cast<const uint8_t*>(pSource);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( 0.f, 0.f, 0.f, static_cast<float>(*sPtr++) / 255.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R1_UNORM:
        if ( size >= sizeof(uint8_t) )
        {
            const uint8_t * __restrict sPtr = reinterpret_cast<const uint8_t*>(pSource);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                for( size_t bcount = 8; bcount > 0; --bcount )
                {
                    if ( dPtr >= ePtr ) break;
                    *(dPtr++) = XMVectorSet( (((*sPtr >> (bcount-1)) & 0x1) ? 1.f : 0.f), 0.f, 0.f, 1.f );
                }
                
                ++sPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R9G9B9E5_SHAREDEXP:
#if DIRECTX_MATH_VERSION >= 306
        LOAD_SCANLINE3( XMFLOAT3SE, XMLoadFloat3SE, g_XMIdentityR3 )
#else
        if ( size >= sizeof(XMFLOAT3SE) )
        {
            const XMFLOAT3SE * __restrict sPtr = reinterpret_cast<const XMFLOAT3SE*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMFLOAT3SE) + 1 ); icount += sizeof(XMFLOAT3SE) )
            {
                union { float f; int32_t i; } fi;
                fi.i = 0x33800000 + (sPtr->e << 23);
                float Scale = fi.f;

                XMVECTORF32 v = {
                    Scale * float( sPtr->xm ),
                    Scale * float( sPtr->ym ),
                    Scale * float( sPtr->zm ),
                    1.0f };

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = v;
            }
            return true;
        }
        return false;
#endif

    case DXGI_FORMAT_R8G8_B8G8_UNORM:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            const XMUBYTEN4 * __restrict sPtr = reinterpret_cast<const XMUBYTEN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                XMVECTOR v = XMLoadUByteN4( sPtr++ );
                XMVECTOR v1 = XMVectorSwizzle<0, 3, 2, 1>( v );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v, g_XMSelect1110 );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v1, g_XMSelect1110 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_G8R8_G8B8_UNORM:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            const XMUBYTEN4 * __restrict sPtr = reinterpret_cast<const XMUBYTEN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                XMVECTOR v = XMLoadUByteN4( sPtr++ );
                XMVECTOR v0 = XMVectorSwizzle<1, 0, 3, 2>( v );
                XMVECTOR v1 = XMVectorSwizzle<1, 2, 3, 0>( v );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v0, g_XMSelect1110 );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v1, g_XMSelect1110 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B5G6R5_UNORM:
        if ( size >= sizeof(XMU565) )
        {
            static XMVECTORF32 s_Scale = { 1.f/31.f, 1.f/63.f, 1.f/31.f, 1.f };
            const XMU565 * __restrict sPtr = reinterpret_cast<const XMU565*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMU565) + 1 ); icount += sizeof(XMU565) )
            {
                XMVECTOR v = XMLoadU565( sPtr++ );
                v = XMVectorMultiply( v, s_Scale );
                v = XMVectorSwizzle<2, 1, 0, 3>( v );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v, g_XMSelect1110 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B5G5R5A1_UNORM:
        if ( size >= sizeof(XMU555) )
        {
            static XMVECTORF32 s_Scale = { 1.f/31.f, 1.f/31.f, 1.f/31.f, 1.f };
            const XMU555 * __restrict sPtr = reinterpret_cast<const XMU555*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMU555) + 1 ); icount += sizeof(XMU555) )
            {
                XMVECTOR v = XMLoadU555( sPtr++ );
                v = XMVectorMultiply( v, s_Scale );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSwizzle<2, 1, 0, 3>( v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B8G8R8A8_UNORM:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            const XMUBYTEN4 * __restrict sPtr = reinterpret_cast<const XMUBYTEN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                XMVECTOR v = XMLoadUByteN4( sPtr++ );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSwizzle<2, 1, 0, 3>( v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B8G8R8X8_UNORM:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            const XMUBYTEN4 * __restrict sPtr = reinterpret_cast<const XMUBYTEN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                XMVECTOR v = XMLoadUByteN4( sPtr++ );
                v = XMVectorSwizzle<2, 1, 0, 3>( v );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSelect( g_XMIdentityR3, v, g_XMSelect1110 );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_AYUV:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            const XMUBYTEN4 * __restrict sPtr = reinterpret_cast<const XMUBYTEN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                int v = int(sPtr->x) - 128;
                int u = int(sPtr->y) - 128;
                int y = int(sPtr->z) - 16;
                unsigned int a = sPtr->w;
                ++sPtr;

                // http://msdn.microsoft.com/en-us/library/windows/desktop/dd206750.aspx

                // Y<>  = Y - 16
                // Cb<43> = Cb - 128
                // Cr<43> = Cr - 128

                // R = 1.1644Y<EFBFBD> + 1.5960Cr<EFBFBD>
                // G = 1.1644Y<EFBFBD> - 0.3917Cb<EFBFBD> - 0.8128Cr<EFBFBD>
                // B = 1.1644Y<EFBFBD> + 2.0172Cb<EFBFBD>

                int r = (298 * y +           409 * v + 128) >> 8;
                int g = (298 * y - 100 * u - 208 * v + 128) >> 8;
                int b = (298 * y + 516 * u           + 128) >> 8;

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 255 ) ) / 255.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 255 ) ) / 255.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 255 ) ) / 255.f,
                                         float( a / 255.f ) );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y410:
        if ( size >= sizeof(XMUDECN4) )
        {
            const XMUDECN4 * __restrict sPtr = reinterpret_cast<const XMUDECN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                int64_t u = int(sPtr->x) - 512;
                int64_t y = int(sPtr->y) - 64;
                int64_t v = int(sPtr->z) - 512;
                unsigned int a = sPtr->w;
                ++sPtr;

                // http://msdn.microsoft.com/en-us/library/windows/desktop/bb970578.aspx

                // Y<>  = Y - 64
                // Cb<43> = Cb - 512
                // Cr<43> = Cr - 512

                // R = 1.1678Y<EFBFBD> + 1.6007Cr<EFBFBD>
                // G = 1.1678Y<EFBFBD> - 0.3929Cb<EFBFBD> - 0.8152Cr<EFBFBD>
                // B = 1.1678Y<EFBFBD> + 2.0232Cb<EFBFBD>

                int r = static_cast<int>( (76533 * y +              104905 * v + 32768) >> 16 );
                int g = static_cast<int>( (76533 * y -  25747 * u -  53425 * v + 32768) >> 16 );
                int b = static_cast<int>( (76533 * y + 132590 * u              + 32768) >> 16 );

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 1023 ) ) / 1023.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 1023 ) ) / 1023.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 1023 ) ) / 1023.f,
                                         float( a / 3.f ) );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y416:
        if ( size >= sizeof(XMUSHORTN4) )
        {
            const XMUSHORTN4 * __restrict sPtr = reinterpret_cast<const XMUSHORTN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUSHORTN4) + 1 ); icount += sizeof(XMUSHORTN4) )
            {
                int64_t u = int64_t(sPtr->x) - 32768;
                int64_t y = int64_t(sPtr->y) - 4096;
                int64_t v = int64_t(sPtr->z) - 32768;
                unsigned int a = sPtr->w;
                ++sPtr;

                // http://msdn.microsoft.com/en-us/library/windows/desktop/bb970578.aspx

                // Y<>  = Y - 4096
                // Cb<43> = Cb - 32768
                // Cr<43> = Cr - 32768

                // R = 1.1689Y<EFBFBD> + 1.6023Cr<EFBFBD>
                // G = 1.1689Y<EFBFBD> - 0.3933Cb<EFBFBD> - 0.8160Cr<EFBFBD>
                // B = 1.1689Y<EFBFBD>+ 2.0251Cb<EFBFBD>

                int r = static_cast<int>( (76607 * y +              105006 * v + 32768) >> 16 );
                int g = static_cast<int>( (76607 * y -  25772 * u -  53477 * v + 32768) >> 16 );
                int b = static_cast<int>( (76607 * y + 132718 * u              + 32768) >> 16 );

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( a, 0 ), 65535 ) ) / 65535.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_YUY2:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            const XMUBYTEN4 * __restrict sPtr = reinterpret_cast<const XMUBYTEN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                int y0 = int(sPtr->x) - 16;
                int u  = int(sPtr->y) - 128;
                int y1 = int(sPtr->z) - 16;
                int v  = int(sPtr->w) - 128;
                ++sPtr;

                // See AYUV
                int r = (298 * y0 +           409 * v + 128) >> 8;
                int g = (298 * y0 - 100 * u - 208 * v + 128) >> 8;
                int b = (298 * y0 + 516 * u           + 128) >> 8;

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 255 ) ) / 255.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 255 ) ) / 255.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 255 ) ) / 255.f,
                                         1.f );
                
                r = (298 * y1 +           409 * v + 128) >> 8;
                g = (298 * y1 - 100 * u - 208 * v + 128) >> 8;
                b = (298 * y1 + 516 * u           + 128) >> 8;

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 255 ) ) / 255.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 255 ) ) / 255.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 255 ) ) / 255.f,
                                         1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y210:
        // Same as Y216 with least significant 6 bits set to zero
        if ( size >= sizeof(XMUSHORTN4) )
        {
            const XMUSHORTN4 * __restrict sPtr = reinterpret_cast<const XMUSHORTN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUSHORTN4) + 1 ); icount += sizeof(XMUSHORTN4) )
            {
                int64_t y0 = int64_t(sPtr->x >> 6) - 64;
                int64_t u  = int64_t(sPtr->y >> 6) - 512;
                int64_t y1 = int64_t(sPtr->z >> 6) - 64;
                int64_t v  = int64_t(sPtr->w >> 6) - 512;
                ++sPtr;

                // See Y410
                int r = static_cast<int>( (76533 * y0 +              104905 * v + 32768) >> 16 );
                int g = static_cast<int>( (76533 * y0 -  25747 * u -  53425 * v + 32768) >> 16 );
                int b = static_cast<int>( (76533 * y0 + 132590 * u              + 32768) >> 16 );

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 1023 ) ) / 1023.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 1023 ) ) / 1023.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 1023 ) ) / 1023.f,
                                         1.f );

                r = static_cast<int>( (76533 * y1 +              104905 * v + 32768) >> 16 );
                g = static_cast<int>( (76533 * y1 -  25747 * u -  53425 * v + 32768) >> 16 );
                b = static_cast<int>( (76533 * y1 + 132590 * u              + 32768) >> 16 );

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 1023 ) ) / 1023.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 1023 ) ) / 1023.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 1023 ) ) / 1023.f,
                                         1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y216:
        if ( size >= sizeof(XMUSHORTN4) )
        {
            const XMUSHORTN4 * __restrict sPtr = reinterpret_cast<const XMUSHORTN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUSHORTN4) + 1 ); icount += sizeof(XMUSHORTN4) )
            {
                int64_t y0 = int64_t(sPtr->x) - 4096;
                int64_t u  = int64_t(sPtr->y) - 32768;
                int64_t y1 = int64_t(sPtr->z) - 4096;
                int64_t v  = int64_t(sPtr->w) - 32768;
                ++sPtr;

                // See Y416
                int r = static_cast<int>( (76607 * y0 +              105006 * v + 32768) >> 16 );
                int g = static_cast<int>( (76607 * y0 -  25772 * u -  53477 * v + 32768) >> 16 );
                int b = static_cast<int>( (76607 * y0 + 132718 * u              + 32768) >> 16 );

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 65535 ) ) / 65535.f,
                                         1.f );

                r = static_cast<int>( (76607 * y1 +              105006 * v + 32768) >> 16 );
                g = static_cast<int>( (76607 * y1 -  25772 * u -  53477 * v + 32768) >> 16 );
                b = static_cast<int>( (76607 * y1 + 132718 * u              + 32768) >> 16 );

                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSet( float( std::min<int>( std::max<int>( r, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( g, 0 ), 65535 ) ) / 65535.f,
                                         float( std::min<int>( std::max<int>( b, 0 ), 65535 ) ) / 65535.f,
                                         1.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B4G4R4A4_UNORM:
        if ( size >= sizeof(XMUNIBBLE4) )
        {
            static XMVECTORF32 s_Scale = { 1.f/15.f, 1.f/15.f, 1.f/15.f, 1.f/15.f };
            const XMUNIBBLE4 * __restrict sPtr = reinterpret_cast<const XMUNIBBLE4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUNIBBLE4) + 1 ); icount += sizeof(XMUNIBBLE4) )
            {
                XMVECTOR v = XMLoadUNibble4( sPtr++ );
                v = XMVectorMultiply( v, s_Scale );
                if ( dPtr >= ePtr ) break;
                *(dPtr++) = XMVectorSwizzle<2, 1, 0, 3>( v );
            }
            return true;
        }
        return false;

    case 116 /* DXGI_FORMAT_R10G10B10_7E3_A2_FLOAT */:
        // Xbox One specific 7e3 format
        if ( size >= sizeof(XMUDECN4) )
        {
            const XMUDECN4 * __restrict sPtr = reinterpret_cast<const XMUDECN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( dPtr >= ePtr ) break;

                XMVECTORF32 vResult = {
                    FloatFrom7e3(sPtr->x),
                    FloatFrom7e3(sPtr->y),
                    FloatFrom7e3(sPtr->z),
                    (float)(sPtr->v >> 30) / 3.0f
                };

                ++sPtr;

                *(dPtr++) = vResult.v;
            }
            return true;
        }
        return false;

    case 117 /* DXGI_FORMAT_R10G10B10_6E4_A2_FLOAT */:
        // Xbox One specific 6e4 format
        if ( size >= sizeof(XMUDECN4) )
        {
            const XMUDECN4 * __restrict sPtr = reinterpret_cast<const XMUDECN4*>(pSource);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( dPtr >= ePtr ) break;

                XMVECTORF32 vResult = {
                    FloatFrom6e4(sPtr->x),
                    FloatFrom6e4(sPtr->y),
                    FloatFrom6e4(sPtr->z),
                    (float)(sPtr->v >> 30) / 3.0f
                };

                ++sPtr;

                *(dPtr++) = vResult.v;
            }
            return true;
        }
        return false;

    // We don't support the planar or palettized formats

    default:
        return false;
    }
}

#undef LOAD_SCANLINE
#undef LOAD_SCANLINE3
#undef LOAD_SCANLINE2


//-------------------------------------------------------------------------------------
// Stores an image row from standard RGBA XMVECTOR (aligned) array
//-------------------------------------------------------------------------------------
#define STORE_SCANLINE( type, func )\
        if ( size >= sizeof(type) )\
        {\
            type * __restrict dPtr = reinterpret_cast<type*>(pDestination);\
            for( size_t icount = 0; icount < ( size - sizeof(type) + 1 ); icount += sizeof(type) )\
            {\
                if ( sPtr >= ePtr ) break;\
                func( dPtr++, *sPtr++ );\
            }\
            return true; \
        }\
        return false;

_Use_decl_annotations_
bool _StoreScanline( LPVOID pDestination, size_t size, DXGI_FORMAT format,
                     const XMVECTOR* pSource, size_t count, float threshold )
{
    assert( pDestination && size > 0 );
    assert( pSource && count > 0 && (((uintptr_t)pSource & 0xF) == 0) );
    assert( IsValid(format) && !IsTypeless(format) && !IsCompressed(format) && !IsPlanar(format) && !IsPalettized(format) );

    const XMVECTOR* __restrict sPtr = pSource;
    if ( !sPtr )
        return false;

    const XMVECTOR* ePtr = pSource + count;

    switch( static_cast<int>(format) )
    {
    case DXGI_FORMAT_R32G32B32A32_FLOAT:
        STORE_SCANLINE( XMFLOAT4, XMStoreFloat4 )

    case DXGI_FORMAT_R32G32B32A32_UINT:
        STORE_SCANLINE( XMUINT4, XMStoreUInt4 )

    case DXGI_FORMAT_R32G32B32A32_SINT:
        STORE_SCANLINE( XMINT4, XMStoreSInt4 )

    case DXGI_FORMAT_R32G32B32_FLOAT:
        STORE_SCANLINE( XMFLOAT3, XMStoreFloat3 )

    case DXGI_FORMAT_R32G32B32_UINT:
        STORE_SCANLINE( XMUINT3, XMStoreUInt3 )

    case DXGI_FORMAT_R32G32B32_SINT:
        STORE_SCANLINE( XMINT3, XMStoreSInt3 )

    case DXGI_FORMAT_R16G16B16A16_FLOAT:
        STORE_SCANLINE( XMHALF4, XMStoreHalf4 )

    case DXGI_FORMAT_R16G16B16A16_UNORM:
        STORE_SCANLINE( XMUSHORTN4, XMStoreUShortN4 ) 

    case DXGI_FORMAT_R16G16B16A16_UINT:
        STORE_SCANLINE( XMUSHORT4, XMStoreUShort4 )

    case DXGI_FORMAT_R16G16B16A16_SNORM:
        STORE_SCANLINE( XMSHORTN4, XMStoreShortN4 )

    case DXGI_FORMAT_R16G16B16A16_SINT:
        STORE_SCANLINE( XMSHORT4, XMStoreShort4 )

    case DXGI_FORMAT_R32G32_FLOAT:
        STORE_SCANLINE( XMFLOAT2, XMStoreFloat2 )

    case DXGI_FORMAT_R32G32_UINT:
        STORE_SCANLINE( XMUINT2, XMStoreUInt2 )

    case DXGI_FORMAT_R32G32_SINT:
        STORE_SCANLINE( XMINT2, XMStoreSInt2 )

    case DXGI_FORMAT_D32_FLOAT_S8X24_UINT:
        {
            const size_t psize = sizeof(float)+sizeof(uint32_t);
            if ( size >= psize )
            {
                float *dPtr = reinterpret_cast<float*>(pDestination);
                for( size_t icount = 0; icount < ( size - psize + 1 ); icount += psize )
                {
                    if ( sPtr >= ePtr ) break;
                    XMFLOAT4 f;
                    XMStoreFloat4( &f, *sPtr++ );
                    dPtr[0] = f.x;
                    uint8_t* ps8 = reinterpret_cast<uint8_t*>( &dPtr[1] );
                    ps8[0] = static_cast<uint8_t>( std::min<float>( 255.f, std::max<float>( 0.f, f.y ) ) );
                    ps8[1] = ps8[2] = ps8[3] = 0;
                    dPtr += 2;
                }
                return true;
            }
        }
        return false;

    case DXGI_FORMAT_R10G10B10A2_UNORM:
        STORE_SCANLINE( XMUDECN4, XMStoreUDecN4 );

    case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
#if DIRECTX_MATH_VERSION >= 306
        STORE_SCANLINE( XMUDECN4, XMStoreUDecN4_XR );
#else
        if ( size >= sizeof(XMUDECN4) )
        {
            static const XMVECTORF32  Scale = { 510.0f, 510.0f, 510.0f, 3.0f };
            static const XMVECTORF32  Bias  = { 384.0f, 384.0f, 384.0f, 0.0f };
            static const XMVECTORF32  C     = { 1023.f, 1023.f, 1023.f, 3.f };

            XMUDECN4 * __restrict dPtr = reinterpret_cast<XMUDECN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( sPtr >= ePtr ) break;

                XMVECTOR N = XMVectorMultiplyAdd( *sPtr++, Scale, Bias );
                N = XMVectorClamp( N, g_XMZero, C );

                XMFLOAT4A tmp;
                XMStoreFloat4A(&tmp, N );

                dPtr->v = ((uint32_t)tmp.w << 30)
                           | (((uint32_t)tmp.z & 0x3FF) << 20)
                           | (((uint32_t)tmp.y & 0x3FF) << 10)
                           | (((uint32_t)tmp.x & 0x3FF));
                ++dPtr;
            }
            return true;
        }
        return false;
#endif

    case DXGI_FORMAT_R10G10B10A2_UINT:
        STORE_SCANLINE( XMUDEC4, XMStoreUDec4 );

    case DXGI_FORMAT_R11G11B10_FLOAT:
        STORE_SCANLINE( XMFLOAT3PK, XMStoreFloat3PK );

    case DXGI_FORMAT_R8G8B8A8_UNORM:
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
        STORE_SCANLINE( XMUBYTEN4, XMStoreUByteN4 )

    case DXGI_FORMAT_R8G8B8A8_UINT:
        STORE_SCANLINE( XMUBYTE4, XMStoreUByte4 )

    case DXGI_FORMAT_R8G8B8A8_SNORM:
        STORE_SCANLINE( XMBYTEN4, XMStoreByteN4 )

    case DXGI_FORMAT_R8G8B8A8_SINT:
        STORE_SCANLINE( XMBYTE4, XMStoreByte4 )

    case DXGI_FORMAT_R16G16_FLOAT:
        STORE_SCANLINE( XMHALF2, XMStoreHalf2 )

    case DXGI_FORMAT_R16G16_UNORM:
        STORE_SCANLINE( XMUSHORTN2, XMStoreUShortN2 )

    case DXGI_FORMAT_R16G16_UINT:
        STORE_SCANLINE( XMUSHORT2, XMStoreUShort2 )

    case DXGI_FORMAT_R16G16_SNORM:
        STORE_SCANLINE( XMSHORTN2, XMStoreShortN2 )

    case DXGI_FORMAT_R16G16_SINT:
        STORE_SCANLINE( XMSHORT2, XMStoreShort2 )

    case DXGI_FORMAT_D32_FLOAT:
    case DXGI_FORMAT_R32_FLOAT:
        if ( size >= sizeof(float) )
        {
            float * __restrict dPtr = reinterpret_cast<float*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(float) + 1 ); icount += sizeof(float) )
            {
                if ( sPtr >= ePtr ) break;
                XMStoreFloat( dPtr++, *(sPtr++) );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R32_UINT:
        if ( size >= sizeof(uint32_t) )
        {
            uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(uint32_t) + 1 ); icount += sizeof(uint32_t) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMConvertVectorFloatToUInt( *(sPtr++), 0 );
                XMStoreInt( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R32_SINT:
        if ( size >= sizeof(int32_t) )
        {
            uint32_t * __restrict dPtr = reinterpret_cast<uint32_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(int32_t) + 1 ); icount += sizeof(int32_t) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMConvertVectorFloatToInt( *(sPtr++), 0 );
                XMStoreInt( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_D24_UNORM_S8_UINT:
        if ( size >= sizeof(uint32_t) )
        {
            static const XMVECTORF32 clamp = { 1.f, 255.f, 0.f, 0.f };
            XMVECTOR zero = XMVectorZero();
            uint32_t *dPtr = reinterpret_cast<uint32_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(uint32_t) + 1 ); icount += sizeof(uint32_t) )
            {
                if ( sPtr >= ePtr ) break;
                XMFLOAT4 f;
                XMStoreFloat4( &f, XMVectorClamp( *sPtr++, zero, clamp ) );
                *dPtr++ = (static_cast<uint32_t>( f.x * 16777215.f ) & 0xFFFFFF)
                          | ((static_cast<uint32_t>( f.y ) & 0xFF) << 24);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8G8_UNORM:
        STORE_SCANLINE( XMUBYTEN2, XMStoreUByteN2 )

    case DXGI_FORMAT_R8G8_UINT:
        STORE_SCANLINE( XMUBYTE2, XMStoreUByte2 )

    case DXGI_FORMAT_R8G8_SNORM:
        STORE_SCANLINE( XMBYTEN2, XMStoreByteN2 )

    case DXGI_FORMAT_R8G8_SINT:
        STORE_SCANLINE( XMBYTE2, XMStoreByte2 )

    case DXGI_FORMAT_R16_FLOAT:
        if ( size >= sizeof(HALF) )
        {
            HALF * __restrict dPtr = reinterpret_cast<HALF*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(HALF) + 1 ); icount += sizeof(HALF) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                *(dPtr++) = XMConvertFloatToHalf(v);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_D16_UNORM:
    case DXGI_FORMAT_R16_UNORM:
        if ( size >= sizeof(uint16_t) )
        {
            uint16_t * __restrict dPtr = reinterpret_cast<uint16_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(uint16_t) + 1 ); icount += sizeof(uint16_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 1.f ), 0.f );
                *(dPtr++) = static_cast<uint16_t>( v*65535.f + 0.5f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R16_UINT:
        if ( size >= sizeof(uint16_t) )
        {
            uint16_t * __restrict dPtr = reinterpret_cast<uint16_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(uint16_t) + 1 ); icount += sizeof(uint16_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 65535.f ), 0.f );
                *(dPtr++) = static_cast<uint16_t>(v);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R16_SNORM:
        if ( size >= sizeof(int16_t) )
        {
            int16_t * __restrict dPtr = reinterpret_cast<int16_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(int16_t) + 1 ); icount += sizeof(int16_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 1.f ), -1.f );
                *(dPtr++) = static_cast<int16_t>( v * 32767.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R16_SINT:
        if ( size >= sizeof(int16_t) )
        {
            int16_t * __restrict dPtr = reinterpret_cast<int16_t*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(int16_t) + 1 ); icount += sizeof(int16_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 32767.f ), -32767.f );
                *(dPtr++) = static_cast<int16_t>(v);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_UNORM:
        if ( size >= sizeof(uint8_t) )
        {
            uint8_t * __restrict dPtr = reinterpret_cast<uint8_t*>(pDestination);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 1.f ), 0.f );
                *(dPtr++) = static_cast<uint8_t>( v * 255.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_UINT:
        if ( size >= sizeof(uint8_t) )
        {
            uint8_t * __restrict dPtr = reinterpret_cast<uint8_t*>(pDestination);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 255.f ), 0.f );
                *(dPtr++) = static_cast<uint8_t>(v);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_SNORM:
        if ( size >= sizeof(int8_t) )
        {
            int8_t * __restrict dPtr = reinterpret_cast<int8_t*>(pDestination);
            for( size_t icount = 0; icount < size; icount += sizeof(int8_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 1.f ), -1.f );
                *(dPtr++) = static_cast<int8_t>( v * 127.f );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8_SINT:
        if ( size >= sizeof(int8_t) )
        {
            int8_t * __restrict dPtr = reinterpret_cast<int8_t*>(pDestination);
            for( size_t icount = 0; icount < size; icount += sizeof(int8_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetX( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 127.f ), -127.f );
                *(dPtr++) = static_cast<int8_t>( v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_A8_UNORM:
        if ( size >= sizeof(uint8_t) )
        {
            uint8_t * __restrict dPtr = reinterpret_cast<uint8_t*>(pDestination);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                if ( sPtr >= ePtr ) break;
                float v = XMVectorGetW( *sPtr++ );
                v = std::max<float>( std::min<float>( v, 1.f ), 0.f );
                *(dPtr++) = static_cast<uint8_t>( v * 255.f);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R1_UNORM:
        if ( size >= sizeof(uint8_t) )
        {
            uint8_t * __restrict dPtr = reinterpret_cast<uint8_t*>(pDestination);
            for( size_t icount = 0; icount < size; icount += sizeof(uint8_t) )
            {
                uint8_t pixels = 0;
                for( size_t bcount = 8; bcount > 0; --bcount )
                {
                    if ( sPtr >= ePtr ) break;
                    float v = XMVectorGetX( *sPtr++ );

                    // Absolute thresholding generally doesn't give good results for all images
                    // Picking the 'right' threshold automatically requires whole-image analysis

                    if ( v > 0.25f )
                        pixels |= 1 << (bcount-1);
                }
                *(dPtr++) = pixels;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R9G9B9E5_SHAREDEXP:
#if DIRECTX_MATH_VERSION >= 306
        STORE_SCANLINE( XMFLOAT3SE, XMStoreFloat3SE )
#else
        if ( size >= sizeof(XMFLOAT3SE) )
        {
            static const float maxf9 = float(0x1FF << 7);
            static const float minf9 = float(1.f / (1 << 16));

            XMFLOAT3SE * __restrict dPtr = reinterpret_cast<XMFLOAT3SE*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMFLOAT3SE) + 1 ); icount += sizeof(XMFLOAT3SE) )
            {
                if ( sPtr >= ePtr ) break;

                XMFLOAT3 rgb;
                XMStoreFloat3( &rgb, *(sPtr++) );

                float r = (rgb.x >= 0.f) ? ( (rgb.x > maxf9) ? maxf9 : rgb.x ) : 0.f;
                float g = (rgb.y >= 0.f) ? ( (rgb.y > maxf9) ? maxf9 : rgb.y ) : 0.f;
                float b = (rgb.z >= 0.f) ? ( (rgb.z > maxf9) ? maxf9 : rgb.z ) : 0.f;

                const float max_rg = (r > g) ? r : g;
                const float max_rgb = (max_rg > b) ? max_rg : b;

                const float maxColor = (max_rgb > minf9) ? max_rgb : minf9;

                union { float f; INT32 i; } fi;
                fi.f = maxColor;
                fi.i &= 0xFF800000; // cut off fraction

                dPtr->e = (fi.i - 0x37800000) >> 23;

                fi.i = 0x83000000 - fi.i;
                float ScaleR = fi.f;

                dPtr->xm = static_cast<uint32_t>( round_to_nearest(r * ScaleR) );
                dPtr->ym = static_cast<uint32_t>( round_to_nearest(g * ScaleR) );
                dPtr->zm = static_cast<uint32_t>( round_to_nearest(b * ScaleR) );
                ++dPtr;
            }
            return true;
        }
        return false;
#endif

    case DXGI_FORMAT_R8G8_B8G8_UNORM:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v0 = *sPtr++;
                XMVECTOR v1 = (sPtr < ePtr) ? XMVectorSplatY( *sPtr++ ) : XMVectorZero();
                XMVECTOR v = XMVectorSelect( v1, v0, g_XMSelect1110 );
                XMStoreUByteN4( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_G8R8_G8B8_UNORM:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            static XMVECTORI32 select1101 = {XM_SELECT_1, XM_SELECT_1, XM_SELECT_0, XM_SELECT_1};

            XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v0 = XMVectorSwizzle<1, 0, 3, 2>( *sPtr++ );
                XMVECTOR v1 = (sPtr < ePtr) ? XMVectorSplatY( *sPtr++ ) : XMVectorZero();
                XMVECTOR v = XMVectorSelect( v1, v0, select1101 );
                XMStoreUByteN4( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B5G6R5_UNORM:
        if ( size >= sizeof(XMU565) )
        {
            static XMVECTORF32 s_Scale = { 31.f, 63.f, 31.f, 1.f };
            XMU565 * __restrict dPtr = reinterpret_cast<XMU565*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMU565) + 1 ); icount += sizeof(XMU565) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( *sPtr++ );
                v = XMVectorMultiply( v, s_Scale );
                XMStoreU565( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B5G5R5A1_UNORM:
        if ( size >= sizeof(XMU555) )
        {
            static XMVECTORF32 s_Scale = { 31.f, 31.f, 31.f, 1.f };
            XMU555 * __restrict dPtr = reinterpret_cast<XMU555*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMU555) + 1 ); icount += sizeof(XMU555) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( *sPtr++ );
                v = XMVectorMultiply( v, s_Scale );
                XMStoreU555( dPtr, v );
                dPtr->w = ( XMVectorGetW( v ) > threshold ) ? 1 : 0;
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B8G8R8A8_UNORM:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( *sPtr++ );
                XMStoreUByteN4( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B8G8R8X8_UNORM:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMVectorPermute<2, 1, 0, 7>( *sPtr++, g_XMIdentityR3 );
                XMStoreUByteN4( dPtr++, v );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_AYUV:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                if ( sPtr >= ePtr ) break;
                
                XMUBYTEN4 rgba;
                XMStoreUByteN4( &rgba, *sPtr++ );

                // http://msdn.microsoft.com/en-us/library/windows/desktop/dd206750.aspx

                // Y  =  0.2568R + 0.5041G + 0.1001B + 16
                // Cb = -0.1482R - 0.2910G + 0.4392B + 128
                // Cr =  0.4392R - 0.3678G - 0.0714B + 128

                int y = ( (  66 * rgba.x + 129 * rgba.y +  25 * rgba.z + 128) >> 8) +  16;
                int u = ( ( -38 * rgba.x -  74 * rgba.y + 112 * rgba.z + 128) >> 8) + 128;
                int v = ( ( 112 * rgba.x -  94 * rgba.y -  18 * rgba.z + 128) >> 8) + 128;

                dPtr->x = static_cast<uint8_t>( std::min<int>( std::max<int>( v, 0 ), 255 ) );
                dPtr->y = static_cast<uint8_t>( std::min<int>( std::max<int>( u, 0 ), 255 ) );
                dPtr->z = static_cast<uint8_t>( std::min<int>( std::max<int>( y, 0 ), 255 ) );
                dPtr->w = rgba.w;
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y410:
        if ( size >= sizeof(XMUDECN4) )
        {
            XMUDECN4 * __restrict dPtr = reinterpret_cast<XMUDECN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( sPtr >= ePtr ) break;
                
                XMUDECN4 rgba;
                XMStoreUDecN4( &rgba, *sPtr++ );

                // http://msdn.microsoft.com/en-us/library/windows/desktop/bb970578.aspx

                // Y  =  0.2560R + 0.5027G + 0.0998B + 64
                // Cb = -0.1478R - 0.2902G + 0.4379B + 512
                // Cr =  0.4379R - 0.3667G - 0.0712B + 512

                int64_t r = rgba.x;
                int64_t g = rgba.y;
                int64_t b = rgba.z;

                int y = static_cast<int>( (  16780 * r + 32942 * g +  6544 * b + 32768) >> 16) + 64;
                int u = static_cast<int>( ( -9683  * r - 19017 * g + 28700 * b + 32768) >> 16) + 512;
                int v = static_cast<int>( (  28700 * r - 24033 * g -  4667 * b + 32768) >> 16) + 512;

                dPtr->x = static_cast<uint32_t>( std::min<int>( std::max<int>( u, 0 ), 1023 ) );
                dPtr->y = static_cast<uint32_t>( std::min<int>( std::max<int>( y, 0 ), 1023 ) );
                dPtr->z = static_cast<uint32_t>( std::min<int>( std::max<int>( v, 0 ), 1023 ) );
                dPtr->w = rgba.w;
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y416:
        if ( size >= sizeof(XMUSHORTN4) )
        {
            XMUSHORTN4 * __restrict dPtr = reinterpret_cast<XMUSHORTN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUSHORTN4) + 1 ); icount += sizeof(XMUSHORTN4) )
            {
                if ( sPtr >= ePtr ) break;
                
                XMUSHORTN4 rgba;
                XMStoreUShortN4( &rgba, *sPtr++ );

                // http://msdn.microsoft.com/en-us/library/windows/desktop/bb970578.aspx

                // Y  =  0.2558R + 0.5022G + 0.0998B + 4096
                // Cb = -0.1476R - 0.2899G + 0.4375B + 32768
                // Cr =  0.4375R - 0.3664G - 0.0711B + 32768

                int64_t r = int64_t(rgba.x);
                int64_t g = int64_t(rgba.y);
                int64_t b = int64_t(rgba.z);

                int y = static_cast<int>( (  16763 * r + 32910 * g +  6537 * b + 32768) >> 16) + 4096;
                int u = static_cast<int>( ( -9674  * r - 18998 * g + 28672 * b + 32768) >> 16) + 32768;
                int v = static_cast<int>( (  28672 * r - 24010 * g -  4662 * b + 32768) >> 16) + 32768;

                dPtr->x = static_cast<uint16_t>( std::min<int>( std::max<int>( u, 0 ), 65535 ) );
                dPtr->y = static_cast<uint16_t>( std::min<int>( std::max<int>( y, 0 ), 65535 ) );
                dPtr->z = static_cast<uint16_t>( std::min<int>( std::max<int>( v, 0 ), 65535 ) );
                dPtr->w = rgba.w;
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_YUY2:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUBYTEN4) + 1 ); icount += sizeof(XMUBYTEN4) )
            {
                if ( sPtr >= ePtr ) break;
                
                XMUBYTEN4 rgb1;
                XMStoreUByteN4( &rgb1, *sPtr++ );

                // See AYUV
                int y0 = ( (  66 * rgb1.x + 129 * rgb1.y +  25 * rgb1.z + 128) >> 8) +  16;
                int u0 = ( ( -38 * rgb1.x -  74 * rgb1.y + 112 * rgb1.z + 128) >> 8) + 128;
                int v0 = ( ( 112 * rgb1.x -  94 * rgb1.y -  18 * rgb1.z + 128) >> 8) + 128;

                XMUBYTEN4 rgb2;
                if(sPtr < ePtr)
                {
                    XMStoreUByteN4( &rgb2, *sPtr++ );
                }
                else
                {
                    rgb2.x = rgb2.y = rgb2.z = rgb2.w = 0;
                }

                int y1 = ( (  66 * rgb2.x + 129 * rgb2.y +  25 * rgb2.z + 128) >> 8) +  16;
                int u1 = ( ( -38 * rgb2.x -  74 * rgb2.y + 112 * rgb2.z + 128) >> 8) + 128;
                int v1 = ( ( 112 * rgb2.x -  94 * rgb2.y -  18 * rgb2.z + 128) >> 8) + 128;

                dPtr->x = static_cast<uint8_t>( std::min<int>( std::max<int>( y0, 0 ), 255 ) );
                dPtr->y = static_cast<uint8_t>( std::min<int>( std::max<int>( (u0 + u1) >> 1, 0 ), 255 ) );
                dPtr->z = static_cast<uint8_t>( std::min<int>( std::max<int>( y1, 0 ), 255 ) );
                dPtr->w = static_cast<uint8_t>( std::min<int>( std::max<int>( (v0 + v1) >> 1, 0 ), 255 ) );
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y210:
        // Same as Y216 with least significant 6 bits set to zero
        if ( size >= sizeof(XMUSHORTN4) )
        {
            XMUSHORTN4 * __restrict dPtr = reinterpret_cast<XMUSHORTN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUSHORTN4) + 1 ); icount += sizeof(XMUSHORTN4) )
            {
                if ( sPtr >= ePtr ) break;

                XMUDECN4 rgb1;
                XMStoreUDecN4( &rgb1, *sPtr++ );

                // See Y410
                int64_t r = rgb1.x;
                int64_t g = rgb1.y;
                int64_t b = rgb1.z;

                int y0 = static_cast<int>( (  16780 * r + 32942 * g +  6544 * b + 32768) >> 16) + 64;
                int u0 = static_cast<int>( ( -9683  * r - 19017 * g + 28700 * b + 32768) >> 16) + 512;
                int v0 = static_cast<int>( (  28700 * r - 24033 * g -  4667 * b + 32768) >> 16) + 512;

                XMUDECN4 rgb2;
                if(sPtr < ePtr)
                {
                    XMStoreUDecN4( &rgb2, *sPtr++ );
                }
                else
                {
                    rgb2.x = rgb2.y = rgb2.z = rgb2.w = 0;
                }

                r = rgb2.x;
                g = rgb2.y;
                b = rgb2.z;

                int y1 = static_cast<int>( (  16780 * r + 32942 * g +  6544 * b + 32768) >> 16) + 64;
                int u1 = static_cast<int>( ( -9683  * r - 19017 * g + 28700 * b + 32768) >> 16) + 512;
                int v1 = static_cast<int>( (  28700 * r - 24033 * g -  4667 * b + 32768) >> 16) + 512;

                dPtr->x = static_cast<uint16_t>( std::min<int>( std::max<int>( y0, 0 ), 1023 ) << 6 );
                dPtr->y = static_cast<uint16_t>( std::min<int>( std::max<int>( (u0 + u1) >> 1, 0 ), 1023 ) << 6 );
                dPtr->z = static_cast<uint16_t>( std::min<int>( std::max<int>( y1, 0 ), 1023 ) << 6 );
                dPtr->w = static_cast<uint16_t>( std::min<int>( std::max<int>( (v0 + v1) >> 1, 0 ), 1023 ) << 6 );
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_Y216:
        if ( size >= sizeof(XMUSHORTN4) )
        {
            XMUSHORTN4 * __restrict dPtr = reinterpret_cast<XMUSHORTN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUSHORTN4) + 1 ); icount += sizeof(XMUSHORTN4) )
            {
                if ( sPtr >= ePtr ) break;

                XMUSHORTN4 rgb1;
                XMStoreUShortN4( &rgb1, *sPtr++ );

                // See Y416
                int64_t r = int64_t(rgb1.x);
                int64_t g = int64_t(rgb1.y);
                int64_t b = int64_t(rgb1.z);

                int y0 = static_cast<int>( ( 16763 * r + 32910 * g +  6537 * b + 32768) >> 16) + 4096;
                int u0 = static_cast<int>( (-9674  * r - 18998 * g + 28672 * b + 32768) >> 16) + 32768;
                int v0 = static_cast<int>( ( 28672 * r - 24010 * g -  4662 * b + 32768) >> 16) + 32768;

                XMUSHORTN4 rgb2;
                if(sPtr < ePtr)
                {
                    XMStoreUShortN4( &rgb2, *sPtr++ );
                }
                else
                {
                    rgb2.x = rgb2.y = rgb2.z = rgb2.w = 0;
                }

                r = int64_t(rgb2.x);
                g = int64_t(rgb2.y);
                b = int64_t(rgb2.z);

                int y1 = static_cast<int>( ( 16763 * r + 32910 * g +  6537 * b + 32768) >> 16) + 4096;
                int u1 = static_cast<int>( (-9674  * r - 18998 * g + 28672 * b + 32768) >> 16) + 32768;
                int v1 = static_cast<int>( ( 28672 * r - 24010 * g -  4662 * b + 32768) >> 16) + 32768;

                dPtr->x = static_cast<uint16_t>( std::min<int>( std::max<int>( y0, 0 ), 65535 ) );
                dPtr->y = static_cast<uint16_t>( std::min<int>( std::max<int>( (u0 + u1) >> 1, 0 ), 65535 ) );
                dPtr->z = static_cast<uint16_t>( std::min<int>( std::max<int>( y1, 0 ), 65535 ) );
                dPtr->w = static_cast<uint16_t>( std::min<int>( std::max<int>( (v0 + v1) >> 1, 0 ), 65535 ) );
                ++dPtr;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B4G4R4A4_UNORM:
        if ( size >= sizeof(XMUNIBBLE4) )
        {
            static XMVECTORF32 s_Scale = { 15.f, 15.f, 15.f, 15.f };
            XMUNIBBLE4 * __restrict dPtr = reinterpret_cast<XMUNIBBLE4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUNIBBLE4) + 1 ); icount += sizeof(XMUNIBBLE4) )
            {
                if ( sPtr >= ePtr ) break;
                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( *sPtr++ );
                v = XMVectorMultiply( v, s_Scale );
                XMStoreUNibble4( dPtr++, v );
            }
            return true;
        }
        return false;

    case 116 /* DXGI_FORMAT_R10G10B10_7E3_A2_FLOAT */:
        // Xbox One specific 7e3 format with alpha
        if ( size >= sizeof(XMUDECN4) )
        {
            static const XMVECTORF32  Scale = { 1.0f, 1.0f, 1.0f, 3.0f };
            static const XMVECTORF32  C     = { 31.875f, 31.875f, 31.875f, 3.f };

            XMUDECN4 * __restrict dPtr = reinterpret_cast<XMUDECN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( sPtr >= ePtr ) break;

                XMVECTOR V = XMVectorMultiply( *sPtr++, Scale );
                V = XMVectorClamp( V, g_XMZero, C );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, V );

                dPtr->x = FloatTo7e3( tmp.x );
                dPtr->y = FloatTo7e3( tmp.y );
                dPtr->z = FloatTo7e3( tmp.z );
                dPtr->w = (uint32_t)tmp.w;
                ++dPtr;
            }
            return true;
        }
        return false;

    case 117 /* DXGI_FORMAT_R10G10B10_6E4_A2_FLOAT */:
        // Xbox One specific 6e4 format with alpha
        if ( size >= sizeof(XMUDECN4) )
        {
            static const XMVECTORF32  Scale = { 1.0f, 1.0f, 1.0f, 3.0f };
            static const XMVECTORF32  C     = { 508.f, 508.f, 508.f, 3.f };

            XMUDECN4 * __restrict dPtr = reinterpret_cast<XMUDECN4*>(pDestination);
            for( size_t icount = 0; icount < ( size - sizeof(XMUDECN4) + 1 ); icount += sizeof(XMUDECN4) )
            {
                if ( sPtr >= ePtr ) break;

                XMVECTOR V = XMVectorMultiply( *sPtr++, Scale );
                V = XMVectorClamp( V, g_XMZero, C );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, V );

                dPtr->x = FloatTo6e4( tmp.x );
                dPtr->y = FloatTo6e4( tmp.y );
                dPtr->z = FloatTo6e4( tmp.z );
                dPtr->w = (uint32_t)tmp.w;
                ++dPtr;
            }
            return true;
        }
        return false;

    // We don't support the planar or palettized formats

    default:
        return false;
    }
}

#undef STORE_SCANLINE


//-------------------------------------------------------------------------------------
// Convert DXGI image to/from GUID_WICPixelFormat128bppRGBAFloat (no range conversions)
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
HRESULT _ConvertToR32G32B32A32( const Image& srcImage, ScratchImage& image )
{
    if ( !srcImage.pixels )
        return E_POINTER;

    HRESULT hr = image.Initialize2D( DXGI_FORMAT_R32G32B32A32_FLOAT, srcImage.width, srcImage.height, 1, 1 );
    if ( FAILED(hr) )
        return hr;

    const Image *img = image.GetImage( 0, 0, 0 );
    if ( !img )
    {
        image.Release();
        return E_POINTER;
    }

    uint8_t* pDest = img->pixels;
    if ( !pDest )
    {
        image.Release();
        return E_POINTER;
    }

    const uint8_t *pSrc = srcImage.pixels;
    for( size_t h = 0; h < srcImage.height; ++h )
    {
        if ( !_LoadScanline( reinterpret_cast<XMVECTOR*>(pDest), srcImage.width, pSrc, srcImage.rowPitch, srcImage.format ) )
        {
            image.Release();
            return E_FAIL;
        }

        pSrc += srcImage.rowPitch;
        pDest += img->rowPitch;
    }

    return S_OK;
}

_Use_decl_annotations_
HRESULT _ConvertFromR32G32B32A32( const Image& srcImage, const Image& destImage )
{
    assert( srcImage.format == DXGI_FORMAT_R32G32B32A32_FLOAT );

    if ( !srcImage.pixels || !destImage.pixels )
        return E_POINTER;

    if ( srcImage.width != destImage.width || srcImage.height != destImage.height )
        return E_FAIL;

    const uint8_t *pSrc = srcImage.pixels;
    uint8_t* pDest = destImage.pixels;

    for( size_t h = 0; h < srcImage.height; ++h )
    {
        if ( !_StoreScanline( pDest, destImage.rowPitch, destImage.format, reinterpret_cast<const XMVECTOR*>(pSrc), srcImage.width ) )
            return E_FAIL;

        pSrc += srcImage.rowPitch;
        pDest += destImage.rowPitch;
    }

    return S_OK;
}

_Use_decl_annotations_
HRESULT _ConvertFromR32G32B32A32( const Image& srcImage, DXGI_FORMAT format, ScratchImage& image )
{
    if ( !srcImage.pixels )
        return E_POINTER;

    HRESULT hr = image.Initialize2D( format, srcImage.width, srcImage.height, 1, 1 );
    if ( FAILED(hr) )
        return hr;

    const Image *img = image.GetImage( 0, 0, 0 );
    if ( !img )
    {
        image.Release();
        return E_POINTER;
    }

    hr = _ConvertFromR32G32B32A32( srcImage, *img );
    if ( FAILED(hr) )
    {
        image.Release();
        return hr;
    }

    return S_OK;
}

_Use_decl_annotations_
HRESULT _ConvertFromR32G32B32A32( const Image* srcImages, size_t nimages, const TexMetadata& metadata, DXGI_FORMAT format, ScratchImage& result )
{
    if ( !srcImages )
        return E_POINTER;

    result.Release();

    assert( metadata.format == DXGI_FORMAT_R32G32B32A32_FLOAT );

    TexMetadata mdata2 = metadata;
    mdata2.format = format;
    HRESULT hr = result.Initialize( mdata2 );
    if ( FAILED(hr) )
        return hr;

    if ( nimages != result.GetImageCount() )
    {
        result.Release();
        return E_FAIL;
    }

    const Image* dest = result.GetImages();
    if ( !dest )
    {
        result.Release();
        return E_POINTER;
    }

    for( size_t index=0; index < nimages; ++index )
    {
        const Image& src = srcImages[ index ];
        const Image& dst = dest[ index ];

        assert( src.format == DXGI_FORMAT_R32G32B32A32_FLOAT );
        assert( dst.format == format );

        if ( src.width != dst.width || src.height != dst.height )
        {
            result.Release();
            return E_FAIL;
        }

        const uint8_t* pSrc = src.pixels;
        uint8_t* pDest = dst.pixels;
        if ( !pSrc || !pDest )
        {
            result.Release();
            return E_POINTER;
        }

        for( size_t h=0; h < src.height; ++h )
        {
            if ( !_StoreScanline( pDest, dst.rowPitch, format, reinterpret_cast<const XMVECTOR*>(pSrc), src.width ) )
            {
                result.Release();
                return E_FAIL;
            }

            pSrc += src.rowPitch;
            pDest += dst.rowPitch;
        }
    }

    return S_OK;
}


//-------------------------------------------------------------------------------------
// Convert from Linear RGB to sRGB
//
// if C_linear <= 0.0031308 -> C_srgb = 12.92 * C_linear
// if C_linear >  0.0031308 -> C_srgb = ( 1 + a ) * pow( C_Linear, 1 / 2.4 ) - a
//                             where a = 0.055
//-------------------------------------------------------------------------------------
#if DIRECTX_MATH_VERSION < 306
static inline XMVECTOR XMColorRGBToSRGB( FXMVECTOR rgb )
{
    static const XMVECTORF32 Cutoff = { 0.0031308f, 0.0031308f, 0.0031308f, 1.f };
    static const XMVECTORF32 Linear = { 12.92f, 12.92f, 12.92f, 1.f };
    static const XMVECTORF32 Scale = { 1.055f, 1.055f, 1.055f, 1.f };
    static const XMVECTORF32 Bias = { 0.055f, 0.055f, 0.055f, 0.f };
    static const XMVECTORF32 InvGamma = { 1.0f/2.4f, 1.0f/2.4f, 1.0f/2.4f, 1.f };

    XMVECTOR V = XMVectorSaturate(rgb);
    XMVECTOR V0 = XMVectorMultiply( V, Linear );
    XMVECTOR V1 = Scale * XMVectorPow( V, InvGamma ) - Bias;
    XMVECTOR select = XMVectorLess( V, Cutoff );
    V = XMVectorSelect( V1, V0, select );
    return XMVectorSelect( rgb, V, g_XMSelect1110 );
}
#endif

_Use_decl_annotations_
bool _StoreScanlineLinear( LPVOID pDestination, size_t size, DXGI_FORMAT format,
                           XMVECTOR* pSource, size_t count, DWORD flags )
{
    assert( pDestination && size > 0 );
    assert( pSource && count > 0 && (((uintptr_t)pSource & 0xF) == 0) );
    assert( IsValid(format) && !IsTypeless(format) && !IsCompressed(format) && !IsPlanar(format) && !IsPalettized(format) );

    switch ( format )
    {
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
        flags |= TEX_FILTER_SRGB;
        break;

    case DXGI_FORMAT_R32G32B32A32_FLOAT:
    case DXGI_FORMAT_R32G32B32_FLOAT:
    case DXGI_FORMAT_R16G16B16A16_FLOAT:
    case DXGI_FORMAT_R16G16B16A16_UNORM:
    case DXGI_FORMAT_R32G32_FLOAT:
    case DXGI_FORMAT_R10G10B10A2_UNORM:
    case DXGI_FORMAT_R11G11B10_FLOAT:
    case DXGI_FORMAT_R8G8B8A8_UNORM:
    case DXGI_FORMAT_R16G16_FLOAT:
    case DXGI_FORMAT_R16G16_UNORM:
    case DXGI_FORMAT_R32_FLOAT:
    case DXGI_FORMAT_R8G8_UNORM:
    case DXGI_FORMAT_R16_FLOAT:
    case DXGI_FORMAT_R16_UNORM:
    case DXGI_FORMAT_R8_UNORM:
    case DXGI_FORMAT_R9G9B9E5_SHAREDEXP:
    case DXGI_FORMAT_R8G8_B8G8_UNORM:
    case DXGI_FORMAT_G8R8_G8B8_UNORM:
    case DXGI_FORMAT_B5G6R5_UNORM:
    case DXGI_FORMAT_B5G5R5A1_UNORM:
    case DXGI_FORMAT_B8G8R8A8_UNORM:
    case DXGI_FORMAT_B8G8R8X8_UNORM:
    case DXGI_FORMAT_B4G4R4A4_UNORM:
        break;

    default:
        // can't treat A8, XR, Depth, SNORM, UINT, or SINT as sRGB
        flags &= ~TEX_FILTER_SRGB;
        break;
    }

    // sRGB output processing (Linear RGB -> sRGB)
    if ( flags & TEX_FILTER_SRGB_OUT )
    {
        // To avoid the need for another temporary scanline buffer, we allow this function to overwrite the source buffer in-place
        // Given the intended usage in the filtering routines, this is not a problem.
        XMVECTOR* ptr = pSource;
        for( size_t i=0; i < count; ++i, ++ptr )
        {
            *ptr = XMColorRGBToSRGB( *ptr );
        }
    }

    return _StoreScanline( pDestination, size, format, pSource, count );
}


//-------------------------------------------------------------------------------------
// Convert from sRGB to Linear RGB
//
// if C_srgb <= 0.04045 -> C_linear = C_srgb / 12.92
// if C_srgb >  0.04045 -> C_linear = pow( ( C_srgb + a ) / ( 1 + a ), 2.4 )
//                         where a = 0.055
//-------------------------------------------------------------------------------------
#if DIRECTX_MATH_VERSION < 306
static inline XMVECTOR XMColorSRGBToRGB( FXMVECTOR srgb )
{
    static const XMVECTORF32 Cutoff = { 0.04045f, 0.04045f, 0.04045f, 1.f };
    static const XMVECTORF32 ILinear = { 1.f/12.92f, 1.f/12.92f, 1.f/12.92f, 1.f };
    static const XMVECTORF32 Scale = { 1.f/1.055f, 1.f/1.055f, 1.f/1.055f, 1.f };
    static const XMVECTORF32 Bias = { 0.055f, 0.055f, 0.055f, 0.f };
    static const XMVECTORF32 Gamma = { 2.4f, 2.4f, 2.4f, 1.f };

    XMVECTOR V = XMVectorSaturate(srgb);
    XMVECTOR V0 = XMVectorMultiply( V, ILinear );
    XMVECTOR V1 = XMVectorPow( (V + Bias) * Scale, Gamma );
    XMVECTOR select = XMVectorGreater( V, Cutoff );
    V = XMVectorSelect( V0, V1, select );
    return XMVectorSelect( srgb, V, g_XMSelect1110 );
}
#endif

_Use_decl_annotations_
bool _LoadScanlineLinear( XMVECTOR* pDestination, size_t count,
                          LPCVOID pSource, size_t size, DXGI_FORMAT format, DWORD flags )
{
    assert( pDestination && count > 0 && (((uintptr_t)pDestination & 0xF) == 0) );
    assert( pSource && size > 0 );
    assert( IsValid(format) && !IsTypeless(format,false) && !IsCompressed(format) && !IsPlanar(format) && !IsPalettized(format) );

    switch ( format )
    {
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
        flags |= TEX_FILTER_SRGB;
        break;

    case DXGI_FORMAT_R32G32B32A32_FLOAT:
    case DXGI_FORMAT_R32G32B32_FLOAT:
    case DXGI_FORMAT_R16G16B16A16_FLOAT:
    case DXGI_FORMAT_R16G16B16A16_UNORM:
    case DXGI_FORMAT_R32G32_FLOAT:
    case DXGI_FORMAT_R10G10B10A2_UNORM:
    case DXGI_FORMAT_R11G11B10_FLOAT:
    case DXGI_FORMAT_R8G8B8A8_UNORM:
    case DXGI_FORMAT_R16G16_FLOAT:
    case DXGI_FORMAT_R16G16_UNORM:
    case DXGI_FORMAT_R32_FLOAT:
    case DXGI_FORMAT_R8G8_UNORM:
    case DXGI_FORMAT_R16_FLOAT:
    case DXGI_FORMAT_R16_UNORM:
    case DXGI_FORMAT_R8_UNORM:
    case DXGI_FORMAT_R9G9B9E5_SHAREDEXP:
    case DXGI_FORMAT_R8G8_B8G8_UNORM:
    case DXGI_FORMAT_G8R8_G8B8_UNORM:
    case DXGI_FORMAT_B5G6R5_UNORM:
    case DXGI_FORMAT_B5G5R5A1_UNORM:
    case DXGI_FORMAT_B8G8R8A8_UNORM:
    case DXGI_FORMAT_B8G8R8X8_UNORM:
    case DXGI_FORMAT_B4G4R4A4_UNORM:
        break;

    default:
        // can't treat A8, XR, Depth, SNORM, UINT, or SINT as sRGB
        flags &= ~TEX_FILTER_SRGB;
        break;
    }

    if ( _LoadScanline( pDestination, count, pSource, size, format ) )
    {
        // sRGB input processing (sRGB -> Linear RGB)
        if ( flags & TEX_FILTER_SRGB_IN )
        {
            XMVECTOR* ptr = pDestination;
            for( size_t i=0; i < count; ++i, ++ptr )
            {
                *ptr = XMColorSRGBToRGB( *ptr );
            }
        }

        return true;
    }

    return false;
}


//-------------------------------------------------------------------------------------
// Convert scanline based on source/target formats
//-------------------------------------------------------------------------------------
struct ConvertData
{
    DXGI_FORMAT format;
    size_t datasize;
    DWORD flags;
};

static const ConvertData g_ConvertTable[] = {
    { DXGI_FORMAT_R32G32B32A32_FLOAT,           32, CONVF_FLOAT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R32G32B32A32_UINT,            32, CONVF_UINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R32G32B32A32_SINT,            32, CONVF_SINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R32G32B32_FLOAT,              32, CONVF_FLOAT | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_R32G32B32_UINT,               32, CONVF_UINT | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_R32G32B32_SINT,               32, CONVF_SINT | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_R16G16B16A16_FLOAT,           16, CONVF_FLOAT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R16G16B16A16_UNORM,           16, CONVF_UNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A }, 
    { DXGI_FORMAT_R16G16B16A16_UINT,            16, CONVF_UINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A }, 
    { DXGI_FORMAT_R16G16B16A16_SNORM,           16, CONVF_SNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A }, 
    { DXGI_FORMAT_R16G16B16A16_SINT,            16, CONVF_SINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A }, 
    { DXGI_FORMAT_R32G32_FLOAT,                 32, CONVF_FLOAT | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R32G32_UINT,                  32, CONVF_UINT | CONVF_R | CONVF_G  },
    { DXGI_FORMAT_R32G32_SINT,                  32, CONVF_SINT | CONVF_R | CONVF_G  },
    { DXGI_FORMAT_D32_FLOAT_S8X24_UINT,         32, CONVF_FLOAT | CONVF_DEPTH | CONVF_STENCIL },
    { DXGI_FORMAT_R10G10B10A2_UNORM,            10, CONVF_UNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R10G10B10A2_UINT,             10, CONVF_UINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R11G11B10_FLOAT,              10, CONVF_FLOAT | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_R8G8B8A8_UNORM,               8, CONVF_UNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R8G8B8A8_UNORM_SRGB,          8, CONVF_UNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R8G8B8A8_UINT,                8, CONVF_UINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R8G8B8A8_SNORM,               8, CONVF_SNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R8G8B8A8_SINT,                8, CONVF_SINT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_R16G16_FLOAT,                 16, CONVF_FLOAT | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R16G16_UNORM,                 16, CONVF_UNORM | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R16G16_UINT,                  16, CONVF_UINT | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R16G16_SNORM,                 16, CONVF_SNORM | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R16G16_SINT,                  16, CONVF_SINT | CONVF_R | CONVF_G },
    { DXGI_FORMAT_D32_FLOAT,                    32, CONVF_FLOAT | CONVF_DEPTH },
    { DXGI_FORMAT_R32_FLOAT,                    32, CONVF_FLOAT | CONVF_R },
    { DXGI_FORMAT_R32_UINT,                     32, CONVF_UINT | CONVF_R },
    { DXGI_FORMAT_R32_SINT,                     32, CONVF_SINT | CONVF_R },
    { DXGI_FORMAT_D24_UNORM_S8_UINT,            32, CONVF_UNORM | CONVF_DEPTH | CONVF_STENCIL },
    { DXGI_FORMAT_R8G8_UNORM,                   8, CONVF_UNORM | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R8G8_UINT,                    8, CONVF_UINT | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R8G8_SNORM,                   8, CONVF_SNORM | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R8G8_SINT,                    8, CONVF_SINT | CONVF_R | CONVF_G },
    { DXGI_FORMAT_R16_FLOAT,                    16, CONVF_FLOAT | CONVF_R },
    { DXGI_FORMAT_D16_UNORM,                    16, CONVF_UNORM | CONVF_DEPTH },
    { DXGI_FORMAT_R16_UNORM,                    16, CONVF_UNORM | CONVF_R },
    { DXGI_FORMAT_R16_UINT,                     16, CONVF_UINT | CONVF_R },
    { DXGI_FORMAT_R16_SNORM,                    16, CONVF_SNORM | CONVF_R },
    { DXGI_FORMAT_R16_SINT,                     16, CONVF_SINT | CONVF_R },
    { DXGI_FORMAT_R8_UNORM,                     8, CONVF_UNORM | CONVF_R },
    { DXGI_FORMAT_R8_UINT,                      8, CONVF_UINT | CONVF_R },
    { DXGI_FORMAT_R8_SNORM,                     8, CONVF_SNORM | CONVF_R },
    { DXGI_FORMAT_R8_SINT,                      8, CONVF_SINT | CONVF_R },
    { DXGI_FORMAT_A8_UNORM,                     8, CONVF_UNORM | CONVF_A },
    { DXGI_FORMAT_R1_UNORM,                     1, CONVF_UNORM | CONVF_R },
    { DXGI_FORMAT_R9G9B9E5_SHAREDEXP,           9, CONVF_SHAREDEXP | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_R8G8_B8G8_UNORM,              8, CONVF_UNORM | CONVF_PACKED | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_G8R8_G8B8_UNORM,              8, CONVF_UNORM | CONVF_PACKED | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_BC1_UNORM,                    8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC1_UNORM_SRGB,               8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC2_UNORM,                    8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC2_UNORM_SRGB,               8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC3_UNORM,                    8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC3_UNORM_SRGB,               8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC4_UNORM,                    8, CONVF_UNORM | CONVF_BC | CONVF_R },
    { DXGI_FORMAT_BC4_SNORM,                    8, CONVF_SNORM | CONVF_BC | CONVF_R },
    { DXGI_FORMAT_BC5_UNORM,                    8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G },
    { DXGI_FORMAT_BC5_SNORM,                    8, CONVF_SNORM | CONVF_BC | CONVF_R | CONVF_G },
    { DXGI_FORMAT_B5G6R5_UNORM,                 5, CONVF_UNORM | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_B5G5R5A1_UNORM,               5, CONVF_UNORM | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_B8G8R8A8_UNORM,               8, CONVF_UNORM | CONVF_BGR | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_B8G8R8X8_UNORM,               8, CONVF_UNORM | CONVF_BGR | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM,   10, CONVF_UNORM | CONVF_XR | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_B8G8R8A8_UNORM_SRGB,          8, CONVF_UNORM | CONVF_BGR | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_B8G8R8X8_UNORM_SRGB,          8, CONVF_UNORM | CONVF_BGR | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_BC6H_UF16,                    16, CONVF_FLOAT | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC6H_SF16,                    16, CONVF_FLOAT | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC7_UNORM,                    8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_BC7_UNORM_SRGB,               8, CONVF_UNORM | CONVF_BC | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_AYUV,                         8, CONVF_UNORM | CONVF_YUV | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_Y410,                         10, CONVF_UNORM | CONVF_YUV | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_Y416,                         16, CONVF_UNORM | CONVF_YUV | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT_YUY2,                         8, CONVF_UNORM | CONVF_YUV | CONVF_PACKED | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_Y210,                         10, CONVF_UNORM | CONVF_YUV | CONVF_PACKED | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_Y216,                         16, CONVF_UNORM | CONVF_YUV | CONVF_PACKED | CONVF_R | CONVF_G | CONVF_B },
    { DXGI_FORMAT_B4G4R4A4_UNORM,               4, CONVF_UNORM | CONVF_BGR | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT(116)
      /* DXGI_FORMAT_R10G10B10_7E3_A2_FLOAT */, 10, CONVF_FLOAT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
    { DXGI_FORMAT(117)
      /* DXGI_FORMAT_R10G10B10_6E4_A2_FLOAT */, 10, CONVF_FLOAT | CONVF_R | CONVF_G | CONVF_B | CONVF_A },
};

#pragma prefast( suppress : 25004, "Signature must match bsearch_s" );
static int __cdecl _ConvertCompare( void *context, const void* ptr1, const void *ptr2 )
{
    UNREFERENCED_PARAMETER(context);
    const ConvertData *p1 = reinterpret_cast<const ConvertData*>(ptr1);
    const ConvertData *p2 = reinterpret_cast<const ConvertData*>(ptr2);
    if ( p1->format == p2->format ) return 0;
    else return (p1->format < p2->format ) ? -1 : 1;
}

_Use_decl_annotations_
DWORD _GetConvertFlags( DXGI_FORMAT format )
{
#ifdef _DEBUG
    // Ensure conversion table is in ascending order
    assert( _countof(g_ConvertTable) > 0 );
    DXGI_FORMAT lastvalue = g_ConvertTable[0].format;
    for( size_t index=1; index < _countof(g_ConvertTable); ++index )
    {
        assert( g_ConvertTable[index].format > lastvalue );
        lastvalue = g_ConvertTable[index].format;
    }
#endif

    ConvertData key = { format, 0 };
    const ConvertData* in = (const ConvertData*) bsearch_s( &key, g_ConvertTable, _countof(g_ConvertTable), sizeof(ConvertData),
                                                            _ConvertCompare, 0 );
    return (in) ? in->flags : 0;
}

_Use_decl_annotations_
void _ConvertScanline( XMVECTOR* pBuffer, size_t count, DXGI_FORMAT outFormat, DXGI_FORMAT inFormat, DWORD flags )
{
    assert( pBuffer && count > 0 && (((uintptr_t)pBuffer & 0xF) == 0) );
    assert( IsValid(outFormat) && !IsTypeless(outFormat) && !IsPlanar(outFormat) && !IsPalettized(outFormat) );
    assert( IsValid(inFormat) && !IsTypeless(inFormat) && !IsPlanar(inFormat) && !IsPalettized(inFormat) );

    if ( !pBuffer )
        return;

#ifdef _DEBUG
    // Ensure conversion table is in ascending order
    assert( _countof(g_ConvertTable) > 0 );
    DXGI_FORMAT lastvalue = g_ConvertTable[0].format;
    for( size_t index=1; index < _countof(g_ConvertTable); ++index )
    {
        assert( g_ConvertTable[index].format > lastvalue );
        lastvalue = g_ConvertTable[index].format;
    }
#endif

    // Determine conversion details about source and dest formats
    ConvertData key = { inFormat, 0 };
    const ConvertData* in = (const ConvertData*) bsearch_s( &key, g_ConvertTable, _countof(g_ConvertTable), sizeof(ConvertData),
                                                            _ConvertCompare, 0 );
    key.format = outFormat;
    const ConvertData* out = (const ConvertData*) bsearch_s( &key, g_ConvertTable, _countof(g_ConvertTable), sizeof(ConvertData),
                                                            _ConvertCompare, 0 );
    if ( !in || !out )
    {
        assert(false);
        return;
    }

    assert( _GetConvertFlags( inFormat ) == in->flags );
    assert( _GetConvertFlags( outFormat ) == out->flags );

    // Handle SRGB filtering modes
    switch ( inFormat )
    {
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
    case DXGI_FORMAT_BC1_UNORM_SRGB:
    case DXGI_FORMAT_BC2_UNORM_SRGB:
    case DXGI_FORMAT_BC3_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
    case DXGI_FORMAT_BC7_UNORM_SRGB:
        flags |= TEX_FILTER_SRGB_IN;
        break;

    case DXGI_FORMAT_A8_UNORM:
    case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
        flags &= ~TEX_FILTER_SRGB_IN;
        break;
    }

    switch ( outFormat )
    {
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
    case DXGI_FORMAT_BC1_UNORM_SRGB:
    case DXGI_FORMAT_BC2_UNORM_SRGB:
    case DXGI_FORMAT_BC3_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
    case DXGI_FORMAT_BC7_UNORM_SRGB:
        flags |= TEX_FILTER_SRGB_OUT;
        break;

    case DXGI_FORMAT_A8_UNORM:
    case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
        flags &= ~TEX_FILTER_SRGB_OUT;
        break;
    }

    if ( (flags & (TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT)) == (TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT) )
    {
        flags &= ~(TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT);
    }

    // sRGB input processing (sRGB -> Linear RGB)
    if ( flags & TEX_FILTER_SRGB_IN )
    {
        if ( !(in->flags & CONVF_DEPTH) && ( (in->flags & CONVF_FLOAT) || (in->flags & CONVF_UNORM) ) )
        {
            XMVECTOR* ptr = pBuffer;
            for( size_t i=0; i < count; ++i, ++ptr )
            {
                *ptr = XMColorSRGBToRGB( *ptr );
            }
        }
    }

    // Handle conversion special cases
    DWORD diffFlags = in->flags ^ out->flags;
    if ( diffFlags != 0)
    {
        if ( out->flags & CONVF_UNORM )
        {
            if ( in->flags & CONVF_SNORM )
            {
                // SNORM -> UNORM
                XMVECTOR* ptr = pBuffer;
                for( size_t i=0; i < count; ++i )
                {
                    XMVECTOR v = *ptr;
                    *ptr++ = XMVectorMultiplyAdd( v, g_XMOneHalf, g_XMOneHalf );
                }
            }
            else if ( in->flags & CONVF_FLOAT )
            {
                // FLOAT -> UNORM
                XMVECTOR* ptr = pBuffer;
                for( size_t i=0; i < count; ++i )
                {
                    XMVECTOR v = *ptr;
                    *ptr++ = XMVectorSaturate( v );
                }
            }
        }
        else if ( out->flags & CONVF_SNORM )
        {
            if ( in->flags & CONVF_UNORM )
            {
                // UNORM -> SNORM
                static XMVECTORF32 two = { 2.0f, 2.0f, 2.0f, 2.0f };
                XMVECTOR* ptr = pBuffer;
                for( size_t i=0; i < count; ++i )
                {
                    XMVECTOR v = *ptr;
                    *ptr++ = XMVectorMultiplyAdd( v, two, g_XMNegativeOne );
                }
            }
            else if ( in->flags & CONVF_FLOAT )
            {
                // FLOAT -> SNORM
                XMVECTOR* ptr = pBuffer;
                for( size_t i=0; i < count; ++i )
                {
                    XMVECTOR v = *ptr;
                    *ptr++ = XMVectorClamp( v, g_XMNegativeOne, g_XMOne );
                }
            }
        }

        // !CONVF_A -> CONVF_A is handled because LoadScanline ensures alpha defaults to 1.0 for no-alpha formats

        // CONVF_PACKED cases are handled because LoadScanline/StoreScanline handles packing/unpacking

        if ( ((out->flags & CONVF_RGBA_MASK) == CONVF_A) && !(in->flags & CONVF_A) )
        {
            // !CONVF_A -> A format
            XMVECTOR* ptr = pBuffer;
            for( size_t i=0; i < count; ++i )
            {
                XMVECTOR v = *ptr;
                *ptr++ = XMVectorSplatX( v );
            }
        }
        else if ( ((in->flags & CONVF_RGBA_MASK) == CONVF_A) && !(out->flags & CONVF_A) )
        {
            // A format -> !CONVF_A
            XMVECTOR* ptr = pBuffer;
            for( size_t i=0; i < count; ++i )
            {
                XMVECTOR v = *ptr;
                *ptr++ = XMVectorSplatW( v );
            }
        }
        else if ( (in->flags & CONVF_RGB_MASK) == CONVF_R )
        {
            if ( (out->flags & CONVF_RGB_MASK) == (CONVF_R|CONVF_G|CONVF_B) )
            {
                // R format -> RGB format
                XMVECTOR* ptr = pBuffer;
                for( size_t i=0; i < count; ++i )
                {
                    XMVECTOR v = *ptr;
                    XMVECTOR v1 = XMVectorSplatX( v );
                    *ptr++ = XMVectorSelect( v, v1, g_XMSelect1110 );
                }
            }
            else if ( (out->flags & CONVF_RGB_MASK) == (CONVF_R|CONVF_G) )
            {
                // R format -> RG format
                XMVECTOR* ptr = pBuffer;
                for( size_t i=0; i < count; ++i )
                {
                    XMVECTOR v = *ptr;
                    XMVECTOR v1 = XMVectorSplatX( v );
                    *ptr++ = XMVectorSelect( v, v1, g_XMSelect1100 );
                }
            }
        }
        else if ( (in->flags & CONVF_RGB_MASK) == (CONVF_R|CONVF_G|CONVF_B) )
        {
            if ( (out->flags & CONVF_RGB_MASK) == CONVF_R )
            {
                // RGB format -> R format
                switch( flags & ( TEX_FILTER_RGB_COPY_RED | TEX_FILTER_RGB_COPY_GREEN | TEX_FILTER_RGB_COPY_BLUE ) )
                {
                case TEX_FILTER_RGB_COPY_RED:
                    // Leave data unchanged and the store will handle this...
                    break;

                case TEX_FILTER_RGB_COPY_GREEN:
                    {
                        XMVECTOR* ptr = pBuffer;
                        for( size_t i=0; i < count; ++i )
                        {
                            XMVECTOR v = *ptr;
                            XMVECTOR v1 = XMVectorSplatY( v );
                            *ptr++ = XMVectorSelect( v, v1, g_XMSelect1110 );
                        }
                    }
                    break;

                case TEX_FILTER_RGB_COPY_BLUE:
                    {
                        XMVECTOR* ptr = pBuffer;
                        for( size_t i=0; i < count; ++i )
                        {
                            XMVECTOR v = *ptr;
                            XMVECTOR v1 = XMVectorSplatZ( v );
                            *ptr++ = XMVectorSelect( v, v1, g_XMSelect1110 );
                        }
                    }
                    break;

                default:
                    {
                        XMVECTOR* ptr = pBuffer;
                        for( size_t i=0; i < count; ++i )
                        {
                            XMVECTOR v = *ptr;
                            XMVECTOR v1 = XMVector3Dot( v, g_Grayscale );
                            *ptr++ = XMVectorSelect( v, v1, g_XMSelect1110 );
                        }
                    }
                    break;
                }
            }
            else if ( (out->flags & CONVF_RGB_MASK) == (CONVF_R|CONVF_G) )
            {
                // RGB format -> RG format
                switch( flags & ( TEX_FILTER_RGB_COPY_RED | TEX_FILTER_RGB_COPY_GREEN | TEX_FILTER_RGB_COPY_BLUE ) )
                {
                case TEX_FILTER_RGB_COPY_RED | TEX_FILTER_RGB_COPY_BLUE:
                    {
                        XMVECTOR* ptr = pBuffer;
                        for( size_t i=0; i < count; ++i )
                        {
                            XMVECTOR v = *ptr;
                            XMVECTOR v1 = XMVectorSwizzle<0,2,0,2>( v );
                            *ptr++ = XMVectorSelect( v, v1, g_XMSelect1100 );
                        }
                    }
                    break;

                case TEX_FILTER_RGB_COPY_GREEN | TEX_FILTER_RGB_COPY_BLUE:
                    {
                        XMVECTOR* ptr = pBuffer;
                        for( size_t i=0; i < count; ++i )
                        {
                            XMVECTOR v = *ptr;
                            XMVECTOR v1 = XMVectorSwizzle<1,2,3,0>( v );
                            *ptr++ = XMVectorSelect( v, v1, g_XMSelect1100 );
                        }
                    }
                    break;

                case TEX_FILTER_RGB_COPY_RED | TEX_FILTER_RGB_COPY_GREEN:
                default:
                    // Leave data unchanged and the store will handle this...
                    break;
                }
            }
        }
    }

    // sRGB output processing (Linear RGB -> sRGB)
    if ( flags & TEX_FILTER_SRGB_OUT )
    {
        if ( !(out->flags & CONVF_DEPTH) && ( (out->flags & CONVF_FLOAT) || (out->flags & CONVF_UNORM) ) )
        {
            XMVECTOR* ptr = pBuffer;
            for( size_t i=0; i < count; ++i, ++ptr )
            {
                *ptr = XMColorRGBToSRGB( *ptr );
            }
        }
    }
}


//-------------------------------------------------------------------------------------
// Dithering
//-------------------------------------------------------------------------------------

// 4X4X4 ordered dithering matrix
static const float g_Dither[] =
{
    // (z & 3) + ( (y & 3) * 8) + (x & 3)
    0.468750f,  -0.031250f, 0.343750f, -0.156250f, 0.468750f, -0.031250f, 0.343750f, -0.156250f,
    -0.281250f,  0.218750f, -0.406250f, 0.093750f, -0.281250f, 0.218750f, -0.406250f, 0.093750f,
    0.281250f,  -0.218750f, 0.406250f, -0.093750f, 0.281250f, -0.218750f, 0.406250f, -0.093750f,
    -0.468750f,  0.031250f, -0.343750f, 0.156250f, -0.468750f, 0.031250f, -0.343750f, 0.156250f,
};

static const XMVECTORF32 g_Scale16pc    = {    65535.f, 65535.f, 65535.f, 65535.f };
static const XMVECTORF32 g_Scale15pc    = {    32767.f, 32767.f, 32767.f, 32767.f };
static const XMVECTORF32 g_Scale10pc    = {     1023.f,  1023.f,  1023.f,     3.f };
static const XMVECTORF32 g_Scale8pc     = {      255.f,   255.f,   255.f,   255.f  };
static const XMVECTORF32 g_Scale7pc     = {      127.f,   127.f,   127.f,   127.f  };
static const XMVECTORF32 g_Scale565pc   = {       31.f,    63.f,    31.f,     1.f  };
static const XMVECTORF32 g_Scale5551pc  = {       31.f,    31.f,    31.f,     1.f  };
static const XMVECTORF32 g_Scale4pc     = {       15.f,    15.f,    15.f,    15.f  };

static const XMVECTORF32 g_ErrorWeight3 = { 3.f/16.f, 3.f/16.f, 3.f/16.f, 3.f/16.f };
static const XMVECTORF32 g_ErrorWeight5 = { 5.f/16.f, 5.f/16.f, 5.f/16.f, 5.f/16.f };
static const XMVECTORF32 g_ErrorWeight1 = { 1.f/16.f, 1.f/16.f, 1.f/16.f, 1.f/16.f };
static const XMVECTORF32 g_ErrorWeight7 = { 7.f/16.f, 7.f/16.f, 7.f/16.f, 7.f/16.f };

#define STORE_SCANLINE( type, scalev, clampzero, norm, itype, mask, row, bgr ) \
        if ( size >= sizeof(type) ) \
        { \
            type * __restrict dest = reinterpret_cast<type*>(pDestination); \
            for( size_t i = 0; i < count; ++i ) \
            { \
                ptrdiff_t index = static_cast<ptrdiff_t>( ( row & 1 ) ? ( count - i - 1 ) : i ); \
                ptrdiff_t delta = ( row & 1 ) ? -2 : 0; \
                \
                XMVECTOR v = sPtr[ index ]; \
                if ( bgr ) { v = XMVectorSwizzle<2, 1, 0, 3>( v ); } \
                if ( norm && clampzero ) v = XMVectorSaturate( v ) ; \
                else if ( clampzero ) v = XMVectorClamp( v, g_XMZero, scalev ); \
                else if ( norm ) v = XMVectorClamp( v, g_XMNegativeOne, g_XMOne ); \
                else v = XMVectorClamp( v, -scalev + g_XMOne, scalev ); \
                v = XMVectorAdd( v, vError ); \
                if ( norm ) v = XMVectorMultiply( v, scalev ); \
                \
                XMVECTOR target; \
                if ( pDiffusionErrors ) \
                { \
                    target = XMVectorRound( v ); \
                    vError = XMVectorSubtract( v, target ); \
                    if (norm) vError = XMVectorDivide( vError, scalev ); \
                    \
                    /* Distribute error to next scanline and next pixel */ \
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError ); \
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError ); \
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError ); \
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 ); \
                } \
                else \
                { \
                    /* Applied ordered dither */ \
                    target = XMVectorAdd( v, ordered[ index & 3 ] ); \
                    target = XMVectorRound( target ); \
                } \
                \
                target = XMVectorMin( scalev, target ); \
                target = XMVectorMax( (clampzero) ? g_XMZero : ( -scalev + g_XMOne ), target ); \
                \
                XMFLOAT4A tmp; \
                XMStoreFloat4A( &tmp, target ); \
                \
                auto dPtr = &dest[ index ]; \
                dPtr->x = static_cast<itype>( tmp.x ) & mask; \
                dPtr->y = static_cast<itype>( tmp.y ) & mask; \
                dPtr->z = static_cast<itype>( tmp.z ) & mask; \
                dPtr->w = static_cast<itype>( tmp.w ) & mask; \
            } \
            return true; \
        } \
        return false;

#define STORE_SCANLINE2( type, scalev, clampzero, norm, itype, mask, row ) \
        /* The 2 component cases are always bgr=false */ \
        if ( size >= sizeof(type) ) \
        { \
            type * __restrict dest = reinterpret_cast<type*>(pDestination); \
            for( size_t i = 0; i < count; ++i ) \
            { \
                ptrdiff_t index = static_cast<ptrdiff_t>( ( row & 1 ) ? ( count - i - 1 ) : i ); \
                ptrdiff_t delta = ( row & 1 ) ? -2 : 0; \
                \
                XMVECTOR v = sPtr[ index ]; \
                if ( norm && clampzero ) v = XMVectorSaturate( v ) ; \
                else if ( clampzero ) v = XMVectorClamp( v, g_XMZero, scalev ); \
                else if ( norm ) v = XMVectorClamp( v, g_XMNegativeOne, g_XMOne ); \
                else v = XMVectorClamp( v, -scalev + g_XMOne, scalev ); \
                v = XMVectorAdd( v, vError ); \
                if ( norm ) v = XMVectorMultiply( v, scalev ); \
                \
                XMVECTOR target; \
                if ( pDiffusionErrors ) \
                { \
                    target = XMVectorRound( v ); \
                    vError = XMVectorSubtract( v, target ); \
                    if (norm) vError = XMVectorDivide( vError, scalev ); \
                    \
                    /* Distribute error to next scanline and next pixel */ \
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError ); \
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError ); \
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError ); \
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 ); \
                } \
                else \
                { \
                    /* Applied ordered dither */ \
                    target = XMVectorAdd( v, ordered[ index & 3 ] ); \
                    target = XMVectorRound( target ); \
                } \
                \
                target = XMVectorMin( scalev, target ); \
                target = XMVectorMax( (clampzero) ? g_XMZero : ( -scalev + g_XMOne ), target ); \
                \
                XMFLOAT4A tmp; \
                XMStoreFloat4A( &tmp, target ); \
                \
                auto dPtr = &dest[ index ]; \
                dPtr->x = static_cast<itype>( tmp.x ) & mask; \
                dPtr->y = static_cast<itype>( tmp.y ) & mask; \
            } \
            return true; \
        } \
        return false;

#define STORE_SCANLINE1( type, scalev, clampzero, norm, mask, row, selectw ) \
        /* The 1 component cases are always bgr=false */ \
        if ( size >= sizeof(type) ) \
        { \
            type * __restrict dest = reinterpret_cast<type*>(pDestination); \
            for( size_t i = 0; i < count; ++i ) \
            { \
                ptrdiff_t index = static_cast<ptrdiff_t>( ( row & 1 ) ? ( count - i - 1 ) : i ); \
                ptrdiff_t delta = ( row & 1 ) ? -2 : 0; \
                \
                XMVECTOR v = sPtr[ index ]; \
                if ( norm && clampzero ) v = XMVectorSaturate( v ) ; \
                else if ( clampzero ) v = XMVectorClamp( v, g_XMZero, scalev ); \
                else if ( norm ) v = XMVectorClamp( v, g_XMNegativeOne, g_XMOne ); \
                else v = XMVectorClamp( v, -scalev + g_XMOne, scalev ); \
                v = XMVectorAdd( v, vError ); \
                if ( norm ) v = XMVectorMultiply( v, scalev ); \
                \
                XMVECTOR target; \
                if ( pDiffusionErrors ) \
                { \
                    target = XMVectorRound( v ); \
                    vError = XMVectorSubtract( v, target ); \
                    if (norm) vError = XMVectorDivide( vError, scalev ); \
                    \
                    /* Distribute error to next scanline and next pixel */ \
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError ); \
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError ); \
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError ); \
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 ); \
                } \
                else \
                { \
                    /* Applied ordered dither */ \
                    target = XMVectorAdd( v, ordered[ index & 3 ] ); \
                    target = XMVectorRound( target ); \
                } \
                \
                target = XMVectorMin( scalev, target ); \
                target = XMVectorMax( (clampzero) ? g_XMZero : ( -scalev + g_XMOne ), target ); \
                \
                dest[ index ] = static_cast<type>( (selectw) ? XMVectorGetW( target ) : XMVectorGetX( target ) ) & mask; \
            } \
            return true; \
        } \
        return false;

#pragma warning(push)
#pragma warning( disable : 4127 )

_Use_decl_annotations_
bool _StoreScanlineDither( LPVOID pDestination, size_t size, DXGI_FORMAT format,
                           XMVECTOR* pSource, size_t count, float threshold, size_t y, size_t z, XMVECTOR* pDiffusionErrors )
{
    assert( pDestination && size > 0 );
    assert( pSource && count > 0 && (((uintptr_t)pSource & 0xF) == 0) );
    assert( IsValid(format) && !IsTypeless(format) && !IsCompressed(format) && !IsPlanar(format) && !IsPalettized(format) );

    XMVECTOR ordered[4];
    if ( pDiffusionErrors )
    {
        // If pDiffusionErrors != 0, then this function performs error diffusion dithering (aka Floyd-Steinberg dithering)

        // To avoid the need for another temporary scanline buffer, we allow this function to overwrite the source buffer in-place
        // Given the intended usage in the conversion routines, this is not a problem.

        XMVECTOR* ptr = pSource;
        const XMVECTOR* err = pDiffusionErrors + 1;
        for( size_t i=0; i < count; ++i )
        {
            // Add contribution from previous scanline
            XMVECTOR v = XMVectorAdd( *ptr, *err++ );
            *ptr++ = v;
        }

        // Reset errors for next scanline
        memset( pDiffusionErrors, 0, sizeof(XMVECTOR)*(count+2) );
    }
    else
    {
        // If pDiffusionErrors == 0, then this function performs ordered dithering

        XMVECTOR dither = XMLoadFloat4( reinterpret_cast<const XMFLOAT4*>( g_Dither + (z & 3) + ( (y & 3) * 8 ) ) );

        ordered[0] = XMVectorSplatX( dither );
        ordered[1] = XMVectorSplatY( dither );
        ordered[2] = XMVectorSplatZ( dither );
        ordered[3] = XMVectorSplatW( dither );
    }

    const XMVECTOR* __restrict sPtr = pSource;
    if ( !sPtr )
        return false;

    XMVECTOR vError = XMVectorZero();

    switch( format )
    {
    case DXGI_FORMAT_R16G16B16A16_UNORM:
        STORE_SCANLINE( XMUSHORTN4, g_Scale16pc, true, true, uint16_t, 0xFFFF, y, false )

    case DXGI_FORMAT_R16G16B16A16_UINT:
        STORE_SCANLINE( XMUSHORT4, g_Scale16pc, true, false, uint16_t, 0xFFFF, y, false )

    case DXGI_FORMAT_R16G16B16A16_SNORM:
        STORE_SCANLINE( XMSHORTN4, g_Scale15pc, false, true, int16_t, 0xFFFF, y, false )

    case DXGI_FORMAT_R16G16B16A16_SINT:
        STORE_SCANLINE( XMSHORT4, g_Scale15pc, false, false, int16_t, 0xFFFF, y, false )

    case DXGI_FORMAT_R10G10B10A2_UNORM:
        STORE_SCANLINE( XMUDECN4, g_Scale10pc, true, true, uint16_t, 0x3FF, y, false )

    case DXGI_FORMAT_R10G10B10A2_UINT:
        STORE_SCANLINE( XMUDEC4, g_Scale10pc, true, false, uint16_t, 0x3FF, y, false )

    case DXGI_FORMAT_R10G10B10_XR_BIAS_A2_UNORM:
        if ( size >= sizeof(XMUDEC4) )
        {
            static const XMVECTORF32  Scale = { 510.0f, 510.0f, 510.0f, 3.0f };
            static const XMVECTORF32  Bias  = { 384.0f, 384.0f, 384.0f, 0.0f };
            static const XMVECTORF32  MinXR = { -0.7529f, -0.7529f, -0.7529f, 0.f };
            static const XMVECTORF32  MaxXR = { 1.2529f, 1.2529f, 1.2529f, 1.0f };

            XMUDEC4 * __restrict dest = reinterpret_cast<XMUDEC4*>(pDestination);
            for( size_t i = 0; i < count; ++i )
            {
                ptrdiff_t index = static_cast<ptrdiff_t>( ( y & 1 ) ? ( count - i - 1 ) : i );
                ptrdiff_t delta = ( y & 1 ) ? -2 : 0;

                XMVECTOR v = XMVectorClamp( sPtr[ index ], MinXR, MaxXR );
                v = XMVectorMultiplyAdd( v, Scale, vError );

                XMVECTOR target;
                if ( pDiffusionErrors )
                {
                    target = XMVectorRound( v );
                    vError = XMVectorSubtract( v, target );
                    vError = XMVectorDivide( vError, Scale );

                    // Distribute error to next scanline and next pixel
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError );
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError );
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError );
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 );
                }
                else
                {
                    // Applied ordered dither
                    target = XMVectorAdd( v, ordered[ index & 3 ] );
                    target = XMVectorRound( target );
                }

                target = XMVectorAdd( target, Bias );
                target = XMVectorClamp( target, g_XMZero, g_Scale10pc );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, target );

                auto dPtr = &dest[ index ];
                dPtr->x = static_cast<uint16_t>( tmp.x ) & 0x3FF;
                dPtr->y = static_cast<uint16_t>( tmp.y ) & 0x3FF;
                dPtr->z = static_cast<uint16_t>( tmp.z ) & 0x3FF;
                dPtr->w = static_cast<uint16_t>( tmp.w );
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8G8B8A8_UNORM:
    case DXGI_FORMAT_R8G8B8A8_UNORM_SRGB:
        STORE_SCANLINE( XMUBYTEN4, g_Scale8pc, true, true, uint8_t, 0xFF, y, false )

    case DXGI_FORMAT_R8G8B8A8_UINT:
        STORE_SCANLINE( XMUBYTE4, g_Scale8pc, true, false, uint8_t, 0xFF, y, false )

    case DXGI_FORMAT_R8G8B8A8_SNORM:
        STORE_SCANLINE( XMBYTEN4, g_Scale7pc, false, true, int8_t, 0xFF, y, false )

    case DXGI_FORMAT_R8G8B8A8_SINT:
        STORE_SCANLINE( XMBYTE4, g_Scale7pc, false, false, int8_t, 0xFF, y, false )

    case DXGI_FORMAT_R16G16_UNORM:
        STORE_SCANLINE2( XMUSHORTN2, g_Scale16pc, true, true, uint16_t, 0xFFFF, y )

    case DXGI_FORMAT_R16G16_UINT:
        STORE_SCANLINE2( XMUSHORT2, g_Scale16pc, true, false, uint16_t, 0xFFFF, y )

    case DXGI_FORMAT_R16G16_SNORM:
        STORE_SCANLINE2( XMSHORTN2, g_Scale15pc, false, true, int16_t, 0xFFFF, y )

    case DXGI_FORMAT_R16G16_SINT:
        STORE_SCANLINE2( XMSHORT2, g_Scale15pc, false, false, int16_t, 0xFFFF, y )

    case DXGI_FORMAT_D24_UNORM_S8_UINT:
        if ( size >= sizeof(uint32_t) )
        {
            static const XMVECTORF32 Clamp  = {       1.f,  255.f, 0.f, 0.f };
            static const XMVECTORF32 Scale  = { 16777215.f,   1.f, 0.f, 0.f };
            static const XMVECTORF32 Scale2 = { 16777215.f, 255.f, 0.f, 0.f };

            uint32_t * __restrict dest = reinterpret_cast<uint32_t*>(pDestination);
            for( size_t i = 0; i < count; ++i )
            {
                ptrdiff_t index = static_cast<ptrdiff_t>( ( y & 1 ) ? ( count - i - 1 ) : i );
                ptrdiff_t delta = ( y & 1 ) ? -2 : 0;

                XMVECTOR v = XMVectorClamp( sPtr[ index ], g_XMZero, Clamp );
                v = XMVectorAdd( v, vError );
                v = XMVectorMultiply( v, Scale );

                XMVECTOR target;
                if ( pDiffusionErrors )
                {
                    target = XMVectorRound( v );
                    vError = XMVectorSubtract( v, target );
                    vError = XMVectorDivide( vError, Scale );

                    // Distribute error to next scanline and next pixel
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError );
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError );
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError );
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 );
                }
                else
                {
                    // Applied ordered dither
                    target = XMVectorAdd( v, ordered[ index & 3 ] );
                    target = XMVectorRound( target );
                }

                target = XMVectorClamp( target, g_XMZero, Scale2 );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, target );

                auto dPtr = &dest[ index ];
                *dPtr = (static_cast<uint32_t>( tmp.x ) & 0xFFFFFF)
                        | ((static_cast<uint32_t>( tmp.y ) & 0xFF) << 24);
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_R8G8_UNORM:
        STORE_SCANLINE2( XMUBYTEN2, g_Scale8pc, true, true, uint8_t, 0xFF, y )

    case DXGI_FORMAT_R8G8_UINT:
        STORE_SCANLINE2( XMUBYTE2, g_Scale8pc, true, false, uint8_t, 0xFF, y )

    case DXGI_FORMAT_R8G8_SNORM:
        STORE_SCANLINE2( XMBYTEN2, g_Scale7pc, false, true, int8_t, 0xFF, y )

    case DXGI_FORMAT_R8G8_SINT:
        STORE_SCANLINE2( XMBYTE2, g_Scale7pc, false, false, int8_t, 0xFF, y )

    case DXGI_FORMAT_D16_UNORM:
    case DXGI_FORMAT_R16_UNORM:
        STORE_SCANLINE1( uint16_t, g_Scale16pc, true, true, 0xFFFF, y, false )

    case DXGI_FORMAT_R16_UINT:
        STORE_SCANLINE1( uint16_t, g_Scale16pc, true, false, 0xFFFF, y, false )

    case DXGI_FORMAT_R16_SNORM:
        STORE_SCANLINE1( int16_t, g_Scale15pc, false, true, 0xFFFF, y, false )

    case DXGI_FORMAT_R16_SINT:
        STORE_SCANLINE1( int16_t, g_Scale15pc, false, false, 0xFFFF, y, false )

    case DXGI_FORMAT_R8_UNORM:
        STORE_SCANLINE1( uint8_t, g_Scale8pc, true, true, 0xFF, y, false )

    case DXGI_FORMAT_R8_UINT:
        STORE_SCANLINE1( uint8_t, g_Scale8pc, true, false, 0xFF, y, false )

    case DXGI_FORMAT_R8_SNORM:
        STORE_SCANLINE1( int8_t, g_Scale7pc, false, true, 0xFF, y, false )

    case DXGI_FORMAT_R8_SINT:
        STORE_SCANLINE1( int8_t, g_Scale7pc, false, false, 0xFF, y, false )

    case DXGI_FORMAT_A8_UNORM:
        STORE_SCANLINE1( uint8_t, g_Scale8pc, true, true, 0xFF, y, true )

    case DXGI_FORMAT_B5G6R5_UNORM:
        if ( size >= sizeof(XMU565) )
        {
            XMU565 * __restrict dest = reinterpret_cast<XMU565*>(pDestination);
            for( size_t i = 0; i < count; ++i )
            {
                ptrdiff_t index = static_cast<ptrdiff_t>( ( y & 1 ) ? ( count - i - 1 ) : i );
                ptrdiff_t delta = ( y & 1 ) ? -2 : 0;

                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( sPtr[ index ] );
                v = XMVectorSaturate( v );
                v = XMVectorAdd( v, vError );
                v = XMVectorMultiply( v, g_Scale565pc );

                XMVECTOR target;
                if ( pDiffusionErrors )
                {
                    target = XMVectorRound( v );
                    vError = XMVectorSubtract( v, target );
                    vError = XMVectorDivide( vError, g_Scale565pc );

                    // Distribute error to next scanline and next pixel
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError );
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError );
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError );
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 );
                }
                else
                {
                    // Applied ordered dither
                    target = XMVectorAdd( v, ordered[ index & 3 ] );
                    target = XMVectorRound( target );
                }

                target = XMVectorClamp( target, g_XMZero, g_Scale565pc );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, target );

                auto dPtr = &dest[ index ];
                dPtr->x = static_cast<uint16_t>( tmp.x ) & 0x1F;
                dPtr->y = static_cast<uint16_t>( tmp.y ) & 0x3F;
                dPtr->z = static_cast<uint16_t>( tmp.z ) & 0x1F;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B5G5R5A1_UNORM:
        if ( size >= sizeof(XMU555) )
        {
            XMU555 * __restrict dest = reinterpret_cast<XMU555*>(pDestination);
            for( size_t i = 0; i < count; ++i )
            {
                ptrdiff_t index = static_cast<ptrdiff_t>( ( y & 1 ) ? ( count - i - 1 ) : i );
                ptrdiff_t delta = ( y & 1 ) ? -2 : 0;

                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( sPtr[ index ] );
                v = XMVectorSaturate( v );
                v = XMVectorAdd( v, vError );
                v = XMVectorMultiply( v, g_Scale5551pc );

                XMVECTOR target;
                if ( pDiffusionErrors )
                {
                    target = XMVectorRound( v );
                    vError = XMVectorSubtract( v, target );
                    vError = XMVectorDivide( vError, g_Scale5551pc );

                    // Distribute error to next scanline and next pixel
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError );
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError );
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError );
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 );
                }
                else
                {
                    // Applied ordered dither
                    target = XMVectorAdd( v, ordered[ index & 3 ] );
                    target = XMVectorRound( target );
                }

                target = XMVectorClamp( target, g_XMZero, g_Scale5551pc );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, target );

                auto dPtr = &dest[ index ];
                dPtr->x = static_cast<uint16_t>( tmp.x ) & 0x1F;
                dPtr->y = static_cast<uint16_t>( tmp.y ) & 0x1F;
                dPtr->z = static_cast<uint16_t>( tmp.z ) & 0x1F;
                dPtr->w = ( XMVectorGetW( target ) > threshold ) ? 1 : 0;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B8G8R8A8_UNORM:
    case DXGI_FORMAT_B8G8R8A8_UNORM_SRGB:
        STORE_SCANLINE( XMUBYTEN4, g_Scale8pc, true, true, uint8_t, 0xFF, y, true )

    case DXGI_FORMAT_B8G8R8X8_UNORM:
    case DXGI_FORMAT_B8G8R8X8_UNORM_SRGB:
        if ( size >= sizeof(XMUBYTEN4) )
        {
            XMUBYTEN4 * __restrict dest = reinterpret_cast<XMUBYTEN4*>(pDestination);
            for( size_t i = 0; i < count; ++i )
            {
                ptrdiff_t index = static_cast<ptrdiff_t>( ( y & 1 ) ? ( count - i - 1 ) : i );
                ptrdiff_t delta = ( y & 1 ) ? -2 : 0;

                XMVECTOR v = XMVectorSwizzle<2, 1, 0, 3>( sPtr[ index ] );
                v = XMVectorSaturate( v );
                v = XMVectorAdd( v, vError );
                v = XMVectorMultiply( v, g_Scale8pc );

                XMVECTOR target;
                if ( pDiffusionErrors )
                {
                    target = XMVectorRound( v );
                    vError = XMVectorSubtract( v, target );
                    vError = XMVectorDivide( vError, g_Scale8pc );

                    // Distribute error to next scanline and next pixel
                    pDiffusionErrors[ index-delta ]   += XMVectorMultiply( g_ErrorWeight3, vError );
                    pDiffusionErrors[ index+1 ]       += XMVectorMultiply( g_ErrorWeight5, vError );
                    pDiffusionErrors[ index+2+delta ] += XMVectorMultiply( g_ErrorWeight1, vError );
                    vError = XMVectorMultiply( vError, g_ErrorWeight7 );
                }
                else
                {
                    // Applied ordered dither
                    target = XMVectorAdd( v, ordered[ index & 3 ] );
                    target = XMVectorRound( target );
                }

                target = XMVectorClamp( target, g_XMZero, g_Scale8pc );

                XMFLOAT4A tmp;
                XMStoreFloat4A( &tmp, target );

                auto dPtr = &dest[ index ];
                dPtr->x = static_cast<uint8_t>( tmp.x ) & 0xFF;
                dPtr->y = static_cast<uint8_t>( tmp.y ) & 0xFF;
                dPtr->z = static_cast<uint8_t>( tmp.z ) & 0xFF;
                dPtr->w = 0;
            }
            return true;
        }
        return false;

    case DXGI_FORMAT_B4G4R4A4_UNORM:
        STORE_SCANLINE( XMUNIBBLE4, g_Scale4pc, true, true, uint8_t, 0xF, y, true )

    default:
        return _StoreScanline( pDestination, size, format, pSource, count, threshold );
    }
}

#pragma warning(pop)

#undef STORE_SCANLINE
#undef STORE_SCANLINE2
#undef STORE_SCANLINE1


//-------------------------------------------------------------------------------------
// Selection logic for using WIC vs. our own routines
//-------------------------------------------------------------------------------------
static inline bool _UseWICConversion( _In_ DWORD filter, _In_ DXGI_FORMAT sformat, _In_ DXGI_FORMAT tformat,
                                      _Out_ WICPixelFormatGUID& pfGUID, _Out_ WICPixelFormatGUID& targetGUID )
{
    memcpy( &pfGUID, &GUID_NULL, sizeof(GUID) );
    memcpy( &targetGUID, &GUID_NULL, sizeof(GUID) );

    if ( filter & TEX_FILTER_FORCE_NON_WIC )
    {
        // Explicit flag indicates use of non-WIC code paths
        return false;
    }

    if ( !_DXGIToWIC( sformat, pfGUID ) || !_DXGIToWIC( tformat, targetGUID ) )
    {
        // Source or target format are not WIC supported native pixel formats
        return false;
    }

    if ( filter & TEX_FILTER_FORCE_WIC )
    {
        // Explicit flag to use WIC code paths, skips all the case checks below
        return true;
    }

    if ( filter & TEX_FILTER_SEPARATE_ALPHA )
    {
        // Alpha is not premultiplied, so use non-WIC code paths
        return false;
    }

    // Check for special cases
    switch ( sformat )
    {
    case DXGI_FORMAT_R32G32B32A32_FLOAT:
    case DXGI_FORMAT_R32G32B32_FLOAT:
    case DXGI_FORMAT_R16G16B16A16_FLOAT:
        switch( tformat )
        {
        case DXGI_FORMAT_R16_FLOAT:
        case DXGI_FORMAT_R32_FLOAT:
        case DXGI_FORMAT_D32_FLOAT:
            // WIC converts via UNORM formats and ends up converting colorspaces for these cases
        case DXGI_FORMAT_A8_UNORM:
            // Conversion logic for these kinds of textures is unintuitive for WIC code paths
            return false;
        }
        break;
    
    case DXGI_FORMAT_R16_FLOAT:
        switch( tformat )
        {
        case DXGI_FORMAT_R32_FLOAT:
        case DXGI_FORMAT_D32_FLOAT:
            // WIC converts via UNORM formats and ends up converting colorspaces for these cases
        case DXGI_FORMAT_A8_UNORM:
            // Conversion logic for these kinds of textures is unintuitive for WIC code paths
            return false;
        }
        break;

    case DXGI_FORMAT_A8_UNORM:
        // Conversion logic for these kinds of textures is unintuitive for WIC code paths
        return false;

    default:
        switch( tformat )
        {
        case DXGI_FORMAT_A8_UNORM:
            // Conversion logic for these kinds of textures is unintuitive for WIC code paths
            return false;
        }
    }
    
    // Check for implicit color space changes
    if ( IsSRGB( sformat ) )
        filter |= TEX_FILTER_SRGB_IN;

    if ( IsSRGB( tformat ) )
        filter |= TEX_FILTER_SRGB_OUT;

    if ( (filter & (TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT)) == (TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT) )
    {
        filter &= ~(TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT);
    }

    DWORD wicsrgb = _CheckWICColorSpace( pfGUID, targetGUID );

    if ( wicsrgb != (filter & (TEX_FILTER_SRGB_IN|TEX_FILTER_SRGB_OUT)) )
    {
        // WIC will perform a colorspace conversion we didn't request
        return false;
    }

    return true;
}


//-------------------------------------------------------------------------------------
// Convert the source image using WIC
//-------------------------------------------------------------------------------------
static HRESULT _ConvertUsingWIC( _In_ const Image& srcImage, _In_ const WICPixelFormatGUID& pfGUID,
                                 _In_ const WICPixelFormatGUID& targetGUID,
                                 _In_ DWORD filter, _In_ float threshold,  _In_ const Image& destImage )
{
    assert( srcImage.width == destImage.width );
    assert( srcImage.height == destImage.height );

    IWICImagingFactory* pWIC = _GetWIC();
    if ( !pWIC )
        return E_NOINTERFACE;

    ComPtr<IWICFormatConverter> FC;
    HRESULT hr = pWIC->CreateFormatConverter( FC.GetAddressOf() );
    if ( FAILED(hr) )
        return hr;

    // Note that WIC conversion ignores the TEX_FILTER_SRGB_IN and TEX_FILTER_SRGB_OUT flags,
    // but also always assumes UNORM <-> FLOAT conversions are changing color spaces sRGB <-> scRGB

    BOOL canConvert = FALSE;
    hr = FC->CanConvert( pfGUID, targetGUID, &canConvert );
    if ( FAILED(hr) || !canConvert )
    {
        // This case is not an issue for the subset of WIC formats that map directly to DXGI
        return E_UNEXPECTED;
    }

    ComPtr<IWICBitmap> source;
    hr = pWIC->CreateBitmapFromMemory( static_cast<UINT>( srcImage.width ), static_cast<UINT>( srcImage.height ), pfGUID,
                                       static_cast<UINT>( srcImage.rowPitch ), static_cast<UINT>( srcImage.slicePitch ),
                                       srcImage.pixels, source.GetAddressOf() );
    if ( FAILED(hr) )
        return hr;

    hr = FC->Initialize( source.Get(), targetGUID, _GetWICDither( filter ), 0, threshold * 100.f, WICBitmapPaletteTypeCustom );
    if ( FAILED(hr) )
        return hr;

    hr = FC->CopyPixels( 0, static_cast<UINT>( destImage.rowPitch ), static_cast<UINT>( destImage.slicePitch ), destImage.pixels );  
    if ( FAILED(hr) )
        return hr;

    return S_OK;
}


//-------------------------------------------------------------------------------------
// Convert the source image (not using WIC)
//-------------------------------------------------------------------------------------
static HRESULT _Convert( _In_ const Image& srcImage, _In_ DWORD filter, _In_ const Image& destImage, _In_ float threshold, _In_ size_t z )
{
    assert( srcImage.width == destImage.width );
    assert( srcImage.height == destImage.height );

    const uint8_t *pSrc = srcImage.pixels;
    uint8_t *pDest = destImage.pixels;
    if ( !pSrc || !pDest )
        return E_POINTER;

    size_t width = srcImage.width;

    if ( filter & TEX_FILTER_DITHER_DIFFUSION )
    {
        // Error diffusion dithering (aka Floyd-Steinberg dithering)
        ScopedAlignedArrayXMVECTOR scanline( reinterpret_cast<XMVECTOR*>( _aligned_malloc( (sizeof(XMVECTOR)*(width*2 + 2)), 16 ) ) );
        if ( !scanline )
            return E_OUTOFMEMORY;

        XMVECTOR* pDiffusionErrors = scanline.get() + width;
        memset( pDiffusionErrors, 0, sizeof(XMVECTOR)*(width+2) );

        for( size_t h = 0; h < srcImage.height; ++h )
        {
            if ( !_LoadScanline( scanline.get(), width, pSrc, srcImage.rowPitch, srcImage.format ) )
                return E_FAIL;

            _ConvertScanline( scanline.get(), width, destImage.format, srcImage.format, filter );

            if ( !_StoreScanlineDither( pDest, destImage.rowPitch, destImage.format, scanline.get(), width, threshold, h, z, pDiffusionErrors ) )
                return E_FAIL;

            pSrc += srcImage.rowPitch;
            pDest += destImage.rowPitch;
        }
    }
    else
    {
        ScopedAlignedArrayXMVECTOR scanline( reinterpret_cast<XMVECTOR*>( _aligned_malloc( (sizeof(XMVECTOR)*width), 16 ) ) );
        if ( !scanline )
            return E_OUTOFMEMORY;

        if ( filter & TEX_FILTER_DITHER )
        {
            // Ordered dithering
            for( size_t h = 0; h < srcImage.height; ++h )
            {
                if ( !_LoadScanline( scanline.get(), width, pSrc, srcImage.rowPitch, srcImage.format ) )
                    return E_FAIL;

                _ConvertScanline( scanline.get(), width, destImage.format, srcImage.format, filter );

                if ( !_StoreScanlineDither( pDest, destImage.rowPitch, destImage.format, scanline.get(), width, threshold, h, z, nullptr ) )
                    return E_FAIL;

                pSrc += srcImage.rowPitch;
                pDest += destImage.rowPitch;
            }
        }
        else
        {
            // No dithering
            for( size_t h = 0; h < srcImage.height; ++h )
            {
                if ( !_LoadScanline( scanline.get(), width, pSrc, srcImage.rowPitch, srcImage.format ) )
                    return E_FAIL;

                _ConvertScanline( scanline.get(), width, destImage.format, srcImage.format, filter );

                if ( !_StoreScanline( pDest, destImage.rowPitch, destImage.format, scanline.get(), width, threshold ) )
                    return E_FAIL;

                pSrc += srcImage.rowPitch;
                pDest += destImage.rowPitch;
            }
        }
    }

    return S_OK;
}


//-------------------------------------------------------------------------------------
static DXGI_FORMAT _PlanarToSingle( _In_ DXGI_FORMAT format )
{
    switch (format)
    {
    case DXGI_FORMAT_NV12:
    case DXGI_FORMAT_NV11:
        return DXGI_FORMAT_YUY2;

    case DXGI_FORMAT_P010:
        return DXGI_FORMAT_Y210;

    case DXGI_FORMAT_P016:
        return DXGI_FORMAT_Y216;

    // We currently do not support conversion for Xbox One specific depth formats

    // We can't do anything with DXGI_FORMAT_420_OPAQUE because it's an opaque blob of bits

    default:
        return DXGI_FORMAT_UNKNOWN;
    }
}


//-------------------------------------------------------------------------------------
// Convert the image from a planar to non-planar image
//-------------------------------------------------------------------------------------
#define CONVERT_420_TO_422( srcType, destType )\
        {\
            size_t rowPitch = srcImage.rowPitch;\
            \
            auto sourceE = reinterpret_cast<const srcType*>( pSrc + srcImage.slicePitch );\
            auto pSrcUV = pSrc + ( srcImage.height * rowPitch );\
            \
            for( size_t y = 0; y < srcImage.height; y+= 2 )\
            {\
                auto sPtrY0 = reinterpret_cast<const srcType*>( pSrc );\
                auto sPtrY2 = reinterpret_cast<const srcType*>( pSrc + rowPitch );\
                auto sPtrUV = reinterpret_cast<const srcType*>( pSrcUV );\
                \
                destType * __restrict dPtr0 = reinterpret_cast<destType*>(pDest);\
                destType * __restrict dPtr1 = reinterpret_cast<destType*>(pDest + destImage.rowPitch);\
                \
                for( size_t x = 0; x < srcImage.width; x+= 2 )\
                {\
                    if ( (sPtrUV+1) >= sourceE ) break;\
                    \
                    srcType u = *(sPtrUV++);\
                    srcType v = *(sPtrUV++);\
                    \
                    dPtr0->x = *(sPtrY0++);\
                    dPtr0->y = u;\
                    dPtr0->z = *(sPtrY0++);\
                    dPtr0->w = v;\
                    ++dPtr0;\
                    \
                    dPtr1->x = *(sPtrY2++);\
                    dPtr1->y = u;\
                    dPtr1->z = *(sPtrY2++);\
                    dPtr1->w = v;\
                    ++dPtr1;\
                }\
                \
                pSrc += rowPitch * 2;\
                pSrcUV += rowPitch;\
                \
                pDest += destImage.rowPitch * 2;\
            }\
        }

static HRESULT _ConvertToSinglePlane( _In_ const Image& srcImage, _In_ const Image& destImage )
{
    assert( srcImage.width == destImage.width );
    assert( srcImage.height == destImage.height );

    const uint8_t *pSrc = srcImage.pixels;
    uint8_t *pDest = destImage.pixels;
    if ( !pSrc || !pDest )
        return E_POINTER;

    switch ( srcImage.format )
    {
    case DXGI_FORMAT_NV12:
        assert( destImage.format == DXGI_FORMAT_YUY2 );
        CONVERT_420_TO_422( uint8_t, XMUBYTEN4 );
        return S_OK;

    case DXGI_FORMAT_P010:
        assert( destImage.format == DXGI_FORMAT_Y210 );
        CONVERT_420_TO_422( uint16_t, XMUSHORTN4 );
        return S_OK;

    case DXGI_FORMAT_P016:
        assert( destImage.format == DXGI_FORMAT_Y216 );
        CONVERT_420_TO_422( uint16_t, XMUSHORTN4 );
        return S_OK;

    case DXGI_FORMAT_NV11:
        assert( destImage.format == DXGI_FORMAT_YUY2 );
        // Convert 4:1:1 to 4:2:2
        {
            size_t rowPitch = srcImage.rowPitch;
            
            const uint8_t* sourceE = pSrc + srcImage.slicePitch;
            const uint8_t* pSrcUV = pSrc + ( srcImage.height * rowPitch );
            
            for( size_t y = 0; y < srcImage.height; ++y )
            {
                const uint8_t* sPtrY = pSrc;
                const uint8_t* sPtrUV = pSrcUV;
                
                XMUBYTEN4 * __restrict dPtr = reinterpret_cast<XMUBYTEN4*>(pDest);
                
                for( size_t x = 0; x < srcImage.width; x+= 4 )
                {
                    if ( (sPtrUV+1) >= sourceE ) break;
                    
                    uint8_t u = *(sPtrUV++);
                    uint8_t v = *(sPtrUV++);
                    
                    dPtr->x = *(sPtrY++);
                    dPtr->y = u;
                    dPtr->z = *(sPtrY++);
                    dPtr->w = v;
                    ++dPtr;
                    
                    dPtr->x = *(sPtrY++);
                    dPtr->y = u;
                    dPtr->z = *(sPtrY++);
                    dPtr->w = v;
                    ++dPtr;
                }
                
                pSrc += rowPitch;
                pSrcUV += (rowPitch >> 1);
                
                pDest += destImage.rowPitch;
            }
        }
        return S_OK;

    default:
        return E_UNEXPECTED;
    }
}

#undef CONVERT_420_TO_422


//=====================================================================================
// Entry-points
//=====================================================================================

//-------------------------------------------------------------------------------------
// Convert image
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
HRESULT Convert( const Image& srcImage, DXGI_FORMAT format, DWORD filter, float threshold, ScratchImage& image )
{
    if ( (srcImage.format == format) || !IsValid( format ) )
        return E_INVALIDARG;

    if ( !srcImage.pixels )
        return E_POINTER;

    if ( IsCompressed(srcImage.format) || IsCompressed(format)
         || IsPlanar(srcImage.format) || IsPlanar(format)
         || IsPalettized(srcImage.format) || IsPalettized(format)
         || IsTypeless(srcImage.format) || IsTypeless(format) )
        return HRESULT_FROM_WIN32( ERROR_NOT_SUPPORTED );

#ifdef _M_X64
    if ( (srcImage.width > 0xFFFFFFFF) || (srcImage.height > 0xFFFFFFFF) )
        return E_INVALIDARG;
#endif

    HRESULT hr = image.Initialize2D( format, srcImage.width, srcImage.height, 1, 1 );
    if ( FAILED(hr) )
        return hr;
   
    const Image *rimage = image.GetImage( 0, 0, 0 );
    if ( !rimage )
    {
        image.Release();
        return E_POINTER;
    }

    WICPixelFormatGUID pfGUID, targetGUID;
    if ( _UseWICConversion( filter, srcImage.format, format, pfGUID, targetGUID ) )
    {
        hr = _ConvertUsingWIC( srcImage, pfGUID, targetGUID, filter, threshold, *rimage );
    }
    else
    {
        hr = _Convert( srcImage, filter, *rimage, threshold, 0 );
    }

    if ( FAILED(hr) )
    {
        image.Release();
        return hr;
    }

    return S_OK;
}


//-------------------------------------------------------------------------------------
// Convert image (complex)
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
HRESULT Convert( const Image* srcImages, size_t nimages, const TexMetadata& metadata,
                 DXGI_FORMAT format, DWORD filter, float threshold, ScratchImage& result )
{
    if ( !srcImages || !nimages || (metadata.format == format) || !IsValid(format) )
        return E_INVALIDARG;

    if ( IsCompressed(metadata.format) || IsCompressed(format)
         || IsPlanar(metadata.format) || IsPlanar(format)
         || IsPalettized(metadata.format) || IsPalettized(format)
         || IsTypeless(metadata.format) || IsTypeless(format) )
        return HRESULT_FROM_WIN32( ERROR_NOT_SUPPORTED );

#ifdef _M_X64
    if ( (metadata.width > 0xFFFFFFFF) || (metadata.height > 0xFFFFFFFF) )
        return E_INVALIDARG;
#endif

    TexMetadata mdata2 = metadata;
    mdata2.format = format;
    HRESULT hr = result.Initialize( mdata2 );
    if ( FAILED(hr) )
        return hr;

    if ( nimages != result.GetImageCount() )
    {
        result.Release();
        return E_FAIL;
    }

    const Image* dest = result.GetImages();
    if ( !dest )
    {
        result.Release();
        return E_POINTER;
    }

    WICPixelFormatGUID pfGUID, targetGUID;
    bool usewic = _UseWICConversion( filter, metadata.format, format, pfGUID, targetGUID );

    switch (metadata.dimension)
    {
    case TEX_DIMENSION_TEXTURE1D:
    case TEX_DIMENSION_TEXTURE2D:
        for( size_t index=0; index < nimages; ++index )
        {
            const Image& src = srcImages[ index ];
            if ( src.format != metadata.format )
            {
                result.Release();
                return E_FAIL;
            }

#ifdef _M_X64
            if ( (src.width > 0xFFFFFFFF) || (src.height > 0xFFFFFFFF) )
                return E_FAIL;
#endif

            const Image& dst = dest[ index ];
            assert( dst.format == format );

            if ( src.width != dst.width || src.height != dst.height )
            {
                result.Release();
                return E_FAIL;
            }

            if ( usewic )
            {
                hr = _ConvertUsingWIC( src, pfGUID, targetGUID, filter, threshold, dst );
            }
            else
            {
                hr = _Convert( src, filter, dst, threshold, 0 );
            }

            if ( FAILED(hr) )
            {
                result.Release();
                return hr;
            }
        }
        break;

    case TEX_DIMENSION_TEXTURE3D:
        {
            size_t index = 0;
            size_t d = metadata.depth;
            for( size_t level = 0; level < metadata.mipLevels; ++level )
            {
                for( size_t slice = 0; slice < d; ++slice, ++index )
                {
                    if ( index >= nimages )
                        return E_FAIL;

                    const Image& src = srcImages[ index ];
                    if ( src.format != metadata.format )
                    {
                        result.Release();
                        return E_FAIL;
                    }

#ifdef _M_X64
                    if ( (src.width > 0xFFFFFFFF) || (src.height > 0xFFFFFFFF) )
                        return E_FAIL;
#endif

                    const Image& dst = dest[ index ];
                    assert( dst.format == format );

                    if ( src.width != dst.width || src.height != dst.height )
                    {
                        result.Release();
                        return E_FAIL;
                    }

                    if ( usewic )
                    {
                        hr = _ConvertUsingWIC( src, pfGUID, targetGUID, filter, threshold, dst );
                    }
                    else
                    {
                        hr = _Convert( src, filter, dst, threshold, slice );
                    }

                    if ( FAILED(hr) )
                    {
                        result.Release();
                        return hr;
                    }
                }

                if ( d > 1 )
                    d >>= 1;
            }
        }
        break;

    default:
        return E_FAIL;
    }

    return S_OK;
}


//-------------------------------------------------------------------------------------
// Convert image from planar to single plane (image)
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
HRESULT ConvertToSinglePlane( const Image& srcImage, ScratchImage& image )
{
    if ( !IsPlanar(srcImage.format) )
        return E_INVALIDARG;

    if ( !srcImage.pixels )
        return E_POINTER;

    DXGI_FORMAT format = _PlanarToSingle( srcImage.format );
    if ( format == DXGI_FORMAT_UNKNOWN )
        return HRESULT_FROM_WIN32( ERROR_NOT_SUPPORTED );

#ifdef _M_X64
    if ( (srcImage.width > 0xFFFFFFFF) || (srcImage.height > 0xFFFFFFFF) )
        return E_INVALIDARG;
#endif

    HRESULT hr = image.Initialize2D( format, srcImage.width, srcImage.height, 1, 1 );
    if ( FAILED(hr) )
        return hr;
   
    const Image *rimage = image.GetImage( 0, 0, 0 );
    if ( !rimage )
    {
        image.Release();
        return E_POINTER;
    }

    hr = _ConvertToSinglePlane( srcImage, *rimage );
    if ( FAILED(hr) )
    {
        image.Release();
        return hr;
    }

    return S_OK;
}


//-------------------------------------------------------------------------------------
// Convert image from planar to single plane (complex)
//-------------------------------------------------------------------------------------
_Use_decl_annotations_
HRESULT ConvertToSinglePlane( const Image* srcImages, size_t nimages, const TexMetadata& metadata,
                              ScratchImage& result )
{
    if ( !srcImages || !nimages )
        return E_INVALIDARG;

    if ( metadata.IsVolumemap() )
    {
        // Direct3D does not support any planar formats for Texture3D
        return HRESULT_FROM_WIN32( ERROR_NOT_SUPPORTED );
    }

    DXGI_FORMAT format = _PlanarToSingle( metadata.format );
    if ( format == DXGI_FORMAT_UNKNOWN )
        return HRESULT_FROM_WIN32( ERROR_NOT_SUPPORTED );

#ifdef _M_X64
    if ( (metadata.width > 0xFFFFFFFF) || (metadata.height > 0xFFFFFFFF) )
        return E_INVALIDARG;
#endif

    TexMetadata mdata2 = metadata;
    mdata2.format = format;
    HRESULT hr = result.Initialize( mdata2 );
    if ( FAILED(hr) )
        return hr;

    if ( nimages != result.GetImageCount() )
    {
        result.Release();
        return E_FAIL;
    }

    const Image* dest = result.GetImages();
    if ( !dest )
    {
        result.Release();
        return E_POINTER;
    }

    for( size_t index=0; index < nimages; ++index )
    {
        const Image& src = srcImages[ index ];
        if ( src.format != metadata.format )
        {
            result.Release();
            return E_FAIL;
        }

#ifdef _M_X64
        if ( (src.width > 0xFFFFFFFF) || (src.height > 0xFFFFFFFF) )
            return E_FAIL;
#endif

        const Image& dst = dest[ index ];
        assert( dst.format == format );

        if ( src.width != dst.width || src.height != dst.height )
        {
            result.Release();
            return E_FAIL;
        }

        hr = _ConvertToSinglePlane( src, dst );
        if ( FAILED(hr) )
        {
            result.Release();
            return hr;
        }
    }

    return S_OK;
}

}; // namespace