1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
|
/*
* Simd Library (http://ermig1979.github.io/Simd).
*
* Copyright (c) 2011-2022 Yermalayeu Ihar.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef __SimdAlphaBlending_h__
#define __SimdAlphaBlending_h__
#include "Simd/SimdMath.h"
#include "Simd/SimdStore.h"
namespace Simd
{
namespace Base
{
SIMD_INLINE int DivideBy255(int value)
{
return (value + 1 + (value >> 8)) >> 8;
}
template<bool argb> void AlphaPremultiply(const uint8_t* src, uint8_t* dst);
template<> SIMD_INLINE void AlphaPremultiply<false>(const uint8_t* src, uint8_t* dst)
{
int alpha = src[3];
dst[0] = DivideBy255(src[0] * alpha);
dst[1] = DivideBy255(src[1] * alpha);
dst[2] = DivideBy255(src[2] * alpha);
dst[3] = alpha;
}
template<> SIMD_INLINE void AlphaPremultiply<true>(const uint8_t* src, uint8_t* dst)
{
int alpha = src[0];
dst[0] = alpha;
dst[1] = DivideBy255(src[1] * alpha);
dst[2] = DivideBy255(src[2] * alpha);
dst[3] = DivideBy255(src[3] * alpha);
}
template<bool argb> void AlphaUnpremultiply(const uint8_t* src, uint8_t* dst);
template<> SIMD_INLINE void AlphaUnpremultiply<false>(const uint8_t* src, uint8_t* dst)
{
float alpha = src[3] ? 255.00001f / src[3] : 0.0f;
dst[0] = RestrictRange(int(src[0] * alpha));
dst[1] = RestrictRange(int(src[1] * alpha));
dst[2] = RestrictRange(int(src[2] * alpha));
dst[3] = src[3];
}
template<> SIMD_INLINE void AlphaUnpremultiply<true>(const uint8_t* src, uint8_t* dst)
{
float alpha = src[0] ? 255.00001f / src[0] : 0.0f;
dst[0] = src[0];
dst[1] = RestrictRange(int(src[1] * alpha));
dst[2] = RestrictRange(int(src[2] * alpha));
dst[3] = RestrictRange(int(src[3] * alpha));
}
}
#ifdef SIMD_SSE41_ENABLE
namespace Sse41
{
SIMD_INLINE __m128i Divide16iBy255(__m128i value)
{
return _mm_mulhi_epi16(_mm_add_epi16(value, K16_0001), K16_0101);
}
SIMD_INLINE __m128i Divide16uBy255(__m128i value)
{
return _mm_mulhi_epu16(_mm_add_epi16(value, K16_0001), K16_0101);
}
SIMD_INLINE __m128i AlphaBlending16i(__m128i src, __m128i dst, __m128i alpha)
{
return Divide16uBy255(_mm_add_epi16(_mm_mullo_epi16(src, alpha), _mm_mullo_epi16(dst, _mm_sub_epi16(K16_00FF, alpha))));
}
template <bool align> SIMD_INLINE void AlphaBlending(const __m128i* src, __m128i* dst, __m128i alpha)
{
__m128i _src = Load<align>(src);
__m128i _dst = Load<align>(dst);
__m128i lo = AlphaBlending16i(_mm_unpacklo_epi8(_src, K_ZERO), _mm_unpacklo_epi8(_dst, K_ZERO), _mm_unpacklo_epi8(alpha, K_ZERO));
__m128i hi = AlphaBlending16i(_mm_unpackhi_epi8(_src, K_ZERO), _mm_unpackhi_epi8(_dst, K_ZERO), _mm_unpackhi_epi8(alpha, K_ZERO));
Store<align>(dst, _mm_packus_epi16(lo, hi));
}
template <bool align> SIMD_INLINE void AlphaBlending2x(const __m128i* src0, __m128i alpha0, const __m128i* src1, __m128i alpha1, __m128i* dst)
{
__m128i _dst = Load<align>(dst);
__m128i lo = _mm_unpacklo_epi8(_dst, K_ZERO);
__m128i hi = _mm_unpackhi_epi8(_dst, K_ZERO);
__m128i _src0 = Load<align>(src0);
lo = AlphaBlending16i(_mm_unpacklo_epi8(_src0, K_ZERO), lo, _mm_unpacklo_epi8(alpha0, K_ZERO));
hi = AlphaBlending16i(_mm_unpackhi_epi8(_src0, K_ZERO), hi, _mm_unpackhi_epi8(alpha0, K_ZERO));
__m128i _src1 = Load<align>(src1);
lo = AlphaBlending16i(_mm_unpacklo_epi8(_src1, K_ZERO), lo, _mm_unpacklo_epi8(alpha1, K_ZERO));
hi = AlphaBlending16i(_mm_unpackhi_epi8(_src1, K_ZERO), hi, _mm_unpackhi_epi8(alpha1, K_ZERO));
Store<align>(dst, _mm_packus_epi16(lo, hi));
}
template <bool align> SIMD_INLINE void AlphaFilling(__m128i* dst, __m128i channelLo, __m128i channelHi, __m128i alpha)
{
__m128i _dst = Load<align>(dst);
__m128i lo = AlphaBlending16i(channelLo, _mm_unpacklo_epi8(_dst, K_ZERO), _mm_unpacklo_epi8(alpha, K_ZERO));
__m128i hi = AlphaBlending16i(channelHi, _mm_unpackhi_epi8(_dst, K_ZERO), _mm_unpackhi_epi8(alpha, K_ZERO));
Store<align>(dst, _mm_packus_epi16(lo, hi));
}
SIMD_INLINE __m128i AlphaPremultiply16i(__m128i value, __m128i alpha)
{
return Divide16uBy255(_mm_mullo_epi16(value, alpha));
}
}
#endif
#ifdef SIMD_AVX2_ENABLE
namespace Avx2
{
SIMD_INLINE __m256i Divide16iBy255(__m256i value)
{
return _mm256_mulhi_epi16(_mm256_add_epi16(value, K16_0001), K16_0101);
}
SIMD_INLINE __m256i Divide16uBy255(__m256i value)
{
return _mm256_mulhi_epu16(_mm256_add_epi16(value, K16_0001), K16_0101);
}
SIMD_INLINE __m256i AlphaBlending16i(__m256i src, __m256i dst, __m256i alpha)
{
return Divide16uBy255(_mm256_add_epi16(_mm256_mullo_epi16(src, alpha), _mm256_mullo_epi16(dst, _mm256_sub_epi16(K16_00FF, alpha))));
}
template <bool align> SIMD_INLINE void AlphaBlending(const __m256i* src, __m256i* dst, __m256i alpha)
{
__m256i _src = Load<align>(src);
__m256i _dst = Load<align>(dst);
__m256i lo = AlphaBlending16i(_mm256_unpacklo_epi8(_src, K_ZERO), _mm256_unpacklo_epi8(_dst, K_ZERO), _mm256_unpacklo_epi8(alpha, K_ZERO));
__m256i hi = AlphaBlending16i(_mm256_unpackhi_epi8(_src, K_ZERO), _mm256_unpackhi_epi8(_dst, K_ZERO), _mm256_unpackhi_epi8(alpha, K_ZERO));
Store<align>(dst, _mm256_packus_epi16(lo, hi));
}
template <bool align> SIMD_INLINE void AlphaBlending2x(const __m256i* src0, __m256i alpha0, const __m256i* src1, __m256i alpha1, __m256i* dst)
{
__m256i _dst = Load<align>(dst);
__m256i lo = _mm256_unpacklo_epi8(_dst, K_ZERO);
__m256i hi = _mm256_unpackhi_epi8(_dst, K_ZERO);
__m256i _src0 = Load<align>(src0);
lo = AlphaBlending16i(_mm256_unpacklo_epi8(_src0, K_ZERO), lo, _mm256_unpacklo_epi8(alpha0, K_ZERO));
hi = AlphaBlending16i(_mm256_unpackhi_epi8(_src0, K_ZERO), hi, _mm256_unpackhi_epi8(alpha0, K_ZERO));
__m256i _src1 = Load<align>(src1);
lo = AlphaBlending16i(_mm256_unpacklo_epi8(_src1, K_ZERO), lo, _mm256_unpacklo_epi8(alpha1, K_ZERO));
hi = AlphaBlending16i(_mm256_unpackhi_epi8(_src1, K_ZERO), hi, _mm256_unpackhi_epi8(alpha1, K_ZERO));
Store<align>(dst, _mm256_packus_epi16(lo, hi));
}
}
#endif
#ifdef SIMD_AVX512BW_ENABLE
namespace Avx512bw
{
SIMD_INLINE __m512i Divide16iBy255(__m512i value)
{
return _mm512_mulhi_epi16(_mm512_add_epi16(value, K16_0001), K16_0101);
}
SIMD_INLINE __m512i Divide16uBy255(__m512i value)
{
return _mm512_mulhi_epu16(_mm512_add_epi16(value, K16_0001), K16_0101);
}
}
#endif
}
#endif
|
