yuv 0.8.13

High performance utilities for YUV format handling and conversion.
Documentation 
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
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214

/*
 * Copyright (c) Radzivon Bartoshyk, 3/2026. All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modification,
 * are permitted provided that the following conditions are met:
 *
 * 1.  Redistributions of source code must retain the above copyright notice, this
 * list of conditions and the following disclaimer.
 *
 * 2.  Redistributions in binary form must reproduce the above copyright notice,
 * this list of conditions and the following disclaimer in the documentation
 * and/or other materials provided with the distribution.
 *
 * 3.  Neither the name of the copyright holder nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
use crate::avx2::avx2_utils::{_mm256_store_interleave_rgb_for_yuv, avx2_pack_u16};
use crate::internals::ProcessedOffset;
use crate::yuv_support::{YuvChromaRange, YuvSourceChannels};
use crate::YuvChromaSubsampling;
use std::arch::x86_64::*;

pub(crate) unsafe fn avx2_ycgco_full_range_to_rgb<
    const DESTINATION_CHANNELS: u8,
    const SAMPLING: u8,
>(
    y_plane: &[u8],
    u_plane: &[u8],
    v_plane: &[u8],
    rgba: &mut [u8],
    width: usize,
    chroma_range: YuvChromaRange,
) -> ProcessedOffset {
    unsafe {
        avx2_ycgco_full_range_to_rgb_impl::<DESTINATION_CHANNELS, SAMPLING>(
            y_plane,
            u_plane,
            v_plane,
            rgba,
            width,
            chroma_range,
        )
    }
}

#[target_feature(enable = "avx2")]
unsafe fn avx2_ycgco_full_range_to_rgb_impl<const DESTINATION_CHANNELS: u8, const SAMPLING: u8>(
    y_plane: &[u8],
    u_plane: &[u8],
    v_plane: &[u8],
    rgba: &mut [u8],
    width: usize,
    chroma_range: YuvChromaRange,
) -> ProcessedOffset {
    let chroma_subsampling: YuvChromaSubsampling = SAMPLING.into();
    let destination_channels: YuvSourceChannels = DESTINATION_CHANNELS.into();
    let channels = destination_channels.get_channels_count();

    let mut cx = 0;
    let mut uv_x = 0;

    let y_ptr = y_plane.as_ptr();
    let u_ptr = u_plane.as_ptr();
    let v_ptr = v_plane.as_ptr();
    let rgba_ptr = rgba.as_mut_ptr();

    let bias_y = _mm256_set1_epi8(chroma_range.bias_y as i8);
    let bias_uv = _mm256_set1_epi8(chroma_range.bias_uv as i8);

    while cx + 32 <= width {
        let y_raw = _mm256_loadu_si256(y_ptr.add(cx) as *const __m256i);
        let y_values = _mm256_subs_epu8(
            y_raw,
            _mm256_broadcastb_epi8(_mm256_castsi256_si128(bias_y)),
        );

        let (u_values, v_values) = match chroma_subsampling {
            YuvChromaSubsampling::Yuv420 | YuvChromaSubsampling::Yuv422 => {
                let u_raw = _mm_loadu_si128(u_ptr.add(uv_x) as *const __m128i);
                let v_raw = _mm_loadu_si128(v_ptr.add(uv_x) as *const __m128i);
                let bias_uv_128 = _mm256_castsi256_si128(bias_uv);
                let u_sub = _mm_sub_epi8(u_raw, bias_uv_128);
                let v_sub = _mm_sub_epi8(v_raw, bias_uv_128);
                // Interleave each byte with itself: [a,b,...] -> [a,a,b,b,...]
                let u_interleaved = _mm_unpacklo_epi8(u_sub, u_sub);
                let u_high = _mm_unpackhi_epi8(u_sub, u_sub);
                let v_interleaved = _mm_unpacklo_epi8(v_sub, v_sub);
                let v_high = _mm_unpackhi_epi8(v_sub, v_sub);
                (
                    _mm256_set_m128i(u_high, u_interleaved),
                    _mm256_set_m128i(v_high, v_interleaved),
                )
            }
            YuvChromaSubsampling::Yuv444 => {
                let u_raw = _mm256_loadu_si256(u_ptr.add(uv_x) as *const __m256i);
                let v_raw = _mm256_loadu_si256(v_ptr.add(uv_x) as *const __m256i);
                (
                    _mm256_sub_epi8(u_raw, bias_uv),
                    _mm256_sub_epi8(v_raw, bias_uv),
                )
            }
        };

        let y_low = _mm256_cvtepu8_epi16(_mm256_castsi256_si128(y_values));
        let y_high = _mm256_cvtepu8_epi16(_mm256_extracti128_si256::<1>(y_values));

        // Sign-reinterpret chroma
        let cg_low = _mm256_cvtepi8_epi16(_mm256_castsi256_si128(u_values));
        let cg_high = _mm256_cvtepi8_epi16(_mm256_extracti128_si256::<1>(u_values));
        let co_low = _mm256_cvtepi8_epi16(_mm256_castsi256_si128(v_values));
        let co_high = _mm256_cvtepi8_epi16(_mm256_extracti128_si256::<1>(v_values));

        // t = Y - Cg
        let t_low = _mm256_sub_epi16(y_low, cg_low);
        let t_high = _mm256_sub_epi16(y_high, cg_high);

        // R = t + Co, B = t - Co, G = Y + Cg
        let r_low = _mm256_add_epi16(t_low, co_low);
        let r_high = _mm256_add_epi16(t_high, co_high);
        let b_low = _mm256_sub_epi16(t_low, co_low);
        let b_high = _mm256_sub_epi16(t_high, co_high);
        let g_low = _mm256_add_epi16(y_low, cg_low);
        let g_high = _mm256_add_epi16(y_high, cg_high);

        // Pack i16 -> u8 with unsigned saturation
        let r_values = avx2_pack_u16(r_low, r_high);
        let g_values = avx2_pack_u16(g_low, g_high);
        let b_values = avx2_pack_u16(b_low, b_high);

        let dst_shift = cx * channels;
        _mm256_store_interleave_rgb_for_yuv::<DESTINATION_CHANNELS>(
            rgba_ptr.add(dst_shift),
            r_values,
            g_values,
            b_values,
            _mm256_set1_epi8(255u8 as i8),
        );

        cx += 32;
        match chroma_subsampling {
            YuvChromaSubsampling::Yuv420 | YuvChromaSubsampling::Yuv422 => uv_x += 16,
            YuvChromaSubsampling::Yuv444 => uv_x += 32,
        }
    }

    if cx < width {
        let diff = width - cx;
        assert!(diff < 32);

        let mut y_buffer: [u8; 32] = [0; 32];
        let mut u_buffer: [u8; 32] = [0; 32];
        let mut v_buffer: [u8; 32] = [0; 32];
        let mut dst_buffer: [u8; 32 * 4] = [0; 32 * 4];

        let ux_diff = match chroma_subsampling {
            YuvChromaSubsampling::Yuv420 | YuvChromaSubsampling::Yuv422 => diff.div_ceil(2),
            YuvChromaSubsampling::Yuv444 => diff,
        };

        std::ptr::copy_nonoverlapping(
            y_plane.get_unchecked(cx..).as_ptr(),
            y_buffer.as_mut_ptr(),
            diff,
        );
        std::ptr::copy_nonoverlapping(
            u_plane.get_unchecked(uv_x..).as_ptr(),
            u_buffer.as_mut_ptr(),
            ux_diff,
        );
        std::ptr::copy_nonoverlapping(
            v_plane.get_unchecked(uv_x..).as_ptr(),
            v_buffer.as_mut_ptr(),
            ux_diff,
        );

        if chroma_subsampling != YuvChromaSubsampling::Yuv444 && diff % 2 != 0 {
            u_buffer[ux_diff] = u_buffer[ux_diff - 1];
            v_buffer[ux_diff] = v_buffer[ux_diff - 1];
        }

        avx2_ycgco_full_range_to_rgb::<DESTINATION_CHANNELS, SAMPLING>(
            &y_buffer,
            &u_buffer,
            &v_buffer,
            &mut dst_buffer,
            32,
            chroma_range,
        );

        let dst_shift = cx * channels;
        std::ptr::copy_nonoverlapping(
            dst_buffer.as_ptr(),
            rgba.get_unchecked_mut(dst_shift..).as_mut_ptr(),
            diff * channels,
        );

        cx += diff;
        uv_x += ux_diff;
    }

    ProcessedOffset { cx, ux: uv_x }
}