zoomvtools 2.0.0

Video motion vector analysis utilities in pure Rust
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

#![allow(clippy::undocumented_unsafe_blocks)]
#![allow(unsafe_op_in_unsafe_fn)]

use std::{
    arch::x86_64::*,
    mem::size_of,
    num::{NonZeroU8, NonZeroUsize},
};

use cpudetect::target_family;
use num_traits::clamp;

use crate::util::{Pixel, round_ties_to_even};
use semisafe::option::unwrap as semisafe_opt_unwrap;
use semisafe::slice::get as semisafe_get;
use semisafe::slice::get_mut as semisafe_get_mut;

#[target_family("x86_64_v3")]
pub(super) unsafe fn float_src_to_pixels<T: Pixel>(
    dst: &mut [T],
    dst_pitch: NonZeroUsize,
    src_dct: &[f32],
    size_x: NonZeroUsize,
    size_y: NonZeroUsize,
    bits_per_sample: NonZeroU8,
    dct_shift: usize,
    dct_shift0: usize,
) {
    match size_of::<T>() {
        1 => float_src_to_pixels_u8(
            dst.as_mut_ptr().cast(),
            dst_pitch,
            src_dct,
            size_x,
            size_y,
            bits_per_sample,
            dct_shift,
        ),
        2 => float_src_to_pixels_u16(
            dst.as_mut_ptr().cast(),
            dst_pitch,
            src_dct,
            size_x,
            size_y,
            bits_per_sample,
            dct_shift,
        ),
        _ => unreachable!(),
    }

    let pixel_max = (1 << bits_per_sample.get() as usize) - 1;
    let pixel_half = 1 << (bits_per_sample.get() as usize - 1);
    let f = *semisafe_get(src_dct, 0) * 0.5;
    let integ = round_ties_to_even(f) as i32;
    *semisafe_get_mut(dst, 0) = semisafe_opt_unwrap(T::from(clamp(
        (integ >> dct_shift0) + pixel_half,
        0,
        pixel_max,
    )));
}

#[target_family("x86_64_v3")]
unsafe fn float_src_to_pixels_u8(
    dst: *mut u8,
    dst_pitch: NonZeroUsize,
    src_dct: &[f32],
    size_x: NonZeroUsize,
    size_y: NonZeroUsize,
    _bits_per_sample: NonZeroU8,
    dct_shift: usize,
) {
    // PERF: u8 only supports 8-bit, so make it constant
    let bits_per_sample = 8usize;

    let size_x = size_x.get();
    let size_y = size_y.get();
    let dst_pitch = dst_pitch.get();

    let sqrt_2_div_2 = _mm256_set1_ps((2f32).sqrt() / 2.0);
    let pixel_half = 1i32 << (bits_per_sample - 1);
    let pixel_max = (1i32 << bits_per_sample) - 1;
    let v_pixel_half = _mm256_set1_epi32(pixel_half);
    let v_pixel_max = _mm256_set1_epi32(pixel_max);
    let v_zero = _mm256_setzero_si256();
    let shift = _mm_cvtsi32_si128(dct_shift as i32);
    let sqrt_2_div_2_scalar = (2f32).sqrt() / 2.0;

    for y in 0..size_y {
        let src_row = semisafe_get(src_dct, y * size_x..y * size_x + size_x);
        let dst_row = dst.add(y * dst_pitch);

        let mut x = 0;
        while x + 8 <= size_x {
            let f = _mm256_loadu_ps(src_row.as_ptr().add(x));
            let scaled = _mm256_mul_ps(f, sqrt_2_div_2);
            let integ = _mm256_cvtps_epi32(scaled);
            let shifted = _mm256_sra_epi32(integ, shift);
            let biased = _mm256_add_epi32(shifted, v_pixel_half);
            let clamped_hi = _mm256_min_epi32(biased, v_pixel_max);
            let clamped = _mm256_max_epi32(clamped_hi, v_zero);
            let packed_u16 = _mm256_packus_epi32(clamped, v_zero);
            let packed_u16 = _mm256_permute4x64_epi64(packed_u16, 0b11_01_10_00);
            let packed_u8 = _mm256_packus_epi16(packed_u16, v_zero);
            let lo = _mm256_castsi256_si128(packed_u8);
            (dst_row.add(x) as *mut u64).write_unaligned(_mm_cvtsi128_si64(lo) as u64);
            x += 8;
        }

        while x < size_x {
            let f = semisafe_get(src_row, x) * sqrt_2_div_2_scalar;
            let integ = round_ties_to_even(f) as i32;
            let clamped = ((integ >> dct_shift) + pixel_half).clamp(0, pixel_max);
            *dst_row.add(x) = clamped as u8;
            x += 1;
        }
    }
}

#[target_family("x86_64_v3")]
unsafe fn float_src_to_pixels_u16(
    dst: *mut u16,
    dst_pitch: NonZeroUsize,
    src_dct: &[f32],
    size_x: NonZeroUsize,
    size_y: NonZeroUsize,
    bits_per_sample: NonZeroU8,
    dct_shift: usize,
) {
    let size_x = size_x.get();
    let size_y = size_y.get();
    let dst_pitch = dst_pitch.get();

    let sqrt_2_div_2 = _mm256_set1_ps((2f32).sqrt() / 2.0);
    let pixel_half = 1i32 << (bits_per_sample.get() as usize - 1);
    let pixel_max = (1i32 << bits_per_sample.get() as usize) - 1;
    let v_pixel_half = _mm256_set1_epi32(pixel_half);
    let v_pixel_max = _mm256_set1_epi32(pixel_max);
    let v_zero = _mm256_setzero_si256();
    let shift = _mm_cvtsi32_si128(dct_shift as i32);
    let sqrt_2_div_2_scalar = (2f32).sqrt() / 2.0;

    for y in 0..size_y {
        let src_row = semisafe_get(src_dct, y * size_x..y * size_x + size_x);
        let dst_row = dst.add(y * dst_pitch);

        let mut x = 0;
        while x + 8 <= size_x {
            let f = _mm256_loadu_ps(src_row.as_ptr().add(x));
            let scaled = _mm256_mul_ps(f, sqrt_2_div_2);
            let integ = _mm256_cvtps_epi32(scaled);
            let shifted = _mm256_sra_epi32(integ, shift);
            let biased = _mm256_add_epi32(shifted, v_pixel_half);
            let clamped_hi = _mm256_min_epi32(biased, v_pixel_max);
            let clamped = _mm256_max_epi32(clamped_hi, v_zero);
            let packed = _mm256_packus_epi32(clamped, v_zero);
            let packed = _mm256_permute4x64_epi64(packed, 0b11_01_10_00);
            _mm_storeu_si128(
                dst_row.add(x) as *mut __m128i,
                _mm256_castsi256_si128(packed),
            );
            x += 8;
        }

        while x < size_x {
            let f = semisafe_get(src_row, x) * sqrt_2_div_2_scalar;
            let integ = round_ties_to_even(f) as i32;
            let clamped = ((integ >> dct_shift) + pixel_half).clamp(0, pixel_max);
            *dst_row.add(x) = clamped as u16;
            x += 1;
        }
    }
}