resampler 0.5.1

A small audio resampling library
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

use core::arch::x86_64::*;

/// Complex multiplication.
#[target_feature(enable = "sse4.2")]
pub(crate) fn complex_mul_sse4_2(left: __m128, right: __m128) -> __m128 {
    let right_re = _mm_moveldup_ps(right);
    let right_im = _mm_movehdup_ps(right);
    let left_swap = _mm_shuffle_ps(left, left, 0b10_11_00_01);
    let prod_re = _mm_mul_ps(left, right_re);
    let prod_im = _mm_mul_ps(left_swap, right_im);
    _mm_addsub_ps(prod_re, prod_im)
}

/// Returns a mask for negating imaginary parts: [-0.0, 0.0, -0.0, 0.0].
#[target_feature(enable = "sse4.2")]
pub(crate) fn load_neg_imag_mask_sse4_2() -> __m128 {
    _mm_set_ps(-0.0, 0.0, -0.0, 0.0)
}

/// Returns the scale constant 1/√2 as a 128-bit SIMD vector.
#[target_feature(enable = "sse4.2")]
pub(crate) fn load_scale_sse4_2() -> __m128 {
    const SQRT_HALF: f32 = core::f32::consts::FRAC_1_SQRT_2;
    _mm_set1_ps(SQRT_HALF)
}

/// Multiplies a complex vector by i (90-degree rotation).
///
/// Performs i * (a + bi) = -b + ai by swapping real/imaginary parts
/// and negating the new imaginary part (which is the old real part).
#[target_feature(enable = "sse4.2")]
pub(crate) fn complex_mul_i_sse4_2(vec: __m128, neg_imag_mask: __m128) -> __m128 {
    let swapped = _mm_shuffle_ps(vec, vec, 0b10_11_00_01);
    _mm_xor_ps(swapped, neg_imag_mask)
}

/// Multiplies a complex vector by sqrt(3)/2 * i.
#[target_feature(enable = "sse4.2")]
pub(crate) fn complex_mul_sqrt3_i_sse4_2(vec: __m128, sqrt3_2: f32) -> __m128 {
    let swapped = _mm_shuffle_ps(vec, vec, 0b10_11_00_01);
    _mm_mul_ps(swapped, _mm_set_ps(-sqrt3_2, sqrt3_2, -sqrt3_2, sqrt3_2))
}

/// Optimized W₈¹ multiplication: (1-i)/√2 * (x+iy).
///
/// Computes ((x+y)/√2, (y-x)/√2) using fewer operations than full complex multiplication.
#[target_feature(enable = "sse4.2")]
pub(crate) fn w8x_sse4_2(xy: __m128, sign_mask: __m128, scale: __m128) -> __m128 {
    let yx = _mm_shuffle_ps(xy, xy, 0b10_11_00_01);
    let ymx = _mm_xor_ps(yx, sign_mask);
    let sum = _mm_add_ps(xy, ymx);
    _mm_mul_ps(scale, sum)
}

/// Optimized W₈³ multiplication: (-1-i)/√2 * (x+iy).
///
/// Computes ((y-x)/√2, -(x+y)/√2) using fewer operations than full complex multiplication.
#[target_feature(enable = "sse4.2")]
pub(crate) fn v8x_sse4_2(xy: __m128, sign_mask: __m128, scale: __m128) -> __m128 {
    let yx = _mm_shuffle_ps(xy, xy, 0b10_11_00_01);
    let ymx = _mm_xor_ps(yx, sign_mask);
    let diff = _mm_sub_ps(ymx, xy);
    _mm_mul_ps(scale, diff)
}