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
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

use core::arch::aarch64::*;

use super::super::ops::{complex_mul, complex_mul_i, load_neg_imag_mask};
use crate::fft::Complex32;

/// Performs a single radix-4 Stockham butterfly stage for stride=1 (out-of-place, NEON).
///
/// This is a specialized version for the stride=1 case (first stage) that uses the zip pattern
/// to enable sequential SIMD stores. For stride=1, output indices are sequential: j=4*i.
#[target_feature(enable = "neon")]
pub(super) unsafe fn butterfly_radix4_stride1_neon(
    src: &[Complex32],
    dst: &mut [Complex32],
    stage_twiddles: &[Complex32],
) {
    let samples = src.len();
    let quarter_samples = samples >> 2;
    let simd_iters = (quarter_samples >> 1) << 1;

    unsafe {
        let neg_imag = load_neg_imag_mask();

        for i in (0..simd_iters).step_by(2) {
            // For stride=1: k=0, so j = 4*i (simplified indexing, outputs are sequential).

            // Load z0 from first quarter (2 complex numbers).
            let z0_ptr = src.as_ptr().add(i) as *const f32;
            let z0 = vld1q_f32(z0_ptr);

            // Load z1, z2, z3 from other quarters using contiguous loads.
            let z1_ptr = src.as_ptr().add(i + quarter_samples) as *const f32;
            let z1 = vld1q_f32(z1_ptr);

            let z2_ptr = src.as_ptr().add(i + quarter_samples * 2) as *const f32;
            let z2 = vld1q_f32(z2_ptr);

            let z3_ptr = src.as_ptr().add(i + quarter_samples * 3) as *const f32;
            let z3 = vld1q_f32(z3_ptr);

            // Identity twiddles: t1 = (1+0i) * z1 = z1, t2 = z2, t3 = z3 (skip twiddle load and multiply)
            let t1 = z1;
            let t2 = z2;
            let t3 = z3;

            // Radix-4 butterfly.
            let a0 = vaddq_f32(z0, t2);
            let a1 = vsubq_f32(z0, t2);
            let a2 = vaddq_f32(t1, t3);

            // a3 = i * (t1 - t3)
            let t1_sub_t3 = vsubq_f32(t1, t3);
            let a3 = complex_mul_i(t1_sub_t3, neg_imag);

            // Final butterfly outputs.
            let out0 = vaddq_f32(a0, a2);
            let out2 = vsubq_f32(a0, a2);
            let out1 = vaddq_f32(a1, a3);
            let out3 = vsubq_f32(a1, a3);

            // Apply zip pattern for sequential stores.
            // We have 4 outputs, each containing 2 complex numbers.
            // Need to interleave: [out0[0], out1[0], out2[0], out3[0], out0[1], out1[1], out2[1], out3[1]]

            let j = 4 * i;

            // Cast to f64 view to treat each complex number as a single 64-bit unit.
            let out0_f64 = vreinterpretq_f64_f32(out0);
            let out1_f64 = vreinterpretq_f64_f32(out1);
            let out2_f64 = vreinterpretq_f64_f32(out2);
            let out3_f64 = vreinterpretq_f64_f32(out3);

            // Perform 2x2 matrix transpose to get the interleaving we need.
            // First level: interleave pairs (out0, out1) and (out2, out3)
            let pair01_lo = vzip1q_f64(out0_f64, out1_f64); // [out0[0], out1[0]]
            let pair01_hi = vzip2q_f64(out0_f64, out1_f64); // [out0[1], out1[1]]
            let pair23_lo = vzip1q_f64(out2_f64, out3_f64); // [out2[0], out3[0]]
            let pair23_hi = vzip2q_f64(out2_f64, out3_f64); // [out2[1], out3[1]]

            // Now combine the lower and upper halves into 128-bit groups
            // We want: [out0[0], out1[0], out2[0], out3[0]] in first two stores
            // and:     [out0[1], out1[1], out2[1], out3[1]] in second two stores

            // Cast back to f32 to access as 128-bit vectors
            let pair01_lo_f32 = vreinterpretq_f32_f64(pair01_lo); // [out0[0].re, out0[0].im, out1[0].re, out1[0].im]
            let pair23_lo_f32 = vreinterpretq_f32_f64(pair23_lo); // [out2[0].re, out2[0].im, out3[0].re, out3[0].im]
            let pair01_hi_f32 = vreinterpretq_f32_f64(pair01_hi); // [out0[1].re, out0[1].im, out1[1].re, out1[1].im]
            let pair23_hi_f32 = vreinterpretq_f32_f64(pair23_hi); // [out2[1].re, out2[1].im, out3[1].re, out3[1].im]

            let dst_ptr = dst.as_mut_ptr().add(j) as *mut f32;
            vst1q_f32(dst_ptr, pair01_lo_f32); // Store out0[0], out1[0]
            vst1q_f32(dst_ptr.add(4), pair23_lo_f32); // Store out2[0], out3[0]
            vst1q_f32(dst_ptr.add(8), pair01_hi_f32); // Store out0[1], out1[1]
            vst1q_f32(dst_ptr.add(12), pair23_hi_f32); // Store out2[1], out3[1]
        }
    }

    super::butterfly_radix4_scalar::<2>(src, dst, stage_twiddles, 1, simd_iters);
}

/// Performs a single radix-4 Stockham butterfly stage for p>1 (out-of-place, NEON).
///
/// Generic version for p>1 cases. Uses direct SIMD stores, accepting non-sequential
/// stores as the shuffle overhead isn't justified for larger strides.
#[target_feature(enable = "neon")]
pub(super) unsafe fn butterfly_radix4_generic_neon(
    src: &[Complex32],
    dst: &mut [Complex32],
    stage_twiddles: &[Complex32],
    stride: usize,
) {
    // We convince the compiler here that stride can't be 0 to optimize better.
    if stride == 0 {
        return;
    }

    let samples = src.len();
    let quarter_samples = samples >> 2;
    let simd_iters = (quarter_samples >> 1) << 1;

    unsafe {
        let neg_imag = load_neg_imag_mask();

        for i in (0..simd_iters).step_by(2) {
            let k = i % stride;
            let k0 = k;
            let k1 = k + 1 - ((k + 1 >= stride) as usize) * stride;

            // Load z0 from first quarter.
            let z0_ptr = src.as_ptr().add(i) as *const f32;
            let z0 = vld1q_f32(z0_ptr);

            // Load z1, z2, z3 from other quarters using contiguous loads.
            let z1_ptr = src.as_ptr().add(i + quarter_samples) as *const f32;
            let z1 = vld1q_f32(z1_ptr);

            let z2_ptr = src.as_ptr().add(i + quarter_samples * 2) as *const f32;
            let z2 = vld1q_f32(z2_ptr);

            let z3_ptr = src.as_ptr().add(i + quarter_samples * 3) as *const f32;
            let z3 = vld1q_f32(z3_ptr);

            // Load pre-packaged twiddles: [w1[i..i+2], w2[i..i+2], w3[i..i+2]]
            let tw_ptr = stage_twiddles.as_ptr().add(i * 3) as *const f32;
            let w1 = vld1q_f32(tw_ptr); // w1[i], w1[i+1]
            let w2 = vld1q_f32(tw_ptr.add(4)); // w2[i], w2[i+1]
            let w3 = vld1q_f32(tw_ptr.add(8)); // w3[i], w3[i+1]

            let t1 = complex_mul(z1, w1);
            let t2 = complex_mul(z2, w2);
            let t3 = complex_mul(z3, w3);

            // Radix-4 butterfly.
            let a0 = vaddq_f32(z0, t2);
            let a1 = vsubq_f32(z0, t2);
            let a2 = vaddq_f32(t1, t3);

            // a3 = i * (t1 - t3)
            let t1_sub_t3 = vsubq_f32(t1, t3);
            let a3 = complex_mul_i(t1_sub_t3, neg_imag);

            // Final butterfly outputs.
            let out0 = vaddq_f32(a0, a2);
            let out2 = vsubq_f32(a0, a2);
            let out1 = vaddq_f32(a1, a3);
            let out3 = vsubq_f32(a1, a3);

            // Calculate output indices.
            let j0 = 4 * i - 3 * k0;
            let j1 = 4 * (i + 1) - 3 * k1;

            // Extract and store each complex number separately using direct f32 stores.
            let dst_ptr = dst.as_mut_ptr() as *mut f32;

            let out0_0 = vget_low_f32(out0);
            let out1_0 = vget_low_f32(out1);
            let out2_0 = vget_low_f32(out2);
            let out3_0 = vget_low_f32(out3);

            vst1_f32(dst_ptr.add(j0 << 1), out0_0);
            vst1_f32(dst_ptr.add((j0 + stride) << 1), out1_0);
            vst1_f32(dst_ptr.add((j0 + stride * 2) << 1), out2_0);
            vst1_f32(dst_ptr.add((j0 + stride * 3) << 1), out3_0);

            let out0_1 = vget_high_f32(out0);
            let out1_1 = vget_high_f32(out1);
            let out2_1 = vget_high_f32(out2);
            let out3_1 = vget_high_f32(out3);

            vst1_f32(dst_ptr.add(j1 << 1), out0_1);
            vst1_f32(dst_ptr.add((j1 + stride) << 1), out1_1);
            vst1_f32(dst_ptr.add((j1 + stride * 2) << 1), out2_1);
            vst1_f32(dst_ptr.add((j1 + stride * 3) << 1), out3_1);
        }
    }

    super::butterfly_radix4_scalar::<2>(src, dst, stage_twiddles, stride, simd_iters);
}