dspk-io 0.0.10

Recording file formats for dspk: raw binary and Zarr traces, format detection, chunk caching
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

//! Little-endian sample decoding / encoding shared by the binary formats.

use dspk_core::{RecordingInfo, SampleFormat};

/// Calls `$f::<B, T>(… , value)` for the stored type of `$format` (`B` bytes, little-endian).
macro_rules! with_format {
    ($format:expr, $f:ident($($arg:expr),*)) => {
        match $format {
            SampleFormat::I8 => $f::<1>($($arg,)* |b| i8::from_le_bytes(b) as f32),
            SampleFormat::I16 => $f::<2>($($arg,)* |b| i16::from_le_bytes(b) as f32),
            SampleFormat::U16 => $f::<2>($($arg,)* |b| u16::from_le_bytes(b) as f32),
            SampleFormat::I32 => $f::<4>($($arg,)* |b| i32::from_le_bytes(b) as f32),
            SampleFormat::F32 => $f::<4>($($arg,)* f32::from_le_bytes),
            SampleFormat::F64 => $f::<8>($($arg,)* |b| f64::from_le_bytes(b) as f32),
        }
    };
}

/// Applies each channel's gain and offset to time-major stored values (`out[t · channels + c]`)
/// of `info`'s channels, giving µV.
pub fn scale_frames(info: &RecordingInfo, out: &mut [f32]) {
    let ch = &info.channels;
    if ch.is_empty() {
        return;
    }
    let (g, o) = (ch[0].gain, ch[0].offset);
    if ch.iter().all(|c| c.gain == g && c.offset == o) {
        if (g, o) != (1.0, 0.0) {
            out.iter_mut().for_each(|v| *v = *v * g + o);
        }
        return;
    }
    let (gains, offsets): (Vec<f32>, Vec<f32>) = ch.iter().map(|c| (c.gain, c.offset)).unzip();
    for frame in out.chunks_exact_mut(ch.len()) {
        for ((v, g), o) in frame.iter_mut().zip(&gains).zip(&offsets) {
            *v = *v * g + o;
        }
    }
}

/// Decodes `channels` (`(stored channel, gain, offset)`) of `n` interleaved frames of `frame_bytes`
/// each into channel-major µV rows of `out` (`out[i · n + s]`). Every frame is read once for all
/// requested channels.
pub fn decode_frames(
    format: SampleFormat,
    frames: &[u8],
    frame_bytes: usize,
    channels: &[(usize, f32, f32)],
    n: usize,
    out: &mut [f32],
) {
    fn run<const B: usize>(
        frames: &[u8],
        frame_bytes: usize,
        channels: &[(usize, f32, f32)],
        n: usize,
        out: &mut [f32],
        value: impl Fn([u8; B]) -> f32,
    ) {
        for (s, frame) in frames.chunks_exact(frame_bytes).take(n).enumerate() {
            for (i, &(ch, gain, offset)) in channels.iter().enumerate() {
                let b: [u8; B] = frame[ch * B..ch * B + B].try_into().expect("sample inside frame");
                out[i * n + s] = value(b) * gain + offset;
            }
        }
    }
    with_format!(format, run(frames, frame_bytes, channels, n, out))
}

/// Copies the stored values of block columns `cols` into channel-major `out` (`bytes` per value).
/// The block is `[n][width]` values when `time_major`, else `[width][n]`.
pub fn select_stored(block: &[u8], time_major: bool, width: usize, n: usize, cols: &[usize], bytes: usize, out: &mut [u8]) {
    fn frames<const B: usize>(block: &[u8], width: usize, n: usize, cols: &[usize], out: &mut [u8]) {
        let (block, _) = block.as_chunks::<B>();
        let (out, _) = out.as_chunks_mut::<B>();
        for (t, frame) in block.chunks_exact(width).take(n).enumerate() {
            for (i, &c) in cols.iter().enumerate() {
                out[i * n + t] = frame[c];
            }
        }
    }
    if time_major {
        match bytes {
            1 => frames::<1>(block, width, n, cols, out),
            2 => frames::<2>(block, width, n, cols, out),
            4 => frames::<4>(block, width, n, cols, out),
            8 => frames::<8>(block, width, n, cols, out),
            other => unreachable!("{other}-byte samples"),
        }
    } else {
        for (i, &c) in cols.iter().enumerate() {
            out[i * n * bytes..(i + 1) * n * bytes].copy_from_slice(&block[c * n * bytes..(c + 1) * n * bytes]);
        }
    }
}

/// Decoded values in memory order to little-endian (a no-op on little-endian hosts).
pub fn native_to_le(values: &mut [u8], bytes: usize) {
    if cfg!(target_endian = "big") {
        for v in values.chunks_exact_mut(bytes) {
            v.reverse();
        }
    }
}

/// Appends `value_uv` to `dst` as a stored sample (integers are rounded and saturated).
#[inline(always)]
pub fn encode(format: SampleFormat, value_uv: f32, gain: f32, offset: f32, dst: &mut Vec<u8>) {
    let stored = (value_uv - offset) / gain;
    match format {
        SampleFormat::I8 => dst.extend_from_slice(&(stored.round() as i8).to_le_bytes()),
        SampleFormat::I16 => dst.extend_from_slice(&(stored.round() as i16).to_le_bytes()),
        SampleFormat::U16 => dst.extend_from_slice(&(stored.round() as u16).to_le_bytes()),
        SampleFormat::I32 => dst.extend_from_slice(&(stored.round() as i32).to_le_bytes()),
        SampleFormat::F32 => dst.extend_from_slice(&stored.to_le_bytes()),
        SampleFormat::F64 => dst.extend_from_slice(&(stored as f64).to_le_bytes()),
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_roundtrip_and_saturation() {
        let mut bytes = Vec::new();
        for v in [-5.0f32, 0.0, 7.4, 1e9] {
            encode(SampleFormat::I16, v, 0.5, 0.0, &mut bytes);
        }
        let mut out = [0.0f32; 4];
        SampleFormat::I16.decode(&bytes, &mut out, 0.5, 0.0);
        assert_eq!(out, [-5.0, 0.0, 7.5, i16::MAX as f32 * 0.5]);

        // Two interleaved frames of three channels; channels 2 and 0 requested.
        let mut frames = Vec::new();
        for v in [1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0] {
            encode(SampleFormat::I16, v, 1.0, 0.0, &mut frames);
        }
        let mut rows = [0.0f32; 4];
        decode_frames(SampleFormat::I16, &frames, 6, &[(2, 2.0, 1.0), (0, 1.0, 0.0)], 2, &mut rows);
        assert_eq!(rows, [7.0, 13.0, 1.0, 4.0]);

        // Every format round-trips through encode / decode
        for format in SampleFormat::ALL {
            let mut b = Vec::new();
            // Values every format holds (uint16 has no negatives, int8 stops at 127)
            for v in [0.0f32, 42.0, 100.0] {
                encode(format, v, 1.0, 0.0, &mut b);
            }
            let mut back = [0.0f32; 3];
            format.decode(&b, &mut back, 1.0, 0.0);
            assert_eq!(back, [0.0, 42.0, 100.0], "{format:?}");
        }

        let mut f = Vec::new();
        encode(SampleFormat::F32, 3.25, 1.0, 0.0, &mut f);
        let mut v = [0.0f32];
        SampleFormat::F32.decode(&f, &mut v, 1.0, 0.0);
        assert_eq!(v, [3.25]);
    }
}