colconv 0.1.0

SIMD-dispatched color-conversion kernels covering the FFmpeg AVPixelFormat space, with a Sink-based API so consumers pick which derived outputs (RGB / Luma / HSV / custom) they want without paying for the ones they don't.
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
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276

//! Scalar kernels for 1-bit-per-pixel monochrome frames (Monoblack / Monowhite).
//!
//! Unpack logic: each byte carries 8 pixels, MSB first (bit 7 = pixel 0, bit 0 =
//! pixel 7). Polarity controlled by `const INVERT: bool`:
//! - INVERT=false (Monoblack): bit=0 → Y=0, bit=1 → Y=255
//! - INVERT=true (Monowhite): bit=0 → Y=255, bit=1 → Y=0

/// Unpack 1-bit-per-pixel buffer to u8 luma array.
/// MSB first: bit 7 of byte[0] is pixel 0, bit 0 of byte[0] is pixel 7, etc.
#[cfg_attr(not(tarpaulin), inline(always))]
fn mono1bit_to_luma_row_generic<const INVERT: bool>(data: &[u8], out: &mut [u8], width: usize) {
  debug_assert!(data.len() >= width.div_ceil(8), "data too short");
  debug_assert!(out.len() >= width, "out too short");
  let mut out_idx = 0;
  for byte_val in data.iter().take(width.div_ceil(8)) {
    for bit_pos in (0..8).rev() {
      if out_idx >= width {
        break;
      }
      let bit = (*byte_val >> bit_pos) & 1;
      let luma = if INVERT { (1 - bit) * 255 } else { bit * 255 };
      out[out_idx] = luma;
      out_idx += 1;
    }
  }
}

// ---- allocation-free pixel-expansion helpers --------------------------------

/// Extract the Y value for one pixel bit, applying polarity.
#[inline(always)]
fn pixel_y<const INVERT: bool>(byte: u8, bit: usize) -> u8 {
  let raw = (byte >> (7 - bit)) & 1;
  if INVERT { (1 - raw) * 255 } else { raw * 255 }
}

// ---- allocation-free RGB/RGBA/u16 generics ----------------------------------

#[cfg_attr(not(tarpaulin), inline(always))]
fn mono1bit_to_rgb_generic<const INVERT: bool>(data: &[u8], out: &mut [u8], width: usize) {
  debug_assert!(data.len() >= width.div_ceil(8));
  debug_assert!(out.len() >= width * 3);
  let full_bytes = width / 8;
  for (byte_idx, &byte) in data[..full_bytes].iter().enumerate() {
    for bit in 0..8usize {
      let y = pixel_y::<INVERT>(byte, bit);
      let i = (byte_idx * 8 + bit) * 3;
      out[i] = y;
      out[i + 1] = y;
      out[i + 2] = y;
    }
  }
  let tail = width % 8;
  if tail > 0 {
    let byte = data[full_bytes];
    for bit in 0..tail {
      let y = pixel_y::<INVERT>(byte, bit);
      let i = (full_bytes * 8 + bit) * 3;
      out[i] = y;
      out[i + 1] = y;
      out[i + 2] = y;
    }
  }
}

#[cfg_attr(not(tarpaulin), inline(always))]
fn mono1bit_to_rgba_generic<const INVERT: bool>(data: &[u8], out: &mut [u8], width: usize) {
  debug_assert!(data.len() >= width.div_ceil(8));
  debug_assert!(out.len() >= width * 4);
  let full_bytes = width / 8;
  for (byte_idx, &byte) in data[..full_bytes].iter().enumerate() {
    for bit in 0..8usize {
      let y = pixel_y::<INVERT>(byte, bit);
      let i = (byte_idx * 8 + bit) * 4;
      out[i] = y;
      out[i + 1] = y;
      out[i + 2] = y;
      out[i + 3] = 0xFF;
    }
  }
  let tail = width % 8;
  if tail > 0 {
    let byte = data[full_bytes];
    for bit in 0..tail {
      let y = pixel_y::<INVERT>(byte, bit);
      let i = (full_bytes * 8 + bit) * 4;
      out[i] = y;
      out[i + 1] = y;
      out[i + 2] = y;
      out[i + 3] = 0xFF;
    }
  }
}

#[cfg_attr(not(tarpaulin), inline(always))]
fn mono1bit_to_rgb_u16_generic<const INVERT: bool>(data: &[u8], out: &mut [u16], width: usize) {
  debug_assert!(data.len() >= width.div_ceil(8));
  debug_assert!(out.len() >= width * 3);
  let full_bytes = width / 8;
  for (byte_idx, &byte) in data[..full_bytes].iter().enumerate() {
    for bit in 0..8usize {
      let y16 = pixel_y::<INVERT>(byte, bit) as u16;
      let i = (byte_idx * 8 + bit) * 3;
      out[i] = y16;
      out[i + 1] = y16;
      out[i + 2] = y16;
    }
  }
  let tail = width % 8;
  if tail > 0 {
    let byte = data[full_bytes];
    for bit in 0..tail {
      let y16 = pixel_y::<INVERT>(byte, bit) as u16;
      let i = (full_bytes * 8 + bit) * 3;
      out[i] = y16;
      out[i + 1] = y16;
      out[i + 2] = y16;
    }
  }
}

#[cfg_attr(not(tarpaulin), inline(always))]
fn mono1bit_to_rgba_u16_generic<const INVERT: bool>(data: &[u8], out: &mut [u16], width: usize) {
  debug_assert!(data.len() >= width.div_ceil(8));
  debug_assert!(out.len() >= width * 4);
  let full_bytes = width / 8;
  for (byte_idx, &byte) in data[..full_bytes].iter().enumerate() {
    for bit in 0..8usize {
      let y16 = pixel_y::<INVERT>(byte, bit) as u16;
      let i = (byte_idx * 8 + bit) * 4;
      out[i] = y16;
      out[i + 1] = y16;
      out[i + 2] = y16;
      out[i + 3] = 0x00FF;
    }
  }
  let tail = width % 8;
  if tail > 0 {
    let byte = data[full_bytes];
    for bit in 0..tail {
      let y16 = pixel_y::<INVERT>(byte, bit) as u16;
      let i = (full_bytes * 8 + bit) * 4;
      out[i] = y16;
      out[i + 1] = y16;
      out[i + 2] = y16;
      out[i + 3] = 0x00FF;
    }
  }
}

#[cfg_attr(not(tarpaulin), inline(always))]
fn mono1bit_to_luma_u16_generic<const INVERT: bool>(data: &[u8], out: &mut [u16], width: usize) {
  debug_assert!(data.len() >= width.div_ceil(8));
  debug_assert!(out.len() >= width);
  let full_bytes = width / 8;
  for (byte_idx, &byte) in data[..full_bytes].iter().enumerate() {
    for bit in 0..8usize {
      out[byte_idx * 8 + bit] = pixel_y::<INVERT>(byte, bit) as u16;
    }
  }
  let tail = width % 8;
  if tail > 0 {
    let byte = data[full_bytes];
    for bit in 0..tail {
      out[full_bytes * 8 + bit] = pixel_y::<INVERT>(byte, bit) as u16;
    }
  }
}

// ---- Monoblack (INVERT=false) -----------------------------------------------

/// Monoblack → RGB u8 (broadcast Y).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_rgb_row(data: &[u8], out: &mut [u8], width: usize) {
  mono1bit_to_rgb_generic::<false>(data, out, width);
}

/// Monoblack → RGBA u8 (broadcast Y, α=0xFF).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_rgba_row(data: &[u8], out: &mut [u8], width: usize) {
  mono1bit_to_rgba_generic::<false>(data, out, width);
}

/// Monoblack → RGB u16 (broadcast Y, zero-extended to u16).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_rgb_u16_row(data: &[u8], out: &mut [u16], width: usize) {
  mono1bit_to_rgb_u16_generic::<false>(data, out, width);
}

/// Monoblack → RGBA u16 (broadcast Y zero-extended to u16, α=0x00FF).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_rgba_u16_row(data: &[u8], out: &mut [u16], width: usize) {
  mono1bit_to_rgba_u16_generic::<false>(data, out, width);
}

/// Monoblack → Luma u8 (pass-through).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_luma_row(data: &[u8], out: &mut [u8], width: usize) {
  mono1bit_to_luma_row_generic::<false>(data, out, width);
}

/// Monoblack → Luma u16 (zero-extended to u16).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_luma_u16_row(data: &[u8], out: &mut [u16], width: usize) {
  mono1bit_to_luma_u16_generic::<false>(data, out, width);
}

/// Monoblack → HSV (H=0, S=0, V=Y, achromatic).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monoblack_to_hsv_row(
  data: &[u8],
  h: &mut [u8],
  s: &mut [u8],
  v: &mut [u8],
  width: usize,
) {
  mono1bit_to_luma_row_generic::<false>(data, v, width);
  for i in 0..width {
    h[i] = 0;
    s[i] = 0;
  }
}

// ---- Monowhite (INVERT=true) ------------------------------------------------

/// Monowhite → RGB u8 (broadcast Y, inverted polarity).
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_rgb_row(data: &[u8], out: &mut [u8], width: usize) {
  mono1bit_to_rgb_generic::<true>(data, out, width);
}

/// Monowhite → RGBA u8.
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_rgba_row(data: &[u8], out: &mut [u8], width: usize) {
  mono1bit_to_rgba_generic::<true>(data, out, width);
}

/// Monowhite → RGB u16.
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_rgb_u16_row(data: &[u8], out: &mut [u16], width: usize) {
  mono1bit_to_rgb_u16_generic::<true>(data, out, width);
}

/// Monowhite → RGBA u16.
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_rgba_u16_row(data: &[u8], out: &mut [u16], width: usize) {
  mono1bit_to_rgba_u16_generic::<true>(data, out, width);
}

/// Monowhite → Luma u8.
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_luma_row(data: &[u8], out: &mut [u8], width: usize) {
  mono1bit_to_luma_row_generic::<true>(data, out, width);
}

/// Monowhite → Luma u16.
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_luma_u16_row(data: &[u8], out: &mut [u16], width: usize) {
  mono1bit_to_luma_u16_generic::<true>(data, out, width);
}

/// Monowhite → HSV.
#[cfg_attr(not(tarpaulin), inline(always))]
pub(crate) fn monowhite_to_hsv_row(
  data: &[u8],
  h: &mut [u8],
  s: &mut [u8],
  v: &mut [u8],
  width: usize,
) {
  mono1bit_to_luma_row_generic::<true>(data, v, width);
  for i in 0..width {
    h[i] = 0;
    s[i] = 0;
  }
}