| 1 |
/* |
| 2 |
* // Copyright (c) Radzivon Bartoshyk 6/2025. All rights reserved. |
| 3 |
* // |
| 4 |
* // Redistribution and use in source and binary forms, with or without modification, |
| 5 |
* // are permitted provided that the following conditions are met: |
| 6 |
* // |
| 7 |
* // 1. Redistributions of source code must retain the above copyright notice, this |
| 8 |
* // list of conditions and the following disclaimer. |
| 9 |
* // |
| 10 |
* // 2. Redistributions in binary form must reproduce the above copyright notice, |
| 11 |
* // this list of conditions and the following disclaimer in the documentation |
| 12 |
* // and/or other materials provided with the distribution. |
| 13 |
* // |
| 14 |
* // 3. Neither the name of the copyright holder nor the names of its |
| 15 |
* // contributors may be used to endorse or promote products derived from |
| 16 |
* // this software without specific prior written permission. |
| 17 |
* // |
| 18 |
* // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
| 19 |
* // AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 20 |
* // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
| 21 |
* // DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE |
| 22 |
* // FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 23 |
* // DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR |
| 24 |
* // SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER |
| 25 |
* // CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, |
| 26 |
* // OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| 27 |
* // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 28 |
*/ |
| 29 |
use crate::conversions::katana::{KatanaFinalStage, KatanaInitialStage}; |
| 30 |
use crate::mlaf::mlaf; |
| 31 |
use crate::safe_math::SafeMul; |
| 32 |
use crate::trc::lut_interp_linear_float; |
| 33 |
use crate::{ |
| 34 |
CmsError, Cube, DataColorSpace, InterpolationMethod, LutMultidimensionalType, MalformedSize, |
| 35 |
Matrix3d, Matrix3f, PointeeSizeExpressible, TransformOptions, Vector3d, Vector3f, |
| 36 |
}; |
| 37 |
use num_traits::AsPrimitive; |
| 38 |
use std::marker::PhantomData; |
| 39 |
|
| 40 |
#[derive(Copy, Clone, Ord, PartialOrd, Eq, PartialEq, Debug)] |
| 41 |
pub(crate) enum MultidimensionalDirection { |
| 42 |
DeviceToPcs, |
| 43 |
PcsToDevice, |
| 44 |
} |
| 45 |
|
| 46 |
struct Multidimensional3x3< |
| 47 |
T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync, |
| 48 |
> { |
| 49 |
a_curves: Option<Box<[Vec<f32>; 3]>>, |
| 50 |
m_curves: Option<Box<[Vec<f32>; 3]>>, |
| 51 |
b_curves: Option<Box<[Vec<f32>; 3]>>, |
| 52 |
clut: Option<Vec<f32>>, |
| 53 |
matrix: Matrix3f, |
| 54 |
bias: Vector3f, |
| 55 |
direction: MultidimensionalDirection, |
| 56 |
options: TransformOptions, |
| 57 |
pcs: DataColorSpace, |
| 58 |
grid_size: [u8; 3], |
| 59 |
_phantom: PhantomData<T>, |
| 60 |
bit_depth: usize, |
| 61 |
} |
| 62 |
|
| 63 |
impl<T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync> |
| 64 |
Multidimensional3x3<T> |
| 65 |
{ |
| 66 |
fn execute_matrix_stage(&self, dst: &mut [f32]) { |
| 67 |
let m = self.matrix; |
| 68 |
let b = self.bias; |
| 69 |
|
| 70 |
if !m.test_equality(Matrix3f::IDENTITY) || !b.eq(&Vector3f::default()) { |
| 71 |
for dst in dst.chunks_exact_mut(3) { |
| 72 |
let x = dst[0]; |
| 73 |
let y = dst[1]; |
| 74 |
let z = dst[2]; |
| 75 |
dst[0] = mlaf(mlaf(mlaf(b.v[0], x, m.v[0][0]), y, m.v[0][1]), z, m.v[0][2]); |
| 76 |
dst[1] = mlaf(mlaf(mlaf(b.v[1], x, m.v[1][0]), y, m.v[1][1]), z, m.v[1][2]); |
| 77 |
dst[2] = mlaf(mlaf(mlaf(b.v[2], x, m.v[2][0]), y, m.v[2][1]), z, m.v[2][2]); |
| 78 |
} |
| 79 |
} |
| 80 |
} |
| 81 |
|
| 82 |
fn execute_simple_curves(&self, dst: &mut [f32], curves: &[Vec<f32>; 3]) { |
| 83 |
let curve0 = &curves[0]; |
| 84 |
let curve1 = &curves[1]; |
| 85 |
let curve2 = &curves[2]; |
| 86 |
|
| 87 |
for dst in dst.chunks_exact_mut(3) { |
| 88 |
let a0 = dst[0]; |
| 89 |
let a1 = dst[1]; |
| 90 |
let a2 = dst[2]; |
| 91 |
let b0 = lut_interp_linear_float(a0, curve0); |
| 92 |
let b1 = lut_interp_linear_float(a1, curve1); |
| 93 |
let b2 = lut_interp_linear_float(a2, curve2); |
| 94 |
dst[0] = b0; |
| 95 |
dst[1] = b1; |
| 96 |
dst[2] = b2; |
| 97 |
} |
| 98 |
} |
| 99 |
|
| 100 |
fn to_pcs_impl<Fetch: Fn(f32, f32, f32) -> Vector3f>( |
| 101 |
&self, |
| 102 |
input: &[T], |
| 103 |
dst: &mut [f32], |
| 104 |
fetch: Fetch, |
| 105 |
) -> Result<(), CmsError> { |
| 106 |
let norm_value = if T::FINITE { |
| 107 |
1.0 / ((1u32 << self.bit_depth) - 1) as f32 |
| 108 |
} else { |
| 109 |
1.0 |
| 110 |
}; |
| 111 |
assert_eq!( |
| 112 |
self.direction, |
| 113 |
MultidimensionalDirection::DeviceToPcs, |
| 114 |
"PCS to device cannot be used on `to pcs` stage" |
| 115 |
); |
| 116 |
|
| 117 |
// A -> B |
| 118 |
// OR B - A A - curves stage |
| 119 |
|
| 120 |
if let (Some(a_curves), Some(clut)) = (self.a_curves.as_ref(), self.clut.as_ref()) { |
| 121 |
if !clut.is_empty() { |
| 122 |
let curve0 = &a_curves[0]; |
| 123 |
let curve1 = &a_curves[1]; |
| 124 |
let curve2 = &a_curves[2]; |
| 125 |
for (src, dst) in input.chunks_exact(3).zip(dst.chunks_exact_mut(3)) { |
| 126 |
let b0 = lut_interp_linear_float(src[0].as_() * norm_value, curve0); |
| 127 |
let b1 = lut_interp_linear_float(src[1].as_() * norm_value, curve1); |
| 128 |
let b2 = lut_interp_linear_float(src[2].as_() * norm_value, curve2); |
| 129 |
let interpolated = fetch(b0, b1, b2); |
| 130 |
dst[0] = interpolated.v[0]; |
| 131 |
dst[1] = interpolated.v[1]; |
| 132 |
dst[2] = interpolated.v[2]; |
| 133 |
} |
| 134 |
} else { |
| 135 |
for (src, dst) in input.chunks_exact(3).zip(dst.chunks_exact_mut(3)) { |
| 136 |
dst[0] = src[0].as_() * norm_value; |
| 137 |
dst[1] = src[1].as_() * norm_value; |
| 138 |
dst[2] = src[2].as_() * norm_value; |
| 139 |
} |
| 140 |
} |
| 141 |
} else { |
| 142 |
for (src, dst) in input.chunks_exact(3).zip(dst.chunks_exact_mut(3)) { |
| 143 |
dst[0] = src[0].as_() * norm_value; |
| 144 |
dst[1] = src[1].as_() * norm_value; |
| 145 |
dst[2] = src[2].as_() * norm_value; |
| 146 |
} |
| 147 |
} |
| 148 |
|
| 149 |
// Matrix stage |
| 150 |
|
| 151 |
if let Some(m_curves) = self.m_curves.as_ref() { |
| 152 |
self.execute_simple_curves(dst, m_curves); |
| 153 |
self.execute_matrix_stage(dst); |
| 154 |
} |
| 155 |
|
| 156 |
// B-curves is mandatory |
| 157 |
if let Some(b_curves) = &self.b_curves.as_ref() { |
| 158 |
self.execute_simple_curves(dst, b_curves); |
| 159 |
} |
| 160 |
|
| 161 |
Ok(()) |
| 162 |
} |
| 163 |
} |
| 164 |
|
| 165 |
impl<T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync> |
| 166 |
KatanaInitialStage<f32, T> for Multidimensional3x3<T> |
| 167 |
{ |
| 168 |
fn to_pcs(&self, input: &[T]) -> Result<Vec<f32>, CmsError> { |
| 169 |
if input.len() % 3 != 0 { |
| 170 |
return Err(CmsError::LaneMultipleOfChannels); |
| 171 |
} |
| 172 |
let fixed_new_clut = Vec::new(); |
| 173 |
let new_clut = self.clut.as_ref().unwrap_or(&fixed_new_clut); |
| 174 |
let lut = Cube::new_checked_cube(new_clut, self.grid_size, 3)?; |
| 175 |
|
| 176 |
let mut new_dst = vec![0f32; input.len()]; |
| 177 |
|
| 178 |
// If PCS is LAB then linear interpolation should be used |
| 179 |
if self.pcs == DataColorSpace::Lab || self.pcs == DataColorSpace::Xyz { |
| 180 |
self.to_pcs_impl(input, &mut new_dst, |x, y, z| lut.trilinear_vec3(x, y, z))?; |
| 181 |
return Ok(new_dst); |
| 182 |
} |
| 183 |
|
| 184 |
match self.options.interpolation_method { |
| 185 |
#[cfg(feature = "options")] |
| 186 |
InterpolationMethod::Tetrahedral => { |
| 187 |
self.to_pcs_impl(input, &mut new_dst, |x, y, z| lut.tetra_vec3(x, y, z))?; |
| 188 |
} |
| 189 |
#[cfg(feature = "options")] |
| 190 |
InterpolationMethod::Pyramid => { |
| 191 |
self.to_pcs_impl(input, &mut new_dst, |x, y, z| lut.pyramid_vec3(x, y, z))?; |
| 192 |
} |
| 193 |
#[cfg(feature = "options")] |
| 194 |
InterpolationMethod::Prism => { |
| 195 |
self.to_pcs_impl(input, &mut new_dst, |x, y, z| lut.prism_vec3(x, y, z))?; |
| 196 |
} |
| 197 |
InterpolationMethod::Linear => { |
| 198 |
self.to_pcs_impl(input, &mut new_dst, |x, y, z| lut.trilinear_vec3(x, y, z))?; |
| 199 |
} |
| 200 |
} |
| 201 |
Ok(new_dst) |
| 202 |
} |
| 203 |
} |
| 204 |
|
| 205 |
impl<T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync> |
| 206 |
Multidimensional3x3<T> |
| 207 |
where |
| 208 |
f32: AsPrimitive<T>, |
| 209 |
{ |
| 210 |
fn to_output_impl<Fetch: Fn(f32, f32, f32) -> Vector3f>( |
| 211 |
&self, |
| 212 |
src: &mut [f32], |
| 213 |
dst: &mut [T], |
| 214 |
fetch: Fetch, |
| 215 |
) -> Result<(), CmsError> { |
| 216 |
let norm_value = if T::FINITE { |
| 217 |
((1u32 << self.bit_depth) - 1) as f32 |
| 218 |
} else { |
| 219 |
1.0 |
| 220 |
}; |
| 221 |
assert_eq!( |
| 222 |
self.direction, |
| 223 |
MultidimensionalDirection::PcsToDevice, |
| 224 |
"Device to PCS cannot be used on `to output` stage" |
| 225 |
); |
| 226 |
|
| 227 |
if let Some(b_curves) = &self.b_curves.as_ref() { |
| 228 |
self.execute_simple_curves(src, b_curves); |
| 229 |
} |
| 230 |
|
| 231 |
// Matrix stage |
| 232 |
|
| 233 |
if let Some(m_curves) = self.m_curves.as_ref() { |
| 234 |
self.execute_matrix_stage(src); |
| 235 |
self.execute_simple_curves(src, m_curves); |
| 236 |
} |
| 237 |
|
| 238 |
if let (Some(a_curves), Some(clut)) = (self.a_curves.as_ref(), self.clut.as_ref()) { |
| 239 |
if !clut.is_empty() { |
| 240 |
let curve0 = &a_curves[0]; |
| 241 |
let curve1 = &a_curves[1]; |
| 242 |
let curve2 = &a_curves[2]; |
| 243 |
for (src, dst) in src.chunks_exact(3).zip(dst.chunks_exact_mut(3)) { |
| 244 |
let b0 = lut_interp_linear_float(src[0], curve0); |
| 245 |
let b1 = lut_interp_linear_float(src[1], curve1); |
| 246 |
let b2 = lut_interp_linear_float(src[2], curve2); |
| 247 |
let interpolated = fetch(b0, b1, b2); |
| 248 |
if T::FINITE { |
| 249 |
dst[0] = (interpolated.v[0] * norm_value) |
| 250 |
.round() |
| 251 |
.max(0.0) |
| 252 |
.min(norm_value) |
| 253 |
.as_(); |
| 254 |
dst[1] = (interpolated.v[1] * norm_value) |
| 255 |
.round() |
| 256 |
.max(0.0) |
| 257 |
.min(norm_value) |
| 258 |
.as_(); |
| 259 |
dst[2] = (interpolated.v[2] * norm_value) |
| 260 |
.round() |
| 261 |
.max(0.0) |
| 262 |
.min(norm_value) |
| 263 |
.as_(); |
| 264 |
} else { |
| 265 |
dst[0] = interpolated.v[0].as_(); |
| 266 |
dst[1] = interpolated.v[1].as_(); |
| 267 |
dst[2] = interpolated.v[2].as_(); |
| 268 |
} |
| 269 |
} |
| 270 |
} else { |
| 271 |
for (src, dst) in src.chunks_exact(3).zip(dst.chunks_exact_mut(3)) { |
| 272 |
if T::FINITE { |
| 273 |
dst[0] = (src[0] * norm_value).round().max(0.0).min(norm_value).as_(); |
| 274 |
dst[1] = (src[1] * norm_value).round().max(0.0).min(norm_value).as_(); |
| 275 |
dst[2] = (src[2] * norm_value).round().max(0.0).min(norm_value).as_(); |
| 276 |
} else { |
| 277 |
dst[0] = src[0].as_(); |
| 278 |
dst[1] = src[1].as_(); |
| 279 |
dst[2] = src[2].as_(); |
| 280 |
} |
| 281 |
} |
| 282 |
} |
| 283 |
} else { |
| 284 |
for (src, dst) in src.chunks_exact(3).zip(dst.chunks_exact_mut(3)) { |
| 285 |
if T::FINITE { |
| 286 |
dst[0] = (src[0] * norm_value).round().max(0.0).min(norm_value).as_(); |
| 287 |
dst[1] = (src[1] * norm_value).round().max(0.0).min(norm_value).as_(); |
| 288 |
dst[2] = (src[2] * norm_value).round().max(0.0).min(norm_value).as_(); |
| 289 |
} else { |
| 290 |
dst[0] = src[0].as_(); |
| 291 |
dst[1] = src[1].as_(); |
| 292 |
dst[2] = src[2].as_(); |
| 293 |
} |
| 294 |
} |
| 295 |
} |
| 296 |
|
| 297 |
Ok(()) |
| 298 |
} |
| 299 |
} |
| 300 |
|
| 301 |
impl<T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync> |
| 302 |
KatanaFinalStage<f32, T> for Multidimensional3x3<T> |
| 303 |
where |
| 304 |
f32: AsPrimitive<T>, |
| 305 |
{ |
| 306 |
fn to_output(&self, src: &mut [f32], dst: &mut [T]) -> Result<(), CmsError> { |
| 307 |
if src.len() % 3 != 0 { |
| 308 |
return Err(CmsError::LaneMultipleOfChannels); |
| 309 |
} |
| 310 |
if dst.len() % 3 != 0 { |
| 311 |
return Err(CmsError::LaneMultipleOfChannels); |
| 312 |
} |
| 313 |
if src.len() != dst.len() { |
| 314 |
return Err(CmsError::LaneSizeMismatch); |
| 315 |
} |
| 316 |
let fixed_new_clut = Vec::new(); |
| 317 |
let new_clut = self.clut.as_ref().unwrap_or(&fixed_new_clut); |
| 318 |
let lut = Cube::new_checked_cube(new_clut, self.grid_size, 3)?; |
| 319 |
|
| 320 |
// If PCS is LAB then linear interpolation should be used |
| 321 |
if self.pcs == DataColorSpace::Lab || self.pcs == DataColorSpace::Xyz { |
| 322 |
return self.to_output_impl(src, dst, |x, y, z| lut.trilinear_vec3(x, y, z)); |
| 323 |
} |
| 324 |
|
| 325 |
match self.options.interpolation_method { |
| 326 |
#[cfg(feature = "options")] |
| 327 |
InterpolationMethod::Tetrahedral => { |
| 328 |
self.to_output_impl(src, dst, |x, y, z| lut.tetra_vec3(x, y, z))?; |
| 329 |
} |
| 330 |
#[cfg(feature = "options")] |
| 331 |
InterpolationMethod::Pyramid => { |
| 332 |
self.to_output_impl(src, dst, |x, y, z| lut.pyramid_vec3(x, y, z))?; |
| 333 |
} |
| 334 |
#[cfg(feature = "options")] |
| 335 |
InterpolationMethod::Prism => { |
| 336 |
self.to_output_impl(src, dst, |x, y, z| lut.prism_vec3(x, y, z))?; |
| 337 |
} |
| 338 |
InterpolationMethod::Linear => { |
| 339 |
self.to_output_impl(src, dst, |x, y, z| lut.trilinear_vec3(x, y, z))?; |
| 340 |
} |
| 341 |
} |
| 342 |
Ok(()) |
| 343 |
} |
| 344 |
} |
| 345 |
|
| 346 |
fn make_multidimensional_3x3< |
| 347 |
T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync, |
| 348 |
>( |
| 349 |
mab: &LutMultidimensionalType, |
| 350 |
options: TransformOptions, |
| 351 |
pcs: DataColorSpace, |
| 352 |
direction: MultidimensionalDirection, |
| 353 |
bit_depth: usize, |
| 354 |
) -> Result<Multidimensional3x3<T>, CmsError> { |
| 355 |
if mab.num_input_channels != 3 && mab.num_output_channels != 3 { |
| 356 |
return Err(CmsError::UnsupportedProfileConnection); |
| 357 |
} |
| 358 |
if mab.b_curves.is_empty() || mab.b_curves.len() != 3 { |
| 359 |
return Err(CmsError::InvalidAtoBLut); |
| 360 |
} |
| 361 |
|
| 362 |
let grid_size = [mab.grid_points[0], mab.grid_points[1], mab.grid_points[2]]; |
| 363 |
|
| 364 |
let clut: Option<Vec<f32>> = if mab.a_curves.len() == 3 && mab.clut.is_some() { |
| 365 |
let clut = mab.clut.as_ref().map(|x| x.to_clut_f32()).unwrap(); |
| 366 |
let lut_grid = (mab.grid_points[0] as usize) |
| 367 |
.safe_mul(mab.grid_points[1] as usize)? |
| 368 |
.safe_mul(mab.grid_points[2] as usize)? |
| 369 |
.safe_mul(mab.num_output_channels as usize)?; |
| 370 |
if clut.len() != lut_grid { |
| 371 |
return Err(CmsError::MalformedCurveLutTable(MalformedSize { |
| 372 |
size: clut.len(), |
| 373 |
expected: lut_grid, |
| 374 |
})); |
| 375 |
} |
| 376 |
Some(clut) |
| 377 |
} else { |
| 378 |
None |
| 379 |
}; |
| 380 |
|
| 381 |
let a_curves: Option<Box<[Vec<f32>; 3]>> = if mab.a_curves.len() == 3 && mab.clut.is_some() { |
| 382 |
let mut arr = Box::<[Vec<f32>; 3]>::default(); |
| 383 |
for (a_curve, dst) in mab.a_curves.iter().zip(arr.iter_mut()) { |
| 384 |
*dst = a_curve.to_clut()?; |
| 385 |
} |
| 386 |
Some(arr) |
| 387 |
} else { |
| 388 |
None |
| 389 |
}; |
| 390 |
|
| 391 |
let b_curves: Option<Box<[Vec<f32>; 3]>> = if mab.b_curves.len() == 3 { |
| 392 |
let mut arr = Box::<[Vec<f32>; 3]>::default(); |
| 393 |
let all_curves_linear = mab.b_curves.iter().all(|curve| curve.is_linear()); |
| 394 |
if all_curves_linear { |
| 395 |
None |
| 396 |
} else { |
| 397 |
for (c_curve, dst) in mab.b_curves.iter().zip(arr.iter_mut()) { |
| 398 |
*dst = c_curve.to_clut()?; |
| 399 |
} |
| 400 |
Some(arr) |
| 401 |
} |
| 402 |
} else { |
| 403 |
return Err(CmsError::InvalidAtoBLut); |
| 404 |
}; |
| 405 |
|
| 406 |
let matrix = mab.matrix.to_f32(); |
| 407 |
|
| 408 |
let m_curves: Option<Box<[Vec<f32>; 3]>> = if mab.m_curves.len() == 3 { |
| 409 |
let all_curves_linear = mab.m_curves.iter().all(|curve| curve.is_linear()); |
| 410 |
if !all_curves_linear |
| 411 |
|| !mab.matrix.test_equality(Matrix3d::IDENTITY) |
| 412 |
|| mab.bias.ne(&Vector3d::default()) |
| 413 |
{ |
| 414 |
let mut arr = Box::<[Vec<f32>; 3]>::default(); |
| 415 |
for (curve, dst) in mab.m_curves.iter().zip(arr.iter_mut()) { |
| 416 |
*dst = curve.to_clut()?; |
| 417 |
} |
| 418 |
Some(arr) |
| 419 |
} else { |
| 420 |
None |
| 421 |
} |
| 422 |
} else { |
| 423 |
None |
| 424 |
}; |
| 425 |
|
| 426 |
let bias = mab.bias.cast(); |
| 427 |
|
| 428 |
let transform = Multidimensional3x3::<T> { |
| 429 |
a_curves, |
| 430 |
b_curves, |
| 431 |
m_curves, |
| 432 |
matrix, |
| 433 |
direction, |
| 434 |
options, |
| 435 |
clut, |
| 436 |
pcs, |
| 437 |
grid_size, |
| 438 |
bias, |
| 439 |
_phantom: PhantomData, |
| 440 |
bit_depth, |
| 441 |
}; |
| 442 |
|
| 443 |
Ok(transform) |
| 444 |
} |
| 445 |
|
| 446 |
pub(crate) fn multi_dimensional_3x3_to_pcs< |
| 447 |
T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync, |
| 448 |
>( |
| 449 |
mab: &LutMultidimensionalType, |
| 450 |
options: TransformOptions, |
| 451 |
pcs: DataColorSpace, |
| 452 |
bit_depth: usize, |
| 453 |
) -> Result<Box<dyn KatanaInitialStage<f32, T> + Send + Sync>, CmsError> { |
| 454 |
let transform = make_multidimensional_3x3::<T>( |
| 455 |
mab, |
| 456 |
options, |
| 457 |
pcs, |
| 458 |
MultidimensionalDirection::DeviceToPcs, |
| 459 |
bit_depth, |
| 460 |
)?; |
| 461 |
Ok(Box::new(transform)) |
| 462 |
} |
| 463 |
|
| 464 |
pub(crate) fn multi_dimensional_3x3_to_device< |
| 465 |
T: Copy + Default + AsPrimitive<f32> + PointeeSizeExpressible + Send + Sync, |
| 466 |
>( |
| 467 |
mab: &LutMultidimensionalType, |
| 468 |
options: TransformOptions, |
| 469 |
pcs: DataColorSpace, |
| 470 |
bit_depth: usize, |
| 471 |
) -> Result<Box<dyn KatanaFinalStage<f32, T> + Send + Sync>, CmsError> |
| 472 |
where |
| 473 |
f32: AsPrimitive<T>, |
| 474 |
{ |
| 475 |
let transform = make_multidimensional_3x3::<T>( |
| 476 |
mab, |
| 477 |
options, |
| 478 |
pcs, |
| 479 |
MultidimensionalDirection::PcsToDevice, |
| 480 |
bit_depth, |
| 481 |
)?; |
| 482 |
Ok(Box::new(transform)) |
| 483 |
} |