filters/layer.hpp - GCC Code Coverage Report

Directory:	avs_core/
Coverage:	low: ≥ 0% medium: ≥ 75.0% high: ≥ 90.0%
	Coverage	Exec / Excl / Total
Lines:	2.6%	192 / 0 / 7481
Functions:	4.2%	11 / 0 / 263
Branches:	1.9%	45 / 0 / 2410
    filters/layer.hpp
    
        Line
        Branch
        Exec
        Source
      
        // Basic utils in C++ language, to be included in both base and processor specific (e.g. avx2) source modules, where they can
      
        // be optimized for the specific instruction set.
      
        // yuv add, subtract, mul, lighten, darken
      
        // Chroma placement helpers (calculate_effective_mask*, prepare_effective_mask_for_row*)
      
        // are now in overlay/blend_common.h, pulled in via layer.h.
      
        // ---------------------------------------------------------------------------
      
        // Rowprep function selectors — defined by the including TU to inject SIMD variants.
      
        // Default: C scalar functions from overlay/blend_common.h.
      
        //
      
        // LAYER_ROWPREP_FN      — spatial mask averaging (+ opacity baking when
      
        //                         full_opacity=false.
      
        //                         Used for the full_opacity=true path: returns only spatial
      
        //                         averages (or maskp directly for MASK444).
      
        //
      
        // Including TU (e.g. layer_avx2.cpp) defines them before this #include:
      
        //   #define LAYER_ROWPREP_FN       prepare_effective_mask_for_row_avx2
      
        // to allow the same C template to compile with arch-spcific different rowprep variants 
      
        // ---------------------------------------------------------------------------
      
        #ifndef LAYER_ROWPREP_FN
      
        #  define LAYER_ROWPREP_FN  prepare_effective_mask_for_row
      
        #endif
      
        DISABLE_WARNING_PUSH
      
        DISABLE_WARNING_UNREFERENCED_LOCAL_VARIABLE
      
        // YUV(A) mul 8-16 bits
      
        // full_opacity == true : rowprep returns mask-average; subsample-aware avg mask is put into effective_mask_buffer[x]. 444 case: returns maskp directly.
      
        // full_opacity == false: rowprep pre-calculates with 'opacity' in the formulae, effective_mask_buffer is filled even for 444.
      
        template<MaskMode maskMode, typename pixel_t, bool is_chroma, bool has_alpha, bool full_opacity>
      
        3
        static void layer_yuv_mul_c_inner(BYTE* dstp8, const BYTE* ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, int opacity_i, int bits_per_pixel) {
      
        3
          pixel_t* dstp = reinterpret_cast<pixel_t*>(dstp8);
      
        3
          const pixel_t* ovrp = reinterpret_cast<const pixel_t*>(ovrp8);
      
        3
          const pixel_t* maskp = has_alpha ? reinterpret_cast<const pixel_t*>(maskp8) : nullptr;
      
        3
          dst_pitch /= sizeof(pixel_t);
      
        3
          overlay_pitch /= sizeof(pixel_t);
      
          if constexpr (has_alpha)
      
        ✗
            mask_pitch /= sizeof(pixel_t);
      
          if constexpr (sizeof(pixel_t) == 1)
      
        3
            bits_per_pixel = 8; // make quasi constexpr
      
        3
          const MagicDiv magic = get_magic_div(bits_per_pixel);
      
        3
          const int max_pixel_value = (1 << bits_per_pixel) - 1;
      
        3
          const int half = max_pixel_value / 2;
      
          // Buffer: needed for subsampled spatial averaging, or MASK444+!full_opacity baking.
      
        3
          std::vector<pixel_t> effective_mask_buffer;
      
          if constexpr (has_alpha && (maskMode != MASK444 || !full_opacity)) {
      
        ✗
            effective_mask_buffer.resize(width);
      
          }
      
          4/130void layer_yuv_mul_c_inner<(MaskMode)0, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 13 → 6 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 21 → 6 not taken.
✗ Branch 21 → 22 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, false, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, false, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✓ Branch 11 → 5 taken 3 times.
✓ Branch 11 → 12 taken 1 time.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, false, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
✗ Branch 13 → 5 not taken.
✗ Branch 13 → 14 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✓ Branch 11 → 5 taken 6 times.
✓ Branch 11 → 12 taken 2 times.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
✗ Branch 13 → 5 not taken.
✗ Branch 13 → 14 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, false, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, false, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, false, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
✗ Branch 13 → 5 not taken.
✗ Branch 13 → 14 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 5 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
✗ Branch 13 → 5 not taken.
✗ Branch 13 → 14 not taken.

        12
          for (int y = 0; y < height; ++y) {
      
        9
            const pixel_t* effective_mask_ptr = nullptr;
      
            // Rowprep: full_opacity path returns spatial avg (or maskp for MASK444).
      
            //          !full_opacity path bakes (avg * opacity_i) / div
      
            if constexpr (has_alpha) {
      
              if constexpr (full_opacity)
      
        ✗
                effective_mask_ptr = LAYER_ROWPREP_FN<maskMode, pixel_t, true>(maskp, mask_pitch, width, effective_mask_buffer);
      
              else
      
                effective_mask_ptr = LAYER_ROWPREP_FN<maskMode, pixel_t, false>(maskp, mask_pitch, width, effective_mask_buffer, opacity_i, half, magic);
      
            }
      
            // Main blending loop — alpha_mask is fully prepared by rowprep.
      
          4/130void layer_yuv_mul_c_inner<(MaskMode)0, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)1, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)2, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)3, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)4, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)5, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)6, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
✗ Branch 19 → 16 not taken.
✗ Branch 19 → 20 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, false, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, false, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✓ Branch 9 → 6 taken 24 times.
✓ Branch 9 → 10 taken 3 times.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, false, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 10 → 7 not taken.
✗ Branch 10 → 11 not taken.
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✓ Branch 9 → 6 taken 48 times.
✓ Branch 9 → 10 taken 6 times.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned char, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 10 → 7 not taken.
✗ Branch 10 → 11 not taken.
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, false, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, false, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, false, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 10 → 7 not taken.
✗ Branch 10 → 11 not taken.
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, true, false, false>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, true, false, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 9 → 6 not taken.
✗ Branch 9 → 10 not taken.
void layer_yuv_mul_c_inner<(MaskMode)7, unsigned short, true, true, true>(unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 10 → 7 not taken.
✗ Branch 10 → 11 not taken.
✗ Branch 11 → 8 not taken.
✗ Branch 11 → 12 not taken.

        81
            for (int x = 0; x < width; ++x) {
      
              // has_alpha=true: rowprep baked level in (!full_opacity) or returned avg (full_opacity).
      
              // has_alpha=false: flat level, no mask.
      
        72
              const uint32_t alpha_mask = has_alpha ? effective_mask_ptr[x] : opacity_i;
      
        72
              const uint32_t inv_alpha = max_pixel_value - alpha_mask;
      
              uint32_t target_pixel;
      
              if constexpr (!is_chroma) {
      
                // luma: blend towards the "Multiplied" product, no rounding is done here.
      
                // Calculate the multiplied product (A*B)/max
      
                // SIMD hint:
      
                // - on Intel for bits_per_pixel == 16, use _mm_mulhi_epu16 to get the high 16 bits of the 16x16->32 product.
      
                //   For lower bit depths, use _mm_mullo_epi16 and shift as needed.
      
                //   This means a separate code path for 10-14 and exact 16-bit pixels.
      
                // - For universal 10-16 bit support on Intel, widen uint16_t to uint32_t, then
      
                //   use _mm_mullo_epi32 for full 32-bit multiplication, then shift right by bits_per_pixel and pack back to uint16_t.
      
                // uint32_t case is needed to hint that the product result is also unsigned.
      
        24
                const pixel_t prod = (pixel_t)(((uint32_t)ovrp[x] * (uint32_t)dstp[x]) >> bits_per_pixel);
      
        24
                target_pixel = prod;
      
                // was: dstp[x] = (pixel_t)(dstp[x] + ((((((calc_t)ovrp[x] * dstp[x]) >> bits_per_pixel) - dstp[x]) * alpha_mask) >> bits_per_pixel));
      
              }
      
              else {
      
                // Note: when use_chrome=false, ovrp is prefilled with neutral, so the same formula applies for both modes
      
                // CHROMA (Normal): Blend towards the overlay chroma (use_chroma=true) or towards neutral (use_chroma=false).
      
                // Instead of dst + ((ovr - dst) * alpha) >> bits we use ((dst * (max - alpha)) + (ovr * alpha)) / max
      
                // U = U + ( ((Uovr - U)*level) >> 8 )
      
                // V = V + ( ((Vovr - V)*level) >> 8 )
      
        48
                target_pixel = (uint32_t)ovrp[x];
      
                // for use_chroma = false, ovrp is prefilled with neutral
      
                // was: dstp[x] = (pixel_t)(dstp[x] + (((calc_t)(ovrp[x] - dstp[x]) * alpha_mask) >> bits_per_pixel));
      
                // The YUY2 MMX heritage (cargo cult programming) of halving the strength is finally eliminated
      
                /*
      
                const int half = 1 << (bits_per_pixel - 1);
      
                dstp[x] = (pixel_t)(dstp[x] + (((calc_t)(half - dstp[x]) * (alpha_mask / 2)) >> bits_per_pixel));
      
                // U = U + ( ((128 - U)*(level/2)) >> 8 )
      
                // V = V + ( ((128 - V)*(level/2)) >> 8 )
      
                */
      
              }
      
        144
              dstp[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp[x] * inv_alpha + target_pixel * alpha_mask + half, magic);
      
            }
      
        9
            dstp += dst_pitch;
      
        9
            ovrp += overlay_pitch;
      
            if constexpr (has_alpha) {
      
              if constexpr (maskMode == MASK420 || maskMode == MASK420_MPEG2 || maskMode == MASK420_TOPLEFT)
      
        ✗
                maskp += mask_pitch * 2;
      
              else
      
        ✗
                maskp += mask_pitch;
      
            }
      
          }
      
        3
        }
      
        // Outer dispatcher: full_opacity branch.
      
        template<MaskMode maskMode, typename pixel_t, bool is_chroma, bool has_alpha>
      
        3
        static void layer_yuv_mul_c(BYTE* dstp8, const BYTE* ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, int opacity_i, int bits_per_pixel) {
      
          if constexpr (!has_alpha) {
      
            // No mask; full_opacity choice is irrelevant; use true to skip dead buffer allocation.
      
        3
            layer_yuv_mul_c_inner<maskMode, pixel_t, is_chroma, false, true>(
      
              dstp8, ovrp8, maskp8, dst_pitch, overlay_pitch, mask_pitch, width, height, opacity_i, bits_per_pixel);
      
          } else {
      
        ✗
            const int max_pixel_value = (1 << bits_per_pixel) - 1;
      
        ✗
            if (opacity_i >= max_pixel_value) // full opacity
      
        ✗
              layer_yuv_mul_c_inner<maskMode, pixel_t, is_chroma, true, true>(
      
                dstp8, ovrp8, maskp8, dst_pitch, overlay_pitch, mask_pitch, width, height, opacity_i, bits_per_pixel);
      
            else
      
        ✗
              layer_yuv_mul_c_inner<maskMode, pixel_t, is_chroma, false, false>(
      
                dstp8, ovrp8, maskp8, dst_pitch, overlay_pitch, mask_pitch, width, height, opacity_i, bits_per_pixel);
      
          }
      
        3
        }
      
        // YUV(A) mul 32 bits
      
        template<MaskMode maskMode, bool is_chroma, bool has_alpha>
      
        ✗
        static void layer_yuv_mul_f_c(BYTE* dstp8, const BYTE* ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, float opacity) {
      
        ✗
          float* dstp = reinterpret_cast<float*>(dstp8);
      
        ✗
          const float* ovrp = reinterpret_cast<const float*>(ovrp8);
      
        ✗
          const float* maskp = reinterpret_cast<const float*>(maskp8);
      
        ✗
          dst_pitch /= sizeof(float);
      
        ✗
          overlay_pitch /= sizeof(float);
      
        ✗
          mask_pitch /= sizeof(float);
      
          // precalculate mask buffer
      
        ✗
          std::vector<float> effective_mask_buffer;
      
          if constexpr (has_alpha && maskMode != MASK444) {
      
        ✗
            effective_mask_buffer.resize(width);
      
          }
      
        ✗
          for (int y = 0; y < height; ++y) {
      
        ✗
            const float* effective_mask_ptr = nullptr;
      
            // precalculate effective mask for this row
      
            if constexpr (has_alpha) {
      
        ✗
              effective_mask_ptr = prepare_effective_mask_for_row_float_c<maskMode>(maskp, mask_pitch, width, effective_mask_buffer);
      
            }
      
            // blending loop
      
        ✗
            for (int x = 0; x < width; ++x) {
      
        ✗
              float effective_mask = has_alpha ? effective_mask_ptr[x] : 0.0f;
      
        ✗
              float alpha_mask = has_alpha ? effective_mask * opacity : opacity;
      
              if constexpr (!is_chroma)
      
        ✗
                dstp[x] = dstp[x] + (ovrp[x] * dstp[x] - dstp[x]) * alpha_mask;
      
              else {
      
                // Note: when use_chrome=false, ovrp is prefilled with neutral, so the same formula applies for both modes
      
        ✗
                dstp[x] = dstp[x] + (ovrp[x] - dstp[x]) * alpha_mask;
      
              }
      
            }
      
        ✗
            dstp += dst_pitch;
      
        ✗
            ovrp += overlay_pitch;
      
            if constexpr (has_alpha) {
      
              if constexpr (maskMode == MASK420 || maskMode == MASK420_MPEG2 || maskMode == MASK420_TOPLEFT)
      
        ✗
                maskp += mask_pitch * 2;
      
              else
      
        ✗
                maskp += mask_pitch;
      
            }
      
          }
      
        ✗
        }
      
        /* Comparison of Multiply (Mul) blend mode: Overlay vs. Layer
      
         * Luma (Y): Compatible. Both use standard (Base * Overlay) / Max logic.
      
         * Chroma (UV): Incompatible. They utilize different "intent" for color:
      
         * - Overlay (Legacy): Dynamic desaturation. Base chroma is pulled toward
      
         * neutral (128/0.0) scaled by the specific Overlay Luma intensity (O_Y).
      
         * Formula: Result_UV = (Base_UV * O_Y + Neutral * (Max - O_Y)) / Max
      
         * - Layer (use_chroma=false): Static desaturation. Ignores overlay pixels
      
         * entirely; blends toward a fixed neutral point via global opacity/mask.
      
         * Formula: Result_UV = Base_UV + (Neutral - Base_UV) * Mask_Opacity
      
         * Summary: Overlay treats darkness as a "color vacuum" (luma-dependent),
      
         * while Layer treats Y and UV as independent channels.
      
         * The function would need all plane parameters and work like lighten/darken instead of
      
         * template<MaskMode maskMode, bool is_chroma, bool has_alpha>
      
         * static void layer_yuv_mul_f_c(BYTE* dstp8, const BYTE* ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, float opacity) {
      
         * like this (?):
      
         * template<int mode, MaskMode maskMode, typename pixel_t, bool lumaonly, bool has_alpha>
      
         * static void layer_yuv_mul_overlaystyle_f_c(
      
         */
      
        // Unlike RGBA version, YUVA does not update destination alpha
      
        // so no "blend_alpha" template parameter like at planar rgb lighten_darken
      
        // Note, that we always call one instance of the filter: the lumaonly==false version,
      
        // since the filter handles U and V first, then recursively calls itself with lumaonly==true;
      
        // the chroma part needs unaltered luma for decision, so we delay Y processing.
      
        // Except Y-only greayscale input, where only luma plane exists.
      
        template<int mode, MaskMode maskMode, typename pixel_t, bool lumaonly, bool has_alpha>
      
        ✗
        static void layer_yuv_lighten_darken_c(
      
          BYTE* dstp8, BYTE* dstp8_u, BYTE* dstp8_v,/* BYTE* dstp8_a,*/
      
          const BYTE* ovrp8, const BYTE* ovrp8_u, const BYTE* ovrp8_v, const BYTE* maskp8,
      
          int dst_pitch, int dst_pitchUV,
      
          int overlay_pitch, int overlay_pitchUV,
      
          int mask_pitch,
      
          int width, int height, int opacity_i, int thresh,
      
          int bits_per_pixel) {
      
        ✗
          pixel_t* dstp = reinterpret_cast<pixel_t*>(dstp8);
      
        ✗
          pixel_t* dstp_u = reinterpret_cast<pixel_t*>(dstp8_u);
      
        ✗
          pixel_t* dstp_v = reinterpret_cast<pixel_t*>(dstp8_v);
      
          // pixel_t* dstp_a = reinterpret_cast<pixel_t *>(dstp8_a); // not destination alpha update
      
        ✗
          const pixel_t* ovrp = reinterpret_cast<const pixel_t*>(ovrp8);
      
        ✗
          const pixel_t* ovrp_u = reinterpret_cast<const pixel_t*>(ovrp8_u);
      
        ✗
          const pixel_t* ovrp_v = reinterpret_cast<const pixel_t*>(ovrp8_v);
      
        ✗
          const pixel_t* maskp = has_alpha ? reinterpret_cast<const pixel_t*>(maskp8) : nullptr;
      
        ✗
          dst_pitch /= sizeof(pixel_t);
      
        ✗
          dst_pitchUV /= sizeof(pixel_t);
      
        ✗
          overlay_pitch /= sizeof(pixel_t);
      
        ✗
          overlay_pitchUV /= sizeof(pixel_t);
      
          if constexpr (has_alpha)
      
            mask_pitch /= sizeof(pixel_t);
      
        ✗
          const int cwidth = (maskMode == MASK444) ? width : (maskMode == MASK411) ? width >> 2 : width >> 1; // 444:/1  420,422:/2  411:/4
      
        ✗
          const int cheight = (maskMode == MASK444 || maskMode == MASK422 || maskMode == MASK422_MPEG2 || maskMode == MASK422_TOPLEFT || maskMode == MASK411) ? height : height >> 1; // 444,422,411:/1  420:/2
      
          // In lighten/darken we need 3 buffers:
      
        ✗
          std::vector<pixel_t> ovr_buffer;
      
        ✗
          std::vector<pixel_t> src_buffer;
      
        ✗
          std::vector<pixel_t> mask_buffer;
      
          // precalculate mask buffer et al.
      
          if constexpr (maskMode != MASK444) {
      
        ✗
            ovr_buffer.resize(cwidth);
      
        ✗
            src_buffer.resize(cwidth);
      
            if constexpr (has_alpha)
      
              mask_buffer.resize(cwidth);
      
          }
      
          if constexpr (sizeof(pixel_t) == 1)
      
        ✗
            bits_per_pixel = 8; // make quasi constexpr
      
        ✗
          const MagicDiv magic = get_magic_div(bits_per_pixel);
      
        ✗
          const int max_pixel_value = (1 << bits_per_pixel) - 1;
      
        ✗
          const int half = max_pixel_value / 2;
      
          // for subsampled color spaces first do chroma, luma is only used for decision
      
          // second pass will do luma only
      
        ✗
          for (int y = 0; y < cheight; ++y) {
      
            // Prepare all three pointers using the helper, full-opacity mode, no opacity_i-magicdiv baking
      
        ✗
            const pixel_t* ovr_ptr = prepare_effective_mask_for_row<maskMode, pixel_t>(ovrp, overlay_pitch, cwidth, ovr_buffer);
      
        ✗
            const pixel_t* src_ptr = prepare_effective_mask_for_row<maskMode, pixel_t>(dstp, dst_pitch, cwidth, src_buffer);
      
        ✗
            const pixel_t* effective_mask_ptr = nullptr;
      
            if constexpr (has_alpha) {
      
              effective_mask_ptr = prepare_effective_mask_for_row<maskMode, pixel_t>(maskp, mask_pitch, cwidth, mask_buffer);
      
            }
      
        ✗
            for (int x = 0; x < cwidth; ++x) {
      
              // opacity_i is in [0..max_pixel_value], the same convention as masked_merge.
      
              uint32_t alpha_eff;
      
              if constexpr (has_alpha) {
      
                const uint32_t alpha_src = (uint32_t)effective_mask_ptr[x];
      
                alpha_eff = (uint32_t)magic_div_rt<pixel_t>(alpha_src * (uint32_t)opacity_i + half, magic);
      
              }
      
              else {
      
        ✗
                alpha_eff = opacity_i;
      
              }
      
        ✗
              int ovr = ovr_ptr[x];
      
        ✗
              int src = src_ptr[x];
      
              // If the threshold isn't met, the overlay weight becomes 0 (no change to dst)
      
              if constexpr (mode == LIGHTEN) {
      
        ✗
                if (!(ovr > (src + thresh))) alpha_eff = 0;
      
              }
      
              else { // DARKEN
      
        ✗
                if (!(ovr < (src - thresh))) alpha_eff = 0;
      
              }
      
        ✗
              const uint32_t inv_alpha = max_pixel_value - alpha_eff;
      
              // formula: dst = (dst * inv_alpha + ovr * alpha_eff) / max_val
      
              if constexpr (!lumaonly)
      
              {
      
                // chroma u,v
      
        ✗
                dstp_u[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_u[x] * inv_alpha + (uint32_t)ovrp_u[x] * alpha_eff + half, magic);
      
        ✗
                dstp_v[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_v[x] * inv_alpha + (uint32_t)ovrp_v[x] * alpha_eff + half, magic);
      
              }
      
              // for 444: update here, width/height is the same as for chroma
      
              if constexpr (maskMode == MASK444)
      
        ✗
                dstp[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp[x] * inv_alpha + (uint32_t)ovrp[x] * alpha_eff + half, magic);
      
            }
      
            if constexpr (maskMode == MASK420 || maskMode == MASK420_MPEG2 || maskMode == MASK420_TOPLEFT) {
      
        ✗
              dstp += dst_pitch * 2; // skip vertical subsampling
      
        ✗
              ovrp += overlay_pitch * 2;
      
              if constexpr (has_alpha) {
      
                //dstp_a += dst_pitch * 2;
      
                maskp += mask_pitch * 2;
      
              }
      
            }
      
            else {
      
        ✗
              dstp += dst_pitch;
      
        ✗
              ovrp += overlay_pitch;
      
              if constexpr (has_alpha) {
      
                //dstp_a += dst_pitch;
      
                maskp += mask_pitch;
      
              }
      
            }
      
            if constexpr (!lumaonly) {
      
        ✗
              dstp_u += dst_pitchUV;
      
        ✗
              dstp_v += dst_pitchUV;
      
        ✗
              ovrp_u += overlay_pitchUV;
      
        ✗
              ovrp_v += overlay_pitchUV;
      
            }
      
          }
      
        ✗
          dst_pitch *= sizeof(pixel_t);
      
        ✗
          dst_pitchUV *= sizeof(pixel_t);
      
        ✗
          overlay_pitch *= sizeof(pixel_t);
      
        ✗
          overlay_pitchUV *= sizeof(pixel_t);
      
        ✗
          mask_pitch *= sizeof(pixel_t);
      
          // Phase #2 recursively call the same function, but with lumaonly=true to update luma plane using the original chroma planes for decision
      
          // make luma
      
          if constexpr (!lumaonly && maskMode != MASK444)
      
        ✗
            layer_yuv_lighten_darken_c<mode, MASK444, pixel_t, true /* lumaonly*/, has_alpha>(
      
              dstp8, dstp8_u, dstp8_v, //dstp8_a,
      
              ovrp8, ovrp8_u, ovrp8_v, maskp8,
      
              dst_pitch, dst_pitchUV, overlay_pitch, overlay_pitchUV, mask_pitch,
      
              width, height, opacity_i, thresh, bits_per_pixel);
      
        ✗
        }
      
        template<int mode, MaskMode maskMode, bool lumaonly, bool has_alpha>
      
        ✗
        static void layer_yuv_lighten_darken_f_c(
      
          BYTE* dstp8, BYTE* dstp8_u, BYTE* dstp8_v /*, BYTE* dstp8_a*/,
      
          const BYTE* ovrp8, const BYTE* ovrp8_u, const BYTE* ovrp8_v, const BYTE* maskp8,
      
          int dst_pitch, int dst_pitchUV,
      
          int overlay_pitch, int overlay_pitchUV,
      
          int mask_pitch,
      
          int width, int height, float opacity, float thresh) {
      
        ✗
          float* dstp = reinterpret_cast<float*>(dstp8);
      
        ✗
          float* dstp_u = reinterpret_cast<float*>(dstp8_u);
      
        ✗
          float* dstp_v = reinterpret_cast<float*>(dstp8_v);
      
          //float* dstp_a = reinterpret_cast<float *>(dstp8_a);
      
        ✗
          const float* ovrp = reinterpret_cast<const float*>(ovrp8);
      
        ✗
          const float* ovrp_u = reinterpret_cast<const float*>(ovrp8_u);
      
        ✗
          const float* ovrp_v = reinterpret_cast<const float*>(ovrp8_v);
      
        ✗
          const float* maskp =  has_alpha ? reinterpret_cast<const float*>(maskp8) : nullptr;
      
        ✗
          dst_pitch /= sizeof(float);
      
        ✗
          dst_pitchUV /= sizeof(float);
      
        ✗
          overlay_pitch /= sizeof(float);
      
        ✗
          overlay_pitchUV /= sizeof(float);
      
          if constexpr (has_alpha)
      
            mask_pitch /= sizeof(float);
      
        ✗
          const int cwidth = (maskMode == MASK444) ? width : (maskMode == MASK411) ? width >> 2 : width >> 1; // 444:/1  420,422:/2  411:/4
      
        ✗
          const int cheight = (maskMode == MASK444 || maskMode == MASK422 || maskMode == MASK422_MPEG2 || maskMode == MASK422_TOPLEFT || maskMode == MASK411) ? height : height >> 1; // 444,422,411:/1  420:/2
      
          // In lighten/darken we need 3 buffers:
      
        ✗
          std::vector<float> ovr_buffer;
      
        ✗
          std::vector<float> src_buffer;
      
        ✗
          std::vector<float> mask_buffer;
      
          // precalculate mask buffer et al.
      
          if constexpr (maskMode != MASK444) {
      
        ✗
            ovr_buffer.resize(cwidth);
      
        ✗
            src_buffer.resize(cwidth);
      
            if constexpr (has_alpha)
      
              mask_buffer.resize(cwidth);
      
          }
      
          // for subsampled color spaces first do chroma, because luma is used for decision
      
          // second pass will do luma only
      
        ✗
          for (int y = 0; y < cheight; ++y) {
      
        ✗
            const float* ovr_ptr = prepare_effective_mask_for_row_float_c<maskMode>(ovrp, overlay_pitch, cwidth, ovr_buffer);
      
        ✗
            const float* src_ptr = prepare_effective_mask_for_row_float_c<maskMode>(dstp, dst_pitch, cwidth, src_buffer);
      
        ✗
            const float* effective_mask_ptr = nullptr;
      
            if constexpr (has_alpha) {
      
              effective_mask_ptr = prepare_effective_mask_for_row_float_c<maskMode>(maskp, mask_pitch, cwidth, mask_buffer);
      
            }
      
        ✗
            for (int x = 0; x < cwidth; ++x) {
      
        ✗
              float alpha_eff = has_alpha ? effective_mask_ptr[x] * opacity : opacity;
      
        ✗
              float ovr = ovr_ptr[x];
      
        ✗
              float src = src_ptr[x];
      
              if constexpr (mode == LIGHTEN) {
      
        ✗
                if (!(ovr > (src + thresh))) alpha_eff = 0;
      
              }
      
              else {// DARKEN
      
        ✗
                if (!(ovr < (src - thresh))) alpha_eff = 0;
      
              }
      
              if constexpr (!lumaonly)
      
              {
      
                // chroma u,v
      
        ✗
                dstp_u[x] = dstp_u[x] + (ovrp_u[x] - dstp_u[x]) * alpha_eff;
      
        ✗
                dstp_v[x] = dstp_v[x] + (ovrp_v[x] - dstp_v[x]) * alpha_eff;
      
                //dstp_a[x] = dstp_a[x] + (maskp[x] - dstp_a[x]) * alpha_eff;
      
              }
      
              // for 444: update here, width/height is the same as for chroma
      
              if constexpr (maskMode == MASK444)
      
        ✗
                dstp[x] = dstp[x] + (ovrp[x] - dstp[x]) * alpha_eff;
      
            }
      
            if constexpr (maskMode == MASK420 || maskMode == MASK420_MPEG2 || maskMode == MASK420_TOPLEFT) {
      
        ✗
              dstp += dst_pitch * 2; // skip vertical subsampling
      
        ✗
              ovrp += overlay_pitch * 2;
      
              if constexpr (has_alpha) {
      
                //dstp_a += dst_pitch * 2;
      
                maskp += mask_pitch * 2;
      
              }
      
            }
      
            else {
      
        ✗
              dstp += dst_pitch;
      
        ✗
              ovrp += overlay_pitch;
      
              if constexpr (has_alpha) {
      
                //dstp_a += dst_pitch;
      
                maskp += mask_pitch;
      
              }
      
            }
      
            if constexpr (!lumaonly) {
      
        ✗
              dstp_u += dst_pitchUV;
      
        ✗
              dstp_v += dst_pitchUV;
      
        ✗
              ovrp_u += overlay_pitchUV;
      
        ✗
              ovrp_v += overlay_pitchUV;
      
            }
      
          }
      
        ✗
          dst_pitch *= sizeof(float);
      
        ✗
          dst_pitchUV *= sizeof(float);
      
        ✗
          overlay_pitch *= sizeof(float);
      
        ✗
          overlay_pitchUV *= sizeof(float);
      
          if constexpr (has_alpha)
      
            mask_pitch *= sizeof(float);
      
          // make luma
      
          // recursively call the same function, but with lumaonly=true to update luma plane using the original chroma planes for decision
      
          if constexpr (!lumaonly && maskMode != MASK444)
      
        ✗
            layer_yuv_lighten_darken_f_c<mode, MASK444, true /* lumaonly*/, has_alpha>(
      
              dstp8, dstp8_u, dstp8_v, //dstp8_a,
      
              ovrp8, ovrp8_u, ovrp8_v, maskp8,
      
              dst_pitch, dst_pitchUV, overlay_pitch, overlay_pitchUV, mask_pitch,
      
              width, height, opacity, thresh);
      
        ✗
        }
      
        DISABLE_WARNING_POP
      
        // Dispatchers
      
        ✗
        static void get_layer_yuv_lighten_darken_functions(bool isLighten, int placement, VideoInfo& vi, int bits_per_pixel,
      
          /*out*/layer_yuv_lighten_darken_c_t** layer_fn,
      
          /*out*/layer_yuv_lighten_darken_f_c_t** layer_f_fn)
      
        {
      
        #define YUV_LIGHTEN_DARKEN_DISPATCH(L_or_D, MaskType, lumaonly, has_alpha) \
      
              { if (bits_per_pixel == 8) \
      
                *layer_fn = layer_yuv_lighten_darken_c<L_or_D, MaskType, uint8_t, lumaonly /*lumaonly*/, has_alpha /*has_alpha*/>; \
      
              else if (bits_per_pixel <= 16) \
      
                *layer_fn = layer_yuv_lighten_darken_c<L_or_D, MaskType, uint16_t, lumaonly /*lumaonly*/, has_alpha /*has_alpha*/>; \
      
              else /* float */ \
      
                *layer_f_fn = layer_yuv_lighten_darken_f_c<L_or_D, MaskType, lumaonly /*lumaonly*/, has_alpha /*has_alpha*/>; \
      
        }
      
          // Note, that we always call one instance of the filter: the lumaonly==false version,
      
          // since the filter handles U and V first, then recursively calls itself with lumaonly==true;
      
          // the chroma part needs unaltered luma for decision, so we delay Y processing.
      
          // Except Y-only greayscale input, where only luma plane exists.
      
        ✗
          if (isLighten) {
      
        ✗
            if (vi.IsYV411())
      
        ✗
              *layer_fn = layer_yuv_lighten_darken_c<LIGHTEN, MASK411, uint8_t, false /*lumaonly*/, false /*has_alpha*/>;
      
        ✗
            else if (vi.Is420())
      
            {
      
        ✗
              if (placement == PLACEMENT_MPEG1)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK420, false, false)
      
        ✗
              else if (placement == PLACEMENT_TOPLEFT)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK420_TOPLEFT, false, false)
      
              else
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK420_MPEG2, false, false)
      
                // PLACEMENT_MPEG2
      
            }
      
        ✗
            else if (vi.Is422())
      
            {
      
        ✗
              if (placement == PLACEMENT_MPEG1)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK422, false, false)
      
        ✗
              else if (placement == PLACEMENT_TOPLEFT)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK422_TOPLEFT, false, false)
      
              else
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK422_MPEG2, false, false)
      
                // PLACEMENT_MPEG2
      
            }
      
        ✗
            else if (vi.Is444())
      
        ✗
              YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK444, false, false)
      
        ✗
            else if (vi.IsY())
      
        ✗
              YUV_LIGHTEN_DARKEN_DISPATCH(LIGHTEN, MASK444, true, false)
      
          }
      
          else {
      
            // darken
      
        ✗
            if (vi.IsYV411())
      
        ✗
              *layer_fn = layer_yuv_lighten_darken_c<DARKEN, MASK411, uint8_t, false /*lumaonly*/, false /*has_alpha*/>;
      
        ✗
            else if (vi.Is420())
      
            {
      
        ✗
              if (placement == PLACEMENT_MPEG1)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK420, false, false)
      
        ✗
              else if (placement == PLACEMENT_TOPLEFT)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK420_TOPLEFT, false, false)
      
              else // PLACEMENT_MPEG2
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK420_MPEG2, false, false)
      
            }
      
        ✗
            else if (vi.Is422())
      
            {
      
        ✗
              if (placement == PLACEMENT_MPEG1)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK422, false, false)
      
        ✗
              else if (placement == PLACEMENT_TOPLEFT)
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK422_TOPLEFT, false, false)
      
              else // PLACEMENT_MPEG2
      
        ✗
                YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK422_MPEG2, false, false)
      
            }
      
        ✗
            else if (vi.Is444())
      
        ✗
              YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK444, false, false)
      
        ✗
            else if (vi.IsY())
      
        ✗
              YUV_LIGHTEN_DARKEN_DISPATCH(DARKEN, MASK444, true, false)
      
          }
      
        #undef YUV_LIGHTEN_DARKEN_DISPATCH
      
        ✗
        }
      
        3
        static void get_layer_yuv_mul_functions(
      
          bool is_chroma, bool hasAlpha,
      
          int placement, VideoInfo& vi, int bits_per_pixel,
      
          /*out*/layer_yuv_mul_c_t** layer_fn,
      
          /*out*/layer_yuv_mul_f_c_t** layer_f_fn)
      
        {
      
        #define YUV_MUL_DISPATCH(MaskType, is_chroma, has_alpha) \
      
          { if (bits_per_pixel == 8) \
      
            *layer_fn = layer_yuv_mul_c<MaskType, uint8_t, is_chroma, has_alpha>; \
      
          else if (bits_per_pixel < 16) \
      
            *layer_fn = layer_yuv_mul_c<MaskType, uint16_t, is_chroma, has_alpha>; \
      
          else if (bits_per_pixel == 16) \
      
            *layer_fn = layer_yuv_mul_c<MaskType, uint16_t, is_chroma, has_alpha>; \
      
          else /* float */ \
      
            *layer_f_fn = layer_yuv_mul_f_c<MaskType, is_chroma, has_alpha>; \
      
          }
      
          2/2✓ Branch 2 → 3 taken 2 times.
✓ Branch 2 → 121 taken 1 time.

        3
          if (is_chroma) // not luma channel
      
          {
      
          1/2✗ Branch 4 → 5 not taken.
✓ Branch 4 → 6 taken 2 times.

        2
            if (vi.IsYV411())
      
            {
      
              // never has Alpha
      
        ✗
              *layer_fn = layer_yuv_mul_c<MASK411, uint8_t, true, false>;
      
            }
      
          1/2✗ Branch 7 → 8 not taken.
✓ Branch 7 → 55 taken 2 times.

        2
            else if (vi.Is420())
      
            {
      
        ✗
              if (placement == PLACEMENT_MPEG1) {
      
        ✗
                if (hasAlpha) {
      
        ✗
                  YUV_MUL_DISPATCH(MASK420, true, true)
      
                }
      
                else {
      
        ✗
                  YUV_MUL_DISPATCH(MASK420, true, false)
      
                }
      
              }
      
        ✗
              else if (placement == PLACEMENT_TOPLEFT) {
      
        ✗
                if (hasAlpha) {
      
        ✗
                  YUV_MUL_DISPATCH(MASK420_TOPLEFT, true, true)
      
                }
      
                else {
      
        ✗
                  YUV_MUL_DISPATCH(MASK420_TOPLEFT, true, false)
      
                }
      
              }
      
              else {
      
        ✗
                if (hasAlpha) {
      
        ✗
                  YUV_MUL_DISPATCH(MASK420_MPEG2, true, true)
      
                }
      
                else {
      
        ✗
                  YUV_MUL_DISPATCH(MASK420_MPEG2, true, false)
      
                }
      
              }
      
            }
      
          1/2✗ Branch 56 → 57 not taken.
✓ Branch 56 → 104 taken 2 times.

        2
            else if (vi.Is422())
      
            {
      
        ✗
              if (placement == PLACEMENT_MPEG1) {
      
        ✗
                if (hasAlpha) {
      
        ✗
                  YUV_MUL_DISPATCH(MASK422, true, true)
      
                }
      
                else {
      
        ✗
                  YUV_MUL_DISPATCH(MASK422, true, false)
      
                }
      
              }
      
        ✗
              else if (placement == PLACEMENT_TOPLEFT) {
      
        ✗
                if (hasAlpha) {
      
        ✗
                  YUV_MUL_DISPATCH(MASK422_TOPLEFT, true, true)
      
                }
      
                else {
      
        ✗
                  YUV_MUL_DISPATCH(MASK422_TOPLEFT, true, false)
      
                }
      
              }
      
              else {
      
        ✗
                if (hasAlpha) {
      
        ✗
                  YUV_MUL_DISPATCH(MASK422_MPEG2, true, true)
      
                }
      
                else {
      
        ✗
                  YUV_MUL_DISPATCH(MASK422_MPEG2, true, false)
      
                }
      
              }
      
            }
      
          1/2✓ Branch 105 → 106 taken 2 times.
✗ Branch 105 → 136 not taken.

        2
            else if (vi.Is444())
      
            {
      
          1/2✗ Branch 106 → 107 not taken.
✓ Branch 106 → 114 taken 2 times.

        2
              if (hasAlpha) {
      
        ✗
                YUV_MUL_DISPATCH(MASK444, true, true)
      
              }
      
              else {
      
          1/6✓ Branch 114 → 115 taken 2 times.
✗ Branch 114 → 116 not taken.
✗ Branch 116 → 117 not taken.
✗ Branch 116 → 118 not taken.
✗ Branch 118 → 119 not taken.
✗ Branch 118 → 120 not taken.

        2
                YUV_MUL_DISPATCH(MASK444, true, false)
      
              }
      
            }
      
          }
      
          else // luma channel
      
          {
      
          1/2✗ Branch 121 → 122 not taken.
✓ Branch 121 → 129 taken 1 time.

        1
            if (hasAlpha)
      
        ✗
              YUV_MUL_DISPATCH(MASK444, false, true)
      
            else
      
          1/6✓ Branch 129 → 130 taken 1 time.
✗ Branch 129 → 131 not taken.
✗ Branch 131 → 132 not taken.
✗ Branch 131 → 133 not taken.
✗ Branch 133 → 134 not taken.
✗ Branch 133 → 135 not taken.

        1
              YUV_MUL_DISPATCH(MASK444, false, false)
      
          }
      
        #undef YUV_MUL_DISPATCH
      
        3
        }
      
        // ---------------------------------------------------------------------------
      
        // mulovr: Overlay-style multiply.
      
        // Overlay luma (Y) drives the effect on all planes:
      
        //   darken_factor = alpha_eff * (max - ovr_Y) / max
      
        //   result_Y  = base_Y  * (max - darken_factor) / max
      
        //   result_UV = base_UV * (max - darken_factor) / max  +  half_pix * darken_factor / max
      
        // For float (neutral UV = 0.0): result = base * (1 - alpha_eff * (1 - ovr_Y)) for all planes.
      
        // Two-pass for subsampled formats (mirrors lighten/darken):
      
        //   Pass 1 (lumaonly=false): UV update using spatially-averaged overlay Y; MASK444 also does Y.
      
        //   Pass 2 (lumaonly=true,  MASK444): Y-only update at full luma resolution.
      
        // greyscale (vi.IsY()): dispatched directly as lumaonly=true — identical to Layer "Mul" luma.
      
        // ---------------------------------------------------------------------------
      
        template<MaskMode maskMode, typename pixel_t, bool lumaonly, bool has_alpha>
      
        2
        static void layer_yuv_mulovr_c(
      
          BYTE* dstp8, BYTE* dstp8_u, BYTE* dstp8_v,
      
          const BYTE* ovrp8,
      
          const BYTE* maskp8,
      
          int dst_pitch, int dst_pitchUV,
      
          int overlay_pitch,
      
          int mask_pitch,
      
          int width, int height, int opacity_i, int bits_per_pixel)
      
        {
      
        2
          pixel_t*       dstp   = reinterpret_cast<pixel_t*>(dstp8);
      
        2
          pixel_t*       dstp_u = reinterpret_cast<pixel_t*>(dstp8_u); // nullptr when lumaonly
      
        2
          pixel_t*       dstp_v = reinterpret_cast<pixel_t*>(dstp8_v); // nullptr when lumaonly
      
        2
          const pixel_t* ovrp   = reinterpret_cast<const pixel_t*>(ovrp8);
      
        2
          const pixel_t* maskp  = has_alpha ? reinterpret_cast<const pixel_t*>(maskp8) : nullptr;
      
        2
          dst_pitch     /= sizeof(pixel_t);
      
        2
          dst_pitchUV   /= sizeof(pixel_t);
      
        2
          overlay_pitch /= sizeof(pixel_t);
      
          if constexpr (has_alpha)
      
        1
            mask_pitch /= sizeof(pixel_t);
      
          if constexpr (sizeof(pixel_t) == 1)
      
        2
            bits_per_pixel = 8;
      
        2
          const MagicDiv magic          = get_magic_div(bits_per_pixel);
      
        2
          const int      max_pixel_value = (1 << bits_per_pixel) - 1;
      
        2
          const int      half            = max_pixel_value / 2; // rounding bias for magic_div
      
        2
          const uint32_t half_pix        = (uint32_t)(max_pixel_value / 2); // integer chroma neutral
      
        2
          const int cwidth  = (maskMode == MASK444) ? width  : (maskMode == MASK411) ? width >> 2 : width >> 1;
      
        2
          const int cheight = (maskMode == MASK444 || maskMode == MASK422 || maskMode == MASK422_MPEG2 || maskMode == MASK422_TOPLEFT || maskMode == MASK411) ? height : height >> 1;
      
        2
          std::vector<pixel_t> ovr_y_buffer;
      
        2
          std::vector<pixel_t> mask_buffer;
      
          if constexpr (maskMode != MASK444) {
      
        ✗
            ovr_y_buffer.resize(cwidth);
      
            if constexpr (has_alpha)
      
        ✗
              mask_buffer.resize(cwidth);
      
          }
      
          4/136void layer_yuv_mulovr_c<(MaskMode)0, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)0, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 7 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 26 → 7 not taken.
✗ Branch 26 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 8 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 38 → 8 not taken.
✗ Branch 38 → 39 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 19 → 6 not taken.
✗ Branch 19 → 20 not taken.
✓ Branch 20 → 6 taken 2 times.
✓ Branch 20 → 21 taken 1 time.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 6 not taken.
✗ Branch 22 → 23 not taken.
✓ Branch 24 → 6 taken 2 times.
✓ Branch 24 → 25 taken 1 time.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 15 → 6 not taken.
✗ Branch 15 → 16 not taken.
✗ Branch 16 → 6 not taken.
✗ Branch 16 → 17 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 6 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 20 → 6 not taken.
✗ Branch 20 → 21 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 19 → 6 not taken.
✗ Branch 19 → 20 not taken.
✗ Branch 20 → 6 not taken.
✗ Branch 20 → 21 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 22 → 6 not taken.
✗ Branch 22 → 23 not taken.
✗ Branch 24 → 6 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 15 → 6 not taken.
✗ Branch 15 → 16 not taken.
✗ Branch 16 → 6 not taken.
✗ Branch 16 → 17 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 18 → 6 not taken.
✗ Branch 18 → 19 not taken.
✗ Branch 20 → 6 not taken.
✗ Branch 20 → 21 not taken.

        6
          for (int y = 0; y < cheight; ++y) {
      
            // Spatially average overlay Y to chroma resolution (full_opacity=true: no opacity baking).
      
          0/68void layer_yuv_mulovr_c<(MaskMode)0, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)0, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 8 → 9 not taken.
✗ Branch 8 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 26 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 19 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 22 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 26 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 19 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 6 → 7 not taken.
✗ Branch 6 → 22 not taken.

        4
            const pixel_t* ovr_y_ptr = LAYER_ROWPREP_FN<maskMode, pixel_t, true>(ovrp, overlay_pitch, cwidth, ovr_y_buffer);
      
        4
            const pixel_t* mask_ptr  = nullptr;
      
            if constexpr (has_alpha)
      
          0/32void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 9 → 10 not taken.
✗ Branch 9 → 27 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 26 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 22 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 26 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 7 → 8 not taken.
✗ Branch 7 → 22 not taken.

        2
              mask_ptr = LAYER_ROWPREP_FN<maskMode, pixel_t, true>(maskp, mask_pitch, cwidth, mask_buffer);
      
          4/136void layer_yuv_mulovr_c<(MaskMode)0, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)0, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)1, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)2, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)3, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)4, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)5, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 24 → 17 not taken.
✗ Branch 24 → 25 not taken.
void layer_yuv_mulovr_c<(MaskMode)6, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 11 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 36 → 27 not taken.
✗ Branch 36 → 37 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 17 → 8 not taken.
✗ Branch 17 → 18 not taken.
✓ Branch 18 → 9 taken 10 times.
✓ Branch 18 → 19 taken 2 times.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 9 not taken.
✗ Branch 20 → 21 not taken.
✓ Branch 22 → 11 taken 10 times.
✓ Branch 22 → 23 taken 2 times.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 13 → 8 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 14 → 9 not taken.
✗ Branch 14 → 15 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned char, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 18 → 11 not taken.
✗ Branch 18 → 19 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 17 → 8 not taken.
✗ Branch 17 → 18 not taken.
✗ Branch 18 → 9 not taken.
✗ Branch 18 → 19 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, false, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 20 → 9 not taken.
✗ Branch 20 → 21 not taken.
✗ Branch 22 → 11 not taken.
✗ Branch 22 → 23 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, false>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 13 → 8 not taken.
✗ Branch 13 → 14 not taken.
✗ Branch 14 → 9 not taken.
✗ Branch 14 → 15 not taken.
void layer_yuv_mulovr_c<(MaskMode)7, unsigned short, true, true>(unsigned char*, unsigned char*, unsigned char*, unsigned char const*, unsigned char const*, int, int, int, int, int, int, int, int):
✗ Branch 16 → 9 not taken.
✗ Branch 16 → 17 not taken.
✗ Branch 18 → 11 not taken.
✗ Branch 18 → 19 not taken.

        24
            for (int x = 0; x < cwidth; ++x) {
      
        20
              const uint32_t alpha_eff = has_alpha
      
        10
                ? magic_div_rt<pixel_t>((uint32_t)mask_ptr[x] * opacity_i + half, magic)
      
                : (uint32_t)opacity_i;
      
              // dark overlay Y → large darken_factor → strong pull toward neutral/black.
      
        20
              const uint32_t darken_factor = magic_div_rt<pixel_t>(alpha_eff * ((uint32_t)max_pixel_value - (uint32_t)ovr_y_ptr[x]) + half, magic);
      
        20
              const uint32_t inv_keep      = (uint32_t)max_pixel_value - darken_factor;
      
              if constexpr (!lumaonly) {
      
                // UV: pull toward half_pix (chroma neutral). Overflow-safe in uint32 (see plan).
      
        20
                dstp_u[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_u[x] * inv_keep + half_pix * darken_factor + half, magic);
      
        20
                dstp_v[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_v[x] * inv_keep + half_pix * darken_factor + half, magic);
      
              }
      
              // For MASK444, luma and chroma share resolution — update Y here.
      
              if constexpr (maskMode == MASK444)
      
        40
                dstp[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp[x] * inv_keep + half, magic);
      
            }
      
            if constexpr (maskMode == MASK420 || maskMode == MASK420_MPEG2 || maskMode == MASK420_TOPLEFT) {
      
        ✗
              dstp  += dst_pitch * 2;
      
        ✗
              ovrp  += overlay_pitch * 2;
      
              if constexpr (has_alpha)
      
        ✗
                maskp += mask_pitch * 2;
      
            }
      
            else {
      
        4
              dstp  += dst_pitch;
      
        4
              ovrp  += overlay_pitch;
      
              if constexpr (has_alpha)
      
        2
                maskp += mask_pitch;
      
            }
      
            if constexpr (!lumaonly) {
      
        4
              dstp_u += dst_pitchUV;
      
        4
              dstp_v += dst_pitchUV;
      
            }
      
          }
      
        2
          dst_pitch     *= sizeof(pixel_t);
      
        2
          dst_pitchUV   *= sizeof(pixel_t);
      
        2
          overlay_pitch *= sizeof(pixel_t);
      
          if constexpr (has_alpha)
      
        1
            mask_pitch *= sizeof(pixel_t);
      
          // Phase 2: luma-only pass at full luma resolution.
      
          if constexpr (!lumaonly && maskMode != MASK444)
      
        ✗
            layer_yuv_mulovr_c<MASK444, pixel_t, true, has_alpha>(
      
              dstp8, nullptr, nullptr,
      
              ovrp8, maskp8,
      
              dst_pitch, 0, overlay_pitch, mask_pitch,
      
              width, height, opacity_i, bits_per_pixel);
      
        2
        }
      
        // Float version — neutral UV = 0.0f in AviSynth+ float YUV, so both luma and chroma
      
        // collapse to the same formula: result = base * (1 - alpha_eff * (1 - ovr_Y)).
      
        template<MaskMode maskMode, bool lumaonly, bool has_alpha>
      
        ✗
        static void layer_yuv_mulovr_f_c(
      
          BYTE* dstp8, BYTE* dstp8_u, BYTE* dstp8_v,
      
          const BYTE* ovrp8,
      
          const BYTE* maskp8,
      
          int dst_pitch, int dst_pitchUV,
      
          int overlay_pitch,
      
          int mask_pitch,
      
          int width, int height, float opacity)
      
        {
      
        ✗
          float*       dstp   = reinterpret_cast<float*>(dstp8);
      
        ✗
          float*       dstp_u = reinterpret_cast<float*>(dstp8_u); // nullptr when lumaonly
      
        ✗
          float*       dstp_v = reinterpret_cast<float*>(dstp8_v); // nullptr when lumaonly
      
        ✗
          const float* ovrp   = reinterpret_cast<const float*>(ovrp8);
      
        ✗
          const float* maskp  = has_alpha ? reinterpret_cast<const float*>(maskp8) : nullptr;
      
        ✗
          dst_pitch     /= sizeof(float);
      
        ✗
          dst_pitchUV   /= sizeof(float);
      
        ✗
          overlay_pitch /= sizeof(float);
      
          if constexpr (has_alpha)
      
        ✗
            mask_pitch /= sizeof(float);
      
        ✗
          const int cwidth  = (maskMode == MASK444) ? width  : (maskMode == MASK411) ? width >> 2 : width >> 1;
      
        ✗
          const int cheight = (maskMode == MASK444 || maskMode == MASK422 || maskMode == MASK422_MPEG2 || maskMode == MASK422_TOPLEFT || maskMode == MASK411) ? height : height >> 1;
      
        ✗
          std::vector<float> ovr_y_buffer;
      
        ✗
          std::vector<float> mask_buffer;
      
          if constexpr (maskMode != MASK444) {
      
        ✗
            ovr_y_buffer.resize(cwidth);
      
            if constexpr (has_alpha)
      
        ✗
              mask_buffer.resize(cwidth);
      
          }
      
        ✗
          for (int y = 0; y < cheight; ++y) {
      
        ✗
            const float* ovr_y_ptr = prepare_effective_mask_for_row_float_c<maskMode>(ovrp, overlay_pitch, cwidth, ovr_y_buffer);
      
        ✗
            const float* mask_ptr  = nullptr;
      
            if constexpr (has_alpha)
      
        ✗
              mask_ptr = prepare_effective_mask_for_row_float_c<maskMode>(maskp, mask_pitch, cwidth, mask_buffer);
      
        ✗
            for (int x = 0; x < cwidth; ++x) {
      
        ✗
              const float alpha_eff = has_alpha ? mask_ptr[x] * opacity : opacity;
      
        ✗
              const float inv_keep  = 1.0f - alpha_eff * (1.0f - ovr_y_ptr[x]);
      
              if constexpr (!lumaonly) {
      
        ✗
                dstp_u[x] *= inv_keep; // neutral UV = 0.0f → same formula as luma
      
        ✗
                dstp_v[x] *= inv_keep;
      
              }
      
              if constexpr (maskMode == MASK444)
      
        ✗
                dstp[x] *= inv_keep;
      
            }
      
            if constexpr (maskMode == MASK420 || maskMode == MASK420_MPEG2 || maskMode == MASK420_TOPLEFT) {
      
        ✗
              dstp  += dst_pitch * 2;
      
        ✗
              ovrp  += overlay_pitch * 2;
      
              if constexpr (has_alpha)
      
        ✗
                maskp += mask_pitch * 2;
      
            }
      
            else {
      
        ✗
              dstp  += dst_pitch;
      
        ✗
              ovrp  += overlay_pitch;
      
              if constexpr (has_alpha)
      
        ✗
                maskp += mask_pitch;
      
            }
      
            if constexpr (!lumaonly) {
      
        ✗
              dstp_u += dst_pitchUV;
      
        ✗
              dstp_v += dst_pitchUV;
      
            }
      
          }
      
        ✗
          dst_pitch     *= sizeof(float);
      
        ✗
          dst_pitchUV   *= sizeof(float);
      
        ✗
          overlay_pitch *= sizeof(float);
      
          if constexpr (has_alpha)
      
        ✗
            mask_pitch *= sizeof(float);
      
          if constexpr (!lumaonly && maskMode != MASK444)
      
        ✗
            layer_yuv_mulovr_f_c<MASK444, true, has_alpha>(
      
              dstp8, nullptr, nullptr,
      
              ovrp8, maskp8,
      
              dst_pitch, 0, overlay_pitch, mask_pitch,
      
              width, height, opacity);
      
        ✗
        }
      
        2
        static void get_layer_yuv_mulovr_functions(
      
          bool has_alpha, int placement, VideoInfo& vi, int bits_per_pixel,
      
          layer_yuv_mulovr_c_t**   layer_fn,
      
          layer_yuv_mulovr_f_c_t** layer_f_fn)
      
        {
      
        #define MULOVR_DISPATCH(MaskType, lumaonly, ha) \
      
          { if (bits_per_pixel == 8) \
      
            *layer_fn   = layer_yuv_mulovr_c  <MaskType, uint8_t,  lumaonly, ha>; \
      
          else if (bits_per_pixel <= 16) \
      
            *layer_fn   = layer_yuv_mulovr_c  <MaskType, uint16_t, lumaonly, ha>; \
      
          else \
      
            *layer_f_fn = layer_yuv_mulovr_f_c<MaskType,           lumaonly, ha>; \
      
          }
      
        #define MULOVR_HA(MaskType, lumaonly) \
      
          { if (has_alpha) { MULOVR_DISPATCH(MaskType, lumaonly, true) } \
      
            else           { MULOVR_DISPATCH(MaskType, lumaonly, false) } }
      
          1/2✗ Branch 3 → 4 not taken.
✓ Branch 3 → 15 taken 2 times.

        2
          if (vi.IsY()) {
      
            // Greyscale: no UV planes — luma-only pass, identical to Layer "Mul" luma.
      
        ✗
            MULOVR_HA(MASK444, true)
      
          }
      
          1/2✗ Branch 16 → 17 not taken.
✓ Branch 16 → 22 taken 2 times.

        2
          else if (vi.IsYV411()) {
      
        ✗
            MULOVR_DISPATCH(MASK411, false, false) // YV411 never has alpha
      
          }
      
          1/2✗ Branch 23 → 24 not taken.
✓ Branch 23 → 59 taken 2 times.

        2
          else if (vi.Is420()) {
      
        ✗
            if      (placement == PLACEMENT_MPEG1)    { MULOVR_HA(MASK420,         false) }
      
        ✗
            else if (placement == PLACEMENT_TOPLEFT)  { MULOVR_HA(MASK420_TOPLEFT, false) }
      
        ✗
            else                                      { MULOVR_HA(MASK420_MPEG2,   false) }
      
          }
      
          1/2✗ Branch 60 → 61 not taken.
✓ Branch 60 → 96 taken 2 times.

        2
          else if (vi.Is422()) {
      
        ✗
            if      (placement == PLACEMENT_MPEG1)    { MULOVR_HA(MASK422,         false) }
      
        ✗
            else if (placement == PLACEMENT_TOPLEFT)  { MULOVR_HA(MASK422_TOPLEFT, false) }
      
        ✗
            else                                      { MULOVR_HA(MASK422_MPEG2,   false) }
      
          }
      
          else { // Is444() — cwidth == width, cheight == height; same-res, single pass
      
          4/10✓ Branch 96 → 97 taken 1 time.
✓ Branch 96 → 102 taken 1 time.
✓ Branch 97 → 98 taken 1 time.
✗ Branch 97 → 99 not taken.
✗ Branch 99 → 100 not taken.
✗ Branch 99 → 101 not taken.
✓ Branch 102 → 103 taken 1 time.
✗ Branch 102 → 104 not taken.
✗ Branch 104 → 105 not taken.
✗ Branch 104 → 106 not taken.

        2
            MULOVR_HA(MASK444, false)
      
          }
      
        #undef MULOVR_HA
      
        #undef MULOVR_DISPATCH
      
        2
        }
      
        // both masked merge (HasAlpha) and plain add (no alpha)
      
        // Though we use it only for masked merge, kept for reference.
      
        ✗
        static void get_layer_yuv_add_masked_functions(
      
          bool is_chroma, bool hasAlpha,
      
          int placement, VideoInfo& vi, int bits_per_pixel,
      
          /*out*/masked_merge_fn_t** layer_fn,
      
          /*out*/masked_merge_float_fn_t** layer_f_fn)
      
        {
      
          // Use the unified (Layer,Overlay) masked merge functions
      
          // Determine MaskMode from format and placement
      
        ✗
          MaskMode maskMode = MASK444;
      
        ✗
          if (is_chroma) {
      
        ✗
            if (vi.IsYV411())
      
        ✗
              maskMode = MASK411;
      
        ✗
            else if (vi.Is420())
      
        ✗
              maskMode = (placement == PLACEMENT_MPEG1) ? MASK420 : (placement == PLACEMENT_TOPLEFT) ? MASK420_TOPLEFT : MASK420_MPEG2;
      
        ✗
            else if (vi.Is422())
      
        ✗
              maskMode = (placement == PLACEMENT_MPEG1) ? MASK422 : (placement == PLACEMENT_TOPLEFT) ? MASK422_TOPLEFT : MASK422_MPEG2;
      
            // Is444() / IsY(): stay MASK444
      
          }
      
          // is_chroma=false (luma): always MASK444
      
        ✗
          *layer_fn = get_overlay_blend_masked_fn_c(is_chroma, maskMode);
      
        ✗
          *layer_f_fn = get_overlay_blend_masked_float_fn_c(is_chroma, maskMode);
      
        ✗
        }
      
        /* planar rgb */
      
        template<int mode, typename pixel_t, bool has_alpha, bool blend_alpha>
      
        ✗
        static void layer_planarrgb_lighten_darken_c(BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, int opacity_i, int thresh, int bits_per_pixel) {
      
        ✗
          pixel_t* dstp_g = reinterpret_cast<pixel_t*>(dstp8[0]);
      
        ✗
          pixel_t* dstp_b = reinterpret_cast<pixel_t*>(dstp8[1]);
      
        ✗
          pixel_t* dstp_r = reinterpret_cast<pixel_t*>(dstp8[2]);
      
          // dstp8[3]: written only when blend_alpha=true (both clips have alpha).
      
          pixel_t* dstp_a;
      
          if constexpr (blend_alpha)
      
        ✗
            dstp_a = reinterpret_cast<pixel_t*>(dstp8[3]);
      
        ✗
          const pixel_t* ovrp_g = reinterpret_cast<const pixel_t*>(ovrp8[0]);
      
        ✗
          const pixel_t* ovrp_b = reinterpret_cast<const pixel_t*>(ovrp8[1]);
      
        ✗
          const pixel_t* ovrp_r = reinterpret_cast<const pixel_t*>(ovrp8[2]);
      
          // maskp: per-pixel blend weight — decoupled from ovrp8[3] so a future mask clip can be wired in.
      
        ✗
          const pixel_t* maskp = reinterpret_cast<const pixel_t*>(maskp8);
      
          // alpha_target: value blended into dstp_a (only used when blend_alpha=true).
      
          // For plain Add this is the same plane as maskp; for Subtract it is the inverted A plane.
      
          const pixel_t* alpha_target;
      
          if constexpr (blend_alpha)
      
        ✗
            alpha_target = reinterpret_cast<const pixel_t*>(ovrp8[3]);
      
        ✗
          dst_pitch /= sizeof(pixel_t);
      
        ✗
          overlay_pitch /= sizeof(pixel_t);
      
        ✗
          mask_pitch /= sizeof(pixel_t);
      
          if constexpr (sizeof(pixel_t) == 1)
      
        ✗
            bits_per_pixel = 8; // make quasi constexpr
      
        ✗
          const MagicDiv magic = get_magic_div(bits_per_pixel);
      
        ✗
          const int max_pixel_value = (1 << bits_per_pixel) - 1;
      
        ✗
          const int half = max_pixel_value / 2;
      
        ✗
          for (int y = 0; y < height; ++y) {
      
        ✗
            for (int x = 0; x < width; ++x) {
      
              // opacity_i is in [0..max_pixel_value], the same convention as masked_merge.
      
              uint32_t alpha_eff;
      
              if constexpr (has_alpha) {
      
        ✗
                const uint32_t alpha_src = (uint32_t)maskp[x];
      
        ✗
                alpha_eff = (uint32_t)magic_div_rt<pixel_t>(alpha_src * (uint32_t)opacity_i + half, magic);
      
              } else {
      
        ✗
                alpha_eff = opacity_i;
      
              }
      
              // standard rgb->luma coefficients, scaled up by 15 bits
      
              // no rounding, not really needed here
      
        ✗
              const int luma_ovr = (cyb * ovrp_b[x] + cyg * ovrp_g[x] + cyr * ovrp_r[x]) >> 15;
      
        ✗
              const int luma_src = (cyb * dstp_b[x] + cyg * dstp_g[x] + cyr * dstp_r[x]) >> 15;
      
              if constexpr (mode == LIGHTEN)
      
        ✗
                alpha_eff = luma_ovr > luma_src + thresh ? alpha_eff : 0;
      
              else // DARKEN
      
        ✗
                alpha_eff = luma_ovr < luma_src - thresh ? alpha_eff : 0;
      
        ✗
              const uint32_t inv_alpha = max_pixel_value - alpha_eff;
      
              // move to magicdiv
      
              //dstp_r[x] = (pixel_t)(dstp_r[x] + (((ovrp_r[x] - dstp_r[x]) * alpha + rounder) >> bits_per_pixel));
      
              //dstp_g[x] = (pixel_t)(dstp_g[x] + (((ovrp_g[x] - dstp_g[x]) * alpha + rounder) >> bits_per_pixel));
      
              //dstp_b[x] = (pixel_t)(dstp_b[x] + (((ovrp_b[x] - dstp_b[x]) * alpha + rounder) >> bits_per_pixel));
      
        ✗
              dstp_r[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_r[x] * inv_alpha + (uint32_t)ovrp_r[x] * alpha_eff + half, magic);
      
        ✗
              dstp_g[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_g[x] * inv_alpha + (uint32_t)ovrp_g[x] * alpha_eff + half, magic);
      
        ✗
              dstp_b[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_b[x] * inv_alpha + (uint32_t)ovrp_b[x] * alpha_eff + half, magic);
      
              if constexpr (blend_alpha)
      
        ✗
                dstp_a[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_a[x] * inv_alpha + (uint32_t)alpha_target[x] * alpha_eff + half, magic);
      
            }
      
        ✗
            dstp_g += dst_pitch;
      
        ✗
            dstp_b += dst_pitch;
      
        ✗
            dstp_r += dst_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              dstp_a += dst_pitch;
      
        ✗
            ovrp_g += overlay_pitch;
      
        ✗
            ovrp_b += overlay_pitch;
      
        ✗
            ovrp_r += overlay_pitch;
      
            if constexpr (has_alpha)
      
        ✗
              maskp += mask_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              alpha_target += overlay_pitch;
      
          }
      
        ✗
        }
      
        template<int mode, bool has_alpha, bool blend_alpha>
      
        ✗
        static void layer_planarrgb_lighten_darken_f_c(BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, float opacity, float thresh) {
      
        ✗
          float* dstp_g = reinterpret_cast<float*>(dstp8[0]);
      
        ✗
          float* dstp_b = reinterpret_cast<float*>(dstp8[1]);
      
        ✗
          float* dstp_r = reinterpret_cast<float*>(dstp8[2]);
      
          // dstp8[3]: written only when blend_alpha=true (both clips have alpha).
      
          float* dstp_a;
      
          if constexpr (blend_alpha)
      
        ✗
            dstp_a = reinterpret_cast<float*>(dstp8[3]);
      
        ✗
          const float* ovrp_g = reinterpret_cast<const float*>(ovrp8[0]);
      
        ✗
          const float* ovrp_b = reinterpret_cast<const float*>(ovrp8[1]);
      
        ✗
          const float* ovrp_r = reinterpret_cast<const float*>(ovrp8[2]);
      
          // maskp: per-pixel blend weight — decoupled from ovrp8[3].
      
        ✗
          const float* maskp = reinterpret_cast<const float*>(maskp8);
      
          const float* alpha_target;
      
          if constexpr (blend_alpha)
      
        ✗
            alpha_target = reinterpret_cast<const float*>(ovrp8[3]);
      
        ✗
          dst_pitch /= sizeof(float);
      
        ✗
          overlay_pitch /= sizeof(float);
      
        ✗
          mask_pitch /= sizeof(float);
      
        ✗
          for (int y = 0; y < height; ++y) {
      
        ✗
            for (int x = 0; x < width; ++x) {
      
        ✗
              float alpha = has_alpha ? maskp[x] * opacity : opacity;
      
        ✗
              float luma_ovr = cyb_f * ovrp_b[x] + cyg_f * ovrp_g[x] + cyr_f * ovrp_r[x];
      
        ✗
              float luma_src = cyb_f * dstp_b[x] + cyg_f * dstp_g[x] + cyr_f * dstp_r[x];
      
              if constexpr (mode == LIGHTEN)
      
        ✗
                alpha = luma_ovr > luma_src + thresh ? alpha : 0;
      
              else // DARKEN
      
        ✗
                alpha = luma_ovr < luma_src - thresh ? alpha : 0;
      
        ✗
              dstp_r[x] = dstp_r[x] + (ovrp_r[x] - dstp_r[x]) * alpha;
      
        ✗
              dstp_g[x] = dstp_g[x] + (ovrp_g[x] - dstp_g[x]) * alpha;
      
        ✗
              dstp_b[x] = dstp_b[x] + (ovrp_b[x] - dstp_b[x]) * alpha;
      
              if constexpr (blend_alpha)
      
        ✗
                dstp_a[x] = dstp_a[x] + (alpha_target[x] - dstp_a[x]) * alpha;
      
            }
      
        ✗
            dstp_g += dst_pitch;
      
        ✗
            dstp_b += dst_pitch;
      
        ✗
            dstp_r += dst_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              dstp_a += dst_pitch;
      
        ✗
            ovrp_g += overlay_pitch;
      
        ✗
            ovrp_b += overlay_pitch;
      
        ✗
            ovrp_r += overlay_pitch;
      
            if constexpr (has_alpha)
      
        ✗
              maskp += mask_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              alpha_target += overlay_pitch;
      
          }
      
        ✗
        }
      
        ✗
        static void get_layer_planarrgb_lighten_darken_functions(bool isLighten, bool hasAlpha, bool blendAlpha, int bits_per_pixel, /*out*/layer_planarrgb_lighten_darken_c_t** layer_fn, /*out*/layer_planarrgb_lighten_darken_f_c_t** layer_f_fn) {
      
        #define PLANARRGB_LD_DISPATCH(LorD, has_alpha, blend_alpha) \
      
              { if (bits_per_pixel == 8) \
      
                *layer_fn = layer_planarrgb_lighten_darken_c<LorD, uint8_t, has_alpha, blend_alpha>; \
      
              else if (bits_per_pixel <= 16) \
      
                *layer_fn = layer_planarrgb_lighten_darken_c<LorD, uint16_t, has_alpha, blend_alpha>; \
      
              else /* float */ \
      
                *layer_f_fn = layer_planarrgb_lighten_darken_f_c<LorD, has_alpha, blend_alpha>; \
      
              }
      
        ✗
          if (isLighten) {
      
        ✗
            if (hasAlpha) {
      
        ✗
              if (blendAlpha) { PLANARRGB_LD_DISPATCH(LIGHTEN, true, true)  }
      
        ✗
              else            { PLANARRGB_LD_DISPATCH(LIGHTEN, true, false) }
      
            }
      
            else {
      
        ✗
              PLANARRGB_LD_DISPATCH(LIGHTEN, false, false)
      
            }
      
          } // lighten end
      
          else {
      
        ✗
            if (hasAlpha) {
      
        ✗
              if (blendAlpha) { PLANARRGB_LD_DISPATCH(DARKEN, true, true)  }
      
        ✗
              else            { PLANARRGB_LD_DISPATCH(DARKEN, true, false) }
      
            }
      
            else {
      
        ✗
              PLANARRGB_LD_DISPATCH(DARKEN, false, false)
      
            }
      
          }
      
        #undef PLANARRGB_LD_DISPATCH
      
        ✗
        }
      
        // subtract=false only: overlay clip is pre-inverted by Layer::Create when op="Subtract".
      
        template<typename pixel_t,bool chroma, bool has_alpha, bool blend_alpha>
      
        ✗
        static void layer_planarrgb_add_c(BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, int opacity_i, int bits_per_pixel) {
      
        ✗
          pixel_t* dstp_g = reinterpret_cast<pixel_t*>(dstp8[0]);
      
        ✗
          pixel_t* dstp_b = reinterpret_cast<pixel_t*>(dstp8[1]);
      
        ✗
          pixel_t* dstp_r = reinterpret_cast<pixel_t*>(dstp8[2]);
      
          // dstp8[3]: written only when blend_alpha=true (both clips have alpha).
      
          pixel_t* dstp_a;
      
          if constexpr (blend_alpha)
      
        ✗
            dstp_a = reinterpret_cast<pixel_t*>(dstp8[3]);
      
        ✗
          const pixel_t* ovrp_g = reinterpret_cast<const pixel_t*>(ovrp8[0]);
      
        ✗
          const pixel_t* ovrp_b = reinterpret_cast<const pixel_t*>(ovrp8[1]);
      
        ✗
          const pixel_t* ovrp_r = reinterpret_cast<const pixel_t*>(ovrp8[2]);
      
          // maskp: per-pixel blend weight — decoupled from ovrp8[3] to support future mask clip param.
      
        ✗
          const pixel_t* maskp = reinterpret_cast<const pixel_t*>(maskp8);
      
          // alpha_target: the value blended into dstp_a.
      
          // For plain Add: same as maskp (original overlay A).
      
          // For Subtract: ovrp8[3] is the inverted A; maskp is the saved pre-invert A.
      
          const pixel_t* alpha_target;
      
          if constexpr (blend_alpha)
      
        ✗
            alpha_target = reinterpret_cast<const pixel_t*>(ovrp8[3]);
      
        ✗
          dst_pitch /= sizeof(pixel_t);
      
        ✗
          overlay_pitch /= sizeof(pixel_t);
      
        ✗
          mask_pitch /= sizeof(pixel_t);
      
          if constexpr (sizeof(pixel_t) == 1)
      
        ✗
            bits_per_pixel = 8; // make quasi constexpr
      
          // use magic-div
      
        ✗
          const MagicDiv magic = get_magic_div(bits_per_pixel);
      
        ✗
          const int max_pixel_value = (1 << bits_per_pixel) - 1;
      
        ✗
          const int half = max_pixel_value / 2;
      
        ✗
          for (int y = 0; y < height; ++y) {
      
        ✗
            for (int x = 0; x < width; ++x) {
      
              uint32_t alpha_eff;
      
              if constexpr (has_alpha) {
      
        ✗
                const uint32_t alpha_src = (uint32_t)maskp[x];
      
        ✗
                alpha_eff = (uint32_t)magic_div_rt<pixel_t>(alpha_src * (uint32_t)opacity_i + half, magic);
      
              }
      
              else {
      
        ✗
                alpha_eff = opacity_i;
      
              }
      
        ✗
              const uint32_t inv_alpha = max_pixel_value - alpha_eff;
      
              if constexpr (chroma) {
      
                //dstp_r[x] = (pixel_t)(dstp_r[x] + (((ovrp_r[x] - dstp_r[x]) * alpha + rounder) >> bits_per_pixel));
      
                //dstp_g[x] = (pixel_t)(dstp_g[x] + (((ovrp_g[x] - dstp_g[x]) * alpha + rounder) >> bits_per_pixel));
      
                //dstp_b[x] = (pixel_t)(dstp_b[x] + (((ovrp_b[x] - dstp_b[x]) * alpha + rounder) >> bits_per_pixel));
      
        ✗
                dstp_r[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_r[x] * inv_alpha + (uint32_t)ovrp_r[x] * alpha_eff + half, magic);
      
        ✗
                dstp_g[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_g[x] * inv_alpha + (uint32_t)ovrp_g[x] * alpha_eff + half, magic);
      
        ✗
                dstp_b[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_b[x] * inv_alpha + (uint32_t)ovrp_b[x] * alpha_eff + half, magic);
      
              }
      
              else { // use luma instead of overlay
      
        ✗
                const int luma = (cyb * ovrp_b[x] + cyg * ovrp_g[x] + cyr * ovrp_r[x]) >> 15; // no rounding not really needed here
      
                //dstp_r[x] = (pixel_t)(dstp_r[x] + (((luma - dstp_r[x]) * alpha + rounder) >> bits_per_pixel));
      
                //dstp_g[x] = (pixel_t)(dstp_g[x] + (((luma - dstp_g[x]) * alpha + rounder) >> bits_per_pixel));
      
                //dstp_b[x] = (pixel_t)(dstp_b[x] + (((luma - dstp_b[x]) * alpha + rounder) >> bits_per_pixel));
      
        ✗
                const uint32_t luma_scaled = (uint32_t)luma * alpha_eff;
      
        ✗
                dstp_r[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_r[x] * inv_alpha + luma_scaled + half, magic);
      
        ✗
                dstp_g[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_g[x] * inv_alpha + luma_scaled + half, magic);
      
        ✗
                dstp_b[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_b[x] * inv_alpha + luma_scaled + half, magic);
      
              }
      
              if constexpr (blend_alpha)
      
        ✗
                dstp_a[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_a[x] * inv_alpha + (uint32_t)alpha_target[x] * alpha_eff + half, magic);
      
            }
      
        ✗
            dstp_g += dst_pitch;
      
        ✗
            dstp_b += dst_pitch;
      
        ✗
            dstp_r += dst_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              dstp_a += dst_pitch;
      
        ✗
            ovrp_g += overlay_pitch;
      
        ✗
            ovrp_b += overlay_pitch;
      
        ✗
            ovrp_r += overlay_pitch;
      
            if constexpr (has_alpha)
      
        ✗
              maskp += mask_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              alpha_target += overlay_pitch;
      
          }
      
        ✗
        }
      
        // subtract=false only: overlay clip is pre-inverted by Layer::Create when op="Subtract".
      
        template<bool chroma, bool has_alpha, bool blend_alpha>
      
        ✗
        static void layer_planarrgb_add_f_c(BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, float opacity) {
      
        ✗
          float* dstp_g = reinterpret_cast<float*>(dstp8[0]);
      
        ✗
          float* dstp_b = reinterpret_cast<float*>(dstp8[1]);
      
        ✗
          float* dstp_r = reinterpret_cast<float*>(dstp8[2]);
      
          // dstp8[3]: written only when blend_alpha=true (both clips have alpha).
      
          float* dstp_a;
      
          if constexpr (blend_alpha)
      
        ✗
            dstp_a = reinterpret_cast<float*>(dstp8[3]);
      
        ✗
          const float* ovrp_g = reinterpret_cast<const float*>(ovrp8[0]);
      
        ✗
          const float* ovrp_b = reinterpret_cast<const float*>(ovrp8[1]);
      
        ✗
          const float* ovrp_r = reinterpret_cast<const float*>(ovrp8[2]);
      
          // maskp: per-pixel blend weight — decoupled from ovrp8[3].
      
        ✗
          const float* maskp = reinterpret_cast<const float*>(maskp8);
      
          const float* alpha_target;
      
          if constexpr (blend_alpha)
      
        ✗
            alpha_target = reinterpret_cast<const float*>(ovrp8[3]);
      
        ✗
          dst_pitch /= sizeof(float);
      
        ✗
          overlay_pitch /= sizeof(float);
      
        ✗
          mask_pitch /= sizeof(float);
      
        ✗
          for (int y = 0; y < height; ++y) {
      
        ✗
            for (int x = 0; x < width; ++x) {
      
        ✗
              float alpha = has_alpha ? maskp[x] * opacity : opacity;
      
              if constexpr (chroma) {
      
        ✗
                dstp_r[x] = dstp_r[x] + (ovrp_r[x] - dstp_r[x]) * alpha;
      
        ✗
                dstp_g[x] = dstp_g[x] + (ovrp_g[x] - dstp_g[x]) * alpha;
      
        ✗
                dstp_b[x] = dstp_b[x] + (ovrp_b[x] - dstp_b[x]) * alpha;
      
              }
      
              else { // use luma instead of overlay
      
        ✗
                float luma = cyb_f * ovrp_b[x] + cyg_f * ovrp_g[x] + cyr_f * ovrp_r[x];
      
        ✗
                dstp_r[x] = dstp_r[x] + (luma - dstp_r[x]) * alpha;
      
        ✗
                dstp_g[x] = dstp_g[x] + (luma - dstp_g[x]) * alpha;
      
        ✗
                dstp_b[x] = dstp_b[x] + (luma - dstp_b[x]) * alpha;
      
              }
      
              if constexpr (blend_alpha)
      
        ✗
                dstp_a[x] = dstp_a[x] + (alpha_target[x] - dstp_a[x]) * alpha;
      
            }
      
        ✗
            dstp_g += dst_pitch;
      
        ✗
            dstp_b += dst_pitch;
      
        ✗
            dstp_r += dst_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              dstp_a += dst_pitch;
      
        ✗
            ovrp_g += overlay_pitch;
      
        ✗
            ovrp_b += overlay_pitch;
      
        ✗
            ovrp_r += overlay_pitch;
      
            if constexpr (has_alpha)
      
        ✗
              maskp += mask_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              alpha_target += overlay_pitch;
      
          }
      
        ✗
        }
      
        // ---------------------------------------------------------------------------
      
        // Planar RGB add — SSE4.1 per-plane wrappers (mirrors AVX2 counterparts).
      
        // All planar RGB planes are MASK444. maskp8 is the per-pixel weight.
      
        // chroma=false and float fall back to C templates.
      
        ✗
        static void layer_planarrgb_add_c_3plane(
      
          BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8,
      
          int dst_pitch, int overlay_pitch, int mask_pitch,
      
          int width, int height, int opacity_i, int bits_per_pixel)
      
        {
      
        ✗
          for (int i = 0; i < 3; i++)
      
        ✗
            masked_merge_c_impl<MASK444>(
      
        ✗
              dstp8[i], ovrp8[i], maskp8,
      
              dst_pitch, overlay_pitch, mask_pitch,
      
              width, height, opacity_i, bits_per_pixel);
      
        ✗
        }
      
        ✗
        static void layer_planarrgb_add_c_4plane(
      
          BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8,
      
          int dst_pitch, int overlay_pitch, int mask_pitch,
      
          int width, int height, int opacity_i, int bits_per_pixel)
      
        {
      
        ✗
          for (int i = 0; i < 4; i++)
      
        ✗
            masked_merge_c_impl<MASK444>(
      
        ✗
              dstp8[i], ovrp8[i], maskp8,
      
              dst_pitch, overlay_pitch, mask_pitch,
      
              width, height, opacity_i, bits_per_pixel);
      
        ✗
        }
      
        ✗
        static void get_layer_planarrgb_add_functions(
      
          bool chroma, bool hasAlpha, bool blendAlpha, int bits_per_pixel,
      
          /*out*/layer_planarrgb_add_c_t** layer_fn,
      
          /*out*/layer_planarrgb_add_f_c_t** layer_f_fn)
      
        {
      
          // chroma is true: Layer can use the unified masked and weighted blend routines
      
          // chroma is false: Layer-specific extension
      
          // Integer + hasAlpha + chroma=true: dispatch per-plane to masked_merge_avx2_impl (MASK444).
      
          // chroma=false (blend toward luma) has a different formula — keep C template.
      
          // float: keep C template (float perf is usually fine; could add later).
      
        ✗
          if (chroma && bits_per_pixel != 32) {
      
        ✗
            if (hasAlpha) {
      
        ✗
              *layer_fn = blendAlpha ? layer_planarrgb_add_c_4plane : layer_planarrgb_add_c_3plane;
      
        ✗
              return;
      
            }
      
            // no alpha: standard weighted merge, to be added later.
      
          }
      
        #define PLANARRGB_ADD_DISPATCH(chroma, has_alpha, blend_alpha) \
      
          { if (bits_per_pixel == 8) \
      
            *layer_fn = layer_planarrgb_add_c<uint8_t, chroma, has_alpha, blend_alpha>; \
      
          else if (bits_per_pixel <= 16) \
      
            *layer_fn = layer_planarrgb_add_c<uint16_t, chroma, has_alpha, blend_alpha>; \
      
          else /* float */ \
      
            *layer_f_fn = layer_planarrgb_add_f_c<chroma, has_alpha, blend_alpha>; \
      
          }
      
        ✗
          if (hasAlpha) {
      
        ✗
            if (chroma) {
      
        ✗
              if (blendAlpha) { PLANARRGB_ADD_DISPATCH(true, true, true)  }
      
        ✗
              else            { PLANARRGB_ADD_DISPATCH(true, true, false) }
      
            }
      
            else {
      
        ✗
              if (blendAlpha) { PLANARRGB_ADD_DISPATCH(false, true, true)  }
      
        ✗
              else            { PLANARRGB_ADD_DISPATCH(false, true, false) }
      
            }
      
          }
      
          else {
      
        ✗
            if (chroma) {
      
        ✗
              PLANARRGB_ADD_DISPATCH(true, false, false)
      
            }
      
            else {
      
        ✗
              PLANARRGB_ADD_DISPATCH(false, false, false)
      
            }
      
          }
      
        #undef PLANARRGB_ADD_DISPATCH
      
        }
      
        template<typename pixel_t, bool chroma, bool has_alpha, bool blend_alpha>
      
        1
        static void layer_planarrgb_mul_c(BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, int opacity_i, int bits_per_pixel) {
      
        1
          pixel_t* dstp_g = reinterpret_cast<pixel_t*>(dstp8[0]);
      
        1
          pixel_t* dstp_b = reinterpret_cast<pixel_t*>(dstp8[1]);
      
        1
          pixel_t* dstp_r = reinterpret_cast<pixel_t*>(dstp8[2]);
      
          // dstp8[3]: written only when blend_alpha=true (both clips have alpha).
      
          pixel_t* dstp_a;
      
          if constexpr (blend_alpha)
      
        1
            dstp_a = reinterpret_cast<pixel_t*>(dstp8[3]);
      
        1
          const pixel_t* ovrp_g = reinterpret_cast<const pixel_t*>(ovrp8[0]);
      
        1
          const pixel_t* ovrp_b = reinterpret_cast<const pixel_t*>(ovrp8[1]);
      
        1
          const pixel_t* ovrp_r = reinterpret_cast<const pixel_t*>(ovrp8[2]);
      
          // maskp: per-pixel blend weight — decoupled from ovrp8[3].
      
        1
          const pixel_t* maskp = reinterpret_cast<const pixel_t*>(maskp8);
      
          // alpha_target: the value multiplied into dstp_a (only when blend_alpha=true).
      
          const pixel_t* alpha_target;
      
          if constexpr (blend_alpha)
      
        1
            alpha_target = reinterpret_cast<const pixel_t*>(ovrp8[3]);
      
        1
          dst_pitch /= sizeof(pixel_t);
      
        1
          overlay_pitch /= sizeof(pixel_t);
      
        1
          mask_pitch /= sizeof(pixel_t);
      
          if constexpr (sizeof(pixel_t) == 1)
      
        ✗
            bits_per_pixel = 8; // make quasi constexpr
      
          // use magic-div
      
        1
          const MagicDiv magic = get_magic_div(bits_per_pixel);
      
        1
          const int max_pixel_value = (1 << bits_per_pixel) - 1;
      
        1
          const int half = max_pixel_value / 2;
      
          2/24void layer_planarrgb_mul_c<unsigned char, false, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 4 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned char, false, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 4 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned char, false, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 18 → 4 not taken.
✗ Branch 18 → 19 not taken.
void layer_planarrgb_mul_c<unsigned char, true, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 4 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned char, true, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 4 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned char, true, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 18 → 4 not taken.
✗ Branch 18 → 19 not taken.
void layer_planarrgb_mul_c<unsigned short, false, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 4 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned short, false, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 4 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned short, false, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 18 → 4 not taken.
✗ Branch 18 → 19 not taken.
void layer_planarrgb_mul_c<unsigned short, true, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 4 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned short, true, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 4 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned short, true, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✓ Branch 18 → 4 taken 2 times.
✓ Branch 18 → 19 taken 1 time.

        3
          for (int y = 0; y < height; ++y) {
      
          2/24void layer_planarrgb_mul_c<unsigned char, false, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
void layer_planarrgb_mul_c<unsigned char, false, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 5 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned char, false, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 5 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned char, true, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
void layer_planarrgb_mul_c<unsigned char, true, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 5 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned char, true, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 5 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned short, false, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
void layer_planarrgb_mul_c<unsigned short, false, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 5 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned short, false, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 16 → 5 not taken.
✗ Branch 16 → 17 not taken.
void layer_planarrgb_mul_c<unsigned short, true, false, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 12 → 5 not taken.
✗ Branch 12 → 13 not taken.
void layer_planarrgb_mul_c<unsigned short, true, true, false>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✗ Branch 14 → 5 not taken.
✗ Branch 14 → 15 not taken.
void layer_planarrgb_mul_c<unsigned short, true, true, true>(unsigned char**, unsigned char const**, unsigned char const*, int, int, int, int, int, int, int):
✓ Branch 16 → 5 taken 12 times.
✓ Branch 16 → 17 taken 2 times.

        14
            for (int x = 0; x < width; ++x) {
      
              uint32_t alpha_eff;
      
              if constexpr (has_alpha) {
      
        12
                const uint32_t alpha_src = (uint32_t)maskp[x];
      
        12
                alpha_eff = (uint32_t)magic_div_rt<pixel_t>(alpha_src * (uint32_t)opacity_i + half, magic);
      
              }
      
              else {
      
        ✗
                alpha_eff = opacity_i;
      
              }
      
        12
              const uint32_t inv_alpha = max_pixel_value - alpha_eff;
      
              if constexpr (chroma) {
      
                // Calculate the multiplied product (A*B)/max
      
        12
                const pixel_t prod_r = (pixel_t)(((uint32_t)ovrp_r[x] * (uint32_t)dstp_r[x]) >> bits_per_pixel);
      
        12
                const pixel_t prod_g = (pixel_t)(((uint32_t)ovrp_g[x] * (uint32_t)dstp_g[x]) >> bits_per_pixel);
      
        12
                const pixel_t prod_b = (pixel_t)(((uint32_t)ovrp_b[x] * (uint32_t)dstp_b[x]) >> bits_per_pixel);
      
        12
                dstp_r[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_r[x] * inv_alpha + (uint32_t)prod_r * alpha_eff + half, magic);
      
        12
                dstp_g[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_g[x] * inv_alpha + (uint32_t)prod_g * alpha_eff + half, magic);
      
        12
                dstp_b[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_b[x] * inv_alpha + (uint32_t)prod_b * alpha_eff + half, magic);
      
              }
      
              else {
      
                // use luma instead of overlay
      
        ✗
                const int luma = (cyb * ovrp_b[x] + cyg * ovrp_g[x] + cyr * ovrp_r[x]) >> 15;
      
        ✗
                const pixel_t prod_r = (pixel_t)(((uint32_t)luma * (uint32_t)dstp_r[x]) >> bits_per_pixel);
      
        ✗
                const pixel_t prod_g = (pixel_t)(((uint32_t)luma * (uint32_t)dstp_g[x]) >> bits_per_pixel);
      
        ✗
                const pixel_t prod_b = (pixel_t)(((uint32_t)luma * (uint32_t)dstp_b[x]) >> bits_per_pixel);
      
        ✗
                dstp_r[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_r[x] * inv_alpha + (uint32_t)prod_r * alpha_eff + half, magic);
      
        ✗
                dstp_g[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_g[x] * inv_alpha + (uint32_t)prod_g * alpha_eff + half, magic);
      
        ✗
                dstp_b[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_b[x] * inv_alpha + (uint32_t)prod_b * alpha_eff + half, magic);
      
              }
      
              if constexpr (blend_alpha) {
      
        12
                const pixel_t prod_a = (pixel_t)(((uint32_t)alpha_target[x] * (uint32_t)dstp_a[x]) >> bits_per_pixel);
      
        24
                dstp_a[x] = (pixel_t)magic_div_rt<pixel_t>((uint32_t)dstp_a[x] * inv_alpha + (uint32_t)prod_a * alpha_eff + half, magic);
      
              }
      
            }
      
        2
            dstp_g += dst_pitch;
      
        2
            dstp_b += dst_pitch;
      
        2
            dstp_r += dst_pitch;
      
            if constexpr (blend_alpha)
      
        2
              dstp_a += dst_pitch;
      
        2
            ovrp_g += overlay_pitch;
      
        2
            ovrp_b += overlay_pitch;
      
        2
            ovrp_r += overlay_pitch;
      
            if constexpr (has_alpha)
      
        2
              maskp += mask_pitch;
      
            if constexpr (blend_alpha)
      
        2
              alpha_target += overlay_pitch;
      
          }
      
        1
        }
      
        template<bool chroma, bool has_alpha, bool blend_alpha>
      
        ✗
        static void layer_planarrgb_mul_f_c(BYTE** dstp8, const BYTE** ovrp8, const BYTE* maskp8, int dst_pitch, int overlay_pitch, int mask_pitch, int width, int height, float opacity) {
      
        ✗
          float* dstp_g = reinterpret_cast<float*>(dstp8[0]);
      
        ✗
          float* dstp_b = reinterpret_cast<float*>(dstp8[1]);
      
        ✗
          float* dstp_r = reinterpret_cast<float*>(dstp8[2]);
      
          // dstp8[3]: written only when blend_alpha=true (both clips have alpha).
      
          float* dstp_a;
      
          if constexpr (blend_alpha)
      
        ✗
            dstp_a = reinterpret_cast<float*>(dstp8[3]);
      
        ✗
          const float* ovrp_g = reinterpret_cast<const float*>(ovrp8[0]);
      
        ✗
          const float* ovrp_b = reinterpret_cast<const float*>(ovrp8[1]);
      
        ✗
          const float* ovrp_r = reinterpret_cast<const float*>(ovrp8[2]);
      
          // maskp: per-pixel blend weight — decoupled from ovrp8[3].
      
        ✗
          const float* maskp = reinterpret_cast<const float*>(maskp8);
      
          const float* alpha_target;
      
          if constexpr (blend_alpha)
      
        ✗
            alpha_target = reinterpret_cast<const float*>(ovrp8[3]);
      
        ✗
          dst_pitch /= sizeof(float);
      
        ✗
          overlay_pitch /= sizeof(float);
      
        ✗
          mask_pitch /= sizeof(float);
      
        ✗
          for (int y = 0; y < height; ++y) {
      
        ✗
            for (int x = 0; x < width; ++x) {
      
        ✗
              float alpha = has_alpha ? maskp[x] * opacity : opacity;
      
              if constexpr (chroma) {
      
        ✗
                dstp_r[x] = dstp_r[x] + (ovrp_r[x] * dstp_r[x] - dstp_r[x]) * alpha;
      
        ✗
                dstp_g[x] = dstp_g[x] + (ovrp_g[x] * dstp_g[x] - dstp_g[x]) * alpha;
      
        ✗
                dstp_b[x] = dstp_b[x] + (ovrp_b[x] * dstp_b[x] - dstp_b[x]) * alpha;
      
              }
      
              else { // use luma instead of overlay
      
        ✗
                float luma = cyb_f * ovrp_b[x] + cyg_f * ovrp_g[x] + cyr_f * ovrp_r[x];
      
        ✗
                dstp_r[x] = dstp_r[x] + (luma * dstp_r[x] - dstp_r[x]) * alpha;
      
        ✗
                dstp_g[x] = dstp_g[x] + (luma * dstp_g[x] - dstp_g[x]) * alpha;
      
        ✗
                dstp_b[x] = dstp_b[x] + (luma * dstp_b[x] - dstp_b[x]) * alpha;
      
              }
      
              if constexpr (blend_alpha)
      
        ✗
                dstp_a[x] = dstp_a[x] + (alpha_target[x] * dstp_a[x] - dstp_a[x]) * alpha;
      
            }
      
        ✗
            dstp_g += dst_pitch;
      
        ✗
            dstp_b += dst_pitch;
      
        ✗
            dstp_r += dst_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              dstp_a += dst_pitch;
      
        ✗
            ovrp_g += overlay_pitch;
      
        ✗
            ovrp_b += overlay_pitch;
      
        ✗
            ovrp_r += overlay_pitch;
      
            if constexpr (has_alpha)
      
        ✗
              maskp += mask_pitch;
      
            if constexpr (blend_alpha)
      
        ✗
              alpha_target += overlay_pitch;
      
          }
      
        ✗
        }
      
        1
        static void get_layer_planarrgb_mul_functions(
      
          bool chroma, bool hasAlpha, bool blendAlpha, int bits_per_pixel,
      
          /*out*/layer_planarrgb_mul_c_t** layer_fn,
      
          /*out*/layer_planarrgb_mul_f_c_t** layer_f_fn)
      
        {
      
        #define PLANARRGB_MUL_DISPATCH(chroma, has_alpha, blend_alpha) \
      
          { if (bits_per_pixel == 8) \
      
            *layer_fn = layer_planarrgb_mul_c<uint8_t, chroma, has_alpha, blend_alpha>; \
      
          else if (bits_per_pixel <= 16) \
      
            *layer_fn = layer_planarrgb_mul_c<uint16_t, chroma, has_alpha, blend_alpha>; \
      
          else /* float */ \
      
            *layer_f_fn = layer_planarrgb_mul_f_c<chroma, has_alpha, blend_alpha>; \
      
          }
      
          1/2✓ Branch 2 → 3 taken 1 time.
✗ Branch 2 → 26 not taken.

        1
          if (hasAlpha) {
      
          1/2✓ Branch 3 → 4 taken 1 time.
✗ Branch 3 → 15 not taken.

        1
            if (chroma) {
      
          3/6✓ Branch 4 → 5 taken 1 time.
✗ Branch 4 → 10 not taken.
✗ Branch 5 → 6 not taken.
✓ Branch 5 → 7 taken 1 time.
✓ Branch 7 → 8 taken 1 time.
✗ Branch 7 → 9 not taken.

        1
              if (blendAlpha) { PLANARRGB_MUL_DISPATCH(true, true, true)  }
      
        ✗
              else            { PLANARRGB_MUL_DISPATCH(true, true, false) }
      
            }
      
            else {
      
        ✗
              if (blendAlpha) { PLANARRGB_MUL_DISPATCH(false, true, true)  }
      
        ✗
              else            { PLANARRGB_MUL_DISPATCH(false, true, false) }
      
            }
      
          }
      
          else {
      
        ✗
            if (chroma) {
      
        ✗
              PLANARRGB_MUL_DISPATCH(true, false, false)
      
            }
      
            else {
      
        ✗
              PLANARRGB_MUL_DISPATCH(false, false, false)
      
            }
      
          }
      
        #undef PLANARRGB_MUL_DISPATCH
      
        1
        }
      
        // ---------------------------------------------------------------------------
      
        // Packed RGBA (RGB32 / RGB64) blend dispatcher — C reference.
      
        // Returns the appropriate masked_blend_packedrgba_c instantiation for the
      
        // given bit depth.  Subtract is handled by pre-inverting the overlay and
      
        // passing a separate maskp8 pointer in the caller (Layer::Create/GetFrame).
      
        // The C reference function selects the alpha source at runtime via the
      
        // maskp8 null-check, so a single function covers both Add and Subtract.
      
        // ---------------------------------------------------------------------------
      
        1
        static void get_layer_packedrgb_blend_functions(
      
          bool has_separate_mask,
      
          int bits_per_pixel,
      
          layer_packedrgb_blend_c_t** fn)
      
        {
      
          1/2✗ Branch 2 → 3 not taken.
✓ Branch 2 → 7 taken 1 time.

        1
          if (bits_per_pixel == 8)
      
        ✗
            *fn = has_separate_mask ?  masked_blend_packedrgba_c<uint8_t, true> : masked_blend_packedrgba_c<uint8_t, false>;
      
          else  // 16-bit (RGB64)
      
          1/2✗ Branch 7 → 8 not taken.
✓ Branch 7 → 9 taken 1 time.

        1
            *fn = has_separate_mask ?  masked_blend_packedrgba_c<uint16_t, true> : masked_blend_packedrgba_c<uint16_t, false>;
      
        1
        }
      
        // Clean up rowprep macros (defined near top of this file or by including TU).
      
        #undef LAYER_ROWPREP_FN