node_composite_filter.cc

Go to the documentation of this file.

/* SPDX-FileCopyrightText: 2006 Blender Authors
 *
 * SPDX-License-Identifier: GPL-2.0-or-later */

 
#include "BLI_math_matrix_types.hh"
#include "BLI_math_vector.hh"
#include "BLI_math_vector_types.hh"
 
#include "UI_interface.hh"
#include "UI_resources.hh"
 
#include "COM_node_operation.hh"
#include "COM_utilities.hh"
 
#include "node_composite_util.hh"
 
/* **************** FILTER  ******************** */
 
namespace blender::nodes::node_composite_filter_cc {
 
static void cmp_node_filter_declare(NodeDeclarationBuilder &b)
{
  b.add_input<decl::Float>("Fac")
      .default_value(1.0f)
      .min(0.0f)
      .max(1.0f)
      .subtype(PROP_FACTOR)
      .compositor_domain_priority(1);
  b.add_input<decl::Color>("Image")
      .default_value({1.0f, 1.0f, 1.0f, 1.0f})
      .compositor_domain_priority(0);
  b.add_output<decl::Color>("Image");
}
 
static void node_composit_buts_filter(uiLayout *layout, bContext * /*C*/, PointerRNA *ptr)
{
  layout->prop(ptr, "filter_type", UI_ITEM_R_SPLIT_EMPTY_NAME, "", ICON_NONE);
}
 
class SocketSearchOp {
 public:
  CMPNodeFilterMethod filter_type = CMP_NODE_FILTER_SOFT;
  void operator()(LinkSearchOpParams &params)
  {
    bNode &node = params.add_node("CompositorNodeFilter");
    node.custom1 = filter_type;
    params.update_and_connect_available_socket(node, "Image");
  }
};
 
static void gather_link_searches(GatherLinkSearchOpParams &params)
{
  const eNodeSocketDatatype from_socket_type = eNodeSocketDatatype(params.other_socket().type);
  if (!params.node_tree().typeinfo->validate_link(from_socket_type, SOCK_RGBA)) {
    return;
  }
 
  params.add_item(IFACE_("Soften"), SocketSearchOp{CMP_NODE_FILTER_SOFT});
  params.add_item(IFACE_("Box Sharpen"), SocketSearchOp{CMP_NODE_FILTER_SHARP_BOX});
  params.add_item(IFACE_("Laplace"), SocketSearchOp{CMP_NODE_FILTER_LAPLACE});
  params.add_item(IFACE_("Sobel"), SocketSearchOp{CMP_NODE_FILTER_SOBEL});
  params.add_item(IFACE_("Prewitt"), SocketSearchOp{CMP_NODE_FILTER_PREWITT});
  params.add_item(IFACE_("Kirsch"), SocketSearchOp{CMP_NODE_FILTER_KIRSCH});
  params.add_item(IFACE_("Shadow"), SocketSearchOp{CMP_NODE_FILTER_SHADOW});
  params.add_item(IFACE_("Diamond Sharpen"), SocketSearchOp{CMP_NODE_FILTER_SHARP_DIAMOND});
}
 
using namespace blender::compositor;
 
class FilterOperation : public NodeOperation {
 public:
  using NodeOperation::NodeOperation;
 
  void execute() override
  {
    const Result &input_image = this->get_input("Image");
    if (input_image.is_single_value()) {
      Result &output_image = this->get_result("Image");
      output_image.share_data(input_image);
      return;
    }
 
    if (this->context().use_gpu()) {
      this->execute_gpu();
    }
    else {
      this->execute_cpu();
    }
  }
 
  void execute_gpu()
  {
    GPUShader *shader = context().get_shader(get_shader_name());
    GPU_shader_bind(shader);
 
    GPU_shader_uniform_mat3_as_mat4(shader, "ukernel", get_filter_kernel().ptr());
 
    const Result &input_image = get_input("Image");
    input_image.bind_as_texture(shader, "input_tx");
 
    const Result &factor = get_input("Fac");
    factor.bind_as_texture(shader, "factor_tx");
 
    const Domain domain = compute_domain();
 
    Result &output_image = get_result("Image");
    output_image.allocate_texture(domain);
    output_image.bind_as_image(shader, "output_img");
 
    compute_dispatch_threads_at_least(shader, domain.size);
 
    input_image.unbind_as_texture();
    factor.unbind_as_texture();
    output_image.unbind_as_image();
    GPU_shader_unbind();
  }
 
  const char *get_shader_name()
  {
    if (this->is_edge_filter()) {
      return "compositor_edge_filter";
    }
    return "compositor_filter";
  }
 
  void execute_cpu()
  {
    const float3x3 kernel = this->get_filter_kernel();
 
    const Result &input = get_input("Image");
    const Result &factor = get_input("Fac");
 
    const Domain domain = compute_domain();
    Result &output = get_result("Image");
    output.allocate_texture(domain);
 
    if (this->is_edge_filter()) {
      parallel_for(domain.size, [&](const int2 texel) {
        /* Compute the dot product between the 3x3 window around the pixel and the edge detection
         * kernel in the X direction and Y direction. The Y direction kernel is computed by
         * transposing the given X direction kernel. */
        float3 color_x = float3(0.0f);
        float3 color_y = float3(0.0f);
        for (int j = 0; j < 3; j++) {
          for (int i = 0; i < 3; i++) {
            float3 color = input.load_pixel_extended<float4>(texel + int2(i - 1, j - 1)).xyz();
            color_x += color * kernel[j][i];
            color_y += color * kernel[i][j];
          }
        }
 
        /* Compute the channel-wise magnitude of the 2D vector composed from the X and Y edge
         * detection filter results. */
        float3 magnitude = math::sqrt(color_x * color_x + color_y * color_y);
 
        /* Mix the channel-wise magnitude with the original color at the center of the kernel using
         * the input factor. */
        float4 color = input.load_pixel<float4>(texel);
        magnitude = math::interpolate(
            color.xyz(), magnitude, factor.load_pixel<float, true>(texel));
 
        /* Store the channel-wise magnitude with the original alpha of the input. */
        output.store_pixel(texel, float4(magnitude, color.w));
      });
    }
    else {
      parallel_for(domain.size, [&](const int2 texel) {
        /* Compute the dot product between the 3x3 window around the pixel and the kernel. */
        float4 color = float4(0.0f);
        for (int j = 0; j < 3; j++) {
          for (int i = 0; i < 3; i++) {
            color += input.load_pixel_extended<float4>(texel + int2(i - 1, j - 1)) * kernel[j][i];
          }
        }
 
        /* Mix with the original color at the center of the kernel using the input factor. */
        color = math::interpolate(
            input.load_pixel<float4>(texel), color, factor.load_pixel<float, true>(texel));
 
        /* Store the color making sure it is not negative. */
        output.store_pixel(texel, math::max(color, float4(0.0f)));
      });
    }
  }
 
  bool is_edge_filter()
  {
    switch (this->get_filter_method()) {
      case CMP_NODE_FILTER_LAPLACE:
      case CMP_NODE_FILTER_SOBEL:
      case CMP_NODE_FILTER_PREWITT:
      case CMP_NODE_FILTER_KIRSCH:
        return true;
      case CMP_NODE_FILTER_SOFT:
      case CMP_NODE_FILTER_SHARP_BOX:
      case CMP_NODE_FILTER_SHADOW:
      case CMP_NODE_FILTER_SHARP_DIAMOND:
        return false;
    }
    return false;
  }
 
  float3x3 get_filter_kernel()
  {
    /* Initialize the kernels as arrays of rows with the top row first. Edge detection kernels
     * return the kernel in the X direction, while the kernel in the Y direction will be computed
     * inside the shader by transposing the kernel in the X direction. */
    switch (get_filter_method()) {
      case CMP_NODE_FILTER_SOFT: {
        const float kernel[3][3] = {{1.0f / 16.0f, 2.0f / 16.0f, 1.0f / 16.0f},
                                    {2.0f / 16.0f, 4.0f / 16.0f, 2.0f / 16.0f},
                                    {1.0f / 16.0f, 2.0f / 16.0f, 1.0f / 16.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_SHARP_BOX: {
        const float kernel[3][3] = {
            {-1.0f, -1.0f, -1.0f}, {-1.0f, 9.0f, -1.0f}, {-1.0f, -1.0f, -1.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_LAPLACE: {
        const float kernel[3][3] = {{-1.0f / 8.0f, -1.0f / 8.0f, -1.0f / 8.0f},
                                    {-1.0f / 8.0f, 1.0f, -1.0f / 8.0f},
                                    {-1.0f / 8.0f, -1.0f / 8.0f, -1.0f / 8.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_SOBEL: {
        const float kernel[3][3] = {{1.0f, 0.0f, -1.0f}, {2.0f, 0.0f, -2.0f}, {1.0f, 0.0f, -1.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_PREWITT: {
        const float kernel[3][3] = {{1.0f, 0.0f, -1.0f}, {1.0f, 0.0f, -1.0f}, {1.0f, 0.0f, -1.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_KIRSCH: {
        const float kernel[3][3] = {
            {5.0f, -3.0f, -2.0f}, {5.0f, -3.0f, -2.0f}, {5.0f, -3.0f, -2.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_SHADOW: {
        const float kernel[3][3] = {{1.0f, 2.0f, 1.0f}, {0.0f, 1.0f, 0.0f}, {-1.0f, -2.0f, -1.0f}};
        return float3x3(kernel);
      }
      case CMP_NODE_FILTER_SHARP_DIAMOND: {
        const float kernel[3][3] = {
            {0.0f, -1.0f, 0.0f}, {-1.0f, 5.0f, -1.0f}, {0.0f, -1.0f, 0.0f}};
        return float3x3(kernel);
      }
      default: {
        const float kernel[3][3] = {{0.0f, 0.0f, 0.0f}, {0.0f, 1.0f, 0.0f}, {0.0f, 0.0f, 0.0f}};
        return float3x3(kernel);
      }
    }
  }
 
  CMPNodeFilterMethod get_filter_method()
  {
    return static_cast<CMPNodeFilterMethod>(bnode().custom1);
  }
};
 
static NodeOperation *get_compositor_operation(Context &context, DNode node)
{
  return new FilterOperation(context, node);
}
 
}  // namespace blender::nodes::node_composite_filter_cc
 
static void register_node_type_cmp_filter()
{
  namespace file_ns = blender::nodes::node_composite_filter_cc;
 
  static blender::bke::bNodeType ntype;
 
  cmp_node_type_base(&ntype, "CompositorNodeFilter", CMP_NODE_FILTER);
  ntype.ui_name = "Filter";
  ntype.ui_description = "Apply common image enhancement filters";
  ntype.enum_name_legacy = "FILTER";
  ntype.nclass = NODE_CLASS_OP_FILTER;
  ntype.declare = file_ns::cmp_node_filter_declare;
  ntype.draw_buttons = file_ns::node_composit_buts_filter;
  ntype.labelfunc = node_filter_label;
  ntype.flag |= NODE_PREVIEW;
  ntype.get_compositor_operation = file_ns::get_compositor_operation;
  ntype.gather_link_search_ops = file_ns::gather_link_searches;
 
  blender::bke::node_register_type(ntype);
}
NOD_REGISTER_NODE(register_node_type_cmp_filter)