virtualx-engine/servers/rendering/renderer_rd/effects/fsr.cpp

129 lines
6.6 KiB
C++
Raw Normal View History

/**************************************************************************/
/* fsr.cpp */
/**************************************************************************/
/* This file is part of: */
/* GODOT ENGINE */
/* https://godotengine.org */
/**************************************************************************/
/* Copyright (c) 2014-present Godot Engine contributors (see AUTHORS.md). */
/* Copyright (c) 2007-2014 Juan Linietsky, Ariel Manzur. */
/* */
/* Permission is hereby granted, free of charge, to any person obtaining */
/* a copy of this software and associated documentation files (the */
/* "Software"), to deal in the Software without restriction, including */
/* without limitation the rights to use, copy, modify, merge, publish, */
/* distribute, sublicense, and/or sell copies of the Software, and to */
/* permit persons to whom the Software is furnished to do so, subject to */
/* the following conditions: */
/* */
/* The above copyright notice and this permission notice shall be */
/* included in all copies or substantial portions of the Software. */
/* */
/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */
/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */
/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. */
/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */
/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */
/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */
/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */
/**************************************************************************/
#include "fsr.h"
#include "../storage_rd/material_storage.h"
#include "../uniform_set_cache_rd.h"
using namespace RendererRD;
FSR::FSR() {
Vector<String> FSR_upscale_modes;
#if defined(MACOS_ENABLED) || defined(IOS_ENABLED)
// MoltenVK does not support some of the operations used by the normal mode of FSR. Fallback works just fine though.
FSR_upscale_modes.push_back("\n#define MODE_FSR_UPSCALE_FALLBACK\n");
#else
// Everyone else can use normal mode when available.
if (RD::get_singleton()->has_feature(RD::SUPPORTS_FSR_HALF_FLOAT)) {
FSR_upscale_modes.push_back("\n#define MODE_FSR_UPSCALE_NORMAL\n");
} else {
FSR_upscale_modes.push_back("\n#define MODE_FSR_UPSCALE_FALLBACK\n");
}
#endif
fsr_shader.initialize(FSR_upscale_modes);
shader_version = fsr_shader.version_create();
pipeline = RD::get_singleton()->compute_pipeline_create(fsr_shader.version_get_shader(shader_version, 0));
}
FSR::~FSR() {
fsr_shader.version_free(shader_version);
}
void FSR::fsr_upscale(Ref<RenderSceneBuffersRD> p_render_buffers, RID p_source_rd_texture, RID p_destination_texture) {
UniformSetCacheRD *uniform_set_cache = UniformSetCacheRD::get_singleton();
ERR_FAIL_NULL(uniform_set_cache);
MaterialStorage *material_storage = MaterialStorage::get_singleton();
ERR_FAIL_NULL(material_storage);
Size2i internal_size = p_render_buffers->get_internal_size();
Size2i target_size = p_render_buffers->get_target_size();
float fsr_upscale_sharpness = p_render_buffers->get_fsr_sharpness();
if (!p_render_buffers->has_texture(SNAME("FSR"), SNAME("upscale_texture"))) {
RD::DataFormat format = p_render_buffers->get_base_data_format();
uint32_t usage_bits = RD::TEXTURE_USAGE_SAMPLING_BIT | RD::TEXTURE_USAGE_STORAGE_BIT | RD::TEXTURE_USAGE_COLOR_ATTACHMENT_BIT;
uint32_t layers = 1; // we only need one layer, in multiview we're processing one layer at a time.
p_render_buffers->create_texture(SNAME("FSR"), SNAME("upscale_texture"), format, usage_bits, RD::TEXTURE_SAMPLES_1, target_size, layers);
}
RID upscale_texture = p_render_buffers->get_texture(SNAME("FSR"), SNAME("upscale_texture"));
FSRUpscalePushConstant push_constant;
memset(&push_constant, 0, sizeof(FSRUpscalePushConstant));
int dispatch_x = (target_size.x + 15) / 16;
int dispatch_y = (target_size.y + 15) / 16;
RD::ComputeListID compute_list = RD::get_singleton()->compute_list_begin();
RD::get_singleton()->compute_list_bind_compute_pipeline(compute_list, pipeline);
push_constant.resolution_width = internal_size.width;
push_constant.resolution_height = internal_size.height;
push_constant.upscaled_width = target_size.width;
push_constant.upscaled_height = target_size.height;
push_constant.sharpness = fsr_upscale_sharpness;
RID shader = fsr_shader.version_get_shader(shader_version, 0);
ERR_FAIL_COND(shader.is_null());
RID default_sampler = material_storage->sampler_rd_get_default(RS::CANVAS_ITEM_TEXTURE_FILTER_LINEAR, RS::CANVAS_ITEM_TEXTURE_REPEAT_DISABLED);
//FSR Easc
RD::Uniform u_source_rd_texture(RD::UNIFORM_TYPE_SAMPLER_WITH_TEXTURE, 0, { default_sampler, p_source_rd_texture });
RD::Uniform u_upscale_texture(RD::UNIFORM_TYPE_IMAGE, 0, { upscale_texture });
push_constant.pass = FSR_UPSCALE_PASS_EASU;
RD::get_singleton()->compute_list_bind_uniform_set(compute_list, uniform_set_cache->get_cache(shader, 0, u_source_rd_texture), 0);
RD::get_singleton()->compute_list_bind_uniform_set(compute_list, uniform_set_cache->get_cache(shader, 1, u_upscale_texture), 1);
RD::get_singleton()->compute_list_set_push_constant(compute_list, &push_constant, sizeof(FSRUpscalePushConstant));
RD::get_singleton()->compute_list_dispatch(compute_list, dispatch_x, dispatch_y, 1);
RD::get_singleton()->compute_list_add_barrier(compute_list);
//FSR Rcas
RD::Uniform u_upscale_texture_with_sampler(RD::UNIFORM_TYPE_SAMPLER_WITH_TEXTURE, 0, { default_sampler, upscale_texture });
RD::Uniform u_destination_texture(RD::UNIFORM_TYPE_IMAGE, 0, { p_destination_texture });
push_constant.pass = FSR_UPSCALE_PASS_RCAS;
RD::get_singleton()->compute_list_bind_uniform_set(compute_list, uniform_set_cache->get_cache(shader, 0, u_upscale_texture_with_sampler), 0);
RD::get_singleton()->compute_list_bind_uniform_set(compute_list, uniform_set_cache->get_cache(shader, 1, u_destination_texture), 1);
RD::get_singleton()->compute_list_set_push_constant(compute_list, &push_constant, sizeof(FSRUpscalePushConstant));
RD::get_singleton()->compute_list_dispatch(compute_list, dispatch_x, dispatch_y, 1);
RD::get_singleton()->compute_list_end();
}