archive/blender-archive
Archived
1
1
Fork 0
Code Pull Requests Packages Activity
This repository has been archived on 2023-10-09. You can view files and clone it, but cannot push or open issues or pull requests.
Files
f937123116abd4ba34ff858fa03415362d6c07cf
blender-archive/source/blender/draw/engines/eevee/eevee_effects.c

475 lines
17 KiB
C
Raw Normal View History

Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
/*
* Copyright 2016, Blender Foundation.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* Contributor(s): Blender Institute
*
*/
/** \file eevee_effects.c
* \ingroup draw_engine
Cleanup: BLI headers first Also move descriptions into doxy header
2017-11-20 14:11:45 +11:00
*
* Gather all screen space effects technique such as Bloom, Motion Blur, DoF, SSAO, SSR, ...
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
*/
#include "DRW_render.h"
#include "eevee_private.h"
Eevee : Add a workaround for bug with AMD RX VEGA Linux + Mesa Driver This bug (explained here https://github.com/dfelinto/opengl-sandbox/blob/downsample/README.md) is breaking eevee beyond the point it's workable. This patch workaround the issue by making sure every fbo have mipmaps that are strictly greater than 16px. This break the bloom visuals a bit but only for this setup.
2017-10-10 17:18:29 +02:00
#include "GPU_texture.h"
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
#include "GPU_extensions.h"
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
static struct {
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
/* Downsample Depth */
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
struct GPUShader *minz_downlevel_sh;
struct GPUShader *maxz_downlevel_sh;
struct GPUShader *minz_downdepth_sh;
struct GPUShader *maxz_downdepth_sh;
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
struct GPUShader *minz_downdepth_layer_sh;
struct GPUShader *maxz_downdepth_layer_sh;
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
struct GPUShader *minz_copydepth_sh;
struct GPUShader *maxz_copydepth_sh;
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
/* Simple Downsample */
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
struct GPUShader *downsample_sh;
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
struct GPUShader *downsample_cube_sh;
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
Eevee: Implement Temporal Anti Aliasing / Super Sampling This adds TAA to eevee. The only thing important to note is that we need to keep the unjittered depth buffer so that the other engines are composited correctly.
2017-09-25 20:14:07 +02:00
/* Theses are just references, not actually allocated */
Eevee: Initial implementation of Volumetrics.
2017-07-03 16:38:14 +02:00
struct GPUTexture *depth_src;
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
struct GPUTexture *color_src;
Eevee: Implement Temporal Anti Aliasing / Super Sampling This adds TAA to eevee. The only thing important to note is that we need to keep the unjittered depth buffer so that the other engines are composited correctly.
2017-09-25 20:14:07 +02:00
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
int depth_src_layer;
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
float cube_texel_size;
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
} e_data = {NULL}; /* Engine data */
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
extern char datatoc_effect_minmaxz_frag_glsl[];
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
extern char datatoc_effect_downsample_frag_glsl[];
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
extern char datatoc_effect_downsample_cube_frag_glsl[];
extern char datatoc_lightprobe_vert_glsl[];
extern char datatoc_lightprobe_geom_glsl[];
Eevee: Fix motion blur when not having set another key.
2017-05-11 14:31:28 +02:00
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
static void eevee_create_shader_downsample(void)
Eevee: SSR: Add two hit option. This option add another raytrace per pixel, clearing some noise. But multiplying the raytrace cost.
2017-07-23 00:03:45 +02:00
{
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
e_data.downsample_sh = DRW_shader_create_fullscreen(datatoc_effect_downsample_frag_glsl, NULL);
Cleanup: line length, right shift
2017-11-20 14:28:24 +11:00
e_data.downsample_cube_sh = DRW_shader_create(
datatoc_lightprobe_vert_glsl,
datatoc_lightprobe_geom_glsl,
datatoc_effect_downsample_cube_frag_glsl, NULL);
e_data.minz_downlevel_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MIN_PASS\n");
e_data.maxz_downlevel_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MAX_PASS\n");
e_data.minz_downdepth_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
Eevee: Remove unecessary defines.
2017-12-08 11:54:15 +01:00
"#define MIN_PASS\n");
Cleanup: line length, right shift
2017-11-20 14:28:24 +11:00
e_data.maxz_downdepth_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
Eevee: Remove unecessary defines.
2017-12-08 11:54:15 +01:00
"#define MAX_PASS\n");
Cleanup: line length, right shift
2017-11-20 14:28:24 +11:00
e_data.minz_downdepth_layer_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MIN_PASS\n"
Eevee: Remove unecessary defines.
2017-12-08 11:54:15 +01:00
"#define LAYERED\n");
Cleanup: line length, right shift
2017-11-20 14:28:24 +11:00
e_data.maxz_downdepth_layer_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MAX_PASS\n"
Eevee: Remove unecessary defines.
2017-12-08 11:54:15 +01:00
"#define LAYERED\n");
Cleanup: line length, right shift
2017-11-20 14:28:24 +11:00
e_data.minz_copydepth_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MIN_PASS\n"
"#define COPY_DEPTH\n");
e_data.maxz_copydepth_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MAX_PASS\n"
"#define COPY_DEPTH\n");
Eevee: SSR: Add two hit option. This option add another raytrace per pixel, clearing some noise. But multiplying the raytrace cost.
2017-07-23 00:03:45 +02:00
}
Eevee: Initial Final Render support. TAA / multiple samples is not working at the moment.
2018-01-29 18:16:11 +01:00
void EEVEE_effects_init(EEVEE_ViewLayerData *sldata, EEVEE_Data *vedata, Object *camera)
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
{
Eevee: Put all constant uniforms in a global UBO. This is an optimization / cleanup commit. The use of a global ubo remove lots of uniform lookups and only transfert data when needed. Lots of renaming for more consistent codestyle.
2018-01-21 17:25:10 +01:00
EEVEE_CommonUniformBuffer *common_data = &sldata->common_data;
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
EEVEE_StorageList *stl = vedata->stl;
EEVEE_FramebufferList *fbl = vedata->fbl;
EEVEE_TextureList *txl = vedata->txl;
Eevee: Initial Depth Of Field commit.
2017-05-09 21:57:13 +02:00
EEVEE_EffectsInfo *effects;
Eevee: Render: Force normal buffer creation if needed.
2018-02-03 03:00:39 +01:00
const DRWContextState *draw_ctx = DRW_context_state_get();
ViewLayer *view_layer = draw_ctx->view_layer;
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
const float *viewport_size = DRW_viewport_size_get();
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
/* Shaders */
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
if (!e_data.downsample_sh) {
eevee_create_shader_downsample();
Eevee: Add Bloom post process. Based on Kino/Bloom v2 - Bloom filter for Unity MIT license.
2017-05-07 14:27:43 +02:00
}
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
if (!stl->effects) {
stl->effects = MEM_callocN(sizeof(EEVEE_EffectsInfo), "EEVEE_EffectsInfo");
}
Eevee: Initial Depth Of Field commit.
2017-05-09 21:57:13 +02:00
effects = stl->effects;
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
effects->enabled_effects = 0;
Eevee: Initial Final Render support. TAA / multiple samples is not working at the moment.
2018-01-29 18:16:11 +01:00
effects->enabled_effects |= EEVEE_motion_blur_init(sldata, vedata, camera);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
effects->enabled_effects |= EEVEE_bloom_init(sldata, vedata);
Eevee: Initial Final Render support. TAA / multiple samples is not working at the moment.
2018-01-29 18:16:11 +01:00
effects->enabled_effects |= EEVEE_depth_of_field_init(sldata, vedata, camera);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
effects->enabled_effects |= EEVEE_temporal_sampling_init(sldata, vedata);
effects->enabled_effects |= EEVEE_occlusion_init(sldata, vedata);
Eevee: Initial Separable Subsurface Scattering implementation. How to use: - Enable subsurface scattering in the render options. - Add Subsurface BSDF to your shader. - Check "Screen Space Subsurface Scattering" in the material panel options. This initial implementation has a few limitations: - only supports gaussian SSS. - Does not support principled shader. - The radius parameters is baked down to a number of samples and then put into an UBO. This means the radius input socket cannot be used. You need to tweak the default vector directly. - The "texture blur" is considered as always set to 1
2017-11-14 00:49:54 +01:00
effects->enabled_effects |= EEVEE_subsurface_init(sldata, vedata);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
effects->enabled_effects |= EEVEE_screen_raytrace_init(sldata, vedata);
effects->enabled_effects |= EEVEE_volumes_init(sldata, vedata);
Eevee: Render: Force normal buffer creation if needed.
2018-02-03 03:00:39 +01:00
/* Force normal buffer creation. */
if (DRW_state_is_image_render() &&
Cleanup: style, warning
2018-02-15 18:03:55 +11:00
(view_layer->passflag & SCE_PASS_NORMAL) != 0)
Eevee: Render: Force normal buffer creation if needed.
2018-02-03 03:00:39 +01:00
{
effects->enabled_effects |= EFFECT_NORMAL_BUFFER;
}
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/**
* Ping Pong buffer
*/
if ((effects->enabled_effects & EFFECT_POST_BUFFER) != 0) {
Eevee: Fix TAA color drifting. This was caused by small float precision being insuficient. The blue component of R11F_G11F_B10F has lower precision than the other 2 components. This resulted in colors drifting towards a yellowish tone. Using RGBA16F for the concerned buffer. This double the memory usage of the framebuffers and add subsequent bandwidth usage.
2017-10-10 18:32:05 +02:00
DRWFboTexture tex = {&txl->color_post, DRW_TEX_RGBA_16, DRW_TEX_FILTER | DRW_TEX_MIPMAP};
Draw Manager: Make use of texture reuse.
2017-05-16 03:03:58 +02:00
DRW_framebuffer_init(&fbl->effect_fb, &draw_engine_eevee_type,
Eevee: Initial Depth Of Field commit.
2017-05-09 21:57:13 +02:00
(int)viewport_size[0], (int)viewport_size[1],
&tex, 1);
}
Eevee: Rework GTAO This includes big improvement: - The horizon search is decoupled from the BSDF evaluation. This means using multiple BSDF nodes have a much lower impact when enbaling AO. - The horizon search is optimized by splitting the search into 4 corners searching similar directions to help which GPU cache coherence. - The AO options are now uniforms and do not trigger shader recompilation (aka. freeze UI). - Include a quality slider similar to the SSR one. - Add a switch for disabling bounce light approximation. - Fix problem with Bent Normals when occlusion get very dark. - Add a denoise option to that takes the neighbors pixel values via glsl derivatives. This reduces noise but exhibit 2x2 blocky artifacts. The downside : Separating the horizon search uses more memory (~3MB for each samples on HD viewport). We could lower the bit depth to 4bit per horizon but it produce noticeable banding (might be fixed with some dithering).
2017-08-18 15:06:51 +02:00
else {
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/* Cleanup to release memory */
DRW_TEXTURE_FREE_SAFE(txl->color_post);
DRW_FRAMEBUFFER_FREE_SAFE(fbl->effect_fb);
Eevee: Ambient Occlusion: Initial implementation. Implement GTAO (Ground Truth Ambient Occlusion) which is a special case of Horizon Based Ambient Occlusion that is more physically accurate. Also add a bent normal option to sample indirect irradiance (diffuse lighting) with the least occluded direction.
2017-06-22 03:10:39 +02:00
}
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/**
* MinMax Pyramid
*/
Eevee: Do not allocate minz buffer. It's not needed for now.
2017-11-01 16:17:33 +01:00
DRWFboTexture texmax = {&txl->maxzbuffer, DRW_TEX_DEPTH_24, DRW_TEX_MIPMAP};
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
if (GPU_type_matches(GPU_DEVICE_INTEL, GPU_OS_ANY, GPU_DRIVER_ANY)) {
/* Intel gpu seems to have problem rendering to only depth format */
texmax.format = DRW_TEX_R_32;
}
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_framebuffer_init(&fbl->downsample_fb, &draw_engine_eevee_type,
Fix "GPUTexture: texture alloc failed. Not enough Video Memory." in area resizing `gpu_texture_try_alloc` invalidates zero-sized textures. The message in the console is not correct in this case (because it is not due to lack of memory).
2018-01-23 22:50:05 -02:00
max_ii((int)viewport_size[0] / 2, 1), max_ii((int)viewport_size[1] / 2, 1),
Eevee: Do not allocate minz buffer. It's not needed for now.
2017-11-01 16:17:33 +01:00
&texmax, 1);
Eevee: Rework GTAO This includes big improvement: - The horizon search is decoupled from the BSDF evaluation. This means using multiple BSDF nodes have a much lower impact when enbaling AO. - The horizon search is optimized by splitting the search into 4 corners searching similar directions to help which GPU cache coherence. - The AO options are now uniforms and do not trigger shader recompilation (aka. freeze UI). - Include a quality slider similar to the SSR one. - Add a switch for disabling bounce light approximation. - Fix problem with Bent Normals when occlusion get very dark. - Add a denoise option to that takes the neighbors pixel values via glsl derivatives. This reduces noise but exhibit 2x2 blocky artifacts. The downside : Separating the horizon search uses more memory (~3MB for each samples on HD viewport). We could lower the bit depth to 4bit per horizon but it produce noticeable banding (might be fixed with some dithering).
2017-08-18 15:06:51 +02:00
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/**
* Compute Mipmap texel alignement.
*/
Eevee: Correct Mipmap texel alignment. Since we are working with non power of 2 textures, the mipmap level UV does not line up perfectly. This resulted in skewed filtering and bad sampling of the min/max depth buffer.
2017-08-09 15:20:18 +02:00
for (int i = 0; i < 10; ++i) {
float mip_size[2] = {viewport_size[0], viewport_size[1]};
for (int j = 0; j < i; ++j) {
mip_size[0] = floorf(fmaxf(1.0f, mip_size[0] / 2.0f));
mip_size[1] = floorf(fmaxf(1.0f, mip_size[1] / 2.0f));
}
Eevee: Put all constant uniforms in a global UBO. This is an optimization / cleanup commit. The use of a global ubo remove lots of uniform lookups and only transfert data when needed. Lots of renaming for more consistent codestyle.
2018-01-21 17:25:10 +01:00
common_data->mip_ratio[i][0] = viewport_size[0] / (mip_size[0] * powf(2.0f, floorf(log2f(floorf(viewport_size[0] / mip_size[0])))));
common_data->mip_ratio[i][1] = viewport_size[1] / (mip_size[1] * powf(2.0f, floorf(log2f(floorf(viewport_size[1] / mip_size[1])))));
Eevee: Correct Mipmap texel alignment. Since we are working with non power of 2 textures, the mipmap level UV does not line up perfectly. This resulted in skewed filtering and bad sampling of the min/max depth buffer.
2017-08-09 15:20:18 +02:00
}
Eevee: SSR: Add fullscreen raytrace.
2017-07-21 15:06:29 +02:00
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/**
* Normal buffer for deferred passes.
*/
if ((effects->enabled_effects & EFFECT_NORMAL_BUFFER) != 0) {
Eevee: Rework GTAO This includes big improvement: - The horizon search is decoupled from the BSDF evaluation. This means using multiple BSDF nodes have a much lower impact when enbaling AO. - The horizon search is optimized by splitting the search into 4 corners searching similar directions to help which GPU cache coherence. - The AO options are now uniforms and do not trigger shader recompilation (aka. freeze UI). - Include a quality slider similar to the SSR one. - Add a switch for disabling bounce light approximation. - Fix problem with Bent Normals when occlusion get very dark. - Add a denoise option to that takes the neighbors pixel values via glsl derivatives. This reduces noise but exhibit 2x2 blocky artifacts. The downside : Separating the horizon search uses more memory (~3MB for each samples on HD viewport). We could lower the bit depth to 4bit per horizon but it produce noticeable banding (might be fixed with some dithering).
2017-08-18 15:06:51 +02:00
if (txl->ssr_normal_input == NULL) {
DRWTextureFormat nor_format = DRW_TEX_RG_16;
txl->ssr_normal_input = DRW_texture_create_2D((int)viewport_size[0], (int)viewport_size[1], nor_format, 0, NULL);
}
/* Reattach textures to the right buffer (because we are alternating between buffers) */
/* TODO multiple FBO per texture!!!! */
DRW_framebuffer_texture_detach(txl->ssr_normal_input);
DRW_framebuffer_texture_attach(fbl->main, txl->ssr_normal_input, 1, 0);
}
else {
/* Cleanup to release memory */
DRW_TEXTURE_FREE_SAFE(txl->ssr_normal_input);
}
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/**
* Setup double buffer so we can access last frame as it was before post processes.
*/
Eevee: SSR: Add double buffer so we can read previous frame color. Also add simple reprojection and screen fade to the SSR resolve pass.
2017-07-19 14:22:03 +02:00
if ((effects->enabled_effects & EFFECT_DOUBLE_BUFFER) != 0) {
Eevee: Fix TAA color drifting. This was caused by small float precision being insuficient. The blue component of R11F_G11F_B10F has lower precision than the other 2 components. This resulted in colors drifting towards a yellowish tone. Using RGBA16F for the concerned buffer. This double the memory usage of the framebuffers and add subsequent bandwidth usage.
2017-10-10 18:32:05 +02:00
DRWFboTexture tex_double_buffer = {&txl->color_double_buffer, DRW_TEX_RGBA_16, DRW_TEX_FILTER | DRW_TEX_MIPMAP};
Eevee: SSR: Add double buffer so we can read previous frame color. Also add simple reprojection and screen fade to the SSR resolve pass.
2017-07-19 14:22:03 +02:00
DRW_framebuffer_init(&fbl->double_buffer, &draw_engine_eevee_type,
(int)viewport_size[0], (int)viewport_size[1],
&tex_double_buffer, 1);
}
else {
/* Cleanup to release memory */
DRW_TEXTURE_FREE_SAFE(txl->color_double_buffer);
DRW_FRAMEBUFFER_FREE_SAFE(fbl->double_buffer);
}
Eevee: Add Bloom post process. Based on Kino/Bloom v2 - Bloom filter for Unity MIT license.
2017-05-07 14:27:43 +02:00
}
Eevee: Put all constant uniforms in a global UBO. This is an optimization / cleanup commit. The use of a global ubo remove lots of uniform lookups and only transfert data when needed. Lots of renaming for more consistent codestyle.
2018-01-21 17:25:10 +01:00
void EEVEE_effects_cache_init(EEVEE_ViewLayerData *sldata, EEVEE_Data *vedata)
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
{
EEVEE_PassList *psl = vedata->psl;
EEVEE_TextureList *txl = vedata->txl;
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
int downsample_write = DRW_STATE_WRITE_DEPTH;
/* Intel gpu seems to have problem rendering to only depth format.
* Use color texture instead. */
if (GPU_type_matches(GPU_DEVICE_INTEL, GPU_OS_ANY, GPU_DRIVER_ANY)) {
downsample_write = DRW_STATE_WRITE_COLOR;
}
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
Gawain API naming refactor Use consistent prefix for gawain API names as well as some abbreviations to avoid over-long names, see: D2678
2017-06-19 20:18:04 +10:00
struct Gwn_Batch *quad = DRW_cache_fullscreen_quad_get();
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
{
psl->color_downsample_ps = DRW_pass_create("Downsample", DRW_STATE_WRITE_COLOR);
DRWShadingGroup *grp = DRW_shgroup_create(e_data.downsample_sh, psl->color_downsample_ps);
DRW_shgroup_uniform_buffer(grp, "source", &e_data.color_src);
Eevee: Put all constant uniforms in a global UBO. This is an optimization / cleanup commit. The use of a global ubo remove lots of uniform lookups and only transfert data when needed. Lots of renaming for more consistent codestyle.
2018-01-21 17:25:10 +01:00
DRW_shgroup_uniform_float(grp, "fireflyFactor", &sldata->common_data.ssr_firefly_fac, 1);
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
DRW_shgroup_call_add(grp, quad, NULL);
}
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
{
static int zero = 0;
DRW: Refactor simple instancing. Instead of creating a new instancing shading group without attrib, we now have instancing calls. The benefits is that they can be culled. They can be used in conjuction with the standard and generate calls but shader must support it (which is generally not the case). We store a pointer to the actual count so that the number can be tweaked between redraw. This will makes multi layer rendering more efficient.
2018-03-01 19:27:38 +01:00
static unsigned int six = 6;
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
psl->color_downsample_cube_ps = DRW_pass_create("Downsample Cube", DRW_STATE_WRITE_COLOR);
DRW: Refactor simple instancing. Instead of creating a new instancing shading group without attrib, we now have instancing calls. The benefits is that they can be culled. They can be used in conjuction with the standard and generate calls but shader must support it (which is generally not the case). We store a pointer to the actual count so that the number can be tweaked between redraw. This will makes multi layer rendering more efficient.
2018-03-01 19:27:38 +01:00
DRWShadingGroup *grp = DRW_shgroup_create(e_data.downsample_cube_sh, psl->color_downsample_cube_ps);
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
DRW_shgroup_uniform_buffer(grp, "source", &e_data.color_src);
DRW_shgroup_uniform_float(grp, "texelSize", &e_data.cube_texel_size, 1);
DRW_shgroup_uniform_int(grp, "Layer", &zero, 1);
DRW: Refactor simple instancing. Instead of creating a new instancing shading group without attrib, we now have instancing calls. The benefits is that they can be culled. They can be used in conjuction with the standard and generate calls but shader must support it (which is generally not the case). We store a pointer to the actual count so that the number can be tweaked between redraw. This will makes multi layer rendering more efficient.
2018-03-01 19:27:38 +01:00
DRW_shgroup_call_instances_add(grp, quad, NULL, &six);
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
}
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
{
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
/* Perform min/max downsample */
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
DRWShadingGroup *grp;
#if 0 /* Not used for now */
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->minz_downlevel_ps = DRW_pass_create("HiZ Min Down Level", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
grp = DRW_shgroup_create(e_data.minz_downlevel_sh, psl->minz_downlevel_ps);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &stl->g_data->minzbuffer);
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#endif
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->maxz_downlevel_ps = DRW_pass_create("HiZ Max Down Level", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
grp = DRW_shgroup_create(e_data.maxz_downlevel_sh, psl->maxz_downlevel_ps);
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &txl->maxzbuffer);
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
/* Copy depth buffer to halfres top level of HiZ */
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#if 0 /* Not used for now */
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->minz_downdepth_ps = DRW_pass_create("HiZ Min Copy Depth Halfres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
grp = DRW_shgroup_create(e_data.minz_downdepth_sh, psl->minz_downdepth_ps);
Eevee: Initial implementation of Volumetrics.
2017-07-03 16:38:14 +02:00
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &e_data.depth_src);
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#endif
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->maxz_downdepth_ps = DRW_pass_create("HiZ Max Copy Depth Halfres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
grp = DRW_shgroup_create(e_data.maxz_downdepth_sh, psl->maxz_downdepth_ps);
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &e_data.depth_src);
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#if 0 /* Not used for now */
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->minz_downdepth_layer_ps = DRW_pass_create("HiZ Min Copy DepthLayer Halfres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
grp = DRW_shgroup_create(e_data.minz_downdepth_layer_sh, psl->minz_downdepth_layer_ps);
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &e_data.depth_src);
DRW_shgroup_uniform_int(grp, "depthLayer", &e_data.depth_src_layer, 1);
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#endif
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->maxz_downdepth_layer_ps = DRW_pass_create("HiZ Max Copy DepthLayer Halfres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
grp = DRW_shgroup_create(e_data.maxz_downdepth_layer_sh, psl->maxz_downdepth_layer_ps);
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &e_data.depth_src);
DRW_shgroup_uniform_int(grp, "depthLayer", &e_data.depth_src_layer, 1);
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
/* Copy depth buffer to halfres top level of HiZ */
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#if 0 /* Not used for now */
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->minz_copydepth_ps = DRW_pass_create("HiZ Min Copy Depth Fullres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
grp = DRW_shgroup_create(e_data.minz_copydepth_sh, psl->minz_copydepth_ps);
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &e_data.depth_src);
DRW_shgroup_call_add(grp, quad, NULL);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#endif
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
Eevee: Fix SSR ans SSAO for intel GPU. On my intel GPU there is some issue when rendering to depth only format. This workaround the issue.
2017-12-08 11:56:48 +01:00
psl->maxz_copydepth_ps = DRW_pass_create("HiZ Max Copy Depth Fullres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
grp = DRW_shgroup_create(e_data.maxz_copydepth_sh, psl->maxz_copydepth_ps);
Eevee: Initial implementation of Volumetrics.
2017-07-03 16:38:14 +02:00
DRW_shgroup_uniform_buffer(grp, "depthBuffer", &e_data.depth_src);
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
DRW_shgroup_call_add(grp, quad, NULL);
}
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
}
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#if 0 /* Not required for now */
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
static void min_downsample_cb(void *vedata, int UNUSED(level))
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
{
EEVEE_PassList *psl = ((EEVEE_Data *)vedata)->psl;
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_draw_pass(psl->minz_downlevel_ps);
}
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#endif
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
static void max_downsample_cb(void *vedata, int UNUSED(level))
{
EEVEE_PassList *psl = ((EEVEE_Data *)vedata)->psl;
DRW_draw_pass(psl->maxz_downlevel_ps);
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
}
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
static void simple_downsample_cb(void *vedata, int UNUSED(level))
{
EEVEE_PassList *psl = ((EEVEE_Data *)vedata)->psl;
DRW_draw_pass(psl->color_downsample_ps);
}
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
static void simple_downsample_cube_cb(void *vedata, int level)
{
EEVEE_PassList *psl = ((EEVEE_Data *)vedata)->psl;
e_data.cube_texel_size = (float)(1 << level) / (float)GPU_texture_width(e_data.color_src);
DRW_draw_pass(psl->color_downsample_cube_ps);
}
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
void EEVEE_create_minmax_buffer(EEVEE_Data *vedata, GPUTexture *depth_src, int layer)
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
{
EEVEE_PassList *psl = vedata->psl;
EEVEE_FramebufferList *fbl = vedata->fbl;
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
EEVEE_TextureList *txl = vedata->txl;
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
Eevee: Initial implementation of Volumetrics.
2017-07-03 16:38:14 +02:00
e_data.depth_src = depth_src;
Eevee: Fix and opimize MinMaxZ generation.
2017-08-12 16:13:08 +02:00
e_data.depth_src_layer = layer;
Eevee: Add AO support in planar reflection. Technically this enables the use of MinmaxZ pyramid inside the probe captures. Also Disable AO for cubemap probes because it shows big discontinuities at cubeface limits.
2017-06-23 21:08:53 +02:00
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#if 0 /* Not required for now */
DRW: Add stats group to eevee and object mode.
2017-07-26 19:58:15 +02:00
DRW_stats_group_start("Min buffer");
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
/* Copy depth buffer to min texture top level */
DRW_framebuffer_texture_attach(fbl->downsample_fb, stl->g_data->minzbuffer, 0, 0);
DRW_framebuffer_bind(fbl->downsample_fb);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
if (layer >= 0) {
DRW_draw_pass(psl->minz_downdepth_layer_ps);
}
else {
DRW_draw_pass(psl->minz_downdepth_ps);
}
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_framebuffer_texture_detach(stl->g_data->minzbuffer);
/* Create lower levels */
DRW_framebuffer_recursive_downsample(fbl->downsample_fb, stl->g_data->minzbuffer, 8, &min_downsample_cb, vedata);
DRW: Add stats group to eevee and object mode.
2017-07-26 19:58:15 +02:00
DRW_stats_group_end();
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
#endif
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW: Add stats group to eevee and object mode.
2017-07-26 19:58:15 +02:00
DRW_stats_group_start("Max buffer");
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
/* Copy depth buffer to max texture top level */
DRW_framebuffer_texture_attach(fbl->downsample_fb, txl->maxzbuffer, 0, 0);
DRW_framebuffer_bind(fbl->downsample_fb);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
if (layer >= 0) {
DRW_draw_pass(psl->maxz_downdepth_layer_ps);
}
else {
DRW_draw_pass(psl->maxz_downdepth_ps);
}
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_framebuffer_texture_detach(txl->maxzbuffer);
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
/* Create lower levels */
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_framebuffer_recursive_downsample(fbl->downsample_fb, txl->maxzbuffer, 8, &max_downsample_cb, vedata);
DRW: Add stats group to eevee and object mode.
2017-07-26 19:58:15 +02:00
DRW_stats_group_end();
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/* Restore */
DRW_framebuffer_bind(fbl->main);
Eevee: Minmax Depth Pyramid. This commit introduce the computation of a depth pyramid containing min and max depth values of the original depth buffer. This is useful for Clustered Light Culling but also for raytracing on the depth buffer (SSR). It's also usefull to have to fetch higher mips in order to improve texture cache usage. As of now, 1st mip (highest res) is half the resolution of the depth buffer, but everything is already done to be able to make a fullres copy of the depth buffer in the 1st mip instead of downsampling. Also, the texture used is RG_32F which is a too much but enough to cover the 24bits of the depth buffer. Reducing the texture size would make things quite faster.
2017-06-22 02:28:49 +02:00
}
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
/**
* Simple downsampling algorithm. Reconstruct mip chain up to mip level.
**/
void EEVEE_downsample_buffer(EEVEE_Data *vedata, struct GPUFrameBuffer *fb_src, GPUTexture *texture_src, int level)
{
e_data.color_src = texture_src;
DRW: Add stats group to eevee and object mode.
2017-07-26 19:58:15 +02:00
DRW_stats_group_start("Downsample buffer");
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
/* Create lower levels */
DRW_framebuffer_recursive_downsample(fb_src, texture_src, level, &simple_downsample_cb, vedata);
DRW: Add stats group to eevee and object mode.
2017-07-26 19:58:15 +02:00
DRW_stats_group_end();
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
}
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
/**
* Simple downsampling algorithm for cubemap. Reconstruct mip chain up to mip level.
**/
void EEVEE_downsample_cube_buffer(EEVEE_Data *vedata, struct GPUFrameBuffer *fb_src, GPUTexture *texture_src, int level)
{
e_data.color_src = texture_src;
DRW_stats_group_start("Downsample Cube buffer");
/* Create lower levels */
DRW_framebuffer_recursive_downsample(fb_src, texture_src, level, &simple_downsample_cube_cb, vedata);
DRW_stats_group_end();
}
Eevee: Fix prev_persmat being used by multiple viewport.
2018-02-26 20:08:48 +01:00
void EEVEE_draw_effects(EEVEE_ViewLayerData *UNUSED(sldata), EEVEE_Data *vedata)
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
{
Eevee: Add Bloom post process. Based on Kino/Bloom v2 - Bloom filter for Unity MIT license.
2017-05-07 14:27:43 +02:00
EEVEE_TextureList *txl = vedata->txl;
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
EEVEE_FramebufferList *fbl = vedata->fbl;
EEVEE_StorageList *stl = vedata->stl;
EEVEE_EffectsInfo *effects = stl->effects;
DefaultTextureList *dtxl = DRW_viewport_texture_list_get();
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/* only once per frame after the first post process */
effects->swap_double_buffer = ((effects->enabled_effects & EFFECT_DOUBLE_BUFFER) != 0);
Eevee: Add Bloom post process. Based on Kino/Bloom v2 - Bloom filter for Unity MIT license.
2017-05-07 14:27:43 +02:00
/* Init pointers */
effects->source_buffer = txl->color; /* latest updated texture */
effects->target_buffer = fbl->effect_fb; /* next target to render to */
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/* Temporal Anti-Aliasing MUST come first */
EEVEE_temporal_sampling_draw(vedata);
Eevee: Implement Temporal Anti Aliasing / Super Sampling This adds TAA to eevee. The only thing important to note is that we need to keep the unjittered depth buffer so that the other engines are composited correctly.
2017-09-25 20:14:07 +02:00
Eevee: Initial Depth Of Field commit.
2017-05-09 21:57:13 +02:00
/* Detach depth for effects to use it */
DRW_framebuffer_texture_detach(dtxl->depth);
Eevee: Effects: Split each effect in it's own file. This also: - make sure to only compile the shader needed by the active effects. - same thing for the shading groups. - disable TAA if motion blur is active (avoid infinite refresh).
2017-11-01 01:03:36 +01:00
/* Post process stack (order matters) */
EEVEE_motion_blur_draw(vedata);
EEVEE_depth_of_field_draw(vedata);
EEVEE_bloom_draw(vedata);
Eevee: Add Bloom post process. Based on Kino/Bloom v2 - Bloom filter for Unity MIT license.
2017-05-07 14:27:43 +02:00
Eevee: Initial Final Render support. TAA / multiple samples is not working at the moment.
2018-01-29 18:16:11 +01:00
/* Save the final texture and framebuffer for final transformation or read. */
effects->final_tx = effects->source_buffer;
effects->final_fb = (effects->target_buffer != fbl->main) ? fbl->main : fbl->effect_fb;
Eevee: SSR: Fixed problem with un-initialized texture. Also add another debug buffer and cleanup in effect_ssr_frag.glsl
2017-07-22 16:22:01 +02:00
/* If no post processes is enabled, buffers are still not swapped, do it now. */
SWAP_DOUBLE_BUFFERS();
Eevee: SSR: Fix Opengl Render. Add a constant number of 4 drawing loop to accumulate 4 "bounce" of light in SSRs.
2017-07-22 20:36:34 +02:00
if (!stl->g_data->valid_double_buffer &&
Cleanup: style, duplicate includes
2017-10-07 15:57:14 +11:00
((effects->enabled_effects & EFFECT_DOUBLE_BUFFER) != 0) &&
(DRW_state_is_image_render() == false))
Eevee: SSR: Fix Opengl Render. Add a constant number of 4 drawing loop to accumulate 4 "bounce" of light in SSRs.
2017-07-22 20:36:34 +02:00
{
Eevee: SSR: Fixed problem with un-initialized texture. Also add another debug buffer and cleanup in effect_ssr_frag.glsl
2017-07-22 16:22:01 +02:00
/* If history buffer is not valid request another frame.
* This fix black reflections on area resize. */
DRW_viewport_request_redraw();
}
Eevee: SSR: Add two hit option. This option add another raytrace per pixel, clearing some noise. But multiplying the raytrace cost.
2017-07-23 00:03:45 +02:00
/* Record pers matrix for the next frame. */
Eevee: Fix prev_persmat being used by multiple viewport.
2018-02-26 20:08:48 +01:00
DRW_viewport_matrix_get(stl->effects->prev_persmat, DRW_MAT_PERS);
Eevee: SSR: Add two hit option. This option add another raytrace per pixel, clearing some noise. But multiplying the raytrace cost.
2017-07-23 00:03:45 +02:00
Eevee: SSR: Fix Opengl Render. Add a constant number of 4 drawing loop to accumulate 4 "bounce" of light in SSRs.
2017-07-22 20:36:34 +02:00
/* Update double buffer status if render mode. */
if (DRW_state_is_image_render()) {
stl->g_data->valid_double_buffer = (txl->color_double_buffer != NULL);
}
Eevee: Simple Camera Motion Blur. Disabled by default. Set ENABLE_EFFECT_MOTION_BLUR to 1 to enable. No fancy motion blur. Use depth and camera matrix to get the motion vectors. Then blur in this direction. Only available in camera view. Only Camera animation is supported, does not take into account the parents motion
2017-05-04 17:36:40 +02:00
}
void EEVEE_effects_free(void)
{
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
DRW_SHADER_FREE_SAFE(e_data.downsample_sh);
Eevee: Fix problem with GPU_texture_generate_mipmap This function was called to recreate the lower mip level of the probe texture. But this is not it's usage and it introduced a stall. This patch add cubemap mipmap level regeneration in eevee_effects.c
2017-08-18 22:31:06 +02:00
DRW_SHADER_FREE_SAFE(e_data.downsample_cube_sh);
Eevee: SSR: Output ssr datas to buffers. Output in 2 buffers Normals, Specular Color and roughness. This way we can raytrace in a defered fashion and blend the exact contribution of the specular lobe on top of the opaque pass.
2017-07-16 23:49:25 +02:00
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_SHADER_FREE_SAFE(e_data.minz_downlevel_sh);
DRW_SHADER_FREE_SAFE(e_data.maxz_downlevel_sh);
DRW_SHADER_FREE_SAFE(e_data.minz_downdepth_sh);
DRW_SHADER_FREE_SAFE(e_data.maxz_downdepth_sh);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
DRW_SHADER_FREE_SAFE(e_data.minz_downdepth_layer_sh);
DRW_SHADER_FREE_SAFE(e_data.maxz_downdepth_layer_sh);
Eevee: HiZ buffer: Split into two 24bit depth buffer This way we don't have float precision issue we had before and we save some bandwidth.
2017-07-21 14:27:16 +02:00
DRW_SHADER_FREE_SAFE(e_data.minz_copydepth_sh);
DRW_SHADER_FREE_SAFE(e_data.maxz_copydepth_sh);
Pass EvaluationContext argument everywhere Note that some little parts of code have been dissabled because eval_ctx was not available there. This should be resolved once DerivedMesh is replaced.
2017-07-21 11:53:13 +02:00
}
Copy Permalink