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This repository has been archived on 2023-10-09.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
b53728b1d25ca56fd88b0a8a013e68cd961e8ab9
blender-archive/source/blender/draw/engines/eevee/eevee_effects.c

574 lines
20 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
/*
* 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.
*
Cleanup: add BEGIN/END to GPL headers
2019-01-23 11:29:18 +11:00
* Copyright 2016, Blender Foundation.
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
*/
doxygen: add newline after \file While \file doesn't need an argument, it can't have another doxy command after it.
2019-02-18 08:08:12 +11:00
/** \file
* \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"
Eevee: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
#include "BKE_global.h" /* for G.debug_value */
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 "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: Fix dot corruption on intel HD5xx/6xx series It is caused by some sync issue apparently. Adding glFlush in these two places fixes it. Caveat: it might have a small perf impact (did not measure it). The fix is limited to the affected driver/operating systems.
2018-10-31 12:32:09 +01:00
#include "GPU_state.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: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
struct GPUShader *maxz_copydepth_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: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
extern char datatoc_common_uniforms_lib_glsl[];
extern char datatoc_common_view_lib_glsl[];
extern char datatoc_bsdf_common_lib_glsl[];
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
Fix crash when rendering a scene first and then goes to eevee preview modes and cleanup The crash only occurs with Intel 3rd generation GPUs. It occurs because the program needs to be used at least once in the opengl context in which it is created.
2018-11-16 13:46:13 -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");
Eevee: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
e_data.maxz_copydepth_layer_sh = DRW_shader_create_fullscreen(
datatoc_effect_minmaxz_frag_glsl,
"#define MAX_PASS\n"
"#define COPY_DEPTH\n"
"#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: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
#define SETUP_BUFFER(tex, fb, fb_color) { \
Cleanup: use eGPU prefix for GPU enum types
2019-01-23 14:15:43 +11:00
eGPUTextureFormat format = (DRW_state_is_scene_render()) ? GPU_RGBA32F : GPU_RGBA16F; \
Cleanup: use lowercase for dimensions in function names Most API's already use this convention.
2019-03-20 18:17:16 +11:00
DRW_texture_ensure_fullscreen_2d(&tex, format, DRW_TEX_FILTER | DRW_TEX_MIPMAP); \
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
GPU_framebuffer_ensure_config(&fb, { \
GPU_ATTACHMENT_TEXTURE(dtxl->depth), \
GPU_ATTACHMENT_TEXTURE(tex), \
}); \
GPU_framebuffer_ensure_config(&fb_color, { \
GPU_ATTACHMENT_NONE, \
GPU_ATTACHMENT_TEXTURE(tex), \
}); \
Cleanup: empty expression statement warning
2019-04-02 17:54:04 +11:00
} ((void)0)
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
#define CLEANUP_BUFFER(tex, fb, fb_color) { \
/* Cleanup to release memory */ \
DRW_TEXTURE_FREE_SAFE(tex); \
GPU_FRAMEBUFFER_FREE_SAFE(fb); \
GPU_FRAMEBUFFER_FREE_SAFE(fb_color); \
Cleanup: empty expression statement warning
2019-04-02 17:54:04 +11:00
} ((void)0)
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
Eevee: Fix crash when baking Was caused by the normal pass being enbaled in the renderlayer settings. Fix T57344
2018-10-23 15:52:04 +02:00
void EEVEE_effects_init(EEVEE_ViewLayerData *sldata, EEVEE_Data *vedata, Object *camera, const bool minimal)
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;
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
DefaultTextureList *dtxl = DRW_viewport_texture_list_get();
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: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
int size_fs[2] = {(int)viewport_size[0], (int)viewport_size[1]};
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: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
effects->enabled_effects |= (G.debug_value == 9) ? EFFECT_VELOCITY_BUFFER : 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. */
Eevee: Fix crash when baking Was caused by the normal pass being enbaled in the renderlayer settings. Fix T57344
2018-10-23 15:52:04 +02:00
if (DRW_state_is_image_render() && !minimal &&
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: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
SETUP_BUFFER(txl->color_post, fbl->effect_fb, fbl->effect_color_fb);
Eevee: Initial Depth Of Field commit.
2017-05-09 21:57:13 +02:00
}
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: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
CLEANUP_BUFFER(txl->color_post, fbl->effect_fb, fbl->effect_color_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: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
const bool half_res_hiz = true;
int size[2], div;
common_data->hiz_mip_offset = (half_res_hiz) ? 1 : 0;
div = (half_res_hiz) ? 2 : 1;
size[0] = max_ii(size_fs[0] / div, 1);
size[1] = max_ii(size_fs[1] / div, 1);
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 */
Cleanup: use lowercase for dimensions in function names Most API's already use this convention.
2019-03-20 18:17:16 +11:00
DRW_texture_ensure_2d(&txl->maxzbuffer, size[0], size[1], GPU_R32F, DRW_TEX_MIPMAP);
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
}
else {
Cleanup: use lowercase for dimensions in function names Most API's already use this convention.
2019-03-20 18:17:16 +11:00
DRW_texture_ensure_2d(&txl->maxzbuffer, size[0], size[1], GPU_DEPTH_COMPONENT24, DRW_TEX_MIPMAP);
EEVEE: Fix bad framebuffer configuration Was causing black / corrupted scene because of broken downsample Add a debug check to not run into this problem again.
2018-03-29 13:43:23 +02:00
}
if (fbl->downsample_fb == NULL) {
fbl->downsample_fb = GPU_framebuffer_create();
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
/**
Merge branch 'master' into blender2.8
2018-09-19 18:19:49 +02:00
* Compute Mipmap texel alignment.
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
*/
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) {
Eevee: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
int mip_size[2];
GPU_texture_get_mipmap_size(txl->color, i, mip_size);
common_data->mip_ratio[i][0] = viewport_size[0] / (mip_size[0] * powf(2.0f, i));
common_data->mip_ratio[i][1] = viewport_size[1] / (mip_size[1] * powf(2.0f, i));
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: 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) {
Cleanup: use lowercase for dimensions in function names Most API's already use this convention.
2019-03-20 18:17:16 +11:00
effects->ssr_normal_input = DRW_texture_pool_query_2d(size_fs[0], size_fs[1], GPU_RG16,
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
&draw_engine_eevee_type);
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
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_framebuffer_texture_attach(fbl->main_fb, effects->ssr_normal_input, 1, 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
}
else {
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
effects->ssr_normal_input = NULL;
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: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
/**
* Motion vector buffer for correct TAA / motion blur.
*/
if ((effects->enabled_effects & EFFECT_VELOCITY_BUFFER) != 0) {
Cleanup: use lowercase for dimensions in function names Most API's already use this convention.
2019-03-20 18:17:16 +11:00
effects->velocity_tx = DRW_texture_pool_query_2d(size_fs[0], size_fs[1], GPU_RG16,
Eevee: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
&draw_engine_eevee_type);
/* TODO output objects velocity during the mainpass. */
// GPU_framebuffer_texture_attach(fbl->main_fb, effects->velocity_tx, 1, 0);
GPU_framebuffer_ensure_config(&fbl->velocity_resolve_fb, {
GPU_ATTACHMENT_NONE,
GPU_ATTACHMENT_TEXTURE(effects->velocity_tx)
});
}
else {
effects->velocity_tx = NULL;
}
Eevee: TAA Reprojection: Initial implementation This "improve" the viewport experience by reducing the noise from random sampling effects (SSAO, Contact Shadows, SSR) when moving the viewport or during playback. This does not do Anti Aliasing because this would conflict with the outline pass. We could enable AA jittering in "only render" mode though. There are many things to improve but this is a solid basis to build upon.
2018-04-20 18:24:14 +02:00
/**
* Setup depth double buffer.
*/
if ((effects->enabled_effects & EFFECT_DEPTH_DOUBLE_BUFFER) != 0) {
Cleanup: use lowercase for dimensions in function names Most API's already use this convention.
2019-03-20 18:17:16 +11:00
DRW_texture_ensure_fullscreen_2d(&txl->depth_double_buffer, GPU_DEPTH24_STENCIL8, 0);
Eevee: TAA Reprojection: Initial implementation This "improve" the viewport experience by reducing the noise from random sampling effects (SSAO, Contact Shadows, SSR) when moving the viewport or during playback. This does not do Anti Aliasing because this would conflict with the outline pass. We could enable AA jittering in "only render" mode though. There are many things to improve but this is a solid basis to build upon.
2018-04-20 18:24:14 +02:00
GPU_framebuffer_ensure_config(&fbl->double_buffer_depth_fb, {
GPU_ATTACHMENT_TEXTURE(txl->depth_double_buffer)
});
}
else {
/* Cleanup to release memory */
DRW_TEXTURE_FREE_SAFE(txl->depth_double_buffer);
GPU_FRAMEBUFFER_FREE_SAFE(fbl->double_buffer_depth_fb);
}
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: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
SETUP_BUFFER(txl->color_double_buffer, fbl->double_buffer_fb, fbl->double_buffer_color_fb);
}
else {
CLEANUP_BUFFER(txl->color_double_buffer, fbl->double_buffer_fb, fbl->double_buffer_color_fb);
}
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
if ((effects->enabled_effects & (EFFECT_TAA | EFFECT_TAA_REPROJECT)) != 0) {
SETUP_BUFFER(txl->taa_history, fbl->taa_history_fb, fbl->taa_history_color_fb);
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
}
else {
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
CLEANUP_BUFFER(txl->taa_history, fbl->taa_history_fb, fbl->taa_history_color_fb);
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
}
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: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
EEVEE_StorageList *stl = vedata->stl;
EEVEE_EffectsInfo *effects = stl->effects;
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
GWN: Port to GPU module: Replace GWN prefix by GPU
2018-07-18 00:12:21 +02:00
struct GPUBatch *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
{
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +02:00
psl->color_downsample_ps = DRW_pass_create(
"Downsample", DRW_STATE_WRITE_COLOR);
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
DRWShadingGroup *grp = DRW_shgroup_create(e_data.downsample_sh, psl->color_downsample_ps);
DRW: Rename DRW_shgroup_uniform_buffer into DRW_shgroup_uniform_texture_ref This is in order to not mix it with the incomming buffer textures.
2018-03-25 19:24:19 +02:00
DRW_shgroup_uniform_texture_ref(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;
Cleanup: use 'uint' in draw manager
2018-05-11 07:48:52 +02:00
static uint six = 6;
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +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);
DRW: Rename DRW_shgroup_uniform_buffer into DRW_shgroup_uniform_texture_ref This is in order to not mix it with the incomming buffer textures.
2018-03-25 19:24:19 +02:00
DRW_shgroup_uniform_texture_ref(grp, "source", &e_data.color_src);
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_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;
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +02:00
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: Rename DRW_shgroup_uniform_buffer into DRW_shgroup_uniform_texture_ref This is in order to not mix it with the incomming buffer textures.
2018-03-25 19:24:19 +02:00
DRW_shgroup_uniform_texture_ref(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: 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
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +02:00
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: Rename DRW_shgroup_uniform_buffer into DRW_shgroup_uniform_texture_ref This is in order to not mix it with the incomming buffer textures.
2018-03-25 19:24:19 +02:00
DRW_shgroup_uniform_texture_ref(grp, "depthBuffer", &e_data.depth_src);
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_call_add(grp, quad, NULL);
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +02:00
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: Rename DRW_shgroup_uniform_buffer into DRW_shgroup_uniform_texture_ref This is in order to not mix it with the incomming buffer textures.
2018-03-25 19:24:19 +02:00
DRW_shgroup_uniform_texture_ref(grp, "depthBuffer", &e_data.depth_src);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
DRW_shgroup_uniform_int(grp, "depthLayer", &e_data.depth_src_layer, 1);
DRW_shgroup_call_add(grp, quad, NULL);
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +02:00
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);
DRW: Rename DRW_shgroup_uniform_buffer into DRW_shgroup_uniform_texture_ref This is in order to not mix it with the incomming buffer textures.
2018-03-25 19:24:19 +02:00
DRW_shgroup_uniform_texture_ref(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: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
psl->maxz_copydepth_layer_ps = DRW_pass_create(
"HiZ Max Copy DepthLayer Halfres", downsample_write | DRW_STATE_DEPTH_ALWAYS);
grp = DRW_shgroup_create(e_data.maxz_copydepth_layer_sh, psl->maxz_copydepth_layer_ps);
DRW_shgroup_uniform_texture_ref(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: 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: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
if ((effects->enabled_effects & EFFECT_VELOCITY_BUFFER) != 0) {
/* This pass compute camera motions to the non moving objects. */
Cleanup: whitespace, long lines, duplicate include
2018-05-26 10:41:25 +02:00
psl->velocity_resolve = DRW_pass_create(
"Velocity Resolve", DRW_STATE_WRITE_COLOR);
Fix crash when rendering a scene first and then goes to eevee preview modes and cleanup The crash only occurs with Intel 3rd generation GPUs. It occurs because the program needs to be used at least once in the opengl context in which it is created.
2018-11-16 13:46:13 -02:00
DRWShadingGroup *grp = DRW_shgroup_create(EEVEE_shaders_velocity_resolve_sh_get(), psl->velocity_resolve);
Eevee: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
DRW_shgroup_uniform_texture_ref(grp, "depthBuffer", &e_data.depth_src);
DRW_shgroup_uniform_block(grp, "common_block", sldata->common_ubo);
DRW_shgroup_uniform_mat4(grp, "currPersinv", effects->velocity_curr_persinv);
DRW_shgroup_uniform_mat4(grp, "pastPersmat", effects->velocity_past_persmat);
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 */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_framebuffer_texture_attach(fbl->downsample_fb, stl->g_data->minzbuffer, 0, 0);
GPU_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);
}
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_framebuffer_texture_detach(stl->g_data->minzbuffer);
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
/* Create lower levels */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_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
Fix T56800: Reflection Plane in Eevee crashes Blender
2018-09-17 12:20:44 +02:00
int minmax_size[3], depth_size[3];
Eevee: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
GPU_texture_get_mipmap_size(depth_src, 0, depth_size);
GPU_texture_get_mipmap_size(txl->maxzbuffer, 0, minmax_size);
bool is_full_res_minmaxz = (minmax_size[0] == depth_size[0] && minmax_size[1] == depth_size[1]);
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 */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_framebuffer_texture_attach(fbl->downsample_fb, txl->maxzbuffer, 0, 0);
GPU_framebuffer_bind(fbl->downsample_fb);
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
if (layer >= 0) {
Eevee: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
if (is_full_res_minmaxz) {
DRW_draw_pass(psl->maxz_copydepth_layer_ps);
}
else {
DRW_draw_pass(psl->maxz_downdepth_layer_ps);
}
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +02:00
}
else {
Eevee: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
if (is_full_res_minmaxz) {
DRW_draw_pass(psl->maxz_copydepth_ps);
}
else {
DRW_draw_pass(psl->maxz_downdepth_ps);
}
Eevee: Make MinmaxZ compatible with textureArray
2017-07-23 20:33:29 +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
/* Create lower levels */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_framebuffer_recursive_downsample(fbl->downsample_fb, 8, &max_downsample_cb, vedata);
EEVEE: Fix bad framebuffer configuration Was causing black / corrupted scene because of broken downsample Add a debug check to not run into this problem again.
2018-03-29 13:43:23 +02:00
GPU_framebuffer_texture_detach(fbl->downsample_fb, txl->maxzbuffer);
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 */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
GPU_framebuffer_bind(fbl->main_fb);
Eevee: Fix dot corruption on intel HD5xx/6xx series It is caused by some sync issue apparently. Adding glFlush in these two places fixes it. Caveat: it might have a small perf impact (did not measure it). The fix is limited to the affected driver/operating systems.
2018-10-31 12:32:09 +01:00
Fix T58266 : Bottom half of meshes are transparent with eevee render
2018-12-11 13:29:32 +01:00
if (GPU_mip_render_workaround() ||
GPU_type_matches(GPU_DEVICE_INTEL_UHD, GPU_OS_WIN, GPU_DRIVER_ANY))
{
/* Fix dot corruption on intel HD5XX/HD6XX series. */
Eevee: Fix dot corruption on intel HD5xx/6xx series It is caused by some sync issue apparently. Adding glFlush in these two places fixes it. Caveat: it might have a small perf impact (did not measure it). The fix is limited to the affected driver/operating systems.
2018-10-31 12:32:09 +01:00
GPU_flush();
}
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.
Cleanup: comment blocks
2019-03-19 15:17:46 +11:00
*/
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
void EEVEE_downsample_buffer(EEVEE_Data *vedata, GPUTexture *texture_src, int level)
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
{
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
EEVEE_FramebufferList *fbl = vedata->fbl;
Eevee: SSR: Add mipmap filtering and bias to reduce noise. Also fix the roughness factors.
2017-07-20 18:40:23 +02:00
e_data.color_src = texture_src;
/* Create lower levels */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
DRW_stats_group_start("Downsample buffer");
GPU_framebuffer_texture_attach(fbl->downsample_fb, texture_src, 0, 0);
GPU_framebuffer_recursive_downsample(fbl->downsample_fb, level, &simple_downsample_cb, vedata);
EEVEE: Fix bad framebuffer configuration Was causing black / corrupted scene because of broken downsample Add a debug check to not run into this problem again.
2018-03-29 13:43:23 +02:00
GPU_framebuffer_texture_detach(fbl->downsample_fb, texture_src);
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.
Cleanup: comment blocks
2019-03-19 15:17:46 +11:00
*/
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
void EEVEE_downsample_cube_buffer(EEVEE_Data *vedata, GPUTexture *texture_src, int level)
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
{
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
EEVEE_FramebufferList *fbl = vedata->fbl;
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
e_data.color_src = texture_src;
/* Create lower levels */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
DRW_stats_group_start("Downsample Cube buffer");
GPU_framebuffer_texture_attach(fbl->downsample_fb, texture_src, 0, 0);
GPU_framebuffer_recursive_downsample(fbl->downsample_fb, level, &simple_downsample_cube_cb, vedata);
EEVEE: Fix bad framebuffer configuration Was causing black / corrupted scene because of broken downsample Add a debug check to not run into this problem again.
2018-03-29 13:43:23 +02:00
GPU_framebuffer_texture_detach(fbl->downsample_fb, texture_src);
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_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 Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
EEVEE_PassList *psl = vedata->psl;
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;
Eevee: Add Velocity pass. This pass create a velocity buffer which is basically a 2D motion vector texture. This is not yet used for rendering but will be usefull for motion blur and temporal reprojection.
2018-04-20 18:18:33 +02:00
/* First resolve the velocity. */
if ((effects->enabled_effects & EFFECT_VELOCITY_BUFFER) != 0) {
DRW_viewport_matrix_get(effects->velocity_curr_persinv, DRW_MAT_PERSINV);
GPU_framebuffer_bind(fbl->velocity_resolve_fb);
DRW_draw_pass(psl->velocity_resolve);
}
DRW_viewport_matrix_get(effects->velocity_past_persmat, DRW_MAT_PERS);
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 */
GPUFramebuffer: Refactor (Part 2) This refactor modernise the use of framebuffers. It also touches a lot of files so breaking down changes we have: - GPUTexture: Allow textures to be attached to more than one GPUFrameBuffer. This allows to create and configure more FBO without the need to attach and detach texture at drawing time. - GPUFrameBuffer: The wrapper starts to mimic opengl a bit closer. This allows to configure the framebuffer inside a context other than the one that will be rendering the framebuffer. We do the actual configuration when binding the FBO. We also Keep track of config validity and save drawbuffers state in the FBO. We remove the different bind/unbind functions. These make little sense now that we have separate contexts. - DRWFrameBuffer: We replace DRW_framebuffer functions by GPU_framebuffer ones to avoid another layer of abstraction. We move the DRW convenience functions to GPUFramebuffer instead and even add new ones. The MACRO GPU_framebuffer_ensure_config is pretty much all you need to create and config a GPUFramebuffer. - DRWTexture: Due to the removal of DRWFrameBuffer, we needed to create functions to create textures for thoses framebuffers. Pool textures are now using default texture parameters for the texture type asked. - DRWManager: Make sure no framebuffer object is bound when doing cache filling. - GPUViewport: Add new color_only_fb and depth_only_fb along with FB API usage update. This let draw engines render to color/depth only target and without the need to attach/detach textures. - WM_window: Assert when a framebuffer is bound when changing context. This balance the fact we are not track ogl context inside GPUFramebuffer. - Eevee, Clay, Mode engines: Update to new API. This comes with a lot of code simplification. This also come with some cleanups in some engine codes.
2018-03-25 17:46:48 +02:00
effects->target_buffer = fbl->effect_color_fb; /* next target to render to */
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: 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: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
EEVEE_temporal_sampling_draw(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
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;
Eevee: Fix AA in render mode. Was due to wrong framebuffer being read because of wrong comparisson caused by b9ec7a92725e1567c12475a9eb59191edca1dc32
2018-04-17 15:32:30 +02:00
effects->final_fb = (effects->target_buffer != fbl->main_color_fb) ? fbl->main_fb : fbl->effect_fb;
Cleanup: style/whitespace Also use 'uint'.
2018-05-30 19:45:03 +02:00
if ((effects->enabled_effects & EFFECT_TAA) &&
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
(effects->source_buffer == txl->taa_history))
Cleanup: style/whitespace Also use 'uint'.
2018-05-30 19:45:03 +02:00
{
Eevee: Depth of Feild: Fix ringing issue of background objects There was an issue caused by Antialiasing being done after DoF. Move TAA after DOF and Motion Blur. Also certain pixel with lower CoC would be spread all over the background because the neighbooring pixel have higher CoC. So we need to apply some bilateral filtering when downsampling. Currently we limit the influence of neighbor pixels with a CoC inside the range [MaxCoC-2, MaxCoC].
2018-09-09 23:24:06 +02:00
effects->final_fb = fbl->taa_history_fb;
EEvEE: LookDev
2018-05-25 08:06:36 +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 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);
Fix build for MSVC: Remove trailing double semicolon Not sure why but MSVC is complaining for some of those. In particular for the struct in BKE_subdiv_ccg.h. Those were the ones crashing here..
2018-09-20 14:47:10 +00:00
stl->g_data->valid_taa_history = (txl->taa_history != NULL);
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: 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: Prepare for fullres tracing
2018-09-11 16:15:34 +02:00
DRW_SHADER_FREE_SAFE(e_data.maxz_copydepth_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