blender
/
blender
75
260
Fork 349
Code Issues 5.9k Pull Requests 543 Projects 19 Wiki Activity

Path Guiding: Adding guiding on glossy surfaces via RIS #107782

Merged
Sebastian Herholz merged 1 commits from sherholz/blender:cycles_path_guiding_RIS into main 2023-05-22 16:47:10 +02:00
Conversation 17 Commits 1 Files Changed 11
Sebastian Herholz commented 2023-05-09 15:13:42 +02:00
Collaborator

This PR adds support for path guiding on glossy materials.
Two new guiding modes are added Re-sampled Importance Sampling(RIS) and Roughness-based.
The first uses Re-Sampling Importance Sampling to with two candidates (one from the BSDF and one from the guiding distribution).
the second adjusts the guiding probability based on material roughness (specular -> no guiding and diffuse -> guiding (50%)).

After some intensive tests by the community (https://blenderartists.org/t/cycles-path-guiding-tests/1388137)
I choose to set RIS as the new default method.

Todo:

  • clean up code in surface_shader.h

Example:
image

UI-changes:
image

This PR adds support for path guiding on glossy materials. Two new guiding modes are added `Re-sampled Importance Sampling`(RIS) and `Roughness-based`. The first uses Re-Sampling Importance Sampling to with two candidates (one from the BSDF and one from the guiding distribution). the second adjusts the guiding probability based on material roughness (specular -> no guiding and diffuse -> guiding (50%)). After some intensive tests by the community (https://blenderartists.org/t/cycles-path-guiding-tests/1388137) I choose to set `RIS` as the new default method. Todo: - clean up code in `surface_shader.h` Example: ![image](/attachments/c65d6ec9-90a1-45b8-b8b9-75094e1fb190) UI-changes: ![image](/attachments/aa9604e2-0ad2-4901-97a4-aebc7e9e4dd1)
bitmap.png
4.0 MiB
image.png
36 KiB
👀 2 ❤️ 3 🎉 3 🚀 3
Brecht Van Lommel was assigned by Sebastian Herholz 2023-05-09 15:13:43 +02:00
Sergey Sharybin was assigned by Sebastian Herholz 2023-05-09 15:13:43 +02:00
Brecht Van Lommel removed their assignment 2023-05-09 15:36:13 +02:00
Sergey Sharybin was unassigned by Brecht Van Lommel 2023-05-09 15:36:13 +02:00
Brecht Van Lommel requested review from Brecht Van Lommel 2023-05-09 15:36:25 +02:00
Brecht Van Lommel requested review from Weizhen Huang 2023-05-09 15:36:25 +02:00
Brecht Van Lommel requested review from Sergey Sharybin 2023-05-09 15:36:35 +02:00
Sergey Sharybin requested changes 2023-05-10 09:49:16 +02:00
Sergey Sharybin left a comment
Owner

It is a bit hard to see the exact difference when images are put on the same canvas. It would be helpful to have them as individual images to help doing pixel-wise comparison.

For the patch itself I do not find the direction of adding more options something that fits the Blender vision, as well as it adds permutations for developers to ensure be working.

If there is a strong case why all of the guiding methods are needed then it needs to be documented, ideally in a way that is clear for artists.
Otherwise I'd propose to stick to one method and focus on ensuring it works the best possible way.

What I am also curios is the performance comparison on GPU.

It is a bit hard to see the exact difference when images are put on the same canvas. It would be helpful to have them as individual images to help doing pixel-wise comparison. For the patch itself I do not find the direction of adding more options something that fits the Blender vision, as well as it adds permutations for developers to ensure be working. If there is a strong case why all of the guiding methods are needed then it needs to be documented, ideally in a way that is clear for artists. Otherwise I'd propose to stick to one method and focus on ensuring it works the best possible way. What I am also curios is the performance comparison on GPU.
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -21,9 +21,33 @@
CCL_NAMESPACE_BEGIN
#define RIS_COSINE
Sergey Sharybin commented 2023-05-10 09:43:24 +02:00
Owner

Is it possible to remove the permutation of the RIS_COSINE and RIS_INCOMMING_RADIANCE?

Is it possible to remove the permutation of the `RIS_COSINE` and `RIS_INCOMMING_RADIANCE`?
Sebastian Herholz commented 2023-05-10 12:05:24 +02:00
Poster
Collaborator

Done,
The other define # define RIS_INCOMING_RADIANCE will be gone after we bumped the Open PGL to 0.5.0

Done, The other define `# define RIS_INCOMING_RADIANCE` will be gone after we bumped the Open PGL to 0.5.0
brecht marked this conversation as resolved
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -22,2 +22,4 @@
CCL_NAMESPACE_BEGIN
#define RIS_COSINE
#if OPENPGL_VERSION_MINOR >= 5
Sergey Sharybin commented 2023-05-10 09:46:28 +02:00
Owner

I think we can make the 0.5 version required.

Other than that, I am not a fan of such implicit assumption that the OPENPGL_VERSION_MAJOR is 0. We really need to have an utility macro like I've shown in the comment in the #106861.

I think we can make the 0.5 version required. Other than that, I am not a fan of such implicit assumption that the `OPENPGL_VERSION_MAJOR` is 0. We really need to have an utility macro like I've shown in the comment in the #106861.
Sebastian Herholz commented 2023-05-10 11:14:30 +02:00
Poster
Collaborator

I already talked with Ray. As soon as every dependency package updated to Open PGL 0.5 I would like to bump the minimal required version to 0.5 and get rid of all these defines.

I already talked with Ray. As soon as every dependency package updated to Open PGL 0.5 I would like to bump the minimal required version to 0.5 and get rid of all these defines.
Brecht Van Lommel commented 2023-05-16 13:50:12 +02:00
Owner

OpenPGL was updated to 0.5 for all platforms, so this can be simplified.

OpenPGL was updated to 0.5 for all platforms, so this can be simplified.
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -530,2 +594,4 @@
return label;
}
ccl_device int surface_shader_bsdf_guided_sample_closure_ris(KernelGlobals kg,
Sergey Sharybin commented 2023-05-10 09:48:02 +02:00
Owner

Is there a link to a paper we can add here to help understanding all the math which is going on in the function?

Is there a link to a paper we can add here to help understanding all the math which is going on in the function?
Sebastian Herholz commented 2023-05-10 11:13:14 +02:00
Poster
Collaborator

This is an extension/variant of the standard RIS where I added some stuff.
A detailed description of the method with all the insides will be part of the Siggraph2023 Course on path guiding, I can add a link as soon as the course notes are online.

This is an extension/variant of the standard RIS where I added some stuff. A detailed description of the method with all the insides will be part of the Siggraph2023 Course on path guiding, I can add a link as soon as the course notes are online.
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -532,0 +639,4 @@
int label_ris[2];
float3 omega_in_ris[2];
float ris_pdfs[2] = {0.f, 0.f};
Sergey Sharybin commented 2023-05-10 09:36:18 +02:00
Owner

Here and everywhere else in the patch: do not omit the 0 suffix (0.0f, not 0.f).

This is per the style guide: https://wiki.blender.org/wiki/Style_Guide/C_Cpp#Value_Literals

Here and everywhere else in the patch: do not omit the 0 suffix (0.0f, not 0.f). This is per the style guide: https://wiki.blender.org/wiki/Style_Guide/C_Cpp#Value_Literals
Sebastian Herholz commented 2023-05-10 14:38:06 +02:00
Poster
Collaborator

Done

Done
brecht marked this conversation as resolved
Sebastian Herholz commented 2023-05-10 11:37:01 +02:00
Poster
Collaborator

It is a bit hard to see the exact difference when images are put on the same canvas. It would be helpful to have them as individual images to help doing pixel-wise comparison.

For the patch itself I do not find the direction of adding more options something that fits the Blender vision, as well as it adds permutations for developers to ensure be working.

If there is a strong case why all of the guiding methods are needed then it needs to be documented, ideally in a way that is clear for artists.
Otherwise I'd propose to stick to one method and focus on ensuring it works the best possible way.

What I am also curios is the performance comparison on GPU.

@Sergey I understand your concerns, ideally I would also prefer to only have one guiding on/off button.
Unfortunately, the is no THE IDEAL WAY to integrate path guiding into a production renderer and I try to keep some option, also for debugging and testing purposes.

It helped a lot to get feedback from the community in nailing down which version seems to work best in practice.

As you can see all these options are only available under the Debug options, which should not be touched be the artists in their daily work. Unfortunately they do not really listen :( and play around with every button they find :).

I was thinking about hiding these options behind the WITH_CYCLES_DEBUG flag so that it is still possible to test and validate different options.

For GPU timings we will see when we have GPU support in Open PGL.
I am currently refactoring the RIS code, I think when porting to the GPU it can be made a little bit more GPU friendly when combined with reservoir sampling.

About the examples, here are some links from the community tests:

https://blenderartists.org/t/cycles-path-guiding-tests/1388137/673
https://blenderartists.org/t/cycles-path-guiding-tests/1388137/659
https://blenderartists.org/t/cycles-path-guiding-tests/1388137/656
https://blenderartists.org/t/cycles-path-guiding-tests/1388137/597

> It is a bit hard to see the exact difference when images are put on the same canvas. It would be helpful to have them as individual images to help doing pixel-wise comparison. > > For the patch itself I do not find the direction of adding more options something that fits the Blender vision, as well as it adds permutations for developers to ensure be working. > > If there is a strong case why all of the guiding methods are needed then it needs to be documented, ideally in a way that is clear for artists. > Otherwise I'd propose to stick to one method and focus on ensuring it works the best possible way. > > What I am also curios is the performance comparison on GPU. @Sergey I understand your concerns, ideally I would also prefer to only have one guiding on/off button. Unfortunately, the is no `THE IDEAL WAY` to integrate path guiding into a production renderer and I try to keep some option, also for debugging and testing purposes. It helped a lot to get feedback from the community in nailing down which version seems to work best in practice. As you can see all these options are only available under the Debug options, which should not be touched be the artists in their daily work. Unfortunately they do not really listen :( and play around with every button they find :). I was thinking about hiding these options behind the `WITH_CYCLES_DEBUG` flag so that it is still possible to test and validate different options. For GPU timings we will see when we have GPU support in Open PGL. I am currently refactoring the RIS code, I think when porting to the GPU it can be made a little bit more GPU friendly when combined with reservoir sampling. About the examples, here are some links from the community tests: https://blenderartists.org/t/cycles-path-guiding-tests/1388137/673 https://blenderartists.org/t/cycles-path-guiding-tests/1388137/659 https://blenderartists.org/t/cycles-path-guiding-tests/1388137/656 https://blenderartists.org/t/cycles-path-guiding-tests/1388137/597
Sergey Sharybin commented 2023-05-11 09:51:21 +02:00
Owner

Thanks for the updates.

The point about the non-existing ideal algorithm applies to other areas of 3d software as well. It is fine to have extra options during development and feedback period, but for the final merge we really try hard to avoid adding them.

As you've said, no matter how well you hide them, artists use all the options, so there is always a maintenance cost (which is not only limited to code side, but also communication with artists).

The extra annoying part in the Cycles kernel is also that unused code paths could affect performance on GPU, so we always look at those with extra care.

Is it possible to limit options to 2: RIS (the officially used one) and some simple implementation to help nailing down issues between Cycles side, OpenPGL side, and weight computation for the guiding?

Also, to confirm, all the changes are purely within surrounding __PATH_GUIDING__ ifdef, so no changes for the current GPUs? (is just something that is not so trivially visible from the diff, so wanted to confirm with you).

Btw, I've also notice you're using std::isfinite. This would not work on all devices/compilers we support as we use -fast-math. The portable way to do so in Cycles is isnan_safe and isfinite_safe.

Thanks for the updates. The point about the non-existing ideal algorithm applies to other areas of 3d software as well. It is fine to have extra options during development and feedback period, but for the final merge we really try hard to avoid adding them. As you've said, no matter how well you hide them, artists use all the options, so there is always a maintenance cost (which is not only limited to code side, but also communication with artists). The extra annoying part in the Cycles kernel is also that unused code paths could affect performance on GPU, so we always look at those with extra care. Is it possible to limit options to 2: RIS (the officially used one) and some simple implementation to help nailing down issues between Cycles side, OpenPGL side, and weight computation for the guiding? Also, to confirm, all the changes are purely within surrounding `__PATH_GUIDING__` ifdef, so no changes for the current GPUs? (is just something that is not so trivially visible from the diff, so wanted to confirm with you). Btw, I've also notice you're using `std::isfinite`. This would not work on all devices/compilers we support as we use `-fast-math`. The portable way to do so in Cycles is `isnan_safe` and `isfinite_safe`.
Sergey Sharybin requested review from Sergey Sharybin 2023-05-11 14:32:53 +02:00
Sergey Sharybin commented 2023-05-11 14:36:23 +02:00
Owner

We talked with Sebastian outside of the code review page.

The important point I understand is that we might want to come back to the MIS/Roughness approaches later on.
We also confirmed all the current changes do not affect current GPU code at all.

The WITH_CYCLES_DEBUG idea seems interesting, but might be better to handle it separately from this patch.

So that pretty much resolves majority of my questions :)

The remaining code aspect I've noted is the finite/nan story. To me it seems easier to just use our "safe" variants, and safe hair-pulling later on.

Sebastian mentioned some cleanup is still planned, so I'll wait for the patch to leave the WIP state to do the final pass of review :)

We talked with Sebastian outside of the code review page. The important point I understand is that we might want to come back to the MIS/Roughness approaches later on. We also confirmed all the current changes do not affect current GPU code at all. The `WITH_CYCLES_DEBUG` idea seems interesting, but might be better to handle it separately from this patch. So that pretty much resolves majority of my questions :) The remaining code aspect I've noted is the finite/nan story. To me it seems easier to just use our "safe" variants, and safe hair-pulling later on. Sebastian mentioned some cleanup is still planned, so I'll wait for the patch to leave the WIP state to do the final pass of review :)
Sebastian Herholz commented 2023-05-15 11:14:58 +02:00
Poster
Collaborator

@blender-bot package

@blender-bot package
Blender Bot commented 2023-05-15 11:15:01 +02:00
Collaborator

Package build started. Download here when ready.

Package build started. [Download here](https://builder.blender.org/download/patch/PR107782) when ready.
Sebastian Herholz force-pushed cycles_path_guiding_RIS from 9a62522681 to aea5dc6ab9 2023-05-15 13:26:08 +02:00 Compare
Sebastian Herholz commented 2023-05-15 13:26:58 +02:00
Poster
Collaborator

@blender-bot package

@blender-bot package
Blender Bot commented 2023-05-15 13:27:00 +02:00
Collaborator

Package build started. Download here when ready.

Package build started. [Download here](https://builder.blender.org/download/patch/PR107782) when ready.
Sebastian Herholz force-pushed cycles_path_guiding_RIS from aea5dc6ab9 to 2f30bd90b1 2023-05-15 16:05:47 +02:00 Compare
Sebastian Herholz commented 2023-05-15 16:06:15 +02:00
Poster
Collaborator

@blender-bot package

@blender-bot package
Blender Bot commented 2023-05-15 16:06:17 +02:00
Collaborator

Package build started. Download here when ready.

Package build started. [Download here](https://builder.blender.org/download/patch/PR107782) when ready.
Sebastian Herholz changed title from WIP: Path Guiding: Adding guiding on glossy surfaces via RIS to Path Guiding: Adding guiding on glossy surfaces via RIS 2023-05-15 16:06:26 +02:00
Sergey Sharybin reviewed 2023-05-15 16:24:43 +02:00
intern/cycles/kernel/integrator/guiding.h Outdated
@ -14,0 +37,4 @@
Spectrum eval{zero_spectrum()};
};
ccl_device_forceinline bool calculate_ris_target(ccl_private GuidingRISSample *risSample,
Sergey Sharybin commented 2023-05-15 16:15:49 +02:00
Owner

We use snake style for variables, so ris_sample .

We use snake style for variables, so `ris_sample `.
Sebastian Herholz commented 2023-05-16 08:43:13 +02:00
Poster
Collaborator

Done

Done
brecht marked this conversation as resolved
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -67,3 +98,2 @@
/* Init guiding (diffuse BSDFs only for now). */
if (!(diffuse_sampling_fraction > 0.0f &&
guiding_bsdf_init(kg, state, sd->P, sd->N, rand_bsdf_guiding)))
if (!fully_opaque || avg_roughness < guiding_roughness_threshold * guiding_roughness_threshold ||
Sergey Sharybin commented 2023-05-15 16:15:04 +02:00
Owner

Units seems a bit odd: comparison of toughness with roughness^2. Is this expected?

Units seems a bit odd: comparison of toughness with roughness^2. Is this expected?
Sebastian Herholz commented 2023-05-16 11:08:39 +02:00
Poster
Collaborator

This is due to the fact that Blender/Cycles uses roughness^2 internally and squared roughness returns alpha.x * alpha.y so it is actually (roughness^2)^2 compared to the user input.
I refactored everything to make this more clear and convert the guiding_roughness_threshold to be guiding_roughness_threshold^2.

This is due to the fact that Blender/Cycles uses `roughness^2` internally and squared roughness returns `alpha.x * alpha.y` so it is actually `(roughness^2)^2` compared to the user input. I refactored everything to make this more clear and convert the `guiding_roughness_threshold` to be `guiding_roughness_threshold^2`.
brecht marked this conversation as resolved
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -532,0 +712,4 @@
sum_ris_weights += risSamples[1].ris_weight;
num_ris_candidates++;
}
kernel_assert(sum_ris_weights >= 0.0f);
Sergey Sharybin commented 2023-05-15 16:23:59 +02:00
Owner

Should this be instead kernel_assert(risSamples[1].ris_weight >= 0.0f) ?
Otherwise if the first RIS sample weight is 0.5, and the second RIS sample is -0.1 we will not catch the issue with this assert.

Should this be instead `kernel_assert(risSamples[1].ris_weight >= 0.0f)` ? Otherwise if the first RIS sample weight is 0.5, and the second RIS sample is -0.1 we will not catch the issue with this assert.
Sebastian Herholz commented 2023-05-16 11:05:45 +02:00
Poster
Collaborator

I agree, I changed it to test in both cases for the ris_weight as well

I agree, I changed it to test in both cases for the `ris_weight` as well
brecht marked this conversation as resolved
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -532,0 +806,4 @@
unguided_bsdf_pdf,
sampled_roughness,
eta);
# if 0
Sergey Sharybin commented 2023-05-15 16:24:28 +02:00
Owner

Is it intended for the merge?

Is it intended for the merge?
Sebastian Herholz commented 2023-05-16 11:10:53 +02:00
Poster
Collaborator

It is mainly there to be able to validate the label, roughness and eta estimation.
this will be useful when the BSDF models change (e.g., after Principle_v2).

Do you have a better idea or suggestion?

It is mainly there to be able to validate the label, roughness and eta estimation. this will be useful when the BSDF models change (e.g., after Principle_v2). Do you have a better idea or suggestion?
Sergey Sharybin commented 2023-05-16 11:23:07 +02:00
Owner

Not really, was just curious if it is something left over from earlier development, or still intended to be useful for debugging.

One idea could be to use #ifdef WITH_CYCLES_DEBUG, but not really sure.

Not really, was just curious if it is something left over from earlier development, or still intended to be useful for debugging. One idea could be to use `#ifdef WITH_CYCLES_DEBUG`, but not really sure.
Brecht Van Lommel commented 2023-05-16 13:52:44 +02:00
Owner

I think it's fine as is, but could use a comment.

I think it's fine as is, but could use a comment.
Sebastian Herholz commented 2023-05-16 14:10:00 +02:00
Poster
Collaborator

done

done
Sebastian Herholz commented 2023-05-16 14:10:01 +02:00
Poster
Collaborator

done

done
Brecht Van Lommel reviewed 2023-05-15 16:59:35 +02:00
Brecht Van Lommel left a comment
Owner

The overall logic and code organization seems reasonable.

It would be nice to do a refactor at some point after this to make a struct like BsdfSample and reduce the number of individual arguments and variables used in code like this.

The overall logic and code organization seems reasonable. It would be nice to do a refactor at some point after this to make a struct like `BsdfSample` and reduce the number of individual arguments and variables used in code like this.
intern/cycles/kernel/integrator/surface_shader.h Outdated
@ -532,0 +691,4 @@
risSamples[1].bsdf_pdf = surface_shader_bsdf_eval_pdfs(
kg, sd, risSamples[1].wo, &risSamples[1].bsdf_eval, unguided_bsdf_pdfs, 0);
risSamples[1].label = risSamples[0].label;
risSamples[1].avg_bsdf_eval = (risSamples[1].bsdf_eval.sum[0] +
Brecht Van Lommel commented 2023-05-15 16:32:33 +02:00
Owner

Use average()

Use `average()`
Sebastian Herholz commented 2023-05-16 08:42:15 +02:00
Poster
Collaborator

Done

Done
sherholz marked this conversation as resolved
Sebastian Herholz commented 2023-05-15 17:03:00 +02:00
Poster
Collaborator

@Sergey I switched to the safe checks and polished the code a little bet.
It is ready for your final review now ;)

@Sergey I switched to the safe checks and polished the code a little bet. It is ready for your final review now ;)
Sergey Sharybin commented 2023-05-15 17:38:52 +02:00
Owner

@sherholz Thanks for the update! There seems to be some out-of-order comments: we checked on the updated code and think the final round of comments is there already :)

@sherholz Thanks for the update! There seems to be some out-of-order comments: we checked on the updated code and think the final round of comments is there already :)
Sergey Sharybin commented 2023-05-16 11:39:33 +02:00
Owner

@blender-bot package

@blender-bot package
Blender Bot commented 2023-05-16 11:39:35 +02:00
Collaborator

Package build started. Download here when ready.

Package build started. [Download here](https://builder.blender.org/download/patch/PR107782) when ready.
Brecht Van Lommel requested changes 2023-05-16 13:54:56 +02:00
Brecht Van Lommel left a comment
Owner

Only thing I would like to see change still is bumping the minimum OpenPGL version.

After that this is ok to go into main for 4.0 (in two days from now).

Only thing I would like to see change still is bumping the minimum OpenPGL version. After that this is ok to go into main for 4.0 (in two days from now).
Sebastian Herholz commented 2023-05-16 14:00:59 +02:00
Poster
Collaborator

Only thing I would like to see change still is bumping the minimum OpenPGL version.

After that this is ok to go into main for 4.0 (in two days from now).

@brecht bumping the version is on my list.
I am pretty much waiting dependency builder and svns to be updated:
#104895

But don't you think that would fit better in a separate PR?
This would also give me the chance to clean up the code and remove older version checks separately.

When you mean 4.0 I assume we are too late for 3.6?

> Only thing I would like to see change still is bumping the minimum OpenPGL version. > > After that this is ok to go into main for 4.0 (in two days from now). @brecht bumping the version is on my list. I am pretty much waiting dependency builder and svns to be updated: https://projects.blender.org/blender/blender/issues/104895 But don't you think that would fit better in a separate PR? This would also give me the chance to clean up the code and remove older version checks separately. When you mean 4.0 I assume we are too late for 3.6?
Brecht Van Lommel commented 2023-05-16 15:49:52 +02:00
Owner

I am pretty much waiting dependency builder and svns to be updated:
#104895

It's done already, all platforms have OpenPGL 0.5.

But don't you think that would fit better in a separate PR?

Ok, it can be done in a separate PR too.

When you mean 4.0 I assume we are too late for 3.6?

Yes, I'd rather not make any bigger changes in 3.6 anymore this close to bcon3.

> I am pretty much waiting dependency builder and svns to be updated: > https://projects.blender.org/blender/blender/issues/104895 It's done already, all platforms have OpenPGL 0.5. > But don't you think that would fit better in a separate PR? Ok, it can be done in a separate PR too. > When you mean 4.0 I assume we are too late for 3.6? Yes, I'd rather not make any bigger changes in 3.6 anymore this close to bcon3.
Brecht Van Lommel approved these changes 2023-05-16 15:50:02 +02:00
Sebastian Herholz commented 2023-05-16 16:37:28 +02:00
Poster
Collaborator

I am pretty much waiting dependency builder and svns to be updated:
#104895

It's done already, all platforms have OpenPGL 0.5.

Ahh OK, I wasn't sure since there are still to points missing:
Add Package file to SVN
and
Update install_linux_packages script

But don't you think that would fit better in a separate PR?

Ok, it can be done in a separate PR too.

I will start preparing a separate one as soon as this PR is merged.

When you mean 4.0 I assume we are too late for 3.6?

Yes, I'd rather not make any bigger changes in 3.6 anymore this close to bcon3.

I completely understand. I just asked so that I can forward the info to the community in the thread.
For 4.0 I will probably also cover translucent materials and have better support for equi-angular sampling.

> > I am pretty much waiting dependency builder and svns to be updated: > > https://projects.blender.org/blender/blender/issues/104895 > > It's done already, all platforms have OpenPGL 0.5. > Ahh OK, I wasn't sure since there are still to points missing: `Add Package file to SVN` and `Update install_linux_packages script` > > But don't you think that would fit better in a separate PR? > > Ok, it can be done in a separate PR too. I will start preparing a separate one as soon as this PR is merged. > > When you mean 4.0 I assume we are too late for 3.6? > > Yes, I'd rather not make any bigger changes in 3.6 anymore this close to bcon3. > I completely understand. I just asked so that I can forward the info to the community in the thread. For 4.0 I will probably also cover translucent materials and have better support for equi-angular sampling.
Brecht Van Lommel commented 2023-05-16 16:49:38 +02:00
Owner

Great, thanks.

Great, thanks.
Sebastian Herholz force-pushed cycles_path_guiding_RIS from 9709cd2348 to 6cce149144 2023-05-22 16:36:37 +02:00 Compare
Sebastian Herholz added 1 commit 2023-05-22 16:45:52 +02:00
Sebastian Herholz merged commit 8d17458569 into main 2023-05-22 16:47:10 +02:00
Howard Trickey referenced this issue from a commit 2023-05-29 02:51:38 +02:00
Merge in orign/main:

Reviewers

Weizhen Huang was requested for review 2023-05-09 15:36:25 +02:00
Sergey Sharybin was requested for review 2023-05-11 14:32:53 +02:00
Brecht Van Lommel approved these changes 2023-05-16 15:50:02 +02:00
Sign in to join this conversation.
Reviewers
Request review
No reviewers
Weizhen Huang
Sergey Sharybin
Brecht Van Lommel
Labels
Apply labels
Clear labels   
Interest
Alembic
   
Interest
Animation & Rigging
   
Interest
Asset Browser
   
Interest
Asset Browser Project Overview
   
Interest
Audio
   
Interest
Automated Testing
   
Interest
Blender Asset Bundle
   
Interest
BlendFile
   
Interest
Collada
   
Interest
Compatibility
This issue affects/is about backward or forward compatibility   
Interest
Compositing
   
Interest
Core
   
Interest
Cycles
   
Interest
Dependency Graph
   
Interest
Development Management
   
Interest
EEVEE & Viewport
   
Interest
Freestyle
   
Interest
Geometry Nodes
   
Interest
Grease Pencil
   
Interest
ID Management
   
Interest
Images & Movies
   
Interest
Import Export
   
Interest
Line Art
   
Interest
Masking
   
Interest
Metal
   
Interest
Modeling
   
Interest
Modifiers
   
Interest
Motion Tracking
   
Interest
Nodes & Physics
   
Interest
OpenGL
   
Interest
Overrides
   
Interest
Performance
   
Interest
Physics
   
Interest
Pipeline, Assets & IO
   
Interest
Platforms, Builds & Tests
   
Interest
Python API
   
Interest
Render & Cycles
   
Interest
Render Pipeline
   
Interest
Sculpt, Paint & Texture
   
Interest
Text Editor
   
Interest
Translations
   
Interest
Triaging
   
Interest
Undo
   
Interest
USD
   
Interest
User Interface
   
Interest
UV Editing
   
Interest
VFX & Video
   
Interest
Video Sequencer
   
Interest
Virtual Reality
   
Interest
Vulkan
  
Interest: Wayland
  
Legacy
Blender 2.8 Project
   
Legacy
Milestone 1: Basic, Local Asset Browser
   
Legacy
OpenGL Error
  
Meta
Good First Issue
   
Meta
Papercut
   
Meta
Retrospective
   
Meta
Security
Issues relating to security: https://wiki.blender.org/wiki/Process/Vulnerability_Reports
  
Module
Animation & Rigging
   
Module
Core
   
Module
Development Management
   
Module
EEVEE & Viewport
   
Module
Grease Pencil
   
Module
Modeling
   
Module
Nodes & Physics
   
Module
Pipeline, Assets & IO
   
Module
Platforms, Builds & Tests
   
Module
Python API
   
Module
Render & Cycles
   
Module
Sculpt, Paint & Texture
   
Module
Triaging
   
Module
User Interface
   
Module
VFX & Video
  
Platform
FreeBSD
   
Platform
Linux
   
Platform
macOS
   
Platform
Windows
  
Priority
High
   
Priority
Low
   
Priority
Normal
   
Priority
Unbreak Now!
  
Status
Archived
   
Status
Confirmed
   
Status
Duplicate
   
Status
Needs Info from Developers
   
Status
Needs Information from User
   
Status
Needs Triage
   
Status
Resolved
  
Type
Bug
   
Type
Design
   
Type
Known Issue
   
Type
Patch
   
Type
Report
   
Type
To Do
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset Browser
Interest
Asset Browser Project Overview
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE & Viewport
Interest
Freestyle
Interest
Geometry Nodes
Interest
Grease Pencil
Interest
ID Management
Interest
Images & Movies
Interest
Import Export
Interest
Line Art
Interest
Masking
Interest
Metal
Interest
Modeling
Interest
Modifiers
Interest
Motion Tracking
Interest
Nodes & Physics
Interest
OpenGL
Interest
Overrides
Interest
Performance
Interest
Physics
Interest
Pipeline, Assets & IO
Interest
Platforms, Builds & Tests
Interest
Python API
Interest
Render & Cycles
Interest
Render Pipeline
Interest
Sculpt, Paint & Texture
Interest
Text Editor
Interest
Translations
Interest
Triaging
Interest
Undo
Interest
USD
Interest
User Interface
Interest
UV Editing
Interest
VFX & Video
Interest
Video Sequencer
Interest
Virtual Reality
Interest
Vulkan
Interest: Wayland
Legacy
Blender 2.8 Project
Legacy
Milestone 1: Basic, Local Asset Browser
Legacy
OpenGL Error
Meta
Good First Issue
Meta
Papercut
Meta
Retrospective
Meta
Security
Module
Animation & Rigging
Module
Core
Module
Development Management
Module
EEVEE & Viewport
Module
Grease Pencil
Module
Modeling
Module
Nodes & Physics
Module
Pipeline, Assets & IO
Module
Platforms, Builds & Tests
Module
Python API
Module
Render & Cycles
Module
Sculpt, Paint & Texture
Module
Triaging
Module
User Interface
Module
VFX & Video
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Priority
High
Priority
Low
Priority
Normal
Priority
Unbreak Now!
Status
Archived
Status
Confirmed
Status
Duplicate
Status
Needs Info from Developers
Status
Needs Information from User
Status
Needs Triage
Status
Resolved
Type
Bug
Type
Design
Type
Known Issue
Type
Patch
Type
Report
Type
To Do
Milestone
Set milestone
Clear milestone
No items
No Milestone
Projects
Set Project
Clear projects
No project
Assignees
Assign users
Clear assignees
No Assignees
4 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: blender/blender#107782
There is no content yet.
Delete Branch "sherholz/blender:cycles_path_guiding_RIS"

Deleting a branch is permanent. Although the deleted branch may exist for a short time before cleaning up, in most cases it CANNOT be undone. Continue?