Log depth buffer #5851

Vineg · 2017-09-28T10:28:11Z

Added option "logDepthBuffer" to Cesium.Viewer. With that option globe is rendered in 1 frustum with logarithmic depth. (fixes #4381)

cesium-concierge · 2017-09-28T10:28:12Z

Welcome to the Cesium community @Vineg!

Can you please send in a Contributor License Agreement (CLA) so that we can review and merge this pull request?

I noticed that CHANGES.md has not been updated. If this change updates the public API in any way, fixes a bug, or makes any non-trivial update, please add a bullet point to CHANGES.md and comment on this pull request so we know it was updated. For more info, see the Pull Request Guidelines.

I am a bot who helps you make Cesium awesome! Thanks again.

Vineg · 2017-09-28T11:54:21Z

@cesium-concierge updated CHANGES.md. I am member of geoscan organization. We've already sent CLA

pjcozzi · 2017-09-28T13:29:50Z

Very cool, thanks @Vineg!

@AnimatedRNG @lilleyse and I have been talking about doing this. I'm pretty sure it will be a win to reduce the total number of frustums requires at the cost of losing early z, but I do not believe it will be robust enough to truly fit all scenes into one frustum without z fighting in the distance.

@lilleyse @bagnell can you review this?

pjcozzi · 2017-09-28T13:30:27Z

@Vineg I'll also add you to our CLA list so you do not get a CLA warning in future pull requests.

pjcozzi · 2017-09-28T13:34:41Z

@Vineg if this requires a lot of changes to get into production, do you have the bandwidth to work with us to make the changes or do prefer that we make them?

Vineg · 2017-09-28T13:56:07Z

@pjcozzi I think I can spend 1-2 more days on this feature. Not sure how much changes it is going to be. I've tested it with 1 frustum in the mountains, with pretty big tile model and didn't see z fighting.

pjcozzi · 2017-09-28T15:59:51Z

Sounds great, thanks for your support.

I've tested it with 1 frustum in the mountains, with pretty big tile model and didn't see z fighting.

lol, we need to test looking at a soda can on a table in a building and looking through a window to the International Space Station with the moon in the background. 😄

I expect we will enable log depth by default with the multi frustum, and for scenes that then only require one frustum since the frustum can be much longer, it will be a great win, and even for massive distance scenes, probably only two frustums will be required. We'll also need to test this with the upcoming depth-based post-processing framework and depth picking. It's a pretty central change to the rendering pipeline so it needs careful attention. @lilleyse and @bagnell can help support this.

Vineg · 2017-09-28T16:27:19Z

I'll be away for 4 days. Will return back to this branch on tuesday.

AnimatedRNG · 2017-09-28T18:10:31Z

@Vineg I talked with @lilleyse about this feature, and we think it would be best implemented with a 32-bit depth buffer rather than the 24-bit depth buffers that we've been using throughout the code right now. In particular the texture internal format DEPTH_COMPONENT32F or DEPTH32F_STENCIL8. This feature is only available in WebGL 2, meaning that the WebGL 1 client will have to use 24-bit depth textures.

You might find this blog post useful, as well as this post (for optimizations).

Vineg · 2017-10-03T09:49:15Z

@AnimatedRNG Why do you want to use floating point values? If I understand it right, logarithmic depth is ought to be used with fixed-point values.

Vineg · 2017-10-03T10:25:51Z

@AnimatedRNG Do we really need 32bit? I've tried to set far plane to 500000000000000 and still don't see any z-fighting in the mountains. I think we have enough precision even with 24bit.

lilleyse · 2017-10-12T13:33:40Z

This does seem promising so far for the globe. In order to test the types of cases like @pjcozzi mentioned we'll need to support log depth throughout the engine - models, primitives, etc. Would you be up for this @Vineg, or would you like someone else to continue this work?

Vineg · 2017-10-12T14:16:52Z

@lilleyse, Yes, I just noticed that GroundPrimitives, Primitives, Billboards, Polylines, PointClouds and 3D Tiles requires additional modifications. If I missed something - let me know, I am now working on it. Also it seems to be pretty difficult to make depth writes optional as I did with globe, so I enable it as long as EXT_frag_depth available.

…into log-depth-buffer

pjcozzi

@bagnell please also see the comments I just left in this PR about the Sandcastle examples that are not part of this code review email.

pjcozzi · 2018-04-03T02:19:17Z

CHANGES.md

+* `Camera.distanceToBoundingSphere` now returns the signed distance to the bounding sphere. Positive values indicate that the bounding sphere is in the positive half-plane of the camera position and view direction while a negative value indicates it is in the negative half-plane.
+
+##### Additions :tada:
+* Added option `logDepthBuffer` to `Viewer`. With this option there is typically a single frustum using logarithmic depth rendered. This increases performance by issuing less draw calls to the GPU and helps to avoid artifacts on the connection of two frustums. [#5851](https://github.com/AnalyticalGraphicsInc/cesium/pull/5851)


Did we also change the default near value?

Late to the party, but should this be logarithmicDepthBuffer in Scene?

pjcozzi · 2018-04-03T02:23:40Z

Source/Scene/OIT.js

@@ -554,12 +554,14 @@ define([

        for (j = 0; j < length; ++j) {
            command = commands[j];
+            command = defined(command.derivedCommands.logDepth) ? command.derivedCommands.logDepth.logDepthCommand : command;


Sorry for the perhaps obvious question here, but is everything based on derived commands? If so, why there is there so much custom code for models, ellipsoid primitives, etc.

The custom code adds the function calls but the function bodies are surrounded by #ifdefs. The derived commands will add the defines when log depth is supported.

pjcozzi · 2018-04-03T02:24:17Z

Source/Scene/EllipsoidPrimitive.js

@@ -398,7 +423,10 @@ define([
            }

            // Recompile shader when material changes
-            if (materialChanged || lightingChanged || !defined(this._pickSP)) {
+            if (materialChanged || lightingChanged || !defined(this._pickSP) || useLogDepthChanged) {


useLogDepthChanged

If it makes things much easier, I don't know that users will want to be able to toggle log depth at runtime.

This wasn't difficult to add since we have to support the fallback anyway.

pjcozzi · 2018-04-03T02:25:54Z

Source/Scene/Scene.js

@@ -287,6 +287,10 @@ define([
        this._primitives = new PrimitiveCollection();
        this._groundPrimitives = new PrimitiveCollection();

+        this._logDepthBuffer = undefined;
+        this._logDepthBufferDirty = true;
+        this.logarithmicDepthBuffer = context.fragmentDepth;


Should this have the same name as the property get/set below?

Yes, this calls the setter below.

pjcozzi · 2018-04-03T02:27:30Z

Source/Scene/Scene.js

+         * @type {Number}
+         * @default 1e9
+         */
+        this.logarithmicDepthFarToNearRatio = 1e9;


Better document the interaction between logarithmicDepthBuffer, farToNearRatio, and logarithmicDepthFarToNearRatio and use @see/@link to link them together. If I am following the boolean controls using one or the other ratios.

pjcozzi · 2018-04-03T02:32:43Z

Source/Scene/Scene.js

+    var vertexlogDepthRegex = /\s+czm_vertexLogDepth\(/;
+    var extensionRegex = /\s*#extension\s+GL_EXT_frag_depth\s*:\s*enable/;
+
+    function getLogDepthShaderProgram(context, shaderProgram) {


This is a lot of new code in Scene.js. Perhaps move these two new log depth shader functions to a new private utility class in a separate file?

pjcozzi · 2018-04-03T02:33:28Z

Source/Scene/SceneTransitioner.js

@@ -870,6 +870,12 @@ define([
                camera.frustum = camera3D.frustum.clone();
            }

+            var frustum = camera.frustum;
+            if (scene._logDepthBuffer && !(frustum instanceof OrthographicFrustum || frustum instanceof OrthographicOffCenterFrustum)) {
+                frustum.near = 0.1;


Same comment here and below.

Same comment as what? That this should be added to CHANGES.md? Done.

Ah, there were some hidden comments.

This has the same problem as in master. When the scene mode switches, we reset to the default near far distances in the frustum constructor. The old ones aren't saved between scene mode changes (should happen, separate PR or open an issue). This just sets to better defaults when log depth is enabled.

pjcozzi · 2018-04-03T02:33:36Z

Source/Scene/SceneTransitioner.js

                Cartesian3.clone(cameraCV.position, camera.position);
                Cartesian3.clone(cameraCV.direction, camera.direction);
                Cartesian3.clone(cameraCV.up, camera.up);
                Cartesian3.cross(camera.direction, camera.up, camera.right);
                Cartesian3.normalize(camera.right, camera.right);
            }

+            var frustum = camera.frustum;
+            if (scene._logDepthBuffer && !(frustum instanceof OrthographicFrustum || frustum instanceof OrthographicOffCenterFrustum)) {
+                frustum.near = 0.1;


pjcozzi · 2018-04-03T02:37:28Z

Source/Shaders/Builtin/Functions/vertexLogDepth.glsl

@@ -0,0 +1,49 @@
+#ifdef LOG_DEPTH
+varying float v_logZ;


Here and in writeLogDepth.glsl, we generally don't add new varyings or uniforms when including a built-in GLSL function (well, I also just noticed writeDepthClampedToFarPlane.glsl). Instead, we try to make the functions standalone and pass in / return anything needed. Is that reasonable here? Or does that require too much shader patching elsewhere?

I opted to put it here instead of making the shader patching code more complicated.

pjcozzi · 2018-04-03T02:40:54Z

Specs/Scene/SceneSpec.js

@@ -251,6 +251,9 @@ defineSuite([
    });

    it('debugCommandFilter does not filter commands', function() {
+        var originalLogDepth = scene.logarithmicDepthBuffer;


Does this file or MultifrustumSpec.js need some new tests where it switches between log depth on/off and sees the number of (redundant) commands executed?

In general, how is code coverage? Does anything need more fine-grained tests? Shouldn't the new built-in GLSL functions have new tests?

The multi-frustum tests have been updated to disable log depth. One test has it enabled to show that log depth has less 1 frustum where the other had 3.

Code coverage seems ok, but it's hard to say. I can't add tests for the new GLSL built-ins since they don't return a value. They update gl_Position and gl_FragDepth.

lilleyse · 2018-04-03T13:21:55Z

Not sure if this is related to this branch, but in the interior 3D Tiles Sandcastle example, if I zoom back the entire model disappears.

This happens on master as well as #6390. It is an edge case related to back faces and skip-lods. When refining the back faces of the parent are rendered, which kind of falls apart for single-sided geometry like this. I'll open an issue.

Without skip-lods:

lilleyse · 2018-04-03T13:36:47Z

In the BIM 3D Tiles Sandcastle example, I get some flickering (looks like a full-screen quad) when inside and then zooming in/out close. Is this related to this branch or is this a skip LOD bug that may be fixed with the new refactor? Seems loading related.

There's a bunch of BIM rendering problems, but they happen on all branches I've tried:

(Geometry z-fighting)

(Distorted flickers - I think a problem with the glTF/shaders)

(skip-lods flicker)

lilleyse · 2018-04-03T13:53:39Z

Not sure if it is related to this branch, but looks like the Draco compressed example in Sandcastle doesn't have depth testing:

It seems like it's just this branch.

pjcozzi · 2018-04-05T19:13:53Z

@bagnell all sounds good.

bagnell · 2018-04-05T19:46:11Z

@pjcozzi @lilleyse This is ready for another look.

pjcozzi · 2018-04-05T19:59:16Z

@lilleyse please merge when ready.

@bagnell don't forget about the blog! 😄

lilleyse · 2018-04-05T21:50:14Z

Sorry for not catching this earlier, but I noticed some problems in the KML example related to billboards:

master vs. log-depth-buffer

lilleyse · 2018-04-05T21:51:30Z

Also check the Cesium Inspector sandcastle. Some of the options aren't working for me.

…rtex shader.

bagnell · 2018-04-09T13:16:31Z

@lilleyse I fixed the billboard issue in 6e9fc3e which you might want to review since it changes the billboard vertex shader to use eye coordinates instead of window coordinates.

lilleyse

Cool, 6e9fc3e fixes the billboard problem and the existing demos still seem to work fine.

lilleyse · 2018-04-10T15:40:22Z

Source/Shaders/BillboardCollectionVS.glsl

-#ifdef LOG_DEPTH
-    czm_vertexLogDepth(czm_projection * positionEC);
-#endif
-
    vec4 positionWC = computePositionWindowCoordinates(positionEC, imageSize, scale, direction, origin, translate, pixelOffset, alignedAxis, validAlignedAxis, rotation, sizeInMeters);


Should positionWC and computePositionWindowCoordinates be renamed?

lilleyse · 2018-04-10T16:07:30Z

Source/Shaders/BillboardCollectionVS.glsl

@@ -71,23 +71,23 @@ vec4 computePositionWindowCoordinates(vec4 positionEC, vec2 imageSize, float sca
        positionEC.xy += halfSize;
    }

-    vec4 positionWC = czm_eyeToWindowCoordinates(positionEC);
+    float mpp = czm_metersPerPixel(positionEC);


Is it possible to only compute this if sizeInMeters is true and still function properly with the new changes? If not, not a big deal.

No, since it's needed regardless of whether sizeInMeters is true or false;

bagnell · 2018-04-10T20:18:26Z

@lilleyse This should be ready.

lilleyse · 2018-04-12T20:30:07Z

Ok I have nothing else to add here. Ready to go!

@bagnell can you open an issue to remove the skinned model workarounds?

geoscan-builder added 5 commits September 26, 2017 16:57

added optional logarithmic depth buffer

80f5976

fixed pickinig issue with log depth buffer

36264f7

removed trailing decimal point to avoid eslint warn

e15adbf

added to CONTRIBUTORS

bd5428f

fallback to multifrustum if no fragDepth extension

ceedc32

Vineg added 2 commits September 28, 2017 14:50

Update CHANGES.md

d70db84

Merge branch 'master' into log-depth-buffer

87f8415

pjcozzi assigned lilleyse and bagnell Sep 28, 2017

pjcozzi requested review from lilleyse and bagnell September 28, 2017 13:32

geoscan-builder added 6 commits October 13, 2017 17:45

added log depth to objects, removed option to turn it off

a79138d

Merge branch 'log-depth-buffer' of https://github.com/geoscan/cesium …

7ee510c

…into log-depth-buffer

Merge branch 'cesium-master' into log-depth-buffer

7133aa5

removed unused variable

4a79331

fix wrong require

50c019c

moved extension to logDepth function

043ebcc

pjcozzi requested changes Apr 3, 2018

View reviewed changes

bagnell added 3 commits April 5, 2018 14:23

Fix Draco compressed models.

4dd219f

Update CHANGES.md

b5f83d9

Merge branch 'master' into log-depth-buffer

9bc6e32

bagnell added 2 commits April 5, 2018 15:22

Move derived commands to own class.

d546e45

Update multi-frustum tests.

e17dbcc

emackey mentioned this pull request Apr 9, 2018

Post processing #5615

Merged

12 tasks

Use eye coordinates instead of window coordinates in the billboard ve…

6e9fc3e

…rtex shader.

Merge branch 'master' into log-depth-buffer

4cf9479

lilleyse reviewed Apr 10, 2018

View reviewed changes

bagnell added 3 commits April 10, 2018 15:39

Fix NaNs returned by barycentricCoordinates.

bed9f0e

Fix debug show globe and pick depth.

3419ad5

Rename from review.

d7d14a5

lilleyse merged commit 50b1d5d into CesiumGS:master Apr 12, 2018

emackey mentioned this pull request Apr 16, 2018

Model clipped inappropriately #6451

Closed

likangning93 mentioned this pull request Apr 18, 2018

shadow volumes broken in master at near distances #6461

Closed

emackey mentioned this pull request Apr 29, 2018

Bear glTF 2.0 model is not displayed in Cesium.js v1.40 or later #6416

Closed

lilleyse mentioned this pull request Apr 6, 2020

Occlusion issue with changes to log depth equation #8725

Closed

Log depth buffer #5851

Log depth buffer #5851

Conversation

Vineg commented Sep 28, 2017 • edited Loading

cesium-concierge commented Sep 28, 2017

Vineg commented Sep 28, 2017 • edited Loading

pjcozzi commented Sep 28, 2017

pjcozzi commented Sep 28, 2017

pjcozzi commented Sep 28, 2017

Vineg commented Sep 28, 2017

pjcozzi commented Sep 28, 2017

Vineg commented Sep 28, 2017

AnimatedRNG commented Sep 28, 2017 • edited Loading

Vineg commented Oct 3, 2017

Vineg commented Oct 3, 2017

lilleyse commented Oct 12, 2017

Vineg commented Oct 12, 2017 • edited Loading

pjcozzi left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lilleyse commented Apr 3, 2018 • edited Loading

lilleyse commented Apr 3, 2018

lilleyse commented Apr 3, 2018

pjcozzi commented Apr 5, 2018

bagnell commented Apr 5, 2018

pjcozzi commented Apr 5, 2018

lilleyse commented Apr 5, 2018

lilleyse commented Apr 5, 2018

bagnell commented Apr 9, 2018

lilleyse left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bagnell commented Apr 10, 2018

lilleyse commented Apr 12, 2018

Vineg commented Sep 28, 2017 •

edited

Loading

Vineg commented Sep 28, 2017 •

edited

Loading

AnimatedRNG commented Sep 28, 2017 •

edited

Loading

Vineg commented Oct 12, 2017 •

edited

Loading

lilleyse commented Apr 3, 2018 •

edited

Loading