[x3d-public] Physically Based Rendering gltf examples in .x3d form; design approach confirmed, encourage X3D4 implementation

Michalis Kamburelis michalis.kambi at gmail.com
Tue Sep 29 15:53:47 PDT 2020


Thank you Andreas!

Yes, we'd like to show in X3D examples some glTF + corresponding X3D
files. That would be a confirmation of my prose on
https://github.com/michaliskambi/x3d-tests/wiki/Converting-glTF-to-X3D
, that describes the conversion with words instead of examples. In
particular some examples from
https://github.com/KhronosGroup/glTF-Sample-Models/ .

As for the particular lizardman from
https://github.com/michaliskambi/x3d-tests/tree/master/pbr/physical_material
: That lizardman demo is more problematic, there is no exact glTF
equivalent to the X3D version there. The lizardman in X3D was created
in the old Blender version, that could export to X3D with textures.
For the new Blender (2.8), and lizardman in glTF, the model was
reworked. Alas, new Blender (since 2.8) has broken X3D exporter that
doesn't export textures at all. In effect, I have 2 different
lizardman versions:

- old one, generated from old Blender to X3D

- new one, that can be opened in new Blender and works and can be
exported to glTF (but not to X3D...)

So the lizardman is not the best case for comparison anymore, sadly.

Regards,
Michalis

śr., 30 wrz 2020 o 00:06 Andreas Plesch <andreasplesch at gmail.com> napisał(a):
>
> Thanks for the report ! Michalis, would it be possible to provide gltf
> versions of the examples in:
>
> https://github.com/michaliskambi/x3d-tests/tree/master/pbr/physical_material ?
>
> It looks like the existing lizardman.gltf is the base for the
> lizardman variations but does not quite match them. The occlusion
> textures is used as the basecolor texture and the lights seem
> different.
>
> Since one of the main use cases for PhysicalMaterial will be glTF
> import and inlining, it would be really useful to have that two way
> ability to confirm high fidelity rendering between glTF and x3d
> renderers. For example, x_ite and x3dom already have gltf renderers
> (but not the PhysicalMaterial node).
>
> Conversely, it would be similarly useful to convert standard glTF
> examples (like the Avocado, they have all liberal licenses) to the new
> PBR nodes. In fact, view3dscene likely already does this internally to
> render gltf. So it may be mostly a matter of exposing the internal
> nodes.
>
> -Thanks, Andreas
>
> On Tue, Sep 29, 2020 at 1:18 PM Don Brutzman <brutzman at nps.edu> wrote:
> >
> > Meeting today: Michalis Dick Don, regrets (and feedback) from Andreas.
> >
> > Michalis has the following .x3dv examples prepared:
> >
> > [6] michaliskambi > x3d-tests > pbr
> >      https://github.com/michaliskambi/x3d-tests/tree/master/pbr
> >
> > He graciously agreed to export XML .x3d versions as well so I don't spend time (or make errors) transcribing them in preparation for regression testing.
> >
> > He will also add screenshots for each model so that everyone knows what expected rendering looks like.
> >
> > We hope that this really facilitates other X3D4 implementations and encourages consistent presentation.
> >
> > The above examples are all authored by Michalis and in open source, some textures are by others and given appropriate credit.
> >
> > On 9/29/2020 7:32 AM, Don Brutzman wrote:
> > > With X3D XML Schema and DOCTYPE complete, and X3D Schematron rules addable without too much difficulty, we're ready for next steps.
> > >
> > > As part of examples improvements, am hoping to put them in our archives so that regression testing can be performed.
> > >
> > > [1] X3D Resources, Example Archives
> > >      https://www.web3d.org/x3d/content/examples/X3dResources.html#Examples
> > >
> > > May I suggest that we create an PhysicallyBasedRendering section directory for these at
> > >
> > > [2] X3D for Advanced Modeling (X3D4AM) Examples Archive
> > >      https://x3dgraphics.com/examples/X3dForAdvancedModeling/
> >
> > We agreed to put copies of them there, matching and in open source (of course).  Thank you Michalis.
> >
> > > Wondering, where might I find the examples you've shown?  Am looking at
> > >
> > > [3] michaliskambi > x3d-tests >
> > >      Include PBR (PhysicalMaterial and related concepts) in the official X3D specification
> > >      https://github.com/michaliskambi/x3d-tests/wiki/Include-PBR-%28PhysicalMaterial-and-related-concepts%29-in-the-official-X3D-specification
> > >
> > > and then
> > >
> > > [4] michaliskambi > x3d-tests > Converting glTF to X3D
> > >      https://github.com/michaliskambi/x3d-tests/wiki/Converting-glTF-to-X3D
> > >
> > > and then
> > >
> > > [5] castle-engine > demo-models > glTF > sample models for Castle Game Engine
> > >      https://github.com/castle-engine/demo-models/tree/master/gltf
> > >
> > > Wondering how to establish a baseline.  Might it be possible to use Castle Engine to Inline load each of these gltf examples and save each one of them back out as X3D?
> >
> > He confirmed that, rather than Inline (which would be preserved verbatim on saving) it is possible to
> > - open a glTF model in view3dscene, then
> > - save as X3D
> >
> > An update to the following converter is in progress to match the latest X3D4 node signatures, will be finished soon.  At that point anyone can perform this step via
> >
> > [7] Castle Game Engine: Convert everything to X3D
> >      https://castle-engine.io/convert.php
> >
> > We will work on those Khronos-provided glTF examples after adapting his existing examples for regression testing.
> >
> > > Primary format in the example archives is XML .x3d file encoding, preforming maximum validations, then all other variants are then converted from that in order to check proper X3D representations in X3DOM, X_ITE, ClassicVRML, JSON, EXI, Java, Python, Turtle, and pretty-print HTML documentation.
> >
> > Interestingly X_ITE and X3DOM versions are automatically provided as part of this Web3D Examples Archive endeavor.  Helpful for checking.
> >
> > > Looking forward to getting into pair-wise production and publication together.
> > Additional implementations: the past week's work on XML Schema and X3DUOM have each provided significant implementation checks that the interface hierarchy is coherent and consistent, without internal contradictions.  X3D Tooltips add clarity and link to various representations.  X3D Validator provides error checking.
> >
> > [8] X3D Specifications: Schema and DOCTYPE Validation
> >      https://www.web3d.org/specifications
> >      https://www.web3d.org/specifications/X3dSchemaDocumentation4.0/x3d-4.0.html
> >      https://www.web3d.org/specifications/X3dDoctypeDocumentation4.0.html
> >
> > [9] X3D Unified Object Model (X3DUOM) v4
> >      https://www.web3d.org/specifications/X3dUnifiedObjectModel-4.0.xml
> >
> > [10] X3DJSAIL: X3D Java Scene Access Interface Library
> >       https://www.web3d.org/specifications/java/X3DJSAIL.html
> >
> > [11] X3D Validator
> >       https://savage.nps.edu/X3dValidator
> >
> > Next: Dick and I will work on clarifying X3D4 specification prose further in a follow-on session this week.
> >
> > The three of us all think it is fair to say that the PBR design is now complete and correct.  All of our major questions are answered satisfactorily. We are next pursuing implementation and evaluation of examples to confirm that conclusion. This work is already implemented in open-source Castle Game Engine / view3dscene.
> >
> > Suggest that other rendering implementations can now proceed with a high degree of confidence that integration of glTF, physically based rendering (PBR) and non-photorealistic rendering (NPR) is feasible and ready for implementation by other X3D4 players.
> >
> > Again thanks to Michalis for superlative design/implementation efforts, again thanks to everyone who helped with many questions critiques and contributions.  Totally impressive team effort.
> >
> > On our way... Have fun with advanced rendering in X3D4!!  8)
> >
> > all the best, Don
> > --
> > Don Brutzman  Naval Postgraduate School, Code USW/Br       brutzman at nps.edu
> > Watkins 270,  MOVES Institute, Monterey CA 93943-5000 USA   +1.831.656.2149
> > X3D graphics, virtual worlds, navy robotics http://faculty.nps.edu/brutzman
>
>
>
> --
> Andreas Plesch
> Waltham, MA 02453



More information about the x3d-public mailing list