[x3d-public] X3D minutes19 June 2020: Mantis issues ongoing resolution - glTF metadata, SIGGRAPH 2020 and release schedule, PBR of point normals

[Don Brutzman]

Thanks for taking time to review + respond.

On 6/19/2020 7:25 PM, Michalis Kamburelis wrote:
>  > 1272            glTF import needs to include metadata import
>  > pending 1269 PBR, review next week!
> 
> This isn't related to PBR work in any way. My PR doesn't explicitly specify how (and if, at all) glTF content is converted to X3D nodes.

Understood and agreed that it is not directly related, rather it merely depends on the glTF work for X3D being a little farther along.   We have simply deferred even looking at it until other glTF-related work is stable.  After Mantis 1269 we expect to at least consider several tangential topics.

> Sure, you can use new X3D PhysicalMaterial to express glTF PBR materials, but that's not something I mandate in any way in the X3Dv4 spec. (Even if it's obvious for the implementors.) And I definitely do not want to mandate any other details about converting glTF->X3D in my PR.
> 
> X3D 4.0 is not the right time to deal with this, IMHO. There are various open questions about how to cooperate with glTF perfectly, and we have now 2 implementations that converge on some ideas, but are not 100% synchronized (CGE and X3DOM). We are not yet at a place where we want to mandate (in X3D spec) any particular conversion scheme glTF -> X3D.

Useful perspective, we also have more work to do demonstrating best practices and patterns with X3D-centric metadata.  I would not expect specification words but am looking forward to our ability to perform further scrutiny.  Step by step, the 3D in glTF is of course the priority.

>  > 1269            glTF physically based rendering PBR, advanced material textures and lighting
>  > next week!
> 
> I expect to still work on my PR next week (including the weekend). I want to finish things by the end of June. I may not be ready with them next Friday (26 June).
> 
> As said in my other email, I would like to present my PR on first Friday in July. So 3rd July. No next week.

Sorry to miss that particular email point or misunderstand our discussion a week ago.  Of course, as when we started, the primary goal above all others is to get it right.  Thanks for your focus on excellence Michalis.

Our goal has been to ship a draft for Web3D board and member review in time for public release before SIGGRAPH 2020, which was originally scheduled for 20-23 JUN 2020.

Checking online... In addition to "going virtual" it looks like SIGGRAPH 2020 is shifting dates to the right:

* SIGGRAPH 2020
   https://s2020.siggraph.org

"Virtual Conference. The August dates for the online conference will be announced soon. Bookmark the conference page for the latest updates."

* SIGGRAPH 2020, Health and Safety
   https://s2020.siggraph.org/attend/health-and-safety

"Coronavirus (COVID-19) Update, 29 May 2020: In a message from ACM SIGGRAPH President Jessica Hodgins, preliminary details were shared about the upcoming virtual SIGGRAPH 2020 conference including a likely shift in conference timing to mid- or late-August. _Click here_ to read her letter. Further, given the move to virtual, our primary focus is now on bringing our community the best virtual conference experience possible."

* What Does it Take to Put on a Virtual Conference?
   A Message from ACM SIGGRAPH President Jessica Hodgins
   https://blog.siggraph.org/2020/05/what-does-it-take-to-put-on-a-virtual-conference.html

We are keen to review your PR8 work together when you are ready.  If it helps to split that across 2 different weeks of meetings, that is fine too.  Plenty of work to do here!

"Act in haste, repent in leisure"   8)

>  > 1252            PointProperties node specification; PointSet/LineSet/IndexedLineSet can contain Normal
>  > Mostly complete.  Sole remaining issue is whether PBR might affect rendering prose.
>  > Need to do a final review when Physically Based rendering (Mantis 1269) is finalized. The current opinion is that PointSet are considered unlit so the rendering will not be affected by PBR.
> 
> I would say that PointSet with normals should be lit -- otherwise, what was the point of specifying normals?
> 
> And all lighting models should be possible (so Material, UnlitMaterial, PhysicalMaterial all make sense on such points).
> 
> The idea, as far as I understand, is that these points represent an underlying object with surfaces, but are represented as points for technical reasons (e.g. were obtained by 3D scanner). Thus you should make these points and lines lit (when they have normals).
> 
> That is also what glTF does, https://github.com/KhronosGroup/glTF/tree/master/specification/2.0#point-and-line-materials . Points/lines with normal vectors are lit.
> 
> Regards,
> Michalis

Very good, seems sensible and similar to our typical X3D approach to rendering.  This issue doesn't appear to be complex, just attention to detail for spec prose.  Have added your feedback here to Mantis 1252.

TODO.  Dick and Don split out whatever items are pending with respect to PBR rendering of points in Mantis 1252 as a new issue, so that the rest of this rather large issue can be confirmed complete and resolved.

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