Improvements to Physically Based Rendering

[Illes Papp]

The Physically Based Rendering Experimental Feature is improved with the following new features:

1. Procedural Sky - The newly introduced Procedural Sky generates a sky dome with an accurately positioned Sun disc which affects not only the background but also the lighting of the scene.
2. Shadow Mapping - Shadow mapping produces much faster results than shadow volumes, with soft shadow edges.
3. Automatic Exposure Control and Bloom Effect - Exposure is automatically controlled, similar to real life cameras. High luminance regions generate a so called “blooming“ effect.
4. Improved Ambient Occlusion - Previously certain areas of 3D scenes introduced some artifacts, and with the improvements these are not appearing any more (available from Archicad 27 Update 2).

For more details, please read the following reference:
Visualization - Graphisoft Community

Looking forward to having your much appreciated feedbacks, questions and insights.

[Eduardo Rolon]

Rhino, which is a far cheaper software package, can do it since basically forever.

GS keeps trying to reinvent the wheel without researching what is already available in other mainstream CAD software.

PBR without contours is not acceptable.

PBR only as a 3D View is not acceptable.

I agree with @Potato Farmer .

This tool should match as a minimum what Rhino already does then improve on it.

The way it is going it will be as useful as AC27's MEP, Paramo, SAF/SAM, etc

[Jim Allen]

I think you might be comparing apples and tomatoes 🙂

I'm a bit taken aback at how what is clearly supposed to be an incremental improvement in the model display has become a subject where people are demanding what could be a major overhaul in the Archicad display engine codebase.

The first problem is that the PBR display isn't even right as currently implemented!

There are lots of things other software does that AC doesn't. Look at how SketchUp's push-pull tool works for example. That has been in existence since it was released and it's so much easier for modelling than the Morph tool. I would rather have that improved than this.

I would rather have the terrain tool rewritten than this.

We all have our pet requests based on what we use the software for.

I don't disagree that it would be much better to have more display styles in AC, that's clearly nice to have!

SketchUp has also had these for well over a decade.

But what do you want AC to do?

If you had to choose between more usable functionality and prettier display options that don't enable us to do anything new - what would you choose?

How many more AC licences do you think would be sold if expanding the feature set was sacrificed for a nicer-looking model?

Software developers have teams working on different aspects of an application. There will be a big matrix of stuff that is complex, stuff that is simple, and stuff that is considered to be urgent vs stuff that is considered to be trivial. These will all be subject to a cost/benefit evaluation.

Have a look at the Roadmap, the idea pool shows the things that are furthest away. Would you prefer better handling of ceilings and areas and terrain https://graphisoft.com/product-roadmap/idea-pool or a prettier display?

What about heat load calculations, embodied carbon, better embodied carbon workflow or skin by skin enhancements for modelling? https://graphisoft.com/product-roadmap/under-research

So I think it's absolutely fine highlight these things as being nice things to have, but I don't think anyone can claim they are essential can they?

Don't you think it would be better to get the feature working properly first before requesting all sorts oaf additional enhancements?

[Potato Farmer]

---

@Jim Allen wrote:

No-one needs that functionality - it's purely a 'nice to have' isn't it?

It absolutely isn't a 'basic requirement' no matter how much you want it.

We don't "need" 3D models at all, we can still draft in autocad or draw our plans on paper. And since I'm paying for this software and have developed workflows to use it in certain ways then I get to decide what is a basic requirement.

The Archicad workflow is built on the notion that the model and documentation are one thing. Plan views are viewpoints from which you change the overall model and the source view for drawings. Sections are the same. (Details are not, but users have been asking tor live details and worksheets for a decade or more.)

The 3D window and 3D documents are no different. Taking screenshots is fine for email but even then we might want higher resolution than offered by the built-in OS tools. And placing auto-update 3D views and 3D documents as drawings in a document set is indeed a requirement, not just nice-to-have, for me and every Archicad user I know. If I had to screenshot, or export to stand-alone rendering software, every 3D view I wanted to include in a drawing set I would dump Archicad and look for something better.

If PBR cannot deliver this then I don't see the point in developing it. It seems in this regard we agree.

[Eduardo Rolon]

I think that we agree on principle.

This is not a pet request I don't need more 3D Styles since I am comfortable using Rhino, C4D, TM, Enscape and multiple rendering packages and I use different tools for what they do best. 

But…

They are asking for input on a new 3D Window style and these are my recommendations as to what is needed.

AC is not for rendering so whatever improvements are in the pipeline they should align with AC's main core and that means contours and availability in floor plan, sections/elevations and 3D Docs.

• Is it nice to have better rendering displays? Yes.
• Should it be a priority over and improved Mesh Tool? Nope
• If they have the resources to do both do I mind? Nope.
• Is PBR ok as is? Nope

---

The issue with the "improvements" over the last 15 years is that they are never finished. PBCs, SEOs, GO, the new MEP, Folders, GO's, the new Keynotes all have essential functionality missing. They are mostly "usable" but everyone has/need different work arounds to handle the missing parts and there is no timeline to fix them AFAIK.

My first comment on PBRs was 2023-12-16 (there were others for AC27 in the old beta site) since then there has not been any improvements or  I can see . So I took the time to show examples.

In order for this feature to work it needs contours and remove the blue tint but if it will only be used as 3D window and will not be integrated in the other views then they should scrap it and invest the time on fixing existing tools.

-----

Based on my experience I should not have commented on this thread. This tool, like the new Keynotes are not usable so I should not waste my time here.

[Jim Allen]

Actually I think we do 'need' 3d models nowadays because they enable much more informed decision-making at the design stage, and they enable a much more open debate to be held with clients about design.

Without 3d models, discussions tend to be about locations and sizes of spaces and cost and specification.

When we present design proposals in a form which is accessible to people not used to trying to build up a 3d mental picture from 2d drawings, it's empowering clients and causing designers to be more analytical and self-critical.

Personally (and others may disagree) simply fixing the model display in its current form (colours in particular) would be acceptable for me for now.

Contours would be good, but they wouldn't be vectors, unless quite a bit of resource was devoted to it. I've worked with this before, it's not trivial.

[Potato Farmer]

I was being snarky about the need for 3D. I would never go back to drafting, and many of my wishes center on reducing the drafting and over-drafting we do now.

If I ran the zoom I would not have spent time on a new 3D engine at all but instead worked on DPI control, and a real-time sun / shadow setting like we just got in BIMx. But this a Tech Preview showcasing the features the great minds at Graphisoft have already deemed worthy, so we play the hand we are dealt. And currently PBR is a pair of threes.

The current 3D engine has contours which are not vectors but look fine with antialiasing enabled. This new engine needs to exceed the current one as an everyday multipurpose modeling + documentation environment or the effort will be for not.