What type of visualization do you do?
We do a lot of simple shaded / white model views to explain the project to clients and builders. We get very productive discussions with clients when they see the live model at meetings in our office, as many people really don’t fully understand plans. We also like making isometric views to explain complex millwork or wall assemblies to builders.
Presentation renders involve Blender/Octane or Enscape.
What is your current workflow for visualization?
All in-house.
For simple shaded views we keep everything inside Archicad as much as possible so that we have live updates of views when models change.
For presentation renders we pull content from different DCC applications based on their particular strengths: render inside Octane Standalone, pulling in the building geometry from Archicad, Blender + plugins for landscape or vegetation, high detail surface/prop assets from Quixel Megascans, and some extra furniture and props from Sketchup’s 3D warehouse.
What are the main challenges or obstacles in your current workflow?
Archicad’s shaded view is frequently disappointing and could use lots more customization. Sketchup for all its limitations makes shaded views that are attractive and easy to adjust with different user styles, it’s popular with architects for good reason, and would be worth studying. A crucial thing that Sketchup does that AC doesn’t is silhouette edges - a round column in AC or a library object with curves never looks convincing in the shaded view. The way that Sketchup has face-me components (that don’t require a rebuild to face the camera) allows them to make effective but low detail billboard representations of trees, people, and other objects. A simple screen-space ambient occlusion pass would also impart a lot of depth to images without a lot of added complexity (as long as certain surfaces like glass can be set as translucent to the AO effect.)
As noted by other users, the appearance of a section or elevation should be associated with the view, not the section marker itself! It’s totally inconsistent with the way everything else is handled in AC.
I don’t consider CineRender even worth keeping in the software as there are much much better alternatives. The new FBX export option is a very very good development, but please allow the export to be limited to a selection rather than just everything visible. When rendering in an outside application you often want to make incremental updates to certain parts of the model, or just render a small chunk of the project.
What third party solutions do you incorporate in your workflow if you do any?
We are using Enscape more and more. The integration with AC is good but Enscape assets that are added to AC look like trash (orange lines and no proper floor plan symbols), so we end up having to maintain a layer for Enscape furniture, another for AC furniture, which is silly. (I prefer adding all these assets in AC so that they can be controlled by renovation filters / design options)
A good solution would be if any AC object could be set as a proxy for a high detail Enscape mesh, just by setting a property of the object. Ideally there would also be some override on the render mesh’s rotation, scale, and position, since render meshes may be at different orientations from AC objects.
I would really recommend that GS avoid trying to build its own high-end renderer, or anything to do with VR/AR. You are not going to out-do Chaos, Epic, NVIDIA or Otoy, and specific workflows are too diverse from office to office. Focus instead on interoperability (GS’s usual strength) and opening up the software API so that it’s easier to build integrations. Look at Rhino - it has a very powerful C# and Python API and its users have contributed mountains of good functionality as a result. Let users build the features they need rather than constantly pleading for something or other from the developers.
