Now that ai is officially out of the bottle and evolving fast, where does ai fit within Enscape? I am curious how the Enscape and Chaos team are discussing how to use this within their software.
What are everyone’s thoughts on this topic?
Now that ai is officially out of the bottle and evolving fast, where does ai fit within Enscape? I am curious how the Enscape and Chaos team are discussing how to use this within their software.
What are everyone’s thoughts on this topic?
Easy, fast, predictable and iterative. Predictable and iterative are missing from most tools - it’s usually all or nothing iteration. For a start, I would desire the ability to use a text prompts across a selection of views and batch export the results. Guiding the designer would help remove complexity; perhaps a “beginner” mode with separate panels asking for useful information 1) Location 2) Time of Day 3) Style 4) Materials. A more advanced mode would only require a single panel for prompts. The ability to save / load / share prompts. The ability to script batch rendering in terms of multiple prompts, view and save the results. Thx!
I think “predictability” or precision is key in the arch viz field. Right now, it’s difficult for ai to know what needs to stay (walls and floors) and what can be adjusted or changed (materials, time, or furniture). Until there is a controllable level of precision or “predictability” on what being iterated, it will be difficult or just time consuming to use ai for arch viz.
Agreed – if you need rendering to be predictable, then you have tools like a Enscape / V-Ray / etc. for visualizing deliberate design decisions. If you need whimsical, fast and beautiful, iteration and ideation during schematic / preliminary design - AI saves time, money and produces emotionally compelling results without having to commit detailed design decisions.
Estimate designers are saving 30-40 person hours per image compared to manual modeling, lighting, materials, set dressing, site dressing, etc. These images are used to solicit client feedback in order to vector in the right direction. Veras is on the right track as it leverages simple massing level models.
Designers are also using Interior AI with simple interior images to solicit client feedback for interior layouts, mood, theme, etc.
AI can improve. But it’s useful now as is. Eager for Enscape to meet and exceed present expectations and workflow.
AI has had difficulty with one of my models for a restaurant. It interpreted the bar to be a bed in each iteration (see attachment, there is supposed to be a bar where the bed is depicted). There’s some advancement that still needs to occur but the technology is moving very quickly. The program I am most excited about is Adobe Firefly which will revolutionize the Photoshop user industry. In their demonstrations they show a photo and with the prompt “create a winter snow scene” it automatically added snow to the photo and changed the sky settings, etc. That immediately made me think of Enscape - one could take an Enscape render and add snow or other effects in a snap. Atmospheric control is one of the selling points for programs like Lumion, which really won’t be needed.