Practical advise would be to look at what people can build using Lumion and how fast they can do it. The only differentiators I can see from there is highly interactive applications which go more into direction of training/simulation/case study than a pure archiviz. Unless you can do it faster and in much higher fidelity than Lumion - just don’t bother. Times when pure game engines are used for this are gone like 5 years ago or more.
I would not rely entirely on ue4 arch-viz to put bread on my table that’s for sure. I do offer standard still imagery made with offline renderers but I have to admit it’s more fun to make interactive scenes!!! I think people are interested in it because it can do more than Lumion if pushed enough. Trying to reach vray-ish quality with unreal is fun stuff to do for me.
But yea, just don’t limit yourself to a single element, like VR with unreal…it’s way too narrow to make a living out of it!
But arch-viz as a whole…I highly doubt it’s a dead business! Future is going to be fun, full of new tech, lots of stuff to learn.
I do agree with you, my point is that whatever feature tech is there it most likely won’t be a game development engine. Simply because pipelines and visual quality requirements are completely different. You already can do a lot if you target solely architecture. I had a real-time GI prototype running (6 years ago) and it was possible because 99% of the scene is static. Tech like this won’t be considered for game engine as only static content is a huge limitation.
Another thing you can do for archiviz is local streaming a very simple idea where you bring your real-time render farm in a back of the van to exhibition center and stream end result over the cable instead of internet. Again this won’t work for game.
This day is coming sooner than we think I suspect. I give UE4 2 years of being viable for Archviz before something else is here that is a far better solution.
Even if it’s not VR, it’s possible to make nice videos with Unreal. Like this one :
Something that would cost $$$ in render farm costs. Sure it’s not perfectly photoreal but it’s enough to make a client VERY happy.
For that reason alone, I think Ue4 is worth using. It looks much better than lumion projects!!!
VR/RT is going to be a niche market for a long time but if your client ask for it, then it’s not much more work to do than a video!
Precisely why I wanted to learn it, it’s good enough for video work. Just have to get the work flow down so it’s efficient.
@Boredengineer - Very interesting perception and it’s a good thing that we all have different reasons and motivations to try new technology.
In my view UE is never meant to replace any arch-viz pipeline that already exits, it meant to extend it by adding a new level of interection/experience. Don’t know why people are taking UE as a replacement, it is not and never will be in coming future.
And I totally disagree with @Boredengineer when said that “whatever feature tech is there it most likely won’t be a game development engine” because Autodesk is thinking otherwise (Stingray is your answer). Game engines are the new poineers in developing technology that breaks orthodoxy. But also this is a sheer truth that time and cost are the main variables when it comes to implementing cost effective pipeline hence we have to take UE as an extension to our existing technological knowledge rather then rejecting it completely.
Not sure why you think this would have a significant cost to render. If you can run it on your PC in realtime you definitely can render it on the same hardware. Content for render and for real-time is two completely different things. It’s a lot of additional work to make it VR/RT. For rendering you make content only for the parts which will be seen on camera. You don’t care about unwrapping your models and polygon budget. You can render in multiple layers at the same time and tune everything in post production. Technical high-poly modelling is way simpler and cheaper than low poly, which you are going to need to have decent frame rate. You are not locked into a single machine and can rent render farm for exactly as much time as needed. Should I go on?
It’s a reality for about 4-5 years already. RTT use Nvidia’s Optix for their software, before that they had a custom solution. The reason why you don’t see something like Lumion running Optix is because whoever can use it, uses it themselves for end products, instead of building tool for competitors. Another issue is being locked into specific hardware and generally small market for this kind of products.
I’m not sure what is your argument with Stingray. Non of us know what Autodesk really thinks and what is the actual strategy behind this product. It can be as simple as monetization one of their previous acquisitions and have nothing to do with the rest of the company.
My point is that technology which is viable for game development isn’t necessary viable for Archiviz and vice versa. These are two very different things and it’s not new story. No of us is making a breakthrough here, niche market for RT Archiviz exist for at least 15 years already.
This is an example of what professional studious who specialize in this market do:
http://3dcapacity.com/
On their older version of website you could see that they make about 50+ RT Archiviz a year and this is relatively small company. But they are in very special place - Netherlands uses RT Archiviz almost like a standard for communicating projects to public.
Look guys, I don’t want to determent your desire to succeed in this market. I was just like you 10 years ago. The truth is, not much changed since then. There are new cool tech out there but most of it is irrelevant to actual business. If you want to build a shiny cool looking real-time applications, don’t look for architects or real-estate developers. Look for advertisement agencies, for whom your application would be a part of the package deal instead of the core delivery. This is where you going to have time and budget necessary to deliver something of higher value then standard visualization. As soon as possible join a studio or a team of similar minded people, you won’t be competitive if you keep working as a freelancer and you need to both specialize and generalize at the same time. There are few things which require extra attention in real-time applications - GUI and sound. Such expertize is non existent in archiviz firms but usually available to marketing and advertizement agencies.
I’m not sure I’m following you? Game engines for Architects seems to me to be relatively new, I’m starting to see more and more Architects requesting UE4 specifically. In terms of expertize in Architecture firms have you ever worked for a world class firm like Gensler, my past employer? Don’t sell these guys short, they didn’t get to be the biggest firm in the world because they aren’t capable. You have to understand the business model of presenting a project in Architecture, it’s not like making a TV commercial, it’s usually an overhead cost or partially subsidized by the client. So with that said budget determines output.
That’s not a new thing - I’ve defended my masters degree in computer science, on this subject, 10 years ago. At that time I was already working for guys who specialized in these products.
I never worked directly in Gensler office but taking into account the amount of projects we did I wouldn’t be surprised if they were end client at some point. To give you an example, there is installation in Heineken museum, software for which was done by me, but I never worked with anyone from the Heineken itself. It’s a normal thing that you have a long chain of intermediaries where you work with media partner of marketing branch of advertisement agency who is hired by real-estate developer. So when you get your floor-plans and sketches they are not even coming from architects directly. It’s very rare to work with the end client directly as it’s not their specialization. Hilarity of this chain becomes obvious when several times you get exactly the same RFP from two different clients.
I’m glad to hear that nowadays they are trying to make such projects themselves. But I wouldn’t be surprised if it’s done more for self promotion than actual service which they sell to the client. Perhaps they do it to get a competitive leverage, which again means that this product is still not a standard in their covered territory. Their capability to sell architecture service doesn’t mean anything with this media. I’ve seen AAA studios failing in this business miserable. It’s a false assumption to think that, for example, if someone like Pixar is good in CG, then they will definitely rock in technical visualization. They can be using similar tools but the core business needs are different.
I don’t know that well business model of architects, most of our clients where real-estate developers, advertisement agencies and “competitors/partners”. Their budget is most of the time provided by investors and we were hired in parallel with architects. The fact that agency gets a 5 mil in marketing budget doesn’t mean that they are ready to finance several month of development for interactive product and spend 80k on it.
It’s about the end product after all. It took us years to find a proper approach to clients so that interactive real-time application wouldn’t be replaceable by cheaper visualization materials. The choice of the engine is the least concern in this area when pretty much everything can be outsourced dirty cheap. If you happen to work for one of the largest companies and they don’t care so much about cost and time to build on interactive projects then good for you. But this is not the situation where 99% of people will end up, this is why products such as Lumion are so popular. The longest development time we ever had was for one of the emirate projects were we spend whooping 3 month! to complete everything 
This was with already nicely optimized production pipeline where things like import of geometry and assigning of materials was automated. Lightning and shadows were standardized and had to be tweaked only for specific mood when necessary. Recording video and taking annotated screenshots could be done in application itself by end user, instead of recording it in editor. This is where you win by having a superior product which is based on superior tools, not on a free engine which everyone in China and India will be using for games and archiviz in a matter of few years.
Just the opposite. In most cases the Architect already has the client it’s provided as a tool for the client and a tool for the design process, the fact that these materials are used to gain further clients comes after the fact as Architects won’t generate a promotional fictitious project in order to gain more clients. They will however generate viz materials if they are competing for a job, size of the project = budget to generate materials. CG Archviz studios on the other hand do this all the time, different business though. If it’s for a developer most of the time that would be outsourced to a viz firm unless the Architect supplies the developer with something already generated in house.
In this day and age everyone cares about cost, the days of unlimited budgets are long gone.
'This is where you win by having a superior product which is based on superior tools, not on a free engine which everyone in China and India will be using for games and archiviz in a matter of few years."
There are very few projects that are large enough in scale to warrant that. If you’re a viz firm with those tools you’ll still be competing with low cost labor especially in the Architecture world as I mentioned the viz materials are considered overhead in most cases.
Sorry I wasn’t clear on that point. what I meant is that it’s not too much more effort to make a vr/app once you’ve built your unreal scene to make a video. But, of course, compared to a vray scene where you need model only what you’ll see in a movie, it’s more work!
So to my understanding we actually agree. I don’t want to argue about specifics as I’m biased by type of clients we were working with.
Returning back to OPs question. In my opinion specifically Archiviz isn’t worth it. But you still could do it if over the years you invest more time, into building a pre-canned functionality, than into making them look as realistic as possible. Later will always cost you more time and will scale linearly, the former will cut that time significantly allowing you to get more projects done in the same amount of time, improving your cash flow. But this means that something like 50% percent of your time will go into programming and later you could start looking into outsourcing or hiring someone for content production. Leaving you more time to work on improving your overal product.
To be continue… Making long posts from the phone is not so comfortable.
I don’t know in what direction this thread is moving on but one thing is certain, it diabolically failing to spark any critical response from major players in community. @Boredengineer is talking about superior tools to target specific clients with specific needs that too with creating custom pipelines (like UE is something cheap and unreliable). Don’t know about the others but I consider this a monumental task - creating custom pipelines. @Boredengineer - why are you rejecting “Game Development Engine” completely as a tool for arch-viz, you reiterated it in every post? It’s a software like others and can be used as desired. Unreal Engine just sparked a possibility in Arch-Viz and it is yet to be tested in market. Nothing can be said with certainty as of now. Lets artists scrutinize it for a while. Definitely it’s not for beginners who are attracted by the newly created hype around RT and VR.
+1. I’m with .
I don’t know where you read that UE4 is something cheap and unreliable. If I would have to choose which game engine to use as a basis, I would choose UE4 at this point of time. But in its current form it’s not well suited to this kind of project. You have to invest plenty of time into tooling it first.
Let me give you a simple example, let’s say we have a model of 5 floor office building. You can easily have about different 30-40 materials and plenty of geometry which should be rather instanced than used as a single mesh. So how does current import pipeline looks like? Are you ready to detach all of your unique door meshes, move them into 0,0,0 in 3dMax and export/import individually just to be able to animate them? What about object which can be easily instanced in modelling packages, like window frames and shutters? You can’t import locations for instances out of the box. There is no pipeline for importing splines either. UE4 is simply not made to tackle such time sink as it rarely happens with game content, but this is a norm for archiviz.
The next question is what are you going to do in 3-5 years when real-time raytracing will be a common thing? Gamedev won’t start using it as easily as there are little benefit from it to gamers - hardware concerns. For archiviz it’s a completely different story as it allows you spend even less time on modelling and cleaning your content, while getting lighting quality similar to offline renders. This was done at least 3 years ago:
I completely agree. After about two months of use I can’t see UE4 in high volume production use, one off projects on the other hand it could be good provided there is a budget for it. Architects don’t have years like game makers to develop viz we have days and weeks.