Selling realtime Arch-Vis

[tammo]

I am very impressed by the new Lumeon Videos.
Question is if the clients are ready for this.
Will they accept a lower visual quality for faster (cheaper) archvis videos?
Do they want to walk around in realtime models?

What do you think?
has anyone done a commercial "realtime" job yet?

[plastic_]

Well I haven't met clients who would be seriously interested in that. Once in a while some architects ask me about animation, but then they don't really want to pay for it because for them it's more like a gimmick. When I tell them that a short animation would be more work than all stills together, they completely lose interest.
Those realtime walk-through things are a dead on arrival. As long as there is no universal web standard to display them (without shady plugin), there's no chance it will ever be of use.
What "Lumeon Videos" are you referring to?

[tammo]

I guess you are right.
I was impressed by the included Library of animated objects.
Talking People, animated water with surf, Trees blowing in the wind, circling birds, animated cats sitting around.
Everything created in realtime with two clicks.

The video was on the lumeon 3 website i think,
or on youtube?

I guess you model the house in max and finish the scene in lumeon.

[super gnu]

for me the rendered results look bad, but that volumetric sky system is pretty damn impressive especially if its realtime..

[Alan Iglesias]

I do have clients that want to be able to walk in real-time through virtual interior sets - with a relatively decent photo-real look of course. I'm looking at baking in textures from a Max/Vray file and then going to Unity with it. Anyone out there doing that?

-Alan

[Neilg]

Anyone out there doing that?

Tried it, it's way more work than you realise.
Realtime is simply not worth it (at least for the kind of work we do), nor do I think it ever will be.

[Alan Iglesias]

Tried it, it's way more work than you realise.
Realtime is simply not worth it (at least for the kind of work we do), nor do I think it ever will be.

He he...actually I messed around some and you're right, it does seem to be a major PITA with no established smooth workflow - that I could find, anyway.

I'll keep researching and if I find a decent way to pull it off I'll post it here...

-Alan

[glorybound]

I have used Autodesk's Showcase

[Alan Iglesias]

I have used Autodesk's Showcase

I'm going to take a look at that, thanks.

-Alan

[trixian]

It mostly depends on what phase of the project you are involved in.
I have used it a few times for projects early on in the planning phase. You just need to make sure they know they will be getting this type of viz, where backgrounds and the extents of the model are a problem (they will not be correct or be lacking).
If your reference materials and base models are good enough it isn't too much hassle, but Lumion is most suitable for small-medium sized projects (area not amount of data), as there are no ways to populate areas with stuff like trees and grass etc, just a crude "place 10 random models at point of click" method of doing this.
Their engine is a mix of rather old and crude ways of rendering mixed with some pretty impressive stuff like the sky\clouds mentioned above, and their ocean\water stuff.
Gaming cards with a descent amount of ram are also much faster at rendering than almost all Quadros, as they have more cuda cores.

Also be aware that the devs seem to have a slightly different view of what is important for a proper work flow, and have some strange misconceptions about how stuff should look when aiming for realism, claiming things like ao\ambient shadows only appear in areas not lit by sunlight. Their forum is full of really strange posts from a few of the devs that made me seriously doubt their ability. One example is the new "pro" version of the program, where they proudly boast of adding render passes\elements like zbuffer, shadow, highlight etc. They just didn't realize that camera effects like noise and chromatic aberration should NOT be included in these passes.
Another example is that up until the latest release, you could not do anything on your computer while rendering, as the lumion widow needed to be at front and active. Now finally, there is a background render option so stuff like windows update or other pop-ups don't halt your rendering on the night before your deadline.

[deflix]

Tried it, it's way more work than you realise.
Realtime is simply not worth it (at least for the kind of work we do), nor do I think it ever will be.

This is partly true.
Real-time will actually end up becoming the only solution and make offline rendering completely redundant eventually.
This is primarily down to the scalability of modern games engines and the convergent functionality of GPU/CPU rendering, combined with the huge market forces pushing development.

Of course this is specific to the evolution of the 5 leading engines, and does not include the plethora of half baked low-end real-time products aimed at us.
The 5 engines are:

Cryengine3
Tech 5
Unreal 4
Source
Frostbite 2

It is still some way off however and for arch vis even farther away because (currently at least) the skills required to build real-time assets differ considerably from those required to make CGis and animations.
Architects are also notoriously boorish and skeptical.

Despite this the results can be impressive when using the right tech, and are not as tricky as you might think to create.
Here is an example of a decent model made for Cryengine3
http://www.youtube.com/watch?v=BnRapBq7h5A

[glorybound]

In my opinion, we are going backwards. Billing rates are going up, but we are charging less per project. So, we are having to find ways to do more with less, which translates to the cutting of corners. We have real-time, it's called Navisworks. I have Autodesk Glue on my iPad, which is free, and allows you to navigate a Revit model in real-time. This, to me, is where it'll be stalled for a decade.

[Neilg]

http://www.youtube.com/watch?v=BnRapBq7h5A

I think you're assuming i've not tried that much. That crysis example is nowhere near good enough to be shown as our work, and neither is anything i've seen real-time outside of the meticulously designed unreal engine demos.
The amount of work involved required to create something which would be good enough is far above and beyond what it currently takes and so long as you need to bake maps, re-build materials and export models from one package to another while keeping poly count into account it will always be more work than it's worth.

[deflix]

and so long as you need to bake maps, re-build materials and export models from one package to another while keeping poly count into account it will always be more work than it's worth.

Agreed that it is currently still some way off in matching top tier quality such as your work . Having said that Ive also seen lots of offline work that is not as good as current engine benchmarks.

The effort required is high because most developers view production speed as a low priority compared to gameplay features and graphics improvements, so tools are rudimentary and definitely not geared for arch vis. Despite this significant work is being done to improve production efficiency with the likes of Mega texture and Mega geometry already on the table. Removing the need to be efficient with polys, create LODS, or meticulously craft multipass textures will be a big deal and is just around the corner. Developers realise that despite the relatively low cost of artists there is a benefit to being able to turn AAA titles out in months as opposed to years.

Having said that, most decent games artists i've worked with can still produce real-time content at a similar time/quality level to a typical offline animation project, whilst providing far greater flexibility in the final product.

Things will only continue to improve so I would never say never to this force of nature - That's my 2 cents.

[Neilg]

Things will only continue to improve so I would never say never to this force of nature

This is true, and I'm very much looking forward to seeing where it ends up - I'm just slightly less optimistic