otoy render offline

This is from a pre-computed movie created for Architectural visualization purposes

In 2008, a company called Otoy has set the goal to revolutionize gaming by offering vastly superior rendering capabilities by replacing your PC or console with a massive cloud (farm of servers). Today, Otoy is still pursuing this goal, but the company has slightly changed its aim. In addition to game developers, Otoy will now provide tools for film makers as well. While powerhouses like Pixar, Dreamworks or ILM have their own rendering farms, Otoy may be an alternative for smaller studios (or schools) who would not otherwise have access to this technology. If it sounds too good to be true, it’s because it partially is. Check this out.

Platform Play

Instead of (only) building “a service” like Onlive, Otoy has  chosen to build a platform, which consists of several tools.

1/ Octane Render is a ray-tracing engine variant that will be available for a low price. Otoy describes it as using a “spectrally correct graphics processing unit path tracer”, which basically means that it is a physically-based image rendering model that should render very pretty images (path tracing is considered to be a generalization of ray-tracing). Here is a video from a student project. Note that this is not rendered in real-time.

WakYak – Architectural visualization from Erik Jansson on Vimeo.

2/ Brigade is a real-time path-tracer. As you can imagine, the quality level is not going to be as high as a pre-computed movie. There are a few demos of Brigade out in the wild but none of the videos show a real mix of amazing rendering and high framerate. And unfortunately, nothing comes close to the recent Battledfield 3 demo (see below). As more developers have access to brigade, we may see higher quality demos.

End Game

Despite making some changes to its plans since 2008, Otoy still has the same end-game in mind: using its video streaming technology to send whatever was rendered on its cloud to the end-user display. By fundamentally using off-device processing, Otoy would allow the user devices to be terminals, rather than computers. This would decrease the hardware requirements for end-user devices, make things less expensive and reach more people, more often.

Nice idea.

Sounds great. What can I do with it?

Unfortunately, as a consumer, there’s not anything that you can play with. As a student or a developer, you may have access to this technology and we can only hope that cool products will come out of it. Keep in mind that this is a very ambitious project that has been talked about for years (since 2004) – but has yet to yield  a single revolutionary game. Actually, a game -any game- would help too at this.

But… will it work?

From my point of view, Otoy is really working on several products here:

1/ The offline rendering tool can function as pitched. This is something that has been done in the past (although with different techniques), and I’m 100% sure that Otoy can pull this one off. Most of the real-world problems related to “off-site rendering” (other than bandwidth) have not been technical problems.

The main reason why companies like Pixar, Dreamworks or ILM will *never* use a public cloud can be described in one word: confidentiality. They will not send any data outside their walls, period. That said, schools, architects or freelancers could be a great customer base for Otoy’s offline rendering product. They just need to find a way to make money with this concept.

2/ The real-time “uber gaming” idea is going to be much harder to pull off. Back in 2008 when Otoy first pitched this idea, I had my reservations and I had described the many issues that Otoy would bump into. If you are curious: read OTOY Server-Side 3D Rendering is Taking the Wrong Path.

Three years later, Otoy still doesn’t have an awesome-looking game in store. Frankly, I don’t blame them, this is a very difficult task, but nothing that I have seen so far from Otoy tells me that this is going to take off. Obviously, with enough work, one could build a good-looking demo, but deploying it on a real-world scale is another problem altogether.

That said, Otoy knows that it has to put developers behind it, and that’s the really hard part. Right now, developers have exactly zero incentive to get on board. Despite the huge technical challenges, this is a classic chicken and egg situation.

At the moment, developers are doing just fine with the current hardware, and a new crop of 3D engines are integrating global illumination at framerates that gamers won’t laugh at. Below: Battlefield 3 game engine demo video, using current PC hardware.

Conclusion

It’s hard to not like the idea of having super-games powered by a powerful cloud, connected to my ultra-light device by an ultra-low latency network link, but I would love to see something tangible even more.

But at the same time, it’s good to see Otoy splitting its efforts into more manageable tasks and start using its technology where it can be useful – today. Offline rendering is something that could work, and streaming current games could work to. Ironically, Onlive did exactly what I thought Otoy should have done in 2008, and they are now leading the streaming game market, and adding features and developers on a regular basis.

As for the “photo-real” gaming, I don’t expect to see anything significant (a megahit title) on Otoy’s platform for the next 3-5 years. It’s basically a few game cycles for titles for the most graphically advanced games in the world, and none of them is screaming for help from a project like Otoy.

Link: This posts adds to the conversation started on Venture beat.

Filed in Computers >Featured >Gaming. Read more about cg.

User Comments