Playing around with Flocking, and removing GC calls, and optimizations for large numbers of independent groupings. I think this is way more than I plan on controlling at any given time, but 3,000 runs pretty smoothly even without the CPU batching planned in Unity 5.6.

A little more contextual example would be 47 Squadrons of 7 ships each with their own dynamic targeting / pathing. This video was captured while playing in Unity on the HTC Vive at 140 fps using an nVidia GTX 980. This also has an extra 200 draw calls in the scene just to make sure I have plenty of headroom for performance once everything is added.

I just migrated Project Neon’s Unity project to Unity 5.5 (from 5.3), and used the new Post Processing by Unity.

The big improvements here are that we can start testing a lot more things with VR (stereoscopic rendering, etc.) The visual difference is pretty stunning. Also was able to get rid of about 10 3rd party solutions we were using for Screen Space effects, and Tonemapping, etc. Smaller project size, and better visuals. Well worth the migration time.

Here’s a sample from some of the renders.

I’ve spent a lot of time getting a lot of trees into the latest Unity project trying to portray Lake Tahoe area and a proposed bike path along the North East portion of the lake. For this project we worked out a number of different workflows. One of them was baking the biking and jogging pedestrians physics into animation paths. This allowed us to setup pedestrians, give them a destination, and then capture them running or biking along the path, and still achieve realtime by pre-calculating all of the physics involved with them running. We additionally worked out how to get the trees and environment to look more realistic than previous projects. Gaia and Tenkoku plugins were a huge help, as well as Aquas plugins for Unity for the water and environment. We ended up going largely with Speed Tree assets, but had to work out a workflow for placing trees adjacent to the path very accurately to reflect the actual trees that would remain within the bike path corridor. Tricky project, with a lot of fun constratins, and I think the renders came out pretty nice.

Finally have a render from Project Neon that can be shared. We’ve made a lot of progress, and have most of the buildings, and vehicles in place for showing off what will finally be built. Took a lot of time building the platform to get these videos as output, but I think these videos look really good for conveying what will be built in Downtown Las Vegas.

GDC this year was a great experience for me. The galvanized opinions around which VR platform would be most successful, or widely adopted were energetic and food which will keep me motivated for a good while. My only regret is not having made more progress on my project, or being more willing to share the progress. I think I’m not quite far enough to share too much, and I’m optimistic that nobody sees my website.

My takeaway is that there are a large number of companies focusing on developing content for these platforms, and they will all have success based on the visibility which is available to the early movers. The only difference between any of them, and any of the independent developers is the location and recognition. Many of the startups are in SF, or Seattle, while many more are located in garages, or bedrooms across the world. It was very motivating realizing that many who are working for the larger teams are successful because of the large group of talent that they’ve gathered. While that seems interesting, I’m currently not sure how relocation would work, and I’m really motivated to find people here in Vegas who have the talent and interest.

The value of Networking is difficult to quantify, but the energy that is within the game industry is undeniable. And it was good to revisit such a condensed form of that energy at GDC.

So – first off – shout out to Real World Terrain.  Totally awesome Unity Plugin that saved my life.  Has a great list of features, but what I needed, was a rough representation of the Las Vegas Valley.   So Their tools allow selecting which aerial mapping server, and specifying the quality level.  So I broke Las Vegas into a 32×32 grid covering roughly 300 square miles, with 128×128 pixel aerials.  I also imported the same grid, with 2048 aerials, and a ton of triangle accuracy.  And of course something in the middle.  I then composited the three sources of data which were all in the same alignment, and ended up with 9 panels of aerials that were very accurate near the center of the project, and decreasing quality and accuracy out for hundreds of miles.  Gives a really nice accurate skyline appearance, similar to what you might see in Google Earth.   But this is Unity.  So let the games begin.

Next up is to get the Aerial topography from the local company, and integrate the more accurate project aerials into the model as well.