Corridor Crew’s Wren uses Render Network on Ants + understanding NeRFs: February 24th, 2023[Behind-the-Network (BTN)]
Over the last few months internet maverick and Render Network user Wren, an artist over at Corridor Crew, has been producing some incredible work. From introducing millions of viewers to the next generation of photorealism with NeRFs, to visualizing the unimaginable (with the help of the Render Network) in a video on the world’s largest population (not humans!), Wren and the team at Corridor Crew continue to show that the limits of VFX are only as limited as our imaginations. With all of the interesting info he’s shared, we at the Render Network wanted to pay the love and attention back, and showcase some of Wren’s awesome work to the Render Network community.
Ants, the Unimaginable, and 2 day Rendering
Ants make up a foundational part of the World’s ecosystem, though the sheer number of ants in the world is nigh unfathomable… at least until Corridor Crew’s recent video, where Wren did just that, rendering out at multiple scales what the entire global population of ants would look and behave like in real life. All of this was powered by the Render Network, and thousands of decentralized GPUs working together to render a cutting edge visualization.
Throughout the video there are hundreds of ants rendered in various forms: massive balls of ants; ants pouring through the streets of a rendered Los Angeles; ants being fired out of a Nerf gun and more. In fact, to render all 20 quadrillion ants as individual instances, at the average size Wren calculates in the video of 4.5 mm, would have been hours of time of constant rendering, hours which Wren did not have. So he turned to the Render Network for the first time, and as he put it after his experience:
“I’m going to be using the Render Network for pretty much all videos going forward from now on.”
This being his first time using the Render Network and being up against a clock, Wren reached out to the Render Network support team for help with getting his scene on the Network. As he readily admitted in the above podcast he was initially hesitant to use the Render Network because “you have to like bake it out to an ORBX file it’s a bit of a thing and it confused me for the longest time”, but with the Render Network support team walking him through it and the glut of new tutorials available recently, he was not only able to get his scene up and rendering, but did so easily.
“there’s a bunch of new tutorials on how to do it these days, but it’s a lot easier than I expected”
Not only was it easier to upload, but the sheer speed astonished him. As he estimated, he uploaded scenes with hundreds of frames that were relatively complex, that would have taken his local 4090 GPU 8–12 hour to render out. Instead, they took the length of a few minutes long conversation to render and be production ready, just enough time for him to explain to a coworker about the Render Network. This issue of time and pushing up against deadlines is such a common occurrence for creators like Wren and those that use the Render Network that has been discussed by the Render Network team ad nauseam, but as Wren emphasized.
“yeah and so that’s always kind of been a bottleneck in a lot of these videos is that I just literally don’t have enough time to do the renders but with this I was able to just upload it out and continue working on a shot”
While he encountered some common issues, with stalled frames needing to be re-rendered or expensive renderings when using the priority Tier, with the help of the Render Network support team he was able to cruise through them and has decided Render Network will continue to be a part of his heavy workflows going forward! We’re more than happy to help bring these amazing videos to life Wren!
NeRFs, the future of photorealism
A few months back Wren released a video detailing a new technique that he believes will be the next step in photorealistic rendering: the use of Neural Radiance Fields, otherwise known as NeRFs. Sharing a view towards the horizon of creation software and tech with Render Network cofounder Jules Urbach, Wren displayed in the above video some of what NeRFs are capable of currently and touched on the immense potential they possess for the future of photorealistic rendering. As Jules discussed in the Render Network Telegram channel last spring,
Photo-scanning has become a common hand tool for creating photorealistic models of real-world objects that can then be retrofitted and edited within editing software. However, it has its limitations — namely when it comes to reflections and transparency, two of the more difficult factors for any software to properly account for. That’s where NeRFs potential to revolutionize creation becomes apparent: similar to a photo-scan, a NeRF uses IRL footage, but then uses AI tools to recreate the space and objects within them as accurately as possible, including highly reflective and transparent surfaces, as Wren demonstrates in the video above. Once the NeRF has been created, camera movements can be made within the scene and the reflections and lighting are accounted for in real-time, changing in time with the digital camera movements.
This is true for additional objects added within the scene as well, as Wren shows by dropping a rendered blanket over the top of the footage of his co-worker within the scene. Once the object is in place, all of the realistic lighting and reflections are carried over to the new object, with texture and reflective surfaces accounted for seamlessly.
While the current state of NeRFs is in its infancy, their incredible potential has been recognized by luminaries like Wren and Jules, who has stated that NeRFs represent a massive step in the field of ray-tracing, and will play a part in the future of the Render Network. We highly recommend watching Wren’s video in its entirety, and to keep an eye on both Corridor Crew and the Render Network for more updates and follow-ups on the ever evolving field of graphics and rendering.
Join us in the Rendering Revolution at:
Website: https://render.x.io
Twitter: https://twitter.com/rendernetwork
Knowledge Base: https://know.rendernetwork.com/
Discord: https://discord.gg/rendernetwork
Render Network Foundation: https://renderfoundation.com/
