What UE4 can do is very impressive. But when it comes to real time photo realism I think path tracing is even more impressive. For example this Brigade 3.0 demo: [this video]
Yup, both of them take at least 1-2 seconds to respond to mouse movements. Could you do WebGL and blend into ray-traced? Have you seen what Octane/Brigade (not exactly same tech, but close) can do in realtime? [this video] (you probably have) Anyways, Brigade is coming to cloud as well on AWS GPU instances, so it would probably be easy for you to just use their solution integrated within your modeling tool then.
In case anyone has about 30-60 minutes and is interested to get a quick glimpse of how easy (I guess meaning free and accessible at the least) it has become to do such graphics: - Download and install Blender 2.71 ( [link] ). On linux (Ubuntu) I did not even have to install it; I just extracted the tarball and ran the blender binary. - Go through this two part ceramic mug tutorial (30-60 minutes): [link] ... [link] As someone who does not have graphics training, I was blown away when I did this. Apparently there is this thing called 'path tracing' based rendering, that takes care of accurate lighting, as long as you give the correct specification of geometry and materials. Some interesting videos: - Octane 2.0 renderer: [link] - Lightwave: [link] - Brigade 3.0: [this video] - Alex Roman, the third and the seventh: [link] Brigade is an effort towards real-time path tracing, and it's predicted that within 2-3 GPU generations, such graphics would be possible in games.
Real-time path tracing effort in an interesting videos list.
There is the Brigade engine[1] that I know of. It's as far as I know a one man project. I haven't heard of an update in a long time though. [1] [this video]
well you COULD use OpenCL to "control every pixel". Here's an example of path tracing in "realtime" (ok, it's not quite there yet) on the GPU. [this video] check out this blog too: [link]
I wonder what convolutional neural networks bring to the table. Maybe you could use them to get rid of the noise, or then you could use ray tracing for the base image and train a CNN to "recolour" that image based on a noisy path traced one of the same scene. This demo is 4 years old, but the noise still seems to be a problem: [this video] If anybody's got more recent impressive demos to link, I'd like to see how things have been developing.
Video list and Hacker News comment labels from HackerTube (prompted by @karpathy), released under ODC-By 1.0, built on the open-index/hacker-news archive. Hacker News data is provided by Y Combinator; comments remain the work of their authors. Video titles, channel names and thumbnails belong to their creators and are shown via YouTube. HackerTube Watch is an unofficial edition. Data export of 2026-10-31.
Plays via YouTube's privacy-enhanced embed
6 Hacker News comments
From the HN archive; up to 20 kept per video, favoring different commenters. Recommendations first.
Citations, criticism and jokes
Read more