David Ariew

I can’t believe this is real!!! 🤯 Getting my art onto the Sphere has been my biggest goal for years and now it’s a reality. I’m so grateful and honored, and stoked that @Abel_Okugawa joined me on this ride. His music is the other half of my art 💛

Yeah I totally agree that closeups on the face should look better because there’s way more underlying data to work with. It’s the same with all the image generators right now too, that the medium and wide shots fail to produce consistent faces compared to tight shots. And yes I also agree about the lighting, where in some cases it feels studio-lit and not matching the scene context very well. Totally an issue but I still prefer the realistic faces to the game faces, even though it’s not perfect yet. I’m sure when this actually releases, people will be toggling it off and on constantly to see what they prefer, and it’ll vary greatly from scene to scene how well it works.

@DavidAriew Lastly, I'm curious to see what happens to faces when you scenes that different focal lengths, dramatically different lighting conditions, maybe differents to the face itself (different makeup, added items like scars, eyepatch, facial hair, etc.) comparing cutscene to in-game

@DavidAriew wanted to revisit the discussion. I think I see where exactly the where issues occur. Comparing the cut scene with model with the filtered one. Lighting difference aside, you can see how much the features are rendered differently. So couple things to note…

Gorgeous work as always! Beautiful motion.

@100PercentRobot No clue but it speeds up the viewport a tonnnn

@DavidAriew Looks a lot like DLSS in Blender. I wonder what model, or if they're just not saying DLSS because of the last couple days

Neural rendering in Octane 😱🤩🤯 Gotta try this immediately!!

Yeah Ian Hubert is absolutely insane. Best one man band 3D artist I’ve ever seen. And yes I agree that AI is stupidly low effort compared to 3D but I’d be willing to bet people would still want to watch one of Ian’s AI films more than someone else’s if he jumped in, because he would polish it to a crazy degree until it’s amazing. The artists driving the tools matter and with this new process you become a director rather than a technical artist, which sucks because we love the technical but it can also be very freeing. You still have to know exactly what you want, you have to direct the camera motion, the environment, the actors, etc and try to piece something together that’s coherent and compelling, which is nearly impossible with AI in its current state, so yeah it’s still a challenge and a lot of work. It’s just different work and a different process. I also think people will always be interested in purely human created art, regardless of the medium, so I don’t see 3D art ever going away, or at least not for a while. But you can’t be mad at artists for jumping in and wanting to see what their vision and creativity looks like with these tools, and we have to adapt to the reality that they’re here to stay. I love the process of 3D but I’ve always been willing to take any and all shortcuts to get to the best looking final product possible, and that has set me apart in my career. It’s allowed me to create great work for clients but also be fast, which is equally important. So if I can create a shot with AI 10x faster and it looks better than what I would do in 3D, then I’ll lean into AI. But even more interesting to me is the prospect of using 3D to create previz quality work quickly and then having AI do all the texturing, lighting and rendering. Or asking AI to rework a section of the shot, add more realism to a character, or micro details that would be painstaking and mind numbing to create in 3D. Facts are that the 3D process is too slow for a single artist to create a feature length film, and AI will allow us to do that soon. It’s a new and different art form and right now it seems like a level playing field where any rando can make something crazy, but I believe as this tech grows in complexity and allows for more art-directability the cream will quickly rise to the top and artists with an understanding of lighting, cinematography, and storytelling will thrive, and their work will be immediately recognizable over the other “slop” Anyway ramble over, probably went off topic. 🤷‍♂️

@DavidAriew ian hubert is a prime example. people love dynamo dream but artsist equally love all the showcases of what goes on to create such work. no one would care if it was just created by throwing promots at a screen untill he found one he liked

its a shame to watch somone who used to be the guy to go to for octane tuts become a ai shill. over the past few months every time i have seen this guy post its been reaffirming ai slop and its just sad to see

I haven’t stopped improving in 3D and I haven’t given up on it. I think the possibilities of AI driven by 3D are limitless and super exciting. I’m also realistic about the actual ceiling on 3D rendering and what it takes to produce the best results. There’s always been a massive gap between what solo artists can achieve and huge teams that create visuals for blockbusters. There are a few outliers, absurdly talented people who manage to do it all themselves, but it’s insanely rare. I see myself as one of those people though I know there are many who are better than me and totally blow my mind… and I can count them on one or two hands. As for AI, every 3D artist I knew got into this to be able to make their own movies on their computers at home. It’s that dream of downloading your ideas directly from your head and into the screen. After so long working in 3D, I’m very aware of its limitations and AI is the only tech I’ve ever seen to begin to turn that dream into a reality.

@DavidAriew i say this with no malice because i respect your hard work and content over the years but thats the ceiling you chose. you made the choice to stop improving and rely on ai. there is countless showcases of artists displaying photorealism without the use of ai or need for it.

I mean it’s not super easy to admit but it adds photorealistic details into scenes to a level that I could never achieve after a 15+ year career in 3D, so it’s not just better than beginners or lazy people, it’s better than I am. Not at abstract scenes and stuff I make in 3D that’s literally never been seen before, but better at stuff that has a ton of references to pull from.

@DavidAriew i forgot to type this when i was writing this out but it does not see patterns better than most humans. human evolution is quite literally pattern recogition. most advanced skills use pattern recognition to the maxium. its just better than sleezy and lazy people

People hated Supersampling Upscaling Dynamic Resolution Gsync HDR OLED Raytracing DLSS Frame Generation Path Tracing VR 3D HD 3 buttons 4 buttons Shoulder Buttons Analog Sticks Dual analog sticks Rumble WiFi 4K Portable PCs/Switch Gyro Controls All new tech gets hated on

It’s giving 10% of that power to AI at best. This is seriously the most nondestructive use of AI I’ve seen in that it’s adding more detail but keeping the structure nearly the same. Based on your other comment it sounds like you’re more worried about being written out of the equation entirely, and none of us know whether that will happen or not, but ima keep creating art forever regardless, with and without AI, and you should too.

@DavidAriew @gonzzzalo_ This argument holds zero weight unless each studio is training this on their own individual work and nothing else. This is LESS art directability, not more. Putting the creative power into the hands of a model instead of the people who actually made the input. Jesus

Everyone is so mad about this but it looks like an upgrade to my eyes. Video game characters look like ass on the whole and this does improve the photorealism, full stop. This wasn’t meant to be a stylistic game. It was trying to be photoreal and ran into limitations. To me this is actual utility for AI, similar to what Magnific can do to add micro details and that extra layer of polish to low res textures. Yes, the animation is still janky and to a degree it’s like putting makeup on a pig, but these are the steps to visual improvement and I’m sure that’s next. As someone who has worked in the 3D industry for over 15 years, I also feel that the tools have plateaued, and what I have access to today isn’t significantly better than what it was a decade ago. This feels like the jump from CPU rendering to GPU rendering and all I’m seeing is hate, why??

@Zants It’s also not in post because the artists and studios have control over where it’s used, how much, and when.

Since when is removing cast shadows and drastically changing mood an improvement? It baffles me how any art director would be okay with this shit happening in post. This looks soulless, and opens up the doors to a ‘design to final image pipeline’ that skips artists entirely. ???

I mean turn it off if you don’t like it. It’s also fully customizable by the devs and they can dial as much or as little in as they want. In the example you shared, the right image looks more photographic and the left more fantasy, so you could argue that it’s not working there, but I’ve seen many examples where it’s a clear improvement. The point imo is this is a jump in tech and it will only improve from here.

And yet it gets closer to the real world than actual render engines, so it’s estimating very very well. The stealing debate is something else entirely, and I believe that AI is actually doing the process of creating art, except that it’s “seen” and understands the patterns in the refs it’s seen far better than any individual artist possibly could. In the same way we’re seeing all coders being replaced by AI and now they’re vibe coding, 3D artists and visual storytellers will be the ones who drive AI best imo. It’s not the future I expected but it’s here in front of us and it’s not going away. I’ve been using AI a lot lately, and to get a great result it’s not easy, and it’s still a lot of work. I end up creating orders of magnitude more content than I do with 3D in the same time period, but I’m still working just as hard as I would on a 3D project. I agree that this is startling in how it democratizes the tech, and the barrier to entry is so small, so there will be a ton more people with access to tools to create amazing visuals, much like when DSLRs hit the scene back in 2008, and it pissed off all the “real artists” and professional photographers. Ultimately I think this will only create more jobs than before (I hope) and to be a well paid artist you will still need all those soft skills and years of experience, which is why I’m not panicking about losing my job.

@DavidAriew ai is not rendering advancements... its esitmation, which in most cases is built by stealing real artists work. people want to become artists without putting in the effort and in most cases devoids any of the original artist intentions.

Reporter: Some are concerned NVIDIA’s DLSS 5 will make games look worse or homogeneous. In our Q&A today, Jensen Huang responded: “They’re completely wrong… you can fine tune the generative AI to your artistic style… if you want cartoon, toon shader, made of glass… it’s up to you. It’s not post processing at the frame level. It’s generative control at the geometry level. This is very different than generative AI. It’s content control generative AI.”

Sure, I agree that it would suck to have a one size fits all process applied and I know that feeling when working in my field of concert visuals, like when I create something pristine with my vision and then the person programming the content completely changes the color and fucks up my work. But I just read an article that says the game devs will have total control over how much is applied to specific characters, environments, objects etc, and there will only be more controls as we progress. Oh and to answer your question about Twitter, you can totally write long posts like this. It'll just include the "show more" twirl down.

@DavidAriew I rather have a negative overreaction than a dismissible one. Video games are a confluence of art and technology but it's worth defending the former from the latter when engineers overstep.

@Dave_MW Agreed! Yeah, the lighting is nicer on the whole but often it's not that well integrated.

It's def the future, it needs to do a better job of simulating the correct lighting of the faces though. It's def adding a more "studio"-lit look. Same with the environments, it adds like an overcast sky to everything. It also seems it's basically a filter that's only controllable with an object based mask, so don't think it's very art directable.

@100PercentRobot Same!! It's not that different from video to video restyling (Kling has a transform shot option that works pretty well) or what Magnific does when upscaling. The tools are still janky but hopefully with the next iterations they'll become a ton more useful.

@DavidAriew Good read! Can't wait for this tech to be implemented in non-real time renderers. Someone has to be working on that, right??

Great article that dives into the complexity of this rendering breakthrough, rather than the anti AI knee-jerk reaction seen all over Twitter.

Those freaking out about AI face filters are obscuring the bigger picture. DLSS5 offers a generational leap in graphics, made possible with software breakthroughs rather than more powerful hardware. Devs will finesse the application for new releases and they’ll look incredible.

I have two functioning eyes. Sorry, I won't pretend this DLSS feature isn't impressive from a technical standpoint.