How Will Games Evolve Beyond Ray Tracing/Path Tracing?
Вставка
- Опубліковано 8 лют 2025
- ► Watch the FULL video here: • DF Direct Weekly #154:...
► Support us on Patreon! bit.ly/3jEGjvx
► Digital Foundry UA-cam: / digitalfoundry
► Digital Foundry Merch: store.digitalf...
► Digital Foundry at Eurogamer: eurogamer.net/...
► Follow on Twitter: / digitalfoundry
What about interactions between materials? Mud sticking on objects, snow melting, cloth getting torn by bullets or caught/torn on branches while walking through a forest. Branches breaking off trees, walls being destroyed (Red Faction style). These are all simple individually, but difficult to implement collectively. I’m sure these would be a big evolution for graphics.
Rdr 2 did a good job with some of this. Love RAGE engine for sim stuff.
We didn't even figure out physx fluids and went straight to RTX. I want realistic fog/atmospherics and liquids. And better enemy AI FFS.
Photorealistic fog and smoke would be amazing
@@AdamMi1but no because we need ray tracing mode and DLSS 6.9 ultra upscaler ( the image is blurry as shit)
I honestly believe physics based objects will be what takes center stage next. I just don't mean waking around and hitting a bucket. I mean smoke, water, objects being wet/burnt/frozen, gas moving around the air, will be as big as Ray tracing in making the world seem more believable.
NPC AI has to be dumbed down otherwise it wouldn't be fun. You are talking about something that would be able to instant head shot as soon as you got within sight range.
@@gothpunkboy89 You can have a challenging and intelligent AI without taking the fun out of it.
Fluid physics. It is really immersive breaking seeing bad water splash, river ripples when the player jump and sea waves
Physics is also harder to do in a multiplayer game and today many games are made for multiplayer first, so it might be one reason why physics has taken the back seat
Its not that hard, and already been done before, since physics is deterministic, syncing over network only means, syncing initial conditions not entire physics simulation, The Finals does this, battlefield 3/ Bad company 2 did it in 2012 with 32/64 players and 1000s of projectiles and 100s of destruction events, modern games are not even close to that level of physics even though we have 10x more CPU power now.
@@niks660097 physics is not deterministic, in fact it takes effort to make it so. In Unreal for example you now have the option to use a separated thread to do the physics computation with its own tick rate. Before you were bound to the game thread and based on the frame rate of the PC you could get different results. But even in this case, based on the performance of the machine, you can get divergent results because the simulation is discreet. To make it clearer, if you run at 30fps instead of 60fps, a falling object will step into the ground more deeply, because the delta between one tick and the other is longer, so the collision will change and therefore the feedback will be different. If you think about BF, the effects were mostly scripted and visual, the debris were not really walkable. The finals is better in that sense, but it requires a special tech that runs server side. Also is very scripted and approximate. So no, it's not easy at all.
@@niks660097Physics is not deterministic when there are people on opposite sides of the connection both interacting with it. You would have to implement some type of rollback across the simulation at a minimum. You also have to worry about how different hardware handles value truncation etc.
@@JohnSmith-sk7cg "Physics is not deterministic when there are people on opposite sides" its still deterministic even then, its always deterministic, unless you are talking about quantum physics, but harder to sync, but that's only a problem when multiple players interact with the physics objects exactly at the same time, even in those cases using priority or FIFO you can resolve it.
@@niks660097 I think you misunderstood what I was saying. Physics is not deterministic on a client side render when the other client is interacting with said physics and changing the state. If you do understand that, then I think you need to reread what I wrote. Again, even if you find a way to state-sync using something like rollback to handle the changes between ticks, it's still not fully deterministic due to the way different hardware will handle value rounding among other things. Making it so that while each set of hardware could follow a deterministic procedure, the end result could be vastly different across different architectures due to chaos.
Alex, how long will we have to wait for your Physx video? I'm really looking forward to it.
after Matrix demo and Marvel, graphics are hitting the top, we don't really need better graphics, now we need better physics, NPCs, and Better Games
Looking at some of the physics stuff in Unreal engine I could see it becoming more common
And less microtransactions.
@@pilouuuu maybe microtransactions are good as long as they don't affect basic game experience.
think about it, if rich people can pay stupid avatar shit, they can make the game better and cheaper for the rest.
This is very true. Just started Horizon: Forbidden West on PC with the graphics pumped up (including DLAA) and the graphical fidelity, image quality and facial animation are INSANE. Still though, some physics have noticeable issues, and I definitely see that as the next area of improvement, as mentioned in the video. A true, fundamental shift to simulated clothing and hair that properly interacts with the characters and environment would be so cool to see
those last two words really sum it up in my opinion.
Look, I love RT and tech in general but looking at my library the games I enjoy and play the most, look like they were made in 1997. I also play the cyberpunks, the god of wars and so on. But at the end of the day they feel like empty shells. I think path tracing is insane from a tech perspective giving a whole new kind of visual fidelity, but for me it doesn't make the game better or more fun, as of now.
in the future we might have games that can be FUN while being impressive from the tech side and not be a soulless corpo cash grab by. Or maybe I am just wrong and am coping. Who knows really.
I couldn't care less about Ray Tracing. I want inventive, creative new ideas, games that are unique, atmospheric, immersive, and FUN, not necessarily realistic. Being well optimized in 2024 is also a nice plus!
Also, one thing that truly fascinated me recently, was the body and facial animations in Alan Wake 2, that was amazing.
Efficient physics would be sweet. Pieces of armor and clothing shifting on a character's body instead of stretching and relying on weight maps. There's a little here and there but we could go much further.
The "inseparable armor" part Alex spoke about, I never even thought about that. Yeah that would be insane if we had tech to animate that properly. Thats a small detail that devs would probably try to improve (Rockstar I could see doing that). Cloth assets should look like its own model atop the character. Should bend and move according to bone movement, independently.
UE5 lately is pushing for cloth simulation in the engine previews.
@@cube2fox that is cool. Cloth physics should definitely take the next step in rendering.
Everybody is guessing what Nintendo's next 'gimmick' might be for the upcoming console, and most people bring up unusual controllers and double screens and stuff like that (which could be the case, I don't know), but after the success of BOTW and TOTK I really hope Nintendo leans into game physics as a main focus of innovation. As the only console that uses Nvidia chips, they could leverage Nvidia's bespoke hardware to push the limits of their PhysX feature set. While PS and XBOX battle over graphics, Nintendo could dominate game physics for an entire console generation. I really hope they do it.
Seems unlikely, they stopped pushing for advanced compute ability after the GameCube.
I think the Switch 2 won't have a gimmick. Because the Switch 1 is already pretty perfect. They'll just bump up the specs.
Nintendo don't give no shit about graphics lol
That's the thing, though - they don't need an expensive graphics-intensive GPU or advanced hardware. They could leverage the advancements that Nvidia has already made with PhysX, maybe work with them to push the chipset's capabilities a bit further with regard to physics (a lot of those features are a decade old or more, after all), and then design their first-party games around that. That could create a clear advantage over other consoles while keeping the overall power of the hardware modest enough to suit the handheld format. Plus, physics fits well with Nintendo's 'games as toys' philosophy. The reason other consoles couldn't compete would not be because of overall hardware power, but because Nvidia is able to design their chips with certain in-house PhysX features in mind. That's why some old PC ports like the Batman Arkham series had Nvidia-exclusive physics features.
AMD will take predictive frame generation AI to the next level by predicting the entire game, rendering it, and playing it back with the title "this is how it would have gone"
Eventually, the games will just play themselves and we can just watch them. I can’t wait.
@@seth4321Mario Party 2 let you set all of the players (even your own) to AI, and then the game would just play itself. Did that a lot as a kid once I got bored
Google showed, last month (I think), a generative AI that does this! So, it's not that far fetched
amd is good at predicting in making black screens for gaming and app crashes as a feature
@@seth4321thats what it called youtube for...watching youtubers playing in misery of bad games
How about stable 60 fps? We seem to be still struggling to get that these days.
Only on consoles
Won't get a "wow" effect when showcased and won't reflect in sales, which is the bottom line.
Get a good mid to high PC and wouldn't have that problem. Consoles, as much as I love them, will always be behind PC in tech. Consoles are good for price, technology, and convenience since they appeal to over half of market share in the gaming industry. It would be wise to use that market to introduce new tech/sw but PC will always expand on it.
By ‘we’ do you actually mean ‘you’?
@@cygnusx7only on hardware that doesn't costs as much as a car and will be artificially obsolete in 2 years because for some reason a game that looks just as good(or worse) as a 5 year old one now requires a NASA shopping list to run at 30fps 1080p
Physics and animation please. World simulation took a giant back step lately. Real life animation and physics are so far ahead of many game even if the textures and lighting are closer now.
Anyone remember how amazing the Crackdown 3 preview videos were with the physics calculated in the cloud hype? But we never did get to play anything quite like that. I imagine a newer console or PC could do that without involving the cloud at all, but everyone involved in creating that tech probably moved onto other things long ago.
Meanwhile, what we're likely to get is even less movement in future 3D games as developers try to shoehorn something pretty but static like Gaussian splatting into them.
As your average gamer I also think the best next step for immersion would be better physics, honestly games look extremly good these days with RT and everything and it feels like the physics got left behind. Maybe there's a way AI can help with that
Yeah, I'd love to see some of that processing power in the future used for cloth, hair, water physics. Also ray-traced sound with accurate bounce and reverb. Also, enemy behavior AI could use some updating. I love visuals and lighting as much as anyone, but it feels like that's gotten all the attention for a decade, time to show sound and simulation some love.
I would imagine more physics not just for clothing, or better rag-doll physics, but character physics 100% of the time. Characters and NPCs actually have to use physics to figure out where to place their feet or interact with objects like doors, stairs or any presumably heavy objects.
What would be amazing is if studios would develop an AI that would tell them not to release a game until they can demonstrate it runs properly on intended hardware.
How about figuring out frame rates
Because the only thing that matters in games is the framerate.
yup can't even fix the frame rates and already looking how to make a features that will kill more frames.
@@istealpopularnamesforlikes3340 yes
Frame rate has already been figured out 20 years ago, it's not an aspect I consider when talking about next gen technology.
Stop whining and just lower your settings
How about that 60fps minimum for all games.
It is for PC
Fps freaks would literally prefer mediocre 60 fps games over groundbreaking 30 fps games
Just lower your graphic setting , we already has even on consoles since PS1 and PC. The problem is people tend to crank up the graph setting
its almost never possible, as better grafix keep pushing the fps back.., but if atleast 60fps was considered base..
...
but than you Switches & Series S.., low end will always exist.
Lmao the way devs and publishers release games now thats a far cry. Upscaling was the worst thing to happen to gaming. Now we have to depend on that for stable fps
7:29 I remember back when I played Dragon Age origin for the first time, and saw their steel armor bending as the character moves and it doesn't felt right, and I hate it when a game character have something like robes or any typed of outfit like for instance a lab coat, and it sticks to the character.
Games are more than just graphics, but characters having clothing move separately from their body (and hair that isn't just a solid clump of textured polygons) is really nice to have.
@@deanchur agree, I don’t care about the graphics or fps as long as it is playable and I would certainly enjoyed the game.
Another big improvement could come via marking certain materials as “not to be DLSS upscaled”. Things like fences or windows could be kept away from the DLSS upscaler, while other objects can be given to the upscaler. While this may seem non-uniform, it could mean developers identify problem-spots and manually fix those spots.
Then you would need to use another AA solution on top. But which one do you use? DLAA is superior in almost any situation except the ghosting artifacts on movement
The Nintendo Switch got me wanting more out of games than just graphics with Zelda BotW and Tears of the Kingdom due to the amazing physics, and Larian Studios with their level of interactivity with Divinity Original Sin 2 and Baldur's Gate 3. Also, VR opened up a lot of unique gameplay possibilities. While great graphics help, they are no longer the system seller for me. I would rather have stronger art direction and presentation and save the extra resources for interesting gameplay.
Another method for AI gen would be to have specific materials be tied to the AI generator. Something like having a cape meant to be generated by AI, so the game does a first-pass over that cape and the AI generator generates a more detailed cape after the first pass.
Agreed. Ive been wanting to see real time destructability and physics in games for a while. Atleast since the 7th gen. Also more realistic and massive crowds to fuck around with in games like GTA, HITMAN series. I assume its hard to implement otherwise we probably wouldve seen more of that by now.
PhysX type stuff certainly need to become a focus again. Cloth, hair, euphoria ragdoll, procedural animation, etc.
Better physics, fluid and gas/particle/atmospheric simulation. CIG is making some interesting headway with atmospheric/weather simulation. As is Microsoft with their latest flight sim. But I would love to see these items become more mainstream.
I think what should be next is voxel-based destruction with accurate breakpoints in destructible environments and objects -- just like game physics became a standard thing. Of course many games would want to leave this out because it may not fit well, but I could see the "next Half-Life" doing something like this to a degree that Red Faction or Enshrouded haven't explored yet.
Yes please more Physik for fog and Fluids..
Fog yes, fluids I doubt that will happen
Game devs should take a look back at physics from Red Faction Guerilla and Battlefield Bad Comany and F.E.A.R. and improve upon it. The race towards a perfect 60fps is useless without proper immersive/fun physics.
I agree. Ten years ago I thought most AA or triple A games would have utilized real time physics by now. Most focus has been on Raytracing, increased resolution and FPS. Can't understand why some gamers crave more FPS above 120 for example. 1000 FPS wouldn't make the gaming experience a lot better. Would like to see swaying strand based hair, cloth, particle physics etc. in upcoming games
We are doing a lot better with hair recently than a decade ago. Most games had static hair while now we are starting to see strand systems
@@crestofhonor2349Yeah, hair simulation has gotten pretty good in recent years. While we haven’t gotten to the point of fine hairs being simulated, we’ve made major progress. As much as I don’t care about Forspoken, I have to admit that the when MC’s hair is long, it looks and animates VERY convincingly.
@@Vorexia Squeenix does have some of the best hair simulation next to Capcom
The reason we went the path of raytracing is because new tech inside more powerful gpus is required. $$$ NVidia chose maximum profit instead of meaningful gaming experience.
I really hope to don't have to look at a 30fps ever again
The Finals has incredible levels of destruction with a AAA quality, and optimized really well. It's like Bad Company 2 times ten. Insane. That kind of stuff would be cool. I don't mean necessarily talking about blowing s__t up, I'm talking about that level of micro and macro physics. Very cool. I think softbody and cloth should continue to improve. With Omniverse level stuff coming out realtime, I assume UE will follow suite in the coming years.
After having seen published the TotK tech, I do hope that Switch 2 will be able to offload the raycasting the game does to the Raytracing cores to have way better sound ambience and sound effects.
When bullets hit stuff, it should damage the area it hits. So the body rigs should be more complex and able to be destroyed bit by bit where the bullet hits.
Basically more complex Far Cry 2 plants applied to more stuff.
Yeah, whatever happend to physx and hairworks, tessfx, etc?
I just want believable soft objects in character animation. I want a character to sit down on a chair and have their thighs deform instead of it looking like someone dropped a granite statue onto a pedestal. I want one character to be able to pick up an object naturally, then hand the object to another character naturally. I want someone to fold their arms in an intimidating manner and actually have their biceps and pecs touch each other without clipping through. I want a man in a business suit to reach forward and have the cuffs of their shirt show as the jacket sleeves go taut.
I agree - I think fidelity wise we don't really need to be pushing so hard and it needs to stabilize at a solid 4k 60fps over the next few years... However in terms of interactions and finer details withing the game worlds, i think expanding Physics and applying it to more objects (think the interaction of all the objects in Skyrim) would add so much to the immersion and role playing/immersive sim potential... (And just for the record, we don't need 8k for another decade at least lol...)
I think the next big thing will be using AI to make more of the environment interactable vs having to hard code rules for everything. For example, AI could just decide that 'this structure is made of wood and can be destroyed with this weapon' or 'this is a chair, I can sit in this' or 'this is a door, I can open it and fill the room with appropriate furnishings'.
The way things are going it will probably evolve by putting settings menus behind a pay wall and any settings above medium will be paid dlc
Next big thing will be NPC AI and also AI assisted and generated games.
And what about ray traced audio? Will we get that?
we already have it in Unreal Engine.
@@tehf00n An approximation yes.
@@nossy232323 But it is a start. Plus both Returnal and the Avatar game features it. Wouldn't be surprised if the next major update of path tracing will include audio.
I think current audio tech is pretty cool. All you need is a 5.1ch home theatre and then you can enjoy 3D/surround audio
Check out audio in Dead Space Remake
physics used to get pushed back in the day and for whatever reason game devs just stopped. this is one of the major things missing in todays games.
I feel like that stupid ragdoll physics came out and stole overall physics limelight. Devs then just implemented ragdoll and said there. Job done.
Myth, Halo and a few other games had phenomenal physics. Some had fully destructable environments even. We definitely started down that path and gamers and media loved it but again devs saw ragdoll, threw it in and never tried again thinking the job was done when ragdoll wasn't even real and was one tiny aspect, character death fall physics.
Oh what could have been. Physics cards were even made. Everyone was excited and that one tiny thing helped kill it, imo.
Hopefully games with good gameplay, that work on release / aren’t buggy rushed messes, and aren’t loaded with predatory money grabbing techniques
Maybe start bringing back physics heavy games? 🤔
How about going back to optimizing games or their engines. Or solve traversal and shader stutter.
Haven't even perfected Ray or Path tracing yet.
Not everyone has a £3/4000 PC to benefit from the 60/120fps Cyberpunk experience
They know, they were speculating about the future.
Yeh seems weird to speculate on something that's only in its infancy. Like where is Harry Kane going next? He just got there!! @@Vorexia
@@MeganeMondeoMX5 idk enjoying every game to it's limits with my 4090 it is just amazing I'm literally over 60fps for cyberpunk path tracing
@@MeganeMondeoMX5 What’s so weird about speculating about future tech we haven’t yet fully realized? That’s kind of what speculation is all about.
@@redclaw72666 The 4090 is also prohibitively expensive, that was kind of their point. But tbh, even the 4070 Super has very respectable RT performance.
Ray Tracing was a mistake, we should have focused on physics simulations and high frame rates, 60 minimum
It was not a mistake.... It was specifically to sell more powerful GPUs.
Whats about A.I. generative physics? Instead of real-time physics.
I just want better physics. Clothing that wrinkles, objects in the world, water that behaves correctly not like jello. I've seen games with amazing water but then the mist/spray is low resolution or the water interaction with your character looks off it completely breaks the realism. I hate when you kill an enemy and they fall as if they weigh a few ounces. Worst still if the bodies move effortlessly if you bump into them. I've been playing Arkham Knight and the game looks incredible to this day. The smoke physics from the Batmobile and smoke/fog in other areas looks great and behaves properly when you walk through it. It's a performance killer though. Literally halves the fps. With the recent advances in AI we're going to be getting better NPC interactions so at least we have that. People have modded AI responses in games and Nvidia showed off their AI NPC stuff recently too.
Lighting innovation: rasterized GI - ray tracing - pathtracing - spectral pathtracing - wave optics (the simulation of literal lightwaves - diffraction, interference, polarization) - the matrix😵💫
Cyberpunk was the first game that impressed me with asset density then horizon. I think nanite level quality will manifest in minor assets and main ones will be photorealistic. Totk level physics will be in all games.
I would like to see innovation with animation tech. Ultra realistic procedural animation on all characters. Photo realistic. Especially with online games.
I feel that we are within a few years from peak Pathtracing optimization, where it will be ran everywhere.
I do feel that audio Pathtracing is next, but will be so similar to light PT that it uses all the same pipelines.
After that, will be something entirely different. Perhaps particle Pathtracing for dirt, water, fog etc?
Will be fun to watch this unravel
Im pretty shure whats going on with nanite. Point cloud distribution/voxels with pathtraced lighting. It will be easiest to include a simulation per dot to real world application using physics engine tied to an AI. Like omniverse.
Before that I think AI will be used first to use a simplified physics-prettained set to simulate and build 3d models from a prompt. The prompting is created by itself in iterative ways depending on set preferences.
Each pass will use a picture to build a 3d model in a physex-lik esetting, where each pass refines and binds the AI generative model to a ’world-concious’ 3d model. The last passes use lighting-pretrained set, then atmosferics-pretrained set, then fluids pretrained set. Each bound to simplified physx-like dedicated simulators.
They bind the flimsy AI to an interactive world. But the generate the world together.
You can create worlds where the physics engine is set to low gravity, different speeds of light, or vlouds and so on.
The last thing that will happen is we put on a hst that is outputting a frequency that modulate a part of the brain found to induce a deeamlike state/resonating the pineal gland.
What i mesn is that the ai-generative solution is bound in each inference step as tokens or depth values. Where the prompt is refined to with those same markers as depth valus in an accelerated way.
You can use both promt and 3d-models to build something the old fashion way but refining it trough this joint accelerated fasion if you want.
Shsders would be bound in this step too before moving to point cloud / voxel. (Accelerated voxels.)
Improved interaction with more physics is what I wanted from this gen. All that CPU talk, and nothing is different to stuff from 2010.
Up to this day, there are only few games who get vegetation right or at least a little bit more authentic. In most cases vegetation looks like additional content which had to be included but is considered more or less unnecessary and the overall quality of it is left in a questionable state. Most vegetation is just randomly placing trees and flat bushes but there is no real system behind it.
No, stacking processors forever isn't the future. Skeptical consumers need to see VR movies firsthand to understand the benefits of interacting with media. This is the future.
I only ever loved the advent of path tracing and other newfangled tech because of the promise that it would allow games to look great, be dynamic, and easier/cheaper to make. Instead, the opposite seems to be happening and I wish we'd pump the breaks.
I feel like graphics are pretty good. Definitely want more realistic lighting.
I would like to see big improvements to NPCs just being more organic and persistent in a game world. Better physics would be nice as well
I think there is a lot to be improved in Ray Tracing and frame generation first. Upscale tech is already in a very good state in my opinion.
that's why i loved TEARDOWN so much
such a different pace than most games these days
want specific AI software & chips, to deliver next-level NPC behaviour & dialogues!
(Chatgpt-Skyrim like)
Sons of the Forest has plenty of good physx stuff but the collision detection is really, really bad and that’s something that should’ve been fixed in game graphics a long time ago. How are we STILL getting guns or swords that when slung on back go right through hair/clothes/armor or through a chair when character sits down. If a game world had objects that reacted properly when they came in contact it would make the world far more convincing
Feel like the 2 newest Zelda games will be a catalyst for increased use in physics in open worlds
I hate that developers completely gave up on physics as part of gameplay. I still, to this day, hope for us being able to shoot pieces of a building off that can fall on top of an enemy and kill them, and they can do the same. Which would create a whole new dynamic to playing games. You would have to watch out for possibilities in the environment that can take you out as well as the enemy. Which makes things much more exciting.
Every encounter a new one and different every time you play. As you said, you barely ever see it anymore. Which makes no sense, considering that it is built into UE5. but nobody is using it. If we are going to move games forward, graphics aren’t the only way to go. While I appreciate incredible graphics, I also appreciate physics more. Gameplay means more to the overall feel of a game. Until we can advance it, we are just playing higher resolution versions of games we played 20 years ago.
Exactly! Game physics has fell off since the days of Red Faction Guerilla and Battlefield Bad Company with honorable mention such as Soldier of Fortune and F.E.A.R.
Is this true? Completely given up? I think there is nuance to this discussion that is being missed.
Better cloth and water physics and better A.I. would be nice.
Alex seems to have missed THE FINALS (I know, they made videos about the game - just funny he didn't bring it up at all. It is basically the only modern example of a strong emphasis being placed on physics, especially in an FPS)
RIP Battlefield
Physics should be number one feature in games.
The major problem with physx was just the performance. But it was pretty fun to play around with the particle and cloth physics in BL2 and Warframe especially was very pretty with their physx effects.
I've been playing through the Crysis series, and Crysis 1 has amazing realtime physics simulation with the trees. Crysis 2 has a very dynamically destructible environment, like every pillar I shot at or rammed with a vehicle crumbles. That stuff is a very nice touch.
physx wasn't really that good. It was just good enough for the time. It had many bad sides including difficult to modify, vehicles were 3-wheeled minimum, reliant on a runtime sdk, fairly awful ragdolls. Chaos Physics in Unreal Engine makes PhysX look dated but does have the one issue of being reliant on the built-in code being exactly how it is and unless you are a UE low level engineer it's not too modifiable. It can be done, I know people who have done it but its not general purpose. Not to mention you need to do a lot of design time asset creation to make it work ideally, but overall it's much easier to work with than PhysX and Apex destruction.
I don't think Ray tracing is even going to take over before completely AI generated images step in to replace all rasterization and Ray tracing methods
I want realistic AI in npcs and enemies.
This will sound dumb but i have no idea what Path Tracing is
It’s hard to say what it really is without going into the academic literature - in games it’s basically just a marketing word.
Smh
It's full ray tracing or let's put it this way: a more demanding form of ray tracing, more accurate. There is a wikipedia page on it.
@@nolram heh fair enough :)
@@Z3t487 ah so it's basically more advanced Ray Tracing, got it :)
Give me a good/fun AI, fun and engaging physics, physics effect gameplay and the look of a game in so many ways. They could just stop focusing ray/path tracing for all I care, it have impact on games, unlike physics and AI. But I understand why they prefer focusing on r/p tracing, it's much easier then AI and physics.
Ryse Son of Rome did the armor thing pretty well in 2013
AI Dungeonmasters. I bet that's where we are heading.
Would love to see a more advanced PhysX make a comeback in games and full destructibles
Squadron 42 seems to be doing some interesting things with cloth simulation.
I love Teardown, sucks that Saber bought it and it started going down hill. There are still some very talented modders still working on it though
Hes what i don't get they give consoles low end ray tracing while performance still sucks btw we only need quality mode at 40fps
Unfortunately physics and gameplay is more difficult to advertise than pretty screenshots, and one sells much easier than the other to the mass.
Hopefully they will tackle world physics and AI again.
Generative graphics is the next real Frontier. The next two generations of graphics hardware will explode the capability of the hardware to generate from generative models. For a published example look at the Blackwell's fp4 capability, what is coming after that is fp1.65 with an even bigger uplift with almost as good output from the models.
We are back at multiplying the capabilites many times of the hardware again every generation.
And the inference models they will run can be front loaded with extremely heavy data processing in massive datacenters to produce efficient models the consumer hardware will run. There is a lot of research that is focused on this now and we already know the way it will go. Just watch, it will happen and it is the next massive step that will change everything. Even more of a change than ray/path tracing was.
I just want games to finally look like WoW cinematics. That’s all I’m waiting for.
The destruction in The Finals is pretty good and realizes what MS was trying to do with the stuff they showed off but never produced years go.
DeezeNutsTracing, TraceDeezeNuts is next. Instead of polygons its just billions of billions of deeze nuts.
The return of 6th and 7th gen style gaming
most these game devs never played Boom Blox on the Nintendo Wii and it shows
While the games industry is trying to figure out what the next "big" thing is ... can we have proper remakes/remasters of ICO, Return To Castle Wolfenstein, Ocarina of Time, Thief - Deadly Shadows, Half-Life 2 et al .. you know, while we're waiting for the NEXT big thing.
Half-Life 2 has an RTX community remaster in development. It will be extremely demanding due to the path tracing, though.
More optimised games engines useing AI, that will give more resources, for other thing like: object physics, rendering, better game mechanics...
After path tracing hopefully 60fps
They’ve honestly overdone the ray tracing thing. I like it personally, but not to the extend of path tracing. Most if not all high end gaming pcs these days aren’t able to handle the highest setting of RT (native fps and res.) with every other setting turned to max unless we go back to 30 fps gaming…. And now they just slap frame gen on it and call it ‘optimised’. We’re headed in a very bad direction sadly… the biggest thing is that most people just look at the shiny numbers and think wow, that’s a high framerate, I need one of those! And it’s not their fault, these companies are just finding new ways to play the numbers game and artificially produce products that seem way better than the last gen, when in actual fact they’re pretty much the same on a raw power level.
It will probably be AI and path-tracing as majority of all of the development. Path-tracing haven't even begun yet and it's only doing 2 light paths (or whatever they call it) to render the light "tracing" and it could be millions. So that puts a perspective to that going forward. The more the better and after a while you can't tell gaming from reality at least graphically.
Also the AI, to build the worlds, to function as the NPC and not be just dumb bots. They could be trained to behave like humans and be exceptionally hard and realistic as NPC and Combat.
Then more with physics as well simulating all aspects of that.
It's amazing how the guy asking thinks they'll begin implementing this after we've figured out Ray Tracing. NVIDIA will be pushing for this despite the poor RT performance because these guys need to pack in features to sell anything because that's the only way anyone is upgrading.
Its always the ugliest, most immersion breaking aspect of game engines that gets addressed first; shadow map pop in lead to ray tracing. For me it really is animation quality and facial geometry. Perhaps a system that allows ai to 'mo cap' video footage, saving the horse from being rigged up in million dollar mo cap gear. Or real world video streaming in game rather than rendering of backgrounds like in cinema.
games are getting super expensive and they get released on a bad state. Maybe we do not need more evolving and instead we need more stable games with cool stories and gameplay -_-
Better tech leads to faster processing and less time needed for optimization, which gives devs more time to polish their games
@@estebanalzate7609You are living in cloud cuckoo land if you really believe that fella
@@estebanalzate7609the past 5 years has evidently proved this to not be true.
I'm all for games being stable. But games should at least evolve each game and generation. Otherwise you would never need to purchase another game/console. Just play the same game over if the new ones not doing anything different.
If games don't evolve then you have to question what's the point in new consoles or newer hardware in general. Both can improve as long as priorities are kept in check. Many AAA games now are trying to appeal to way too broad of an audience and it's hurting them. A more focused vision would help a ton
Wave tracing is the next rendering paradigm. Certain effects of light can only be simulated through waves.
And particles? I'm assuming.
Seems a little niche. What effects would this include? Thin-film interference? Defraction? These are rarely visible in most scenes.
Take a look at this paper arxiv.org/pdf/2303.15762.pdf
Richard looking posh $$$$$
I imagine AI completely rendering the scene. It takes the scenery instructions and "interprets" it. Benefits would be to not rely on huge textures anymore but being able to interact with this AI that works as a playmaker, like in a D&D or DSA classic roleplaying game. It will - depending on the system power - generate the most detailed, realistic or stylised scenes and modify it frame by frame.
While I like the idea and I think that concept could be interesting to a certain gamer demographic, I'd wager most people wouldn't want to design or have too much input on their gameplay to get something worthwile out of it. In my experience, most working people when they come back from work they want to play, not put in too much effort and have some "fast food" style gaming to conclude their day. And since these people are the ones the current market is already focusing on I can't see your idea find large success with the AAA market in the near or distant future without some heavy guardrails in place that will guide the player to an appropriately nice experience.
P.S. Once implemented it's going to be perverted in the most asinine, microtransactional infested, psychologically optimized way possible to force you to spend more money.