Could potentially make it right now.... technically. Maybe not 8k, but certainly super high pixel density panels exist. I saw a tech demo with so many LEDs that you'd think the panel was an LCD, but it wasn't. That said, what would make a mess is the driver tech, it did look bulky, you'd need to somehow move all that down under, behind the rotating connection.
@@TheRubiksrocks Things like the crazy demos on this channel are a reminder at just how beginner-level I am in all my understanding of electronics and tech lol
I feel like we are back in the 70s, early 80s where 3D graphics was so limited it was only capable of rendering wireframe models, even without occlusion. There must be a viewer headtracking for the system to figure out what to render and display and what not. It's pretty dope!
@@lievenvv It is truly volumetric, no doubt. Question is: do we want to render occluded objects? Look at the truck, do we want to render the left side if we see it from the right? It would enhance one-user experience sacrificing 360-viewing ability. Hard call.
id love to see it with a speaker, i know its not the point of the experiment but hearing the environment and things like driving into objects would be an easy addition to help so much in making out whats going on. like that illusion where balls look like they pass through when its silent and bounce when you hear the impact
My approach so far has been to first download a 3D model of a classic game map and see if it's even recognisable displayed at some sort of playable scale. Doom and GTA looked good, which is why I went ahead with them. Minecraft was surprisingly disappointing which is why I didn't. I'll see what Silent Hill looks like. (It also needs to have some sort of source available, or be simple enough to be a quick remake)
@@ancientjames Try Half Life (and the others, afaik, they are open). Heck, shoot a message to Valve and they'll probably give you the stuff. They like stuff like this.
@@aserta Core engine functionality is not open for GoldSrc or Source enigne. That being said, I believe this should techincally be feasible to adapt to it, still, but I couldn't say for certain, as I never had to touch base rendering functions.
@@ancientjames you can try implementing it in one of the emulators (like PS1), that would allow you to run many games with one change. Also might work great with 3DS emulation, since it had small screen and most games supported 3D effect and were tuned to have small scenes with mostly close objects. All objects are large, since screen was low-res. Something like Super Mario 3D Land might look great on volumetric display
Love this so much - I *think* I have thought of a way of increasing the resolution and the light through put, allowing for physical interaction and a more stable platform needing less stabilising mass. It does, however, introduce another limiting factor to the resolution. It's taken from another youtuber, I will link below. He is taking an LED panel, bouncing the light off a half mirror, onto a retroreflective panel, back through the half mirror and onto a "plane of cohesion" floating mid air. The quality of the retroreflector is the limiting factor. Imagine a glass table, the underside half-mirrored. At the foot, pointing upwards, an array of white LEDs. The light is columnated before passing through a platter rotating about an axis at the center of the glass table. Embedded, flat into holes cutinto the platter, are two equally spaced LCDs. Above each is a mirror, angled closer to 90 degrees than 45, such that the reflection of the LCD hits the half mirrored underside of the glass table, before hitting a retroreflector - held normal to the apparent reflection of the image, forming a cohesive plane, in space, floating in air, above the table. You would have to match the reflected distance to the center of the table, and another slightly forward, using the first pair of mirrors, but their distance from the center of rotation would match, just the height and angle differing... More stable to rotate. It might be more practible to use multiple stationary retro reflectors mounted at the radius, rather than rotate 2 - but that just might not work - I dunno - you probably want 2 of higher quality, and they're inherently quite low mass. The middle of the table is unused, so you could put a phased array of high frequency sounders there, so not only could you touch it, it can touch you. Below that, may be enough room to mount the motor for the rotating platter - which might be useful. If the table top had a hole in the center, you could brace a central shaft against the exterior of the table at both ends, but that might not be necessary - it might be as easy as holding it stable at the right height, under a half mirror table to which it is not attached. opg.optica.org/oe/fulltext.cfm?uri=oe-22-22-26919&id=303352
Wicked good stuff. Imagine this with a 200ppi display density (I guess it would be 40,000 ppi^3, pixels per cubic inch) and its already mind blowing. Volumetric displays have the potential to revolutionize the visualization and interaction with a whole slew of things. Imagine integrating this with a Leap Motion style of finger tracking or an IR/depth sensing camera as your control input - then put a map on the display and interact with it. Something we see in video games and sci-fi movies all the time and you're RIGHT THERE making it happen. Gives me the goosebumps. Tell me when you are taking investments - i'll be your first investor, hands down. I've got a stack of cash waiting for you my man. Seriously, awesome stuff.
There are already a few companies specializing in this technique and selling these kind of displays. I'm looking forward to this becoming open source. I would like to build one myself.
thats actually really cool. part of me wants to see google earth on that thing. it would make it look like its a high tech world map almost like it belongs on a starship thats plotting a landing zone or something
I wonder if you could get Cocoon running on there visibly enough. That kind of top down game with woshy effects would look cool I bet, but it might be an issue if its isometric
Amazing stuff! I wonder what spinning a transparent OLED with high resolution and refresh rate would look like?? I wonder if that is a likely way to obtain more detail.
I'd like to try grabbing the depth buffer to turn the scene into a point cloud - that would probably make a bunch of games playable without hand crafting each of them individually. But really the game needs to target this type of display explicitly. A first person perspective view doesn't work well. Top down or isometric is good, or something with a really close far clip plane - it needs to fit in a little constrained bubble of space.
@@ancientjames you can move the camera/projection/etcif you're clever with it. There are multiple injector mods that modify the camera and projection (not to mention Nvidia's Ansel, though they have the advantage of driver and hardware access). It might not be worth your time, but it's likely possible to get good support across a wide range of games with a single solution. You could also implement a 3d buffer and get the results back directly. All of this might be too much effort, but if you plan on selling these it might be worth it.
@@ancientjamesI think you could look into VR injections for different games, there should be bunch of open-source ones. But I have no idea how hard it is to turn this stuff into voxel cloud renderer
The issue isn't so much the practicality of modifying the camera, as the fact that lots of games just aren't designed with all the action contained in a small volume. If you imagine something like Elite, you want your ship to be large enough to see but then anything you're interacting with (shooting at, landing on) is way outside the bubble. GTA works well because it can be played top down (and the first two installments were). Doom doesn't work very well because a fair bit of the action happens over large distances - you can shoot a rocket right across the map.
Do you have one of these running 24/7? I'd be really curious to see an estimate on the lifetime of one of these things. Also, and more importantly: I'd like to see the failure mode. I can imagine a few pretty dangerous one if it suddenly went cattywampus at full speed
Looks like a magician’s sphere. Have you considered connecting this device with VR? I would love to play a game similar to ‘Voidness’ using this device.
My guess it's a modified "re3", which is an open-source GTA3 engine. But maybe this change could be ported to other RW games, if they have an open-source engine implementation
Is this actually rendering it in real-life 3d space of volumetric display? Meaning if the character is facing me, can I walk around the back of the display and see the back of the character? Or is it the same 2d space rendered from all directions of the display?
@@ancientjames I would be interested in a video to see how the game was changed to render it this way. I would guess there are a lot of challenges doing this. I.e. if theres a building between the camera and player, games usually make the building transparent or zoom in the camera until building is not visible. How does the renderer behave if there's no camera in this case? Also - does the game run faster since you don't need to render a lot (just the immediate area around the player)? Is there still technically a player camera for stuff like screen shakes? How would this render the UI? Nice demo either way
GTA III doesn't have any occlusion culling - everything in the camera frustum gets sent to the renderer. That made it pretty simple to get going - by using the top down view, everything I want to render into the display is visible to the camera that the game thinks is being used. I centre my render on the location of the camera's target entity and scale it according to the camera's height, and it ends up matching pretty closely to what you'd see on a monitor. I haven't implemented the UI, and I'm not sure how far I'll go with it given that the maximum install base is currently 1. With Doom I intercepted all the 3D stuff to do my own render, but let the 2D elements draw as normal to a 2D buffer which I then wrapped around the the outside of the display. It's conceptually a neat solution, but it ignores a weakness of this display which is that the central outer region is sparser and dimmer, and hard to read. I'll probably at least add the map to GTA, but put it flat on the axis to one side.
How many years until we get 8K spinning in that little contraption?? Looks amazing.
Can't believe Jeff is following this dude. That's wild.
Hey hi... Big fan Jeff
Could potentially make it right now.... technically. Maybe not 8k, but certainly super high pixel density panels exist. I saw a tech demo with so many LEDs that you'd think the panel was an LCD, but it wasn't. That said, what would make a mess is the driver tech, it did look bulky, you'd need to somehow move all that down under, behind the rotating connection.
@@aserta I have no doubt there is higher resolution panels, but I am pretty sure the issue with bigger panels is the refresh rate.
@@TheRubiksrocks Things like the crazy demos on this channel are a reminder at just how beginner-level I am in all my understanding of electronics and tech lol
We got gta in a crystal ball before gta 6
Lmao
Sorry to break it to you, but that isn't lead glass.
If you look into it hard enough, you’ll see into the future and get some more leaks
@@MattIsTheCatif you want to be an annoying pedant, achktually lead glass is not a crystal!
Huh it kinda looks like one of those visualizations of a self-driving car’s LIDAR point cloud
I feel like we are back in the 70s, early 80s where 3D graphics was so limited it was only capable of rendering wireframe models, even without occlusion. There must be a viewer headtracking for the system to figure out what to render and display and what not. It's pretty dope!
I'm pretty sure this is truly volumetric and doesn't need a specific viewpoint
@@lievenvv It is truly volumetric, no doubt. Question is: do we want to render occluded objects? Look at the truck, do we want to render the left side if we see it from the right? It would enhance one-user experience sacrificing 360-viewing ability. Hard call.
We got Dreamcast and now volumetric GTA 3 at the same time holy cow
Impressive. The water effect actually looks pretty cool.
When there's nothing on the orb to ponder so you fire up the good ol' gta.
really amazing proof of concept. hope we get true 3d displays one day. dunno why humanity stopped innovating laterally
When I first saw this device I clicked 100% thinking it was a cool animation proof-of-concept I still cant believe its rea
id love to see it with a speaker, i know its not the point of the experiment but hearing the environment and things like driving into objects would be an easy addition to help so much in making out whats going on. like that illusion where balls look like they pass through when its silent and bounce when you hear the impact
cool and brought me lots of joy tbh, thanks for deciding to do this gamer
i wanna put this in my car and sync its rotation to a top-mounted LIDAR
Thats amazing, great job! The framerste and clarity is a huge advancement even from the last game displayed. This looks truly amazing. Great job!
You made a homebrew-Voxon display. Really amazing btw.
Reminds me of the holo-display in Matrix (1999) when the 'squidies' were chasing them
can you do silent hill?
i think snow would look pretty cool on that
My approach so far has been to first download a 3D model of a classic game map and see if it's even recognisable displayed at some sort of playable scale. Doom and GTA looked good, which is why I went ahead with them. Minecraft was surprisingly disappointing which is why I didn't. I'll see what Silent Hill looks like. (It also needs to have some sort of source available, or be simple enough to be a quick remake)
@@ancientjames why was Minecraft bad?
@@ancientjames Try Half Life (and the others, afaik, they are open). Heck, shoot a message to Valve and they'll probably give you the stuff. They like stuff like this.
@@aserta Core engine functionality is not open for GoldSrc or Source enigne. That being said, I believe this should techincally be feasible to adapt to it, still, but I couldn't say for certain, as I never had to touch base rendering functions.
@@ancientjames you can try implementing it in one of the emulators (like PS1), that would allow you to run many games with one change. Also might work great with 3DS emulation, since it had small screen and most games supported 3D effect and were tuned to have small scenes with mostly close objects. All objects are large, since screen was low-res. Something like Super Mario 3D Land might look great on volumetric display
Love this so much - I *think* I have thought of a way of increasing the resolution and the light through put, allowing for physical interaction and a more stable platform needing less stabilising mass. It does, however, introduce another limiting factor to the resolution. It's taken from another youtuber, I will link below. He is taking an LED panel, bouncing the light off a half mirror, onto a retroreflective panel, back through the half mirror and onto a "plane of cohesion" floating mid air. The quality of the retroreflector is the limiting factor.
Imagine a glass table, the underside half-mirrored. At the foot, pointing upwards, an array of white LEDs. The light is columnated before passing through a platter rotating about an axis at the center of the glass table. Embedded, flat into holes cutinto the platter, are two equally spaced LCDs.
Above each is a mirror, angled closer to 90 degrees than 45, such that the reflection of the LCD hits the half mirrored underside of the glass table, before hitting a retroreflector - held normal to the apparent reflection of the image, forming a cohesive plane, in space, floating in air, above the table.
You would have to match the reflected distance to the center of the table, and another slightly forward, using the first pair of mirrors, but their distance from the center of rotation would match, just the height and angle differing... More stable to rotate. It might be more practible to use multiple stationary retro reflectors mounted at the radius, rather than rotate 2 - but that just might not work - I dunno - you probably want 2 of higher quality, and they're inherently quite low mass.
The middle of the table is unused, so you could put a phased array of high frequency sounders there, so not only could you touch it, it can touch you. Below that, may be enough room to mount the motor for the rotating platter - which might be useful. If the table top had a hole in the center, you could brace a central shaft against the exterior of the table at both ends, but that might not be necessary - it might be as easy as holding it stable at the right height, under a half mirror table to which it is not attached.
opg.optica.org/oe/fulltext.cfm?uri=oe-22-22-26919&id=303352
ruclips.net/video/EctawhlzZEw/видео.htmlsi=YyoqH7xFE0V1BHBW
every one of these is just cooler and cooler
That skin of Claude looks like the alpha colors. Awesome!
Would you be able to add a vacuum inside the sphere and reduce the sound emitted from the display?
Wonder if you could put a screen behind it to show the background and sky
Any backlighting will make the display harder to see
@ yes that’s true but at least you would be able to see the background and not just the stuff very near you
I'd love to see GTA2 running on this thing man. But also, think of all the other RenderWare possibilities!
Wicked good stuff. Imagine this with a 200ppi display density (I guess it would be 40,000 ppi^3, pixels per cubic inch) and its already mind blowing. Volumetric displays have the potential to revolutionize the visualization and interaction with a whole slew of things. Imagine integrating this with a Leap Motion style of finger tracking or an IR/depth sensing camera as your control input - then put a map on the display and interact with it. Something we see in video games and sci-fi movies all the time and you're RIGHT THERE making it happen. Gives me the goosebumps. Tell me when you are taking investments - i'll be your first investor, hands down. I've got a stack of cash waiting for you my man. Seriously, awesome stuff.
There are already a few companies specializing in this technique and selling these kind of displays.
I'm looking forward to this becoming open source. I would like to build one myself.
it's patented technology, patent law is so broken something such simple and essential shouldn't be able to patented but here we are.
jesus. this is amazing
thats actually really cool. part of me wants to see google earth on that thing. it would make it look like its a high tech world map almost like it belongs on a starship thats plotting a landing zone or something
I wonder if you could get Cocoon running on there visibly enough. That kind of top down game with woshy effects would look cool I bet, but it might be an issue if its isometric
briliant, keep makin
So this is what they did with Palantiri instead of listening to orders from Sauron.
This is what i dreaming for years
Amazing stuff! I wonder what spinning a transparent OLED with high resolution and refresh rate would look like?? I wonder if that is a likely way to obtain more detail.
This is exactly why I don’t trust self-driving cars.
This looks so old school, and still ... Requires bleeding edge graphic card for the pre reqs computations ...
It's running on one core of a Raspberry Pi 4!
Glorp watches as homeless junkie he probed yesterday drowns himself in a truck
Sam, you will need to connect this area to the Chiral network before you can print the *_Volumetric Oracle_*
Have you considered modifying DX/OpenGL/Vulkan so that most games are just directly compatible?
I'd like to try grabbing the depth buffer to turn the scene into a point cloud - that would probably make a bunch of games playable without hand crafting each of them individually. But really the game needs to target this type of display explicitly. A first person perspective view doesn't work well. Top down or isometric is good, or something with a really close far clip plane - it needs to fit in a little constrained bubble of space.
You can do a 3rd person patch + depth buffer to make it work maybe.
@@ancientjames you can move the camera/projection/etcif you're clever with it. There are multiple injector mods that modify the camera and projection (not to mention Nvidia's Ansel, though they have the advantage of driver and hardware access). It might not be worth your time, but it's likely possible to get good support across a wide range of games with a single solution.
You could also implement a 3d buffer and get the results back directly. All of this might be too much effort, but if you plan on selling these it might be worth it.
@@ancientjamesI think you could look into VR injections for different games, there should be bunch of open-source ones. But I have no idea how hard it is to turn this stuff into voxel cloud renderer
The issue isn't so much the practicality of modifying the camera, as the fact that lots of games just aren't designed with all the action contained in a small volume. If you imagine something like Elite, you want your ship to be large enough to see but then anything you're interacting with (shooting at, landing on) is way outside the bubble. GTA works well because it can be played top down (and the first two installments were). Doom doesn't work very well because a fair bit of the action happens over large distances - you can shoot a rocket right across the map.
I'm surprised you haven't made a volumetric snow globe yet
pls god give us instructions or a kit or something
You're making vaakum? I realize it can be tricky, but it can improve the rotation
more pixels please
Do you have one of these running 24/7?
I'd be really curious to see an estimate on the lifetime of one of these things.
Also, and more importantly: I'd like to see the failure mode.
I can imagine a few pretty dangerous one if it suddenly went cattywampus at full speed
Could you do... Silent Hill on PS1?
How do you do these ports? Would it be possible to do something like Luanti (formerly Minetest engine)?
Now you should attempt a 100% run on it x'D
Looks like a magician’s sphere. Have you considered connecting this device with VR?
I would love to play a game similar to ‘Voidness’ using this device.
How were you able to get the engine to display 3d from all sides? Could you share info or make a video?
So cool!
Is this GTA III
Software video coming?
I think we’re gonna need more pixels, boss.
Looks cool
Wow, that render distance is ridiculously small! 😯 How the heck can you even navigate? 🤔
It's the same view distance you get with the top down camera. ruclips.net/video/CwbKXtWpzv8/видео.html
Are you selling kits?
Does this mean all renderware games can be ported to this???
My guess it's a modified "re3", which is an open-source GTA3 engine. But maybe this change could be ported to other RW games, if they have an open-source engine implementation
Original Silent Hill should be accurate enough.
May I know what is the hz of the screen?
I trade off colour depth against refresh rate. With 1 bit colour it can manage 6000 Hz, but here I'm running in 2 bit colour - 3000 Hz.
Is this actually rendering it in real-life 3d space of volumetric display? Meaning if the character is facing me, can I walk around the back of the display and see the back of the character? Or is it the same 2d space rendered from all directions of the display?
It’s fully 3D. There are other videos where I move the camera around to show that off ruclips.net/video/pcAEqbYwixU/видео.html (skip to about 16;55)
@@ancientjames I would be interested in a video to see how the game was changed to render it this way. I would guess there are a lot of challenges doing this. I.e. if theres a building between the camera and player, games usually make the building transparent or zoom in the camera until building is not visible. How does the renderer behave if there's no camera in this case? Also - does the game run faster since you don't need to render a lot (just the immediate area around the player)? Is there still technically a player camera for stuff like screen shakes? How would this render the UI?
Nice demo either way
GTA III doesn't have any occlusion culling - everything in the camera frustum gets sent to the renderer. That made it pretty simple to get going - by using the top down view, everything I want to render into the display is visible to the camera that the game thinks is being used. I centre my render on the location of the camera's target entity and scale it according to the camera's height, and it ends up matching pretty closely to what you'd see on a monitor.
I haven't implemented the UI, and I'm not sure how far I'll go with it given that the maximum install base is currently 1.
With Doom I intercepted all the 3D stuff to do my own render, but let the 2D elements draw as normal to a 2D buffer which I then wrapped around the the outside of the display. It's conceptually a neat solution, but it ignores a weakness of this display which is that the central outer region is sparser and dimmer, and hard to read.
I'll probably at least add the map to GTA, but put it flat on the axis to one side.
My monitor is 4200RPM! (but 8 fps)
Impressive!
Awww, should've played DOOM! :))
he already did!
amazing
Can it run doom?
Meh. All this for 8k views.
Hope you have a real job.
Awesome 💪🔥-🔥