Cool new rendering technology... maybe?

[Satirik]

I think video hardware is not made to display this kind of voxel space in high resolutions.
So this would be software rendering if there is no special hardware for it, right?

There should be a way to use the rendering power of the gfx card, it would be useless if all that would be rendered by your cpu only.

Anyways I would like a demo to run on my pc to see if it really works that well.

http://video.google.com/videosearch?q=gigavoxels

[PicassoCT]

As everybody is allowed to fanboi his favoured method in this thread - i think splines are the future. They allow to scale up and down (in theory) pretty seemless in max from 32 polys to milliontrillionkazillions

http://www.youtube.com/watch?v=5pp64FgMMyU
http://www.youtube.com/watch?v=EBYQUR_vMZc

One off the big downfalls of every eeevvvaaamoooooaaaaar detailed technology is that it is eating up even moar worktime- shrinking the companys with abilitys list.

A fukken GranTourismo Car takes a modeller 2 months today, fukken two months, with the whole interior inside.

So i think the most intersting thing would be the mentioned before autodownscaler, which still leaves the dilema of how far up you should be able to scale.

[SwiftSpear]

Their aim is probably to generate a bit of hype.. thats why the retard tone. It seems to work well.

The fact they are not searchable means nothing.. you can find your university or company by googling your name? I cannot and i'm happy for that :)

Don't get me wrong, I'd like to believe it's true too, but the whole thing just doesn't sit right with me.

[aegis]

There should be a way to use the rendering power of the gfx card, it would be useless if all that would be rendered by your cpu only.

you can raytrace in CUDA/OpenCL

[PicassoCT]

i think there hook is the fixed estatica 2 look- they limit the number of dots, be sizing them up, so the dots can hide moar dots behind them- at a distance one big dot might hide a whole statue, but you see that the whole thing looks mighty artifishyal- and the view flirted everytime the cam moves. Also, i dont think you can import polygonmodels into that thing- if it works, the should reproduce a comon polygame with there engine, they have to proof they are workin not the poly faction.

In Additon: I m not against Voxels, i loved Outcast, so inb4 gamsav

http://www.youtube.com/watch?v=QlMCToxlt1c

The Narrator of those vids is pissing me off to no ends, even if he trys to sound less douchebaggypanty in the second vid.

[Soul]

I would probably most likely to fall under the category of amatur when it comes alot of these things, but i'm still curious in how they made that water using "point cloud data system".

As i understand the strength with this system is that it is alot of "atom-like" points, this makes geometry of ex. a wall very "easy" to make realistic.

But wouldn't this "strength" become a weakness when it comes to the water surface?
Wouldn't you need endlessly amounts of these round atoms to make the water surface look any good?

And wouldn't the reflection in the water become distorted if the water surface is made out of millions of round attoms? wouldn't a large amount of the reflection be reflected in other directions?

If anyone has more insight in this, please try to explain it in layman's terms for me.

Picture of water

I'm just guessing, but i don't think that the water is made in the point cloud data system, i think that the water is just in a 2d screen with shaders.

Could i be right?


Btw they do show some animation of some scorpionlike animal at 1:43->

[knorke]

But wouldn't this "strength" become a weakness when it comes to the water surface?
Wouldn't you need endlessly amounts of these round atoms to make the water surface look any good?

I dont have much clue about this but why would water need more atoms than normal ground? You just have to figure out what points get reflected. And if some blur/waterwave effect is applied it might actually be possible to use less points because they get blurred anyway.

One off the big downfalls of every eeevvvaaamoooooaaaaar detailed technology is that it is eating up even moar worktime- shrinking the companys with abilitys list.

There will probally be new modelling technics and stuff, like 3D scanning objects.
Like the laser scanning here: http://www.lfs.net/?page=rockingham but more detailled and without the step of converting to polygons.

A fukken GranTourismo Car takes a modeller 2 months today, fukken two months, with the whole interior inside.

Yes but they have to do things like optimize polygon counts. Now if they would not have to worry about that?

they should reproduce a comon polygame with there engine

No, what a waste! They should do something that really can use the advantages of the engine. Don't know, but one thing i always liked about voxel techdemos was how you could destroy everything and put holes in walls. Not sure if possible with their presorted points thing though.

[PicassoCT]

handson or hyperhoax- that is the question!


There engine looks a little like LandsofLore2, they should redo that game with highrespointcloud..

Im also willing to admit, that i have some fear with my modelling knowledge getting outdated, my equipment insufficient, all my little tricks and turns - worthless.

[Satirik]

handson or hyperhoax- that is the question!


There engine looks a little like LandsofLore2, they should redo that game with highrespointcloud..

not an hoax but a lot of marketing crap (unlimited = lies ofc, etc)

[Soul]

But wouldn't this "strength" become a weakness when it comes to the water surface?
Wouldn't you need endlessly amounts of these round atoms to make the water surface look any good?

I dont have much clue about this but why would water need more atoms than normal ground? You just have to figure out what points get reflected. And if some blur/waterwave effect is applied it might actually be possible to use less points because they get blurred anyway.

No the ground needs alot less atoms then the water.
Since the ground made in these video is just lots of round rocks, making a surface that is very rough.
The excact oposite of the water surface.

I'm not sure what the blur/waterwave effects does to something like a vast ocean of round atoms, but i'm quite surten that you will need alot of these atoms to be able to make any useful reflection at all.

Ill make a example and edit post it soon.

Edit:example

The circles are the "water", the gray line is self reflection, the blue is sky reflection, the green is the reflection you want to see and the red lines are the the "extreme" reflections.

Edit: damn, one of the blue reflectionlines should ofc be green, its not hard to figure out wich (the one to the furthest left), ill edit change that when i get the time to do so ¬_¬

Edit: fixed

So watching reflections on water made out of alot of round atoms would give some quite nasty side effects.

Atleast that is what my simple mind tells me, perhaps you can make a workaround this, but i still think that the water in the screens are faked, its probable just a screen made with polygons, or just shaders on a 2d screen.

But maybe they have programmed som epic filter to get rid of the unwanted reflections...

I can however see the benefits of using point cloud in making water, it could make swimming or drinking water/liquids very realistic in future games.

[Teutooni]

The water looks like it's just one point and normal (plane).

[Soul]

I just though of a way to sidestep the problem i just presented, if you only have one gigantic cloud point, as big as a planet in comparision to the enviorment, then the whole reflection problem would be solved, since this would give a smooth surface.

But this seams like overkill, and dynamic water effects might look weird.

The water looks like it's just one point and normal (plane).

Yeah, but a normal plane is classified as polgon related right?
Or am i just confused and am mixing up plane with a plane polygon?

Point cloud doesn't use polygons, atleast not if we are to belive that page/video.
So its safe to say they are not using a normal plane right?


But besides that we are quite of one mind i think. ^^

[Hobo Joe]

I think you're a little hung up on thinking that these points are little balls, but you do make a point, I wonder how well something like this would animate. I can't imagine it would be memory efficient. But as for the 'water' he has in his videos, looks like it's completely flat, which means he either has a method for interpolating points, or is using a completely unrelated method, possibly even polygons.

[AF]

I dont really see much evidence outside of the giant pyramids that there is really unlimited detail, and that they havent just rendered it on a renderfarm. Like the bit when he says theres no model transitions despite visible but subtle jumps ( could be video codec artefacts mind )

The pebble ground thing could be adv lighting effects too, if he'd physically shifted them around so we could see they were all real geometry maybe..

[Google_Frog]

Judging from his tone and failure to cover issues such as animation I would be surprised if this wasn't a hoax.

[SwiftSpear]

I just though of a way to sidestep the problem i just presented, if you only have one gigantic cloud point, as big as a planet in comparision to the enviorment, then the whole reflection problem would be solved, since this would give a smooth surface.

But this seams like overkill, and dynamic water effects might look weird.

The water looks like it's just one point and normal (plane).

Yeah, but a normal plane is classified as polgon related right?
Or am i just confused and am mixing up plane with a plane polygon?

Point cloud doesn't use polygons, atleast not if we are to belive that page/video.
So its safe to say they are not using a normal plane right?


But besides that we are quite of one mind i think. ^^

Does the use of point clouds necessitate the total disuse of polygons? It seems to me fairly trivial to just throw one polygon in for a surface like that. Also, I'm not sure point cloud technology significantly alters the capability of lighting renderers. You can make a surface out of surface points and tell the engine to treat it like a mirror surface based on the points the material is composed of.

There is animation shown in the videos, but only animation so far as to say the camera moves around the scene (which does force the scene to render differently than before, and thus, to some degree, animate).

Who knows how a system like point clouds would work for something like skeletal animation in regard to how the polygonal system allows stretching and morphing and other techniques which can do things like make a small surface grow into a large surface fairly seamlessly.

Most likely they just haven't addressed those kind of issues yet, I see no technical reason they should be impossible, just not well realized at the moment. It's more convincing from a technical standpoint if they can show off a scene that should contain several trillion polies rendered in real time, which is what they have done.

Still, this being said, my problem is the problem that AF is pointing out. There's nothing they've shown that proves they haven't just made a few complicated scenes render through a render farm and said "oh hey, look at this awesome real time video!"

[Teutooni]

The water looks like it's just one point and normal (plane).

Yeah, but a normal plane is classified as polgon related right?
Or am i just confused and am mixing up plane with a plane polygon?

Point cloud doesn't use polygons, atleast not if we are to belive that page/video.
So its safe to say they are not using a normal plane right?


But besides that we are quite of one mind i think. ^^

I'm obviously only guessing, but I know 3D rendering relies on hacks to boost performance, and plane equations are fairly simple. To me it makes no sense to use gazillion points (or even a polygon) to define a reflective plane when all you need to do is define one point and a normal vector and make the point-search/raytrace/whatever algorithm they have to take that into account. Dynamic water would be another story...

[==Troy==]

Reflection is easy to do, simple physics, just reflect the WHOLE set of points in the mirror (i.e. creating another searchable set of points), and do the search only for that window. That is why the water in his image was plain, you would run into trouble if you had more than 1 plane to reflect your whole scene in.


The image hick-ups is more laptop slow-down Id say, if this is all true anyway, rather than anything else. The whole presentation is very poor and very basic, so IMO, something to keep an eye on, and wait until more serious stuff will happen.


As of the main idea : draw the only thing that is shown on the screen, is a very good approach. When I read how the current graphics does that (draw the whole scene, beginning from the back of it to front), a LOT (huge amount) of drawn "pixels" is "wasted" and never actually seen. Tesselation technology and similar try to approach that by not drawing polies which are smaller than 1 pixel, but still there is a lot of waste and a lot of place to improve.

[Soul]

Now, i may just be even more confused then before, but if you aren't rendering things you can't see what would hapend if a creature was going to attack you from behind?

As i see it, these points that don't render when not on screen is all good when it comes to static unchanging things, but things that move would still need to be processed right?

It might not need to be rendered, but all the "search points" of the unit and then the enviorment the unit is moving on would have to be processed i think.

This might imply shorter time in the loadingscreen, but it feels like the ingame fps would be drasticly lowered, since you would be rendering everything infront of your in each frame.
Instead of prerendering everything before gamestart.

And now for a newbie question, doesn't the game physics screw up if stuff collide without being rendered?
I'm guessing no, cause the rendered models aren't realy effected by game physics right?

I guess it's not the model in it self that are effected, i assume that it is all the prerendered points that are effected by the physics?

But still you would require a monstereous amout of RAM wouldn't you?
So okay, you don't need to buy the best grafic cards for the next 20 year as he said in the video, but you would instead have to buy four of the latest RAM-sticks and continuously upgrade them and your motherboard wouldn't you?

I think its safe to assume that it will atleast be as costly as getting new grafics cards, if not more expensive even.

[==Troy==]

Now, i may just be even more confused then before, but if you aren't rendering things you can't see what would hapend if a creature was going to attack you from behind?

As i see it, these points that don't render when not on screen is all good when it comes to static unchanging things, but things that move would still need to be processed right?

It might not need to be rendered, but all the "search points" of the unit and then the enviorment the unit is moving on would have to be processed i think.

Processed, yes, rendered, no way. I.e. you do not care what state your model is, and you do not need to animate it, as long as player does not see it, all you care about is "physics" and "sound", NOT under any circumstances "vision"


Why in Q3 if you stare at wall your FPS can just as high as triple the average?