NERDGASM ALERT: Detailed Rendering of CG just got infinately better. The polygon is dead

More from John Carmack, who is probably the #1 expert in the world regarding real-time rendering for games:

nwn.blogs.com...

"I don't disagree with anything Notch said.

"I do not think non-polygonal renderers will be the winning choice in the next five years, but someone will probably have a go at it and learn something.

"I write a voxel tracer every five years or so..."

"You can real time ray trace a static world on high end hardware today as a demo, but there is a long path between a demo and something that is competitive with rasterization in a real product. My plan for such technologies has always been to emit a depth buffer as well as color from the voxel/point cloud renderer and continue to use existing technologies for characters/particles/etc.

"It took us five years to go from a megatexture demo to a (almost) shipping game."

Originally posted by john_bmth
reply to post by BIGPoJo

 

You still do not understand the criticisms that have been levelled at this technology. No one is saying that the vids are a "hoax", rather the technology has severe limitations that are not mentioned by the inventors. Their technology is old and certainty not "the end of polygons". Seriously, stop grasping at straws.

When people like John Carmack level criticisms at technology, you'd be very wise to listen.

edit on 4-8-2011 by john_bmth because: (no reason given)

It’s a scam! Perhaps you’ve seen the videos about some groundbreaking “unlimited detail” rendering technology? If not, check it out here, then get back to this post: www.youtube.com... Well, it is a scam.

Well, he said scam instead of hoax so I was off on that statement.

They’re hyping this as something new and revolutionary because they want funding.It’s a scam. Don’t get excited. Or, more correctly, get excited about voxels, but not about the snake oil salesmen.

And here he is vilifying Mr. Dell instead of digging deeper.

SOURCE

reply to post by BIGPoJo

A scam to get money is not the same as a hoax video. People are saying it's a scam because the promo vids and "information" detailed in said vids was clearly meant for laymen who are ignorant of computer graphics, probably to get more investment from said people ignorant people. People like you who are championing a cause you do not understand in a field you are completely ignorant of. Clearly they've attained their goals. You really are clasping at straws and I'm still baffled how you can fight something to the hilt when you are completely ignorant of the underlying concepts (and clearly not interested in learning).

Regardless, people calling it a scam or a hoax does not in any way change the rather serious limitations of the technology so arguing semantics is a lost cause.

Originally posted by john_bmth
reply to post by BIGPoJo

 

A scam to get money is not the same as a hoax video. People are saying it's a scam because the promo vids and "information" detailed in said vids was clearly meant for laymen who are ignorant of computer graphics, probably to get more investment from said people ignorant people. People like you who are championing a cause you do not understand in a field you are completely ignorant of. Clearly they've attained their goals. You really are clasping at straws and I'm still baffled how you can fight something to the hilt when you are completely ignorant of the underlying concepts (and clearly not interested in learning).

Regardless, people calling it a scam or a hoax does not in any way change the rather serious limitations of the technology so arguing semantics is a lost cause.

edit on 4-8-2011 by john_bmth because: (no reason given)

Calling me an ignoramus is not a valid argument. Instead of making arguments why the tech is a scam you are attacking my intelligence. Good day sir.

I sent this video to a friend who works at NAMCO Bandai Games as an artist. He said it was already creating a storm at his office. They feel its a hoax due to no motion of water and plants. It can easily be CGI.

Just my 2 cents.

reply to post by BIGPoJo


It is a valid argument:

ig·no·rant/ˈignərənt/Adjective 1. Lacking knowledge or awareness in general; uneducated or unsophisticated. 2. Lacking knowledge, information, or awareness about something in particular: "ignorant of astronomy".

Emphasis mine. You are ignorant of the field of computer graphics. To argue otherwise is nonsense. I am not using that word as an insult, I am using that word because it accurately describes your grasp of computer graphics.

Originally posted by john_bmth
reply to post by BIGPoJo

 

It is a valid argument:

ig·no·rant/ˈignərənt/Adjective 1. Lacking knowledge or awareness in general; uneducated or unsophisticated. 2. Lacking knowledge, information, or awareness about something in particular: "ignorant of astronomy".

Emphasis mine. You are ignorant of the field of computer graphics. To argue otherwise is nonsense. I am not using that word as an insult, I am using that word because it accurately describes your grasp of computer graphics.

edit on 4-8-2011 by john_bmth because: (no reason given)

I am not ignorant in that field, just not an expert like you claim to be. I am a developer that started as a network engineer. I am a computer scientist. I know what I am talking about. To argue otherwise is nonsense.

Originally posted by BIGPoJo

And here he is vilifying Mr. Dell instead of digging deeper.

SOURCE

Er, that links to an article where he IS digging deeper, he explains

notch.tumblr.com...

And this isn't vilifying, it's an honest assessment:

If these guys were being honest with the drawbacks and weaknesses of their system, I’d be their biggest fan. As it is now, it’s almost like they’re trying NOT to be trustworthy.

All this said, voxels are amazing. So is raytracing and raycasting. As computers get more powerful, and storage gets faster and cheaper, we will see amazing things happen.

And a final word to the engineers who worked on this: Great job, I am impressed! But please tell your marketing department to stop lying.

reply to post by BIGPoJo


The field of networking is completely different to the field of computer graphics. I'm completely ignorant about the field of networking, the difference is I don't fight the corner of people who's fishy "new" networking protocol claims to be the be-all and end-all when their claims fly in the face of logic and industry opinion. Imagine if the tables were reversed and I was championing a dubious cause in your field. You'd be scratching your head and thinking "WTF is he doing this for?". You would be completely justified in holding my view to task and calling me ignorant, because my views would be ignorant.

Like if someone came in as said "Ha, we can network with LIGHTBULBS now, ethernet is dead!", you might have some reasonable objections.

it just doesn't add up to me

I would like to see a patent of the technology.
If there is no patent and this is so groundbreaking/world changing would they be stupid not to patent it ?

If it is existing tech surely they can't patent it ?
That rings bells to me as a scam or at least something is being hidden/fishy or it is using existing tech (SVO)?.
Without a patent we can't really see how it works and would solve a lot if not all of the issues.

There's claims about it being a voxel engine. I am beginning to wonder how?
He claims to render point data sets on screen and if you look at his demo you see there is little visible
aliasing. At some point he is going to have to super sample many points together to lower aliasing problems.
An octree would address this but then we come back to the claims it is a sparse octree algorithm (SVO)
Does he construct some kind of volume (to reconstruct the original mesh) from the point data sets
so he can sample the correct pixel to minimize aliasing but then again aliasing would happen with this.?

He claims he picks only one point per pixel which ideally suits the SVO algorithm and lays weight to the
thing being voxel based but he claims it isn't a voxel engine so what does he really mean.?.

We've seen no figures about the data size of non instanced data so we can't really gauge how
much of a memory footprint his tech has (This nullifies some of Notch's argument).
There's no skinned/animated entities and there's evidence it was dropped originally because of the memory footprint of animations.

EDIT : MOOT POINT His search loop contains no multiplies or divides which contradicts what I know about the SVO algorithm.

Blah what a mess!.
Limbo

(Some tech on mega meshes to keep you drooling + their paper.)
www.youtube.com...

Part of Dennis Bautembach's excellent thesis on „Animated Sparse Voxel Octrees

bautembach.de...
bautembach.de...
(I really recommend reading his paper )

Originally posted by john_bmth
reply to post by BIGPoJo

 

The field of networking is completely different to the field of computer graphics. I'm completely ignorant about the field of networking, the difference is I don't fight the corner of people who's fishy "new" networking protocol claims to be the be-all and end-all when their claims fly in the face of logic and industry opinion. Imagine if the tables were reversed and I was championing a dubious cause in your field. You'd be scratching your head and thinking "WTF is he doing this for?". You would be completely justified in holding my view to task and calling me ignorant, because my views would be ignorant.

Would you completely dismiss a new idea that was too unbelievable to be true or would find out for yourself? You have chosen the path of dismissing the unbelievable instead of imagining how it could be achieved.

If a new piece of hardware came out that you could plug inline into your router which would cause it to transmit 4 times at much data most networking guys would call it a scam too. Well this actually happened back in 2000 and I was lucky enough to be able to test this magic box. It was indeed real. Since then I have decided to never dismiss the impossible because the next innovation is being manufactured everyday.

I am making the argument that the video presented by the OP is indeed being rendered on a laptop in software mode. I am making the argument that it is indeed possible and is not a scam.

Originally posted by Uncinus
Like if someone came in as said "Ha, we can network with LIGHTBULBS now, ethernet is dead!", you might have some reasonable objections.

You CAN network with light bulbs!

network with leds

I have also heard of people using florescent light bulbs for a similar effect like wifi.

Originally posted by Limbo
His search loop contains no multiplies or divides which contradicts what I know about the SVO algorithm.

SVO (Sparse Voxel Octree) searching is finding the first solid voxel at the appropriate resolution along a line. One you know what the line is you can use a line rendering algorithm to perform the stepping though the octree at the various levels. Bresenhams's algorithm can do this without multiplies or divides:

en.wikipedia.org...

By stepping though the tree at voxel levels from the root node (the entire space), then progressively though root nodes, you are essentially doing a binary or dichotomic search to find the right voxel.

en.wikipedia.org...

Originally posted by BIGPoJo

Originally posted by Uncinus
Like if someone came in as said "Ha, we can network with LIGHTBULBS now, ethernet is dead!", you might have some reasonable objections.

You CAN network with light bulbs!

network with leds

I have also heard of people using florescent light bulbs for a similar effect like wifi.

edit on 4-8-2011 by BIGPoJo because: edit for proof

Exactly, and you can render with voxels. The problem is an over-enthusiasm for the technology. The problem is not saying "we can network with lightbulbs" it's saying "ethernet is dead".

Originally posted by BIGPoJo
I am making the argument that the video presented by the OP is indeed being rendered on a laptop in software mode. I am making the argument that it is indeed possible and is not a scam.

Of course it's possibly. I'm sure the videos are real. That does not stop it being a scam. The scam is saying this is actually a practical and useful technology that will revolutionize game engines. The scam is saying it has "unlimited" detail, when all it can do is render unlimited copies of a very small number of objects.

Originally posted by Uncinus

Originally posted by BIGPoJo
I am making the argument that the video presented by the OP is indeed being rendered on a laptop in software mode. I am making the argument that it is indeed possible and is not a scam.

Of course it's possibly. I'm sure the videos are real. That does not stop it being a scam. The scam is saying this is actually a practical and useful technology that will revolutionize game engines. The scam is saying it has "unlimited" detail, when all it can do is render unlimited copies of a very small number of objects.

Yeah but when you add some overlapping patterns things start to look more random. You could use this tech to create huge terrains with realistic features. For NPCs you could still use polygons to do the hard work. To say that it will not be used is almost laughable.

Originally posted by Uncinus

Originally posted by Limbo
His search loop contains no multiplies or divides which contradicts what I know about the SVO algorithm.

SVO (Sparse Voxel Octree) searching is finding the first solid voxel at the appropriate resolution along a line. One you know what the line is you can use a line rendering algorithm to perform the stepping though the octree at the various levels. Bresenhams's algorithm can do this without multiplies or divides:

en.wikipedia.org...

By stepping though the tree at voxel levels from the root node (the entire space), then progressively though root nodes, you are essentially doing a binary or dichotomic search to find the right voxel.

en.wikipedia.org...

Yep but I realise is isnt really an issue now since on the vast majority of modern CPUs mult/divs are not so expensive. I don't know why he made this point - seems to add another argument that his tech or him being in the past. (It might have been when he first started working on the tech.)

Originally posted by BIGPoJo

Originally posted by Uncinus

Originally posted by BIGPoJo
I am making the argument that the video presented by the OP is indeed being rendered on a laptop in software mode. I am making the argument that it is indeed possible and is not a scam.

Of course it's possibly. I'm sure the videos are real. That does not stop it being a scam. The scam is saying this is actually a practical and useful technology that will revolutionize game engines. The scam is saying it has "unlimited" detail, when all it can do is render unlimited copies of a very small number of objects.

Yeah but when you add some overlapping patterns things start to look more random. You could use this tech to create huge terrains with realistic features. For NPCs you could still use polygons to do the hard work. To say that it will not be used is almost laughable.

Then he would have to use a zbuffer and we are back to polygons for the animated/skinned models!

Originally posted by Limbo
Yep but I realise is isnt really an issue now since on the vast majority of modern CPUs mult/divs are not so expensive. I don't know why he made this point - seems to add another argument that his tech or him being in the past. (It might have been when he first started working on the tech.)

I think he was actually talking about ray tracing against an arbitrary polygonal soup, vs ray casting into a octree. Ray tracing uses more math