I'm not sure if anyone's posted this before (a search for "Euclideon" came up with no results), but it's pretty impressive to me. Unlimited 3D detail and still only running on the CPU. Makes me wonder again what BYOND is doing wrong... | |
Listening to it as well as this interview clears up a couple things that I can repeat. At least, this is my understanding of it.
Forum_account wrote: I didn't see much shading (it looked like they were showing ambient occlusion at the end but I didn't see any proper shadows)It's far from complete, and what they were showing with ambient occlusion is that they already have better shading that they're not going to show until the final product is released. and I don't remember seeing much animation.The interview shows ugly animations from 7-8 years ago that have been made much better, but like with shading they're refusing to show anything until they're finished. The fact that it's running on a CPU could be a statement of speed (that they don't need to use a GPU) or it could mean that the drawing isn't parallelizable and wouldn't benefit from being run on a GPU.It actually is a statement of speed, and using the GPU would only make it faster. Polygonal rendering doesn't have to be as limited as the video shows. You can use techniques to increase the polygon count of a model - that doesn't mean you're increasing the detail, but this video seemed to say that the problem with polygonal rendering is that there are cases where objects look obviously polygonal (they showed many examples of this) but this is something that can be avoided. You can create a lower polygon count model, animate it, then use subdivision surfaces (like this) to increase the quality of the model.I don't have much to say on polygons and their apparent impact on... rendering speed...? But it's irrelevant to what they're doing, since they're not actually using polygons at all. They use a "search algorithm" that picks one atom for every pixel displayed on the screen. I don't think anyone knows exactly what search algorithm they're using, but apparently it allows for unlimited amounts of detail in the world while only displaying the atoms that can be seen. This sounds a lot like raytracing to me, but he says in the interview that it doesn't use any kind of rays. | |
There is merit to this approach. Rendering polygons has a cost associated with each polygon and a cost associated with each pixel you draw. With other methods it's possible to just have a cost associated with each pixel you draw - letting you achieve realtime graphics even with complex geometry. What they're claiming isn't entirely outlandish, but it's not nearly as impressive as they want you to think it is.
Kaiochao wrote: But it's irrelevant to what they're doing, since they're not actually using polygons at all. The problem is that they're not showing what can be done with polygons. They're showing an example of a low polygon count palm tree and are ignoring existing techniques that can be used to address that problem. For example:  The right side is the result of applying Catmull-Clark subdivision to the left side (I'm guessing that's the method they used). Given the reduced geometry a game can compute the detailed geometry as needed. You could say that's infinite detail. Their video is essentially showing just the left side and saying "see, polygons look bad". | |
i can imagine a litte, cute, 3D graphic option in byond!
| |
I watched both videos, and it doesn't really explain what they're actually doing. If it is real, it certainly looks promising and would change the way 3D graphics are rendered everywhere.
It just doesn't seem possible. They seem to explain it as 'we simply grab one atom per one pixel on the screen' but this doesn't explain the process in detail at all. There must be something more complicated going on here, and they're either not very good at explaining and marketing, or it's fake. I understand why there are so many critics, because their claims are certainly very unbelievable. | |
In the interview he does his best to refute Notch's accusations, saying how wrong everyone is who disliked the first demo. It does sound pretty awesome to me, but people can only say so much about a "small, secretive bunch" who are keeping all the awesomeness a secret until a public demo.
The public demo is supposed to be released some time this year, so I'm hoping soon. I'm hoping a lot of people are proven wrong, because I like the idea of using a ton of tiny "atoms" to make a big thing, as it is in the real world, but it just doesn't seem like it could fit in one little laptop. | |
When a guy tells you that he's come up with a revolutionary idea/invention/whatever but then proceeds to act super secretive about it and refuses to offer any hard evidence to support his case, he's full of it. People who have come up with legitimate breakthroughs do not go public until they have something that can be tested or proven through hard data.
What you have here is someone running a hoax for attention. You really shouldn't get your hopes up or waste any of your time letting him string you along. | |
Silk Games wrote:
When a guy tells you that he's come up with a revolutionary idea/invention/whatever but then proceeds to act super secretive about it and refuses to offer any hard evidence to support his case, I assume he is full of it. Normally people who have come up with legitimate breakthrough/s do not go public until they have something that can be tested or proven through hard data. I fixed that post Silk. | |
A.T.H.K wrote:
Silk Games wrote: I fixed that fixing post A.T.H.K. | |
Lynch this heathen at once.
| |
EmpirezTeam wrote:
Lynch this heathen at once. EmpirezTeam why don't you just explain it instead of being rude? | |
 | |
A.T.H.K wrote:
I fixed that post Silk. Every statement that a person makes doesn't need to be qualified with "in my opinion" or "from my perspective". You'd have to be astonishingly insecure to read a forum post and become upset by the author expressing his point of view without reminding you every other sentence that what he is writing is just his opinion. Most importantly, what you're suggesting makes for incredibly poor writing: http://www.articlesbase.com/writing-articles/ in-my-opinion-i-think-that-i-believe-this-is-bad-writing-992 050.html | |
We learned that in 11th grade. My teacher said saying "I think" or "in my opinion" is redundant because you wouldn't be saying it if you didn't think it or if it wasn't your opinion, so there's no need to start the sentence off that way. You're supposed to just write your statement without things like "If you ask me," or "The way I see it,".
Although I do still do that from time to time out of habit because I used to write like that all the time before 11th grade. | |
Polygonal rendering doesn't have to be as limited as the video shows. You can use techniques to increase the polygon count of a model - that doesn't mean you're increasing the detail, but this video seemed to say that the problem with polygonal rendering is that there are cases where objects look obviously polygonal (they showed many examples of this) and this is something that can be avoided. You can create a lower polygon count model, animate it, then use subdivision surfaces (here's an example) to increase the quality of the model.
Kaiochao wrote:
Agreed. BYOND won't ever offer graphical capabilities on par with modern games and I get the impression that the staff uses that as a cop out. It'll always have limitations but BYOND could be so much better of a 2D engine and the staff just doesn't seem interested.