Advertisement
Guest User

on euclideon voxel raytracing + calcuable curves

a guest
Jan 20th, 2017
232
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 16.69 KB | None | 0 0
  1. 6:48 PM: THE COMPUTER ONLY HAS TO CALCULATE/DRAW THE OUTLINES BASED ON WHAT IT SEES AS AN EDGE IN 3D SPACE GIVEN THAT IT HAS BEEN BEEN PRE FED THE 3D MAP OF LASER LENGTHS FROM MULTIPLE ANGLES
  2.  
  3. 7:03 PM: COMBINE MOVEMENT ON RAYTRACED MAP TO BE LINEAR WITH TRADITIONAL POLYGONAL OBJECTS TO CREATE EVEN BETTER GRAPHICS
  4.  
  5.  
  6. Never tell your password to anyone.
  7. 4:24 PM - euclideon raytracing: do you teach math? I have a faint memory of you teaching math when we played highlander
  8. 4:24 PM - Fire Dog: yeah i taught math and physics stuff
  9. 4:25 PM - euclideon raytracing: what is a computational color buffer
  10. 4:26 PM - euclideon raytracing: ?
  11. 4:26 PM - euclideon raytracing: I read this phrase online and need to understand it
  12. 4:26 PM - Fire Dog: can you link me what you read?
  13. 4:28 PM - euclideon raytracing: http://www.codersnotes.com/notes/euclideon-explained/
  14. 4:29 PM - euclideon raytracing: does any of this make sense to you?
  15. 4:29 PM - Fire Dog: yeah i'm just reading it
  16. 4:30 PM - euclideon raytracing: when I take showers I think about a lot of questions that relate to this subject
  17. 4:30 PM - euclideon raytracing: and I've brought it up to my parents but I never have read it online until today
  18. 4:31 PM - euclideon raytracing: I didn't there were people thinking the way I do
  19. 4:31 PM - euclideon raytracing: didn't think*
  20. 4:31 PM - Fire Dog: I've never heard that exact term before but i think i know what it means
  21. 4:31 PM - Fire Dog: so it seems this is an attempt to do 3-d rendering but like the article says it can't actually do animation so it has to do stop mmotion using prebuilt trees
  22. 4:32 PM - Fire Dog: a buffer is a temporary storage space in memmory for things to just wait until they get moved to where they need to go
  23. 4:33 PM - euclideon raytracing: imagine if you replace an entire map's static polygons with raytracing and an extremely simple color system
  24. 4:33 PM - euclideon raytracing: is this possible
  25. 4:33 PM - Fire Dog: I think a computational color buffer is just a piece of memory where color information is stored until it's used to actually render the scene
  26. 4:34 PM - Fire Dog: I mean you could but raytracing is way more computationally expensive
  27. 4:34 PM - Fire Dog: you'd be playing at like less than an fps with current graphics cards I'm pretty sure
  28. 4:34 PM - Fire Dog: if you're talking about video games or something like that
  29. 4:35 PM - euclideon raytracing: couldn't a color buffer be as simple as overlaying an hd video from the pov of the raytracer camera
  30. 4:35 PM - euclideon raytracing: ?
  31. 4:35 PM - euclideon raytracing: maybe you can write something
  32. 4:36 PM - euclideon raytracing: that combines the raytracing and video
  33. 4:37 PM - Fire Dog: maybe but what would be the benefit
  34. 4:37 PM - Fire Dog: in order for it to be fast you would have to have minimal raytracing
  35. 4:38 PM - euclideon raytracing: easier special effects
  36. 4:37 PM - Fire Dog: so would it even have a worthwhile benefit
  37. 4:37 PM - Fire Dog: oh wait
  38. 4:38 PM - Fire Dog: you mean like render different things with different methods
  39. 4:38 PM - euclideon raytracing: I mean what I said yes
  40. 4:39 PM - euclideon raytracing: pieces of pencil led are useless unless you have a mechanical pencil
  41. 4:39 PM - euclideon raytracing: lead*
  42. 4:39 PM - Fire Dog: yeah it sounds plausible but I haven't really done any serious work with this stuff so I don't know the implementation
  43. 4:40 PM - euclideon raytracing: if you simply think about the logical implications it becomes clear that the uncanny valley is deeper than we might have once thought
  44. 4:41 PM - euclideon raytracing: the camera movement in these euclideon youtube videos unnerves me
  45. 4:41 PM - euclideon raytracing: on an evolutionary level
  46. 4:41 PM - Fire Dog: I think part of it is that they use this stop motion method
  47. 4:42 PM - Fire Dog: it does look very strange
  48. 4:42 PM - euclideon raytracing: explain stop motion method
  49. 4:43 PM - euclideon raytracing: also does a very large data base decrease the need for computational power in some applications?
  50. 4:43 PM - Fire Dog: stop motion is just linking still images together
  51. 4:43 PM - Fire Dog: so they render each image ahead of time and then chain them together
  52. 4:44 PM - Fire Dog: versus what people usually use which is creating an actual model and animating it
  53. 4:45 PM - Fire Dog: also alarge database would increase computational time depending on how it's accessed
  54. 4:45 PM - euclideon raytracing: if you take multiple pictures of an object from all angles with lasers can you extrapolate a computable 3d space
  55. 4:45 PM - euclideon raytracing: by simply analyzing the pictures
  56. 4:45 PM - Fire Dog: yeah you should be able to
  57. 4:46 PM - Fire Dog: that's basically how we perceive 3-d anyway
  58. 4:46 PM - Fire Dog: our eyes only capture a 2 dimensional image that we interpret as 3-d
  59. 4:49 PM - euclideon raytracing: if you can make a bunch of these real life equivalents to tf2 smooths could you add an extremely realistic bob and head turning then place the smooth in a 3d environment where the camera was above the ground at the average height of a human
  60. 4:50 PM - euclideon raytracing: we will be able to replace voice actors soon
  61. 4:51 PM - euclideon raytracing: that's not related to anything I just thought about it
  62. 4:50 PM - Fire Dog: lol
  63. 4:50 PM - Fire Dog: i mean
  64. 4:51 PM - Fire Dog: to really create a perfect 3-d space from 2d images
  65. 4:51 PM - Fire Dog: you would need an infinite amount to be sure
  66. 4:51 PM - Fire Dog: but you really only need to make sure you know the distances between things
  67. 4:51 PM - Fire Dog: to get a decent idea
  68. 4:53 PM - euclideon raytracing: does an artist need to know the superposition of every atom's particles to draw a picture of it?
  69. 4:53 PM - Fire Dog: nah which is whhat I'm saying
  70. 4:53 PM - Fire Dog: you don't need unlimited detail to get a good representation
  71. 4:54 PM - Fire Dog: so you would probably need like
  72. 4:54 PM - euclideon raytracing: you need very good detail
  73. 4:54 PM - Fire Dog: idk
  74. 4:54 PM - euclideon raytracing: infinite and very good are infinitely apart
  75. 4:55 PM - Fire Dog: you would have to quantify that somehow
  76. 4:55 PM - Fire Dog: like do you want to know sutff on the millimeter scale or what
  77. 4:55 PM - Fire Dog: that would determine how many 2-d images you would need to ensure that much detail
  78. 4:56 PM - euclideon raytracing: how good are laser camera
  79. 4:57 PM - euclideon raytracing: cameras?
  80. 4:56 PM - Fire Dog: well theoretically
  81. 4:57 PM - Fire Dog: they use photons
  82. 4:57 PM - Fire Dog: depending on the type of photon
  83. 4:57 PM - Fire Dog: the wavelength specifically
  84. 4:57 PM - Fire Dog: it would affect the resolution
  85. 4:57 PM - euclideon raytracing: I thought you couldn't send information at light speed
  86. 4:58 PM - euclideon raytracing: shining laser back at a singular point to be interpreted is not sending information
  87. 4:58 PM - euclideon raytracing: lasers*
  88. 4:59 PM - euclideon raytracing: isn't that like seeing with eyeballs
  89. 4:59 PM - euclideon raytracing: but I don't see and interpret laser light
  90. 4:59 PM - euclideon raytracing: and why can't you just take close up pictures
  91. 4:59 PM - euclideon raytracing: and store them
  92. 5:00 PM - euclideon raytracing: if memories are in the past could they not simply be prerendered?
  93. 5:01 PM - euclideon raytracing: can we make electricity move in a simulation exactly how it would move through a flesh brain?
  94. 5:02 PM - euclideon raytracing: sorry if you don't like my questions
  95. 5:02 PM - Fire Dog: simulations are not really exact at that level
  96. 5:02 PM - euclideon raytracing: not trying to be autistic
  97. 5:02 PM - Fire Dog: the processes are too complicated
  98. 5:03 PM - euclideon raytracing: what is so special about a live process that it needs to happen in real time
  99. 5:03 PM - euclideon raytracing: what if it took an extremely long time to compute something
  100. 5:04 PM - euclideon raytracing: but we used tricks
  101. 5:03 PM - Fire Dog: you could and it would be more accurrate but there's virtually always apprroximations
  102. 5:04 PM - Fire Dog: except in extremely simple cases
  103. 5:05 PM - euclideon raytracing: would caching calcuable curves benefit the process if you want it to run faster
  104. 5:05 PM - euclideon raytracing: then you assign a curve to a magnification level
  105. 5:05 PM - euclideon raytracing: and put them together
  106. 5:05 PM - Fire Dog: people already use a sort of similar idea
  107. 5:05 PM - Fire Dog: there are all sorts of problems
  108. 5:05 PM - Fire Dog: like
  109. 5:06 PM - Fire Dog: you can use those curves but yourr algorithmm might not mamp them properly to the actual situation
  110. 5:07 PM - euclideon raytracing: what if every spot of the computer that needs the cached curve has that data available to it pysically at all times is that the entire purpose of caching something?
  111. 5:07 PM - Fire Dog: it still is problematic
  112. 5:08 PM - Fire Dog: because how can you determine if your cahced curve is close to what the real thing should be
  113. 5:08 PM - Fire Dog: that's the really hard question
  114. 5:08 PM - euclideon raytracing: cubes
  115. 5:09 PM - euclideon raytracing: determine what curve to apply to your cube based on the meta data extrapolated from 2d pictures
  116. 5:09 PM - euclideon raytracing: does calculating a curve instead of a line need too much space?
  117. 5:10 PM - euclideon raytracing: should computations limits paint by the rules of a photo realist or that of picasso
  118. 5:11 PM - Fire Dog: i mean how things would work is you would have some cached equation/curve
  119. 5:11 PM - Fire Dog: you would have to map that onto what your pictures give
  120. 5:12 PM - Fire Dog: which is a very complicated problem
  121. 5:12 PM - Fire Dog: that doesn't even work some of the time
  122. 5:12 PM - euclideon raytracing: change the curve based on an interpretation of the meta data
  123. 5:13 PM - euclideon raytracing: combine the multiple angle pictures in a smart way
  124. 5:13 PM - euclideon raytracing: like a character reference sheet
  125. 5:14 PM - Fire Dog: again the question is what is that smart way
  126. 5:14 PM - Fire Dog: it's not an easy question to answer
  127. 5:14 PM - Fire Dog: most methods for doing stuff like that don't have a 100% success rate
  128. 5:14 PM - euclideon raytracing: you can make the assumption for something static
  129. 5:14 PM - euclideon raytracing: it will stay there
  130. 5:15 PM - euclideon raytracing: multiple angles of something static
  131. 5:15 PM - euclideon raytracing: you can write assumptions for the curves to follow
  132. 5:15 PM - euclideon raytracing: if this flower pot
  133. 5:16 PM - euclideon raytracing: has the same width at all angles
  134. 5:16 PM - euclideon raytracing: it is circular
  135. 5:16 PM - euclideon raytracing: whether or not something is circular is defined by it's shading
  136. 5:16 PM - Fire Dog: yeah you could do something like that for basica objects but if you really want high definition it will be extremely difficult
  137. 5:16 PM - Fire Dog: there aren't easy rules like that to follow for stuff
  138. 5:17 PM - euclideon raytracing: shading comes from your color buffer
  139. 5:16 PM - Fire Dog: like if you want to model the bark of a tree or something
  140. 5:17 PM - euclideon raytracing: bark follows patterns
  141. 5:17 PM - Fire Dog: in the short term it does
  142. 5:17 PM - Fire Dog: but what if the tree got hit by lightning
  143. 5:17 PM - Fire Dog: or something
  144. 5:17 PM - Fire Dog: all of a sudden the pattern is broken
  145. 5:19 PM - euclideon raytracing: well idk I'm not thinking about simulating things that are super interactive
  146. 5:19 PM - euclideon raytracing: if the game calls for that you can just write it in the story
  147. 5:19 PM - Fire Dog: that sounds more plausible then
  148. 5:20 PM - Fire Dog: you could predefine some patterns and force your program to follow them to generate the environment
  149. 5:21 PM - euclideon raytracing: can you combine polygon computing and raytracing in the same 3d game engine?
  150. 5:21 PM - euclideon raytracing: and make them be consistent when you move around it
  151. 5:22 PM - Fire Dog: you probably could but idk about how well it would function
  152. 5:22 PM - euclideon raytracing: assign a value to hammer units let's say 1 for every hammer unit equals a single raytrace unit
  153. 5:23 PM - euclideon raytracing: when you move along them they progress together
  154. 5:23 PM - Fire Dog: you could probably do somemthing like that but the raycasting is so expensive commputationally
  155. 5:23 PM - euclideon raytracing: why
  156. 5:23 PM - Fire Dog: you would have to get a huge benefit out of it for it to be worth it, and it still wouldn't be good on modern hardware
  157. 5:23 PM - Fire Dog: raycasting just uses a ton of resources
  158. 5:24 PM - euclideon raytracing: how many lasers need powered
  159. 5:24 PM - euclideon raytracing: what energy is required
  160. 5:25 PM - Fire Dog: do you mean computational raycasting?
  161. 5:25 PM - Fire Dog: it just needs to calculate things a lot
  162. 5:25 PM - Fire Dog: so it takes foreverr
  163. 5:26 PM - euclideon raytracing: computational?
  164. 5:26 PM - euclideon raytracing: no like cameras
  165. 5:27 PM - euclideon raytracing: if you want to make ultra realistic stop motion animation
  166. 5:27 PM - euclideon raytracing: can you do it just by looking at 2d pictures
  167. 5:27 PM - euclideon raytracing: from multiple angles
  168. 5:27 PM - euclideon raytracing: hi def
  169. 5:28 PM - euclideon raytracing: also at the cellular level you can you ask the computer to draw pattern of what a cell looks like close up
  170. 5:29 PM - euclideon raytracing: you can ask*
  171. 5:29 PM - euclideon raytracing: BUT
  172. 5:29 PM - euclideon raytracing: you would need a lot of pictures for the right colors
  173. 5:29 PM - euclideon raytracing: so you can compare them all
  174. 5:31 PM - Fire Dog: BASICALLY
  175. 5:31 PM - Fire Dog: sorry caps
  176. 5:32 PM - euclideon raytracing: basically wat
  177. 5:34 PM - euclideon raytracing: the ability to amass light speed distance measurements into a large data center like if the government gives you a lot of money
  178. 5:34 PM - Fire Dog: i mean you could use 2d pictures like you said to create an environment
  179. 5:34 PM - euclideon raytracing: what if it was the United Nations and they sponsored this idea with 5 billion dollars
  180. 5:35 PM - euclideon raytracing: instead of just the australian government
  181. 5:35 PM - Fire Dog: to measure the distance of what exactly
  182. 5:36 PM - euclideon raytracing: anything that's near you and bounces light off itself
  183. 5:36 PM - euclideon raytracing: can the human eye evolve to be better?
  184. 5:37 PM - euclideon raytracing: do I happen to have the most perfect eyes an intelligent species could ever have?
  185. 5:37 PM - euclideon raytracing: what if we made computers that simply used pure visible light as input
  186. 5:38 PM - euclideon raytracing: is this possible
  187. 5:38 PM - euclideon raytracing: ?
  188. 5:38 PM - euclideon raytracing: what meta data is not stored in pictures but can be found in light
  189. 5:39 PM - euclideon raytracing: is the only thing maybe material type?
  190. 5:39 PM - euclideon raytracing: that's how we know what the contents of space are made of
  191. 5:41 PM - euclideon raytracing: if something can pass a turing test with an entirely consistent personality does it become more than an illusion?
  192. 5:41 PM - euclideon raytracing: why are robots bad at being self consistent
  193. 5:42 PM - euclideon raytracing: with their personalities
  194. 5:42 PM - euclideon raytracing: one could say having a self consistent personality is the only thing that separates you from the average human profile
  195. 5:44 PM - euclideon raytracing: my questions are probably retarded im sorry
  196. 5:44 PM - euclideon raytracing: no one wants to talk about that stuff with me ever
  197. 5:46 PM - euclideon raytracing: if you wanted to create a photorealistic painting would you do it by making tiny lines
  198. 5:46 PM - euclideon raytracing: as a literal question I'd say no to that
  199. 5:47 PM - Fire Dog is now Online.
  200. 5:48 PM - euclideon raytracing: caching pictures with different focal lenses but maybe in the same place
  201. 5:49 PM - euclideon raytracing: same place in the 3d area
  202. 5:49 PM - euclideon raytracing: that gives you more data to use
  203. 5:50 PM - euclideon raytracing: I suppose lasers don't have that problem
  204. 5:51 PM - euclideon raytracing: what if the key is knowing where the camera is at what time
  205. 5:54 PM - euclideon raytracing: if you can tween in flash apply that math to your raytrace model
  206. 5:55 PM - euclideon raytracing: you can fake the inbetween who says it has to actually happen
  207. 6:28 PM - euclideon raytracing: if the computer can separate the background from the foreground definitively using the length of each laser you probably need a minimum 4 laser picture angles because that is the number of surfaces needed to enclose a 3d space
  208. 6:30 PM - euclideon raytracing: I wonder if they're taking something closer to a bunch of laser pictures or a bunch of laser videos
  209.  
  210. IS PRECACHING A DATA BASE OF CURVES or "TURNS" THAT ACT ON A 2D GRID BUT AT THE SAME TIME CAN MOVE IN 3D SPACE DATA [(extrapolated by taking many laser measurements from at least 4 angles looking inward toward the object like encasing something within a triangular pyramid)] LIKE A CAMERA THE FUTURE OF GRAPHICAL COMPUTING?
  211.  
  212. QUESTION: DO MORE THAN 4 ANGLES GIVE A BETTER OR MORE ACCURATE 3D MAP?
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement