Search the Community

Showing results for tags 'R&D'.



More search options

  • Search By Tags

    Type tags separated by commas.
  • Search By Author

Content Type


Categories

  • Audio
    • Music and Sound FX
  • Business
    • Business and Law
    • Career Development
    • Production and Management
  • Game Design
    • Game Design and Theory
    • Writing for Games
    • UX for Games
  • Industry
    • Interviews
    • Event Coverage
  • Programming
    • Artificial Intelligence
    • General and Gameplay Programming
    • Graphics and GPU Programming
    • Engines and Middleware
    • Math and Physics
    • Networking and Multiplayer
  • Visual Arts
  • Archive

Categories

  • News

Categories

  • Audio
  • Visual Arts
  • Programming
  • Writing

Categories

  • Audio Jobs
  • Business Jobs
  • Game Design Jobs
  • Programming Jobs
  • Visual Arts Jobs

Categories

  • GameDev Unboxed

Forums

  • Audio
    • Music and Sound FX
  • Business
    • Games Career Development
    • Production and Management
    • Games Business and Law
  • Game Design
    • Game Design and Theory
    • Writing for Games
  • Programming
    • Artificial Intelligence
    • Engines and Middleware
    • General and Gameplay Programming
    • Graphics and GPU Programming
    • Math and Physics
    • Networking and Multiplayer
  • Visual Arts
    • 2D and 3D Art
    • Critique and Feedback
  • Topical
    • Virtual and Augmented Reality
    • News
  • Community
    • GameDev Challenges
    • For Beginners
    • GDNet+ Member Forum
    • GDNet Lounge
    • GDNet Comments, Suggestions, and Ideas
    • Coding Horrors
    • Your Announcements
    • Hobby Project Classifieds
    • Indie Showcase
    • Article Writing
  • Affiliates
    • NeHe Productions
    • AngelCode
  • Workshops
    • C# Workshop
    • CPP Workshop
    • Freehand Drawing Workshop
    • Hands-On Interactive Game Development
    • SICP Workshop
    • XNA 4.0 Workshop
  • Archive
    • Topical
    • Affiliates
    • Contests
    • Technical

Calendars

  • Community Calendar
  • Games Industry Events
  • Game Jams

Blogs

There are no results to display.

There are no results to display.

Marker Groups

  • Members

Developers


Group


About Me


Website


Industry Role


Twitter


Github


Twitch


Steam

Found 10 results

  1. My SDF font looks great at large sizes, but not when I draw it at smaller sizes. I have my orthogonal projection matrix setup so that each unit is a 1x1 pixel. The text is rendered from Freetype2 to a texture atlas @ 56px with a spread of 8 pixels (the multiplier is 8x and scaled down). I'm drawing @ 18px in the screenshot attached to this post. The way I calculate the size of the text quads is by dividing the desired size (18px in the screenshot) by the size of the glyphs in the atlas (56px in this case), and scaling the glyph sprite by that factor. So: 18/56 = ~0.32, and I multiply the rect's size vector by that when it comes to vertex placement (this obviously doesn't apply to the vertices' texture coords). Now, I made sure that all metrics stored in my SDF font files are whole numbers (rect position/size, bearing amounts, advance, etc), but when I scale the font, vertex positions are almost always not going to be whole numbers. I increase the "edge" smoothstep shader parameter for smaller text as well, but it doesn't seem to help all that much.
  2. Hi guys, So I have an AI game in mind and I was wondering what are the best ways or techniques to sell my idea of my prototype and proof of concept. Should I make a trailer? Should make a magazine style book? Should I make a video in talking about my game like they do in Kickstarter campaigns? Any feedback would be highly appreciated!
  3. Hello guys, I just registered this site and heard from my lecturer that this a good site to talk about certain topics since my research topic are mostly programmer who are experienced with AI can answer the survey. The reason of the survey below is to understand which is suitable solution for 2d platformer pathfinding for AI and which one is easier to implement for 2D platformer. I would appreciate if you guys give your responses for the survey link shared and thank you for spending time answering the survey. Sorry if the survey is a bit hard to understand, I tried to make it understandable as best as I can. Again, thank you! https://goo.gl/forms/S0etAlAAHL6S5kTI2
  4. NVIDIA has developed a new machine learning methodology for generating unique and realistic looking faces using GAN i-e Generative Adversarial Network. The GAN is not a new technology, but where NVIDIA differentiates is through the progressive training method it developed. The key idea is to grow both the generator and discriminator progressively: starting from a low resolution, they add new layers that model increasingly fine details as training progresses. This both speeds the training up and greatly stabilizes it, allowing them to produce images of unprecedented quality. NVIDIA took a database of photographs of famous people and used that to train its system. By working together, the neural networks were able to produce fake images that are nearly indistinguishable from real human photographs. https://www.youtube.com/watch?v=2edOMMREazo
  5. Hi, Tile based renderers are quite popular nowadays, like tiled deferred, forward+ and clustered renderers. There is a presentation about GPU based particle systems from AMD. What particularly interest me is the tile based rendering part. The basic idea is, that leave the rasterization pipeline when rendering billboards and do it in a compute shader instead, much like Forward+. You determine tile frustums, cull particles, sort front to back, then render them until the accumulated alpha value is below 1. The performance results at the end of the slides seems promising. Has anyone ever implemented this? Was it a success, is it worth doing? The front to back rendering is the most interesting part in my opinion, because overdraw can be eliminated for alpha blending. The demo is sadly no longer available..
  6. Hi everyone, I'm new to learning AI pathfinding, although R&D for 3 month... so here is the problem that I had, hope there are someone can give me some advice. I was trying to develop my Flight Agent in 3D space. and store all the obstacles reference into octree. so I got the space & obstacle distribution represent in octree already, however I wanted to implement 1) Navigation network graph region (to define the space in region) 2) HPA* , to implement the high level path finding based on simplify network graph the major issue : I know how to collect all remain space from octree node, but I didn't know how to segment the space, and build up a network graph based on it.
  7. Hi, i’m trying to build an effective AI for the Buraco card game (2 and 4 players). I want to avoid the heuristic approach : i’m not an expert of the game and for the last games i’ve developed this way i obtained mediocre results with that path. I know the montecarlo tree search algorithm, i’ve used it for a checkers game with discrete result but I’m really confused by the recent success of other Machine Learning options. For example i found this answer in stack overflow that really puzzles me, it says : "So again: build a bot which can play against itself. One common basis is a function Q(S,a) which assigns to any game state and possible action of the player a value -- this is called Q-learning. And this function is often implemented as a neural network ... although I would think it does not need to be that sophisticated here.” I’m very new to Machine Learning (this should be Reinforcement Learning, right?) and i only know a little of Q-learning but it sounds like a great idea: i take my bot, making play against itself and then it learns from its results… the problem is that i have no idea how to start! (and neither if this approach could be good or not). Could you help me to get the right direction? Is the Q-learning strategy a good one for my domain? Is the Montecarlo still the best option for me? Would it work well in a 4 players game like Buraco (2 opponents and 1 team mate)? Is there any other method that i’m ignoring? PS: My goal is to develop an enjoyable AI for a casual application, i can even consider the possibility to make the AI cheating for example by looking at the players hands or deck. Even with this, ehm, permission i would not be able to build a good heuristic, i think Thank you guys for your help!
  8. Hello, I have been here about 2 months ago with my Dynamic Foam project and got help from a developer but we ran into some problems connecting centroids to establish a network between the 'foam bubbles'. Here is some extra information for why those centroids need to be connected and form a network along with the problem explained at the end. I. Dynamic Medium II. Fields and Boids III. From Boids to a PBD Network IV. From Curling Motion to Strings V. Position Based Dynamics (PBD) VI. The Problem -- I. Dynamic Medium The goal of this project is to create a kind on a Zero-player game, a bit like Conway's Game of Life, where cells fluctuate based on surrounding cells, and where the goal to generate Knots within a Dynamic Medium. To get a sense of such a dynamic network think of a foam with bubbles (volumes) where a micro-fluid runs through the edges (currents). These currents move from high to low pressure, forming circuits. The intensity of the fluid passing the bubbles can make the bubbles shrink or expand. Some currents will be able to line up and form closed circuits. In 2D these structures are simple loops, in 3D these loops can form strings, and in a next step these strings form again closed-circuits -> knots. -- II. Fields and Boids One way to get a model of bubbles and currents going was by using Boid-particles and Fields, which was tried in Processing. Here are short clips to show the interaction: • http://imgur.com/a/Mp0SG • https://vimeo.com/user37290268 A. In the first clip the Fields (red circles) keep their size while the small particles flow in between and form a circuit. B. In the second version the size of the Fields is influenced by the number of particles within. This gives rise to a dynamic foam with Fields fluctuating. Expand <-> Shrink -- III. From Boids to a PBD Network The problem with using Boids is that they have their limitations whereby the quick expansion of Fields causes the particles to be splattered around, interrupting the steady current. (stable flow vs. splashed around) Keeping the Boid-particles flowing is important, because a flow-circuit gives rise to organisational rules such as flows going against each other block each other, others can go along and strengthen each other, similar to an electronic circuit: So to fix these problems a move to PBD was made and hoping to create a network model (nodes/edges) that replaces the small particles. The flow replaced by edges: Fields expanding and shrinking depending on the amount of flow through the edges: • Having two Fields A.B. and in between from point a. to b. an edge. • This edges (a.b.) can replace all the boids moving in between (A.B.) • Edge (a.b.) can represents 1 or 1 million small particles or more, simplification. • Between Field A. and B. a measurement of tension •-VVVVV-• (A/B) • The more tension between (A.B.) the less flow there can be in (a.b.) • Inverse the more flow there is in (a.b.) the less tension between (A.B.) • The edge (a.b.) runs through the Fields A. and B, cooling or heating up Fields A. and B. • Through edge (a.b.) runs a current the larger this current the bigger the Fields A. and B. become, expansion vs. shrinking, hot vs. cold, condense vs. vaporise. The inspiration for the move to 'Position Based Dynamics' came from a talk by Jos Stam where he explained a method explicit on position and using springs: http://www.birs.ca/events/2014/5-day-workshops/14w5147/videos/watch/201402191414-Stam.html http://www.birs.ca//workshops//2014/14w5147/files/Stam-BIRS-2014.pdf -- IV. From Curling Motion to Strings At the basis of this Dynamic Foam is the idea to get the bubbles moving through the medium based on self regulated currents between the bubbles: The mechanics are like a sliding-puzzle where the parts can move to where there is space (created): Space can be created by letting the bubbles shrink on one side and expand on the other, so they can move into that created space, and as a result we get fluctuations and a rolling-curling-motion: Here is a fun animation that is similar to the idea: https://imgur.com/gallery/xIUGOd9 But for those lemons the input-force is gravity, in my model it are the micro-fluid-currents that effect the size. If the fluid is colder than the bubble than it will condensate and the field will expand; when it is warmer the field will vaporise and shrink. The dynamic curing motion from above, could also turn around and close-loop into knots: In 3D the wind up curls would be like these folk dansers waving strings: http://www.youtube.com/watch?v=uvE5yt83WPU (at 1:44) These strings can turn into knots etc. etc. -- V. Position Based Dynamics (PBD) For a look at the current situation you can check this short clip, where a small network is in place based on centroids, and where the volumes can fluctuate: http://imgur.com/QSG71hW Here is are the project files on GitHub if you would like to have a closer look: • https://github.com/VirtualOrganics/PBD_DynamicFoam_Files • https://github.com/InteractiveComputerGraphics/PositionBasedDynamics (The PDB software) -- VI. The Problem We took the PBD software from GitHub and modified the Demo example called 'GenericConstraintsDemos'. All constraints are taken out except the distance constraint , and the simulation is limited to work on a 2D plane with the intention of extending to 3D eventually. Check out the graph: As you can see in the example, there are nine circles. The simulation allows for any number of circles. The green dots represent the centers of the circles; the red dots represent the centroids of the three adjacent/connected circles; the black dots represent some (outer) intersections between circles; and the blue lines connect either centroids with other centroids, or centroids to intersections in a particular pattern. In the simulation, the radii of the circles can change and, since a distance constraint is applied, the varying sizes of the circles causes the locations (and existence) of the dots and lines to change. The code at this point is able to identify/calculate the locations of the black dots and the red dots, but he had trouble coming up with an algorithm that correctly connects the red and black dots. The diagram above shows the correct connections, but coming up with an algorithm to do this automatically and as the simulation runs stalled our progress. So I'm now looking for a developer who can take it to the next level. It doesn't necessarily need to be in PBD an other Physics Engine (Bullet?) might do the job, one where on top of the physics-interactions a network-formula can balance out the weights and regulate the system. All suggestions and questions are more than welcome! Best, m.
  9. I was reworking on my LightProbe filter, and I wrote some code to generate the Reference Cubemap, but then I noticed some discontinuous on the border of each face.(Top:CPU implementaion, Bottom: GPU implementation, the contrast has been adjusted on the right side) At first I think it maybe caused by the interpolation, but then I tried the same algorithm in 2D (like a slice in the normal light probe prefiltering) for better visualization, and the result really confused me. See the attachments, the top half is the Prefiltered Color value, displayed per channel, it's upside down because I used the ColorValue directly as the y coordinate. The bottom half is the differential of the color, it's very clearly there is a discontinuous, and the position is where the border should be. And as the roughness goes higher, the plot gets stranger . So, I am kinda of stuck in here, what's happening and what to do to remove this artifact? Anybody have any idea? and here is my code inline FVector2D Map(int32 FaceIndex, int32 i, int32 FaceSize, float& SolidAngle) { float u = 2 * (i + 0.5) / (float)FaceSize - 1; FVector2D Return; switch (FaceIndex) { case 0: Return = FVector2D(-u, -1); break; case 1: Return = FVector2D(-1, u); break; case 2: Return = FVector2D(u, 1); break; case 3: Return = FVector2D(1, -u); break; } SolidAngle = 1.0f / FMath::Pow(Return.SizeSquared(), 3.0f / 2.0f); return Return.SafeNormal(); } void Test2D() { const int32 Res = 256; const int32 MipLevel = 8; TArray<FLinearColor> Source; TArray<FLinearColor> Prefiltered; Source.AddZeroed(Res * 4); Prefiltered.AddZeroed(Res * 4); for (int32 i = 0; i < Res; ++i) { Source[i] = FLinearColor(1, 0, 0); Source[Res + i] = FLinearColor(0, 1, 0); Source[Res * 2 + i] = FLinearColor(0, 0, 1); Source[Res * 3 + i] = FLinearColor(0, 0, 0); } const float Roughness = MipLevel / 8.0f; const float a = Roughness * Roughness; const float a2 = a * a; // Brute force sampling with GGX kernel for (int32 FaceIndex = 0; FaceIndex < 4; ++FaceIndex) { for (int32 i = 0; i < Res; ++i) { float SolidAngle = 0; FVector2D N = Map(FaceIndex, i, Res, SolidAngle); double TotalColor[3] = {}; double TotalWeight = 0; for (int32 SampleFace = 0; SampleFace < 4; ++SampleFace) { for (int32 j = 0; j < Res; ++j) { float SampleJacobian = 0; FVector2D L = Map(SampleFace, j, Res, SampleJacobian); const float NoL = (L | N); if (NoL <= 0) continue; const FVector2D H = (N + L).SafeNormal(); const float NoH = (N | H); float D = a2 * NoL * SampleJacobian / FMath::Pow(NoH*NoH * (a2 - 1) + 1, 2.0f) ; TotalWeight += D; FLinearColor Sample = Source[SampleFace * Res + j] * D; TotalColor[0] += Sample.R; TotalColor[1] += Sample.G; TotalColor[2] += Sample.B; } } if (TotalWeight > 0) { Prefiltered[FaceIndex * Res + i] = FLinearColor( TotalColor[0] / TotalWeight, TotalColor[1] / TotalWeight, TotalColor[2] / TotalWeight); } } } // Save to bmp const int32 Width = 4 * Res; const int32 Height = 768; TArray<FColor> Bitmap; Bitmap.SetNum(Width * Height); // Prefiltered Color curve per channel float MaxDelta = 0; for (int32 x = 0; x < Width; ++x) { FColor SourceColor = Source[x].ToFColor(false); Bitmap[x] = SourceColor; FColor Sample = Prefiltered[x].ToFColor(false); check(Sample.R < 256); check(Sample.G < 256); check(Sample.B < 256); Bitmap[Sample.R * Width + x] = FColor(255, 0, 0); Bitmap[Sample.G * Width + x] = FColor(0, 255, 0); Bitmap[Sample.B * Width + x] = FColor(0, 0, 255); if (x > 0) { const FLinearColor Delta = Prefiltered[x] - Prefiltered[x - 1]; MaxDelta = FMath::Max(MaxDelta, FMath::Max3(FMath::Abs(Delta.R), FMath::Abs(Delta.G), FMath::Abs(Delta.B))); } } // Differential per channel const float Scale = 128 / MaxDelta; for (int32 x = 1; x < Width; ++x) { const FLinearColor Delta = Prefiltered[x] - Prefiltered[x - 1]; Bitmap[int32(512 + Delta.R * Scale) * Width + x] = FColor(255, 0, 0); Bitmap[int32(512 + Delta.G * Scale) * Width + x] = FColor(0, 255, 0); Bitmap[int32(512 + Delta.B * Scale) * Width + x] = FColor(0, 0, 255); } FFileHelper::CreateBitmap(TEXT("Test"), Width, Height, Bitmap.GetData()); } Roughness 0.5.bmp Roughness 1.bmp
  10. Hello, I'd like to ask your take on Lagarde's renormalization of the Disney BRDF for the diffuse term, but applied to Lambert. Let me explain. In this document: https://seblagarde.files.wordpress.com/2015/07/course_notes_moving_frostbite_to_pbr_v32.pdf (page 10, listing 1) we see that he uses 1/1.51 * percetualRoughness as a factor to renormalize the diffuse part of the lighting function. Ok. Now let's take Karis's assertion at the beginning of his famous document: http://blog.selfshadow.com/publications/s2013-shading-course/karis/s2013_pbs_epic_notes_v2.pdf Page 2, diffuse BRDF: I think his premise applies and is enough reason to use Lambert (at least in my case). But from Lagarde's document page 11 figure 10, we see that Lambert looks frankly equivalent to Disney. From that observation, the question that naturally comes up is, if Disney needs renormalization, doesn't Lambert too ? And I'm not talking about 1/π (this one is obvious), but that roughness related factor. A wild guess would tell me that because there is no Schlick in Lambert. and no dependence on roughness, and as long as 1/π is there, in all cases Lambert albedo is inferior to 1, so it shouldn't need further renormalization. So then, where does that extra energy appear in Disney ? According to the graph, it's high view angle and high roughness zone, so that would mean, here: (cf image) This is super small of a difference. This certainly doesn't justify in my eyes the need for the huge darkening introduced by the 1/1.51 factor that enters in effect on a much wider range of the function. But this could be perceptual, or just my stupidity. Looking forward to be educated Bests