gimymblert
|
|
« Reply #300 on: July 31, 2018, 12:01:04 AM » |
|
Roguelikes and the Modular Genre
|
|
|
Logged
|
|
|
|
|
gimymblert
|
|
« Reply #302 on: July 31, 2018, 12:59:58 AM » |
|
Building The AI For BioShock Infinite's Elizabeth
Creating Bioshock Infinite's Elizabeth
|
|
|
Logged
|
|
|
|
|
gimymblert
|
|
« Reply #304 on: July 31, 2018, 01:03:35 AM » |
|
Helping It All Emerge: Managing Crowd AI in Watch Dogs 2 Free Range AI: Creating Compelling Characters for Open World Games https://www.youtube.com/watch?v=RFWrKHM0vAgModeling AI Perception and Awareness in Splinter Cell: Blacklist
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #305 on: July 31, 2018, 01:38:24 AM » |
|
AI For Generated Worlds
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #306 on: August 01, 2018, 11:30:05 PM » |
|
|
|
« Last Edit: August 01, 2018, 11:45:08 PM by gimymblert »
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #307 on: August 02, 2018, 12:32:07 AM » |
|
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #308 on: August 03, 2018, 04:36:33 AM » |
|
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #309 on: August 17, 2018, 12:18:03 AM » |
|
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #310 on: September 10, 2018, 08:57:26 AM » |
|
|
|
|
Logged
|
|
|
|
JobLeonard
|
|
« Reply #311 on: September 10, 2018, 11:26:14 AM » |
|
The explanation of how to project tiles onto spheres and whatnot is good, but the repetitive 2D tiles in the end looks horrible and I thought the article was going to fix that issue. But there already is a fix for that (that surely has been linked here before, because it's from 2006), namely Wang tileshttp://johanneskopf.de/publications/blue_noise/
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #312 on: September 10, 2018, 12:49:18 PM » |
|
To check if something has been posted, click print on the top or back of the page, then do a "crtl f" to use your browser to search for a particular link. That's how I do it, then decide if it's worth reposting or not depending on context (like a thematic dump of links).
That's interesting, I never thought of using that for noise.
Also I don't post based on merit, I post to archive to have option to look at different takes. People will have different requirement depending on goal, technical proficiency or optimization need. EXAMPLE: So sometimes the less optimal visually might just be good enough if it's simple to do.
|
|
|
Logged
|
|
|
|
JobLeonard
|
|
« Reply #313 on: September 10, 2018, 01:49:46 PM » |
|
I'm sorry, I wasn't clear: it wasn't a bad article, it's just that based on the title I went in with different expectations of what it would be explaining
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #314 on: September 17, 2018, 06:10:19 PM » |
|
You have to be a little careful with predictable pseudo-random sequences, because even if the sequence of numbers is predictable and repeatable, different parts of your code need to be sure to ask for the "next number" in the same temporal order every time.
IOW, you don't want 20 GOs each asking for their number in an order that depends on the GOs' execution order during init as decided by Unity.
Dungeon Architect solves this problem by having one external integer determine things predictably for the entire model, and if you have more than one DA model in your game, you track which numbers go to which models somewhere outside DA itself. This is a very solvable problem, but one that needs to be reflected in the code design. https://forum.unity.com/threads/archimatix-pro-node-based-parametric-modeling-for-unity-unity-awards-finalist.330869/page-55#post-3688477
|
|
|
Logged
|
|
|
|
JobLeonard
|
|
« Reply #315 on: September 18, 2018, 06:32:49 AM » |
|
How does that work with asynchronous code, like threaded stuff?
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #316 on: September 18, 2018, 06:42:22 AM » |
|
surely the same way, the seed are fix in a specific structure I think. It's also depend on your generative architecture. Personally I would recommend hash on a number series, easier to manage because number have intrinsic order. But random hash tend to be more expensive, especially perfect random hash I think.
|
|
|
Logged
|
|
|
|
BorisTheBrave
|
|
« Reply #317 on: September 18, 2018, 01:52:24 PM » |
|
Just to be clear, the trick is very simple and goes like this (pseudo code): Thing generateAThing(int seed) { ... }
Thing[] generate20Things(int seed) { var prng = new PRNG(seed); int[] seeds = prng.getInts(20); parallel for(int i=0;i<20;i++) { generateAThing(seeds[i]); } }
generateAThing always gives the same thing for the same seed, so the exeuction order of the for loop is irrelevant. Basically, every time you want to do things in parallel, or lazily, and so on, make a new seed or PRNG sequence, non-randomly based on the input seed / PRNG.
|
|
|
Logged
|
|
|
|
JobLeonard
|
|
« Reply #318 on: September 19, 2018, 03:41:47 AM » |
|
BTW, this made me remember the PCG family of RNGS: http://www.pcg-random.org/Supposedly really good/fast (search didn't turn up any results so not posted here before).
|
|
|
Logged
|
|
|
|
gimymblert
|
|
« Reply #319 on: September 19, 2018, 10:36:59 AM » |
|
Good catch, I knew about it, i'm surprise I didn't put it, I think I put the video of the creator instead :/
I'm looking in the future to understand how the party trick work (when it start spouting shakespear stuff) To see if I can actually encode/compress stuff into a seed, how much and at what cost.
edit: I share it once october 25 2016, but a bis is never hurting! Going to reshare the video as well: it's very accessible and great education
Stanford Seminar - PCG: A Family of Better Random Number Generators
|
|
|
Logged
|
|
|
|
|