|
Hanpan posted:Are there any shader buffs itt? I'm struggling to find an answer to what I'm hoping is relatively simple problem. If I understand correctly, one way to do it would be to set the stencil buffer when you render the water. Then you render the objects with a two pass shader, one for stencil pass, one for fail. When it passes the stencil comparison, your object shader does its normal thing. When it fails the stencil comparison, you render red(using your example) or whatever. You'll also need to make sure you render your water before the other things in your scene. The stencil buffer is Pro only until Unity 4.6.
|
# ? Sep 5, 2014 17:45 |
|
|
# ? May 28, 2024 08:43 |
|
So I've been trying to figure out exactly how dual contouring works and make my own implementation of it. I think I get the gist of what's going on with the algorithm, but I'm having trouble wrapping my head around the "Adaptive" part of it. The original dual contoring paper posted:The previous algorithm for dual contouring has the obvious disadvantage of being formulated for uniform grids. In practice, most of a uniform grid is devoted to storing homogeneous cubes (i.e; cubes whose vertices all have the same sign). Only a small fraction of the cubes are heterogeneous and, thus, intersect the contour. One way to avoid this waste of space is to replace the uniform grid by an octree. In this section, we describe an adaptive version of dual contouring based on simplifying an octree whose leaves contain QEFs. This method is essential an adaptive variant of a uniform simplification method. Our method has three steps. This reads to me like, "Instead of just a 3D grid, we simplify the grid by putting it into an octree and "collapsing" the tree for large stretches of space that are completely solid or completely empty" So it's not really an octree in the sense that you can't just throw random vertices into it because then you wouldn't really have a way to figure out where the "edge" of the surface is. So your data is still in a uniform grid but it's been "compacted" into the octree? I think this is what the paper means but I'm not 100% that I'm missing something.
|
# ? Sep 5, 2014 18:09 |
|
Obsurveyor posted:If I understand correctly, one way to do it would be to set the stencil buffer when you render the water. Then you render the objects with a two pass shader, one for stencil pass, one for fail. When it passes the stencil comparison, your object shader does its normal thing. When it fails the stencil comparison, you render red(using your example) or whatever. You'll also need to make sure you render your water before the other things in your scene. The stencil buffer is Pro only until Unity 4.6.
|
# ? Sep 5, 2014 18:30 |
|
OneEightHundred posted:For this type of thing, you only need depth test, and the problem in this case is probably that the depth test is the reverse of what it should be. Obsurveyor posted:If I understand correctly, one way to do it would be to set the stencil buffer when you render the water. Then you render the objects with a two pass shader, one for stencil pass, one for fail. When it passes the stencil comparison, your object shader does its normal thing. When it fails the stencil comparison, you render red(using your example) or whatever. You'll also need to make sure you render your water before the other things in your scene. The stencil buffer is Pro only until Unity 4.6. Hmm - I guess the issue here is going to be making it work on mobile. Feel like I must be missing a trick here, or at least the terminology to describe what I'm after, because you see this kind of thing quite a lot in games like Fez and Monument valley.
|
# ? Sep 5, 2014 19:46 |
|
Hanpan posted:Hmm - I guess the issue here is going to be making it work on mobile. It's nothing complex and I don't know why stencil popped into my head before depth but I think either would work would work on mobile fine, they've been used in 3D graphics for decades now. quote:Feel like I must be missing a trick here, or at least the terminology to describe what I'm after, because you see this kind of thing quite a lot in games like Fez and Monument valley. Maybe you should try describing what you're after again. It sounded to me like you wanted to show some kind shadow for the object under water instead of its actual color.
|
# ? Sep 5, 2014 20:03 |
|
Obsurveyor posted:
Basically this exact effect: http://imgur.com/mfn5Q0O See how the water turns the building a flat colour? That.
|
# ? Sep 5, 2014 21:42 |
|
Hanpan posted:Basically this exact effect: Do you really need a shader to do that? Not that shaders aren't cool, but you could do the same with unlit geometry set to some color.
|
# ? Sep 5, 2014 21:47 |
|
dupersaurus posted:Do you really need a shader to do that? Not that shaders aren't cool, but you could do the same with unlit geometry set to some color. I'd like it to be a shader, as not shown is how the water level rises and falls.
|
# ? Sep 5, 2014 22:14 |
|
There's usually a bunch of ways to achieve an effect like that. I haven't tried any of these but they are what I would try. This is the simplest one I can think of but involves modifying the shader of all your models and assumes that the water is flat. code:
code:
code:
code:
Flownerous fucked around with this message at 03:08 on Sep 6, 2014 |
# ? Sep 6, 2014 02:34 |
|
Anyone have a suggestion for improving my collision detection (floors) routine? I started playing around with a Javascript Donkey Kong clone. My sprite sheet (5x8) looks like this: Levels (by tile index) Works here: Doesn't work here: My code looks like this: code:
|
# ? Sep 7, 2014 16:05 |
|
Bob Morales posted:Anyone have a suggestion for improving my collision detection (floors) routine? In some platformer code I've seen there is a very thin invisible "skin" that goes on top of all the floors to sort of "help" the collision detection. Your code would something like: code:
http://gameprogrammingpatterns.com/state.html
|
# ? Sep 7, 2014 16:59 |
|
I have to be missing something completely trivial. Any ideas?quote:Error 5 error C2783: 'boost::weak_ptr<T> Entity::GetComponent(const ComponentId)' : could not deduce template argument for 'ComponentType' c:\users\floorislava\dropbox\workspace\ces\ces\src\main.cpp 44 1 CES Inside the entity class as a public member function: code:
|
# ? Sep 9, 2014 03:35 |
|
Which line is the error message indicating?
|
# ? Sep 9, 2014 04:26 |
|
floor is lava posted:I have to be missing something completely trivial. Any ideas? Since you didn't post the line that actually triggers the error, I assume you're trying to do "entity.GetComponent(id)" when you need to do "entity.GetComponent<ComponentTypeIWant>(id)" because the compiler can't deduce what type is appropriate.
|
# ? Sep 9, 2014 04:29 |
|
SupSuper posted:Since you didn't post the line that actually triggers the error, I assume you're trying to do "entity.GetComponent(id)" when you need to do "entity.GetComponent<ComponentTypeIWant>(id)" because the compiler can't deduce what type is appropriate. That did it. Thanks.
|
# ? Sep 9, 2014 11:23 |
|
Been working over the past few days on implementing a procedural dungeon generator as detailed here. Main algorithm is finally complete and I'm liking the results so far:
|
# ? Sep 9, 2014 14:19 |
|
ahmeni posted:Been working over the past few days on implementing a procedural dungeon generator as detailed here. Main algorithm is finally complete and I'm liking the results so far: Ah! That's my favorite dungeon generation algorithm because it's so unique. I like your take on it, and the realtime visualisation is really drat neat!
|
# ? Sep 9, 2014 14:24 |
|
ahmeni posted:Been working over the past few days on implementing a procedural dungeon generator as detailed here. Main algorithm is finally complete and I'm liking the results so far: This is sweet.
|
# ? Sep 9, 2014 15:00 |
|
Bob Morales posted:Anyone have a suggestion for improving my collision detection (floors) routine? http://higherorderfun.com/blog/2012/05/20/the-guide-to-implementing-2d-platformers/
|
# ? Sep 9, 2014 20:28 |
|
ahmeni posted:Been working over the past few days on implementing a procedural dungeon generator as detailed here. Main algorithm is finally complete and I'm liking the results so far: The best is those two rooms on the left that are right next to each other, but require a trek all the way around to get to between the two.
|
# ? Sep 9, 2014 22:41 |
|
SynthOrange posted:The best is those two rooms on the left that are right next to each other, but require a trek all the way around to get to between the two. Yeah, minimum spanning trees can unfortunately have that effect as the only condition is that the graph is entirely reachable. There's a bit of logic that throws some of the discarded edges back in and I may tweak it to prefer adding low cost routes. Though on the other hand you can do things like taking the longest chain and turn them into boss rooms or treasure rooms, etc. so it wouldn't so much be a long trek as two different objective paths.
|
# ? Sep 9, 2014 23:06 |
|
I don't think there's anything wrong with it, I kinda like it. Feels more unique to have that kind of thing occasionally. Very cool animation by the way. I love seeing algorithms depicted visually.
|
# ? Sep 10, 2014 00:15 |
ahmeni posted:Been working over the past few days on implementing a procedural dungeon generator as detailed here. Main algorithm is finally complete and I'm liking the results so far: This is seriously awesome looking. Did you do the visualization in-engine, or is this just a neat thing you produced with outside tools?
|
|
# ? Sep 10, 2014 03:49 |
|
Jo posted:This is seriously awesome looking. Did you do the visualization in-engine, or is this just a neat thing you produced with outside tools? At the moment this is all the engine does, with the underlying data being a bunch of room objects with some iteration logic to slow things down. With the amount of positive feedback I've heard about the animation it would be neat to integrate it as an intro to a level. The next step is translation into Phaser's tile map support.
|
# ? Sep 10, 2014 08:32 |
|
The reason I am so into game development and computers is that when you put in the time and theory you get awesome poo poo like that. You cna acheive your dreams
|
# ? Sep 10, 2014 09:00 |
|
ahmeni posted:At the moment this is all the engine does, with the underlying data being a bunch of room objects with some iteration logic to slow things down. With the amount of positive feedback I've heard about the animation it would be neat to integrate it as an intro to a level. The next step is translation into Phaser's tile map support. Yes it would be a pretty awesome opening but of course could give away a great deal of information about the map. If your game is exploration device you could show everything but the final version (up to the direct lines, but not the right angled lines), but it might also be awesome to let them know the room layout but have no idea what's in each room. You would get a much higher tolerance for waiting if you let people see that for their own levels. What did you program the generation in?
|
# ? Sep 10, 2014 10:01 |
|
So I got an Unreal Engine account this last weekend and have spent a few hours playing around with it. First and foremost, I can not believe how good the lighting in UE4 is. It is incredible. Secondly, Blueprint is pretty freaking awesome. I had originally planned on doing more in C++, but Blueprint is powerful enough that it looks like it can do practically everything I would ever need to do, quicker and easier than I could do it in C++ (I am not a professional programmer). Finally, I find UE4's pricing to be more attractive. I don't really need Unity3d's PRO features, but I feel better knowing that I have a legit license for everything with UE4. And I don't think I am ever going to have to worry about UE4 5% fee since I am mostly planning on free games and just screwing around. I only really have two complaints about UE4 when compared to Unity3d 1: The editor has more than a couple of bugs. So far, they mostly seem to be stability bugs. I have had a few instances where I was able to cause the editor to crash by repeating certain steps in a certain order. 2: WebGL is still in beta and is kind of a pain to get set up (also requires a 64bit browser for now). I know that Unity3d doesn't really have WebGL yet, but they do have a pretty decent browser plugin that gets you more or less to the same place.
|
# ? Sep 10, 2014 15:03 |
|
Hey chaps, I was wondering if anyone had any experience with speech recognition and specifically if there are good packages available for somehow deriving intent from speech that isn't just recognising single words regardless of context, I remember stuff involving POS tagging and hidden markov models from a rudimentary AI class I had years ago, but I was hoping there would be some sort of prebuilt solution for this sort of thing. It's basically for directing a character to do things like running to objects and interacting them and so on.
|
# ? Sep 10, 2014 18:19 |
|
brian posted:Hey chaps, I was wondering if anyone had any experience with speech recognition and specifically if there are good packages available for somehow deriving intent from speech that isn't just recognising single words regardless of context, I remember stuff involving POS tagging and hidden markov models from a rudimentary AI class I had years ago, but I was hoping there would be some sort of prebuilt solution for this sort of thing. It's basically for directing a character to do things like running to objects and interacting them and so on. It's been a while since I looked at this stuff as part of my master's thesis but as far as I know there is no ready solution. For general-purpose stuff like POS etc. you could look at the Python NLTK or the Stanford NLP tools but you should be aware that you'll need to do some post-processing to transform from POS to your specific intent. You'd also need something for the speech recognition part and you can probably find similar libraries for that part; looking around, I found CMU Sphinx but I can't comment on the quality. Still, dealing with the general case of extracting semantics from speech/text is a really hard problem and I'd recommend to make things easier by going with a more limited solution, either by writing your own natural language DSL or adapting a tool implementing subset of natural language like Attempto, which is what I ended up doing. Someone else may know better, game development-specific solutions but this is what I'd recommend.
|
# ? Sep 10, 2014 18:56 |
|
Ah brilliant, i'll look into all the links and see whether I'll ever have the time to implement it well enough to be worthwhile, a limited solution will almost definitely be fine given the context of just commanding a thing I imagine, I just want it to be as intuitive for the player as possible. Thanks!
|
# ? Sep 10, 2014 20:21 |
|
Thankfully the devs are solid working away improving things! The lighting is amazing when static but not as such when it's dynamic. i find the mix of the two unsettling at times. also the default temporal AA doesn't work for my game and there aren't many other options for tuning it! blueprints are amazing though and also the way you can use them, make a object that is a wall with lights built in and have the wall build it's own colours and yeah I am stoked l have made rockets that fire and follow my pawn and they run by themselves! I am exite
|
# ? Sep 10, 2014 22:27 |
|
Stick100 posted:Yes it would be a pretty awesome opening but of course could give away a great deal of information about the map. If your game is exploration device you could show everything but the final version (up to the direct lines, but not the right angled lines), but it might also be awesome to let them know the room layout but have no idea what's in each room. The generation itself is all currently in javascript with Phaser. I don't really like it but being pure js means I can hack away at it on breaks at work without the wind turbine that is Unity kicking in. I'll likely port it over to Unity/Futile once it's nearer to completion.
|
# ? Sep 11, 2014 06:54 |
|
Lowen SoDium posted:1: The editor has more than a couple of bugs. So far, they mostly seem to be stability bugs. I have had a few instances where I was able to cause the editor to crash by repeating certain steps in a certain order. 1. Which editor are you using, Windows Mac or OSX? 2. WebGL doesn't "require" 64 bit exactly. If you make a shipping build instead of a development build it will likely work on 32 bit browsers. There is also a setting you can change (it changes the heap size) that will make it much more likely to deploy to browsers. They asked for feed back on the HTML5 builds and I started a thread and it was 1. Get it working on 32 bit browsers, or give a dip switch to help. 2. Get it working on the launcher builds so you don't have to download the source. There were some other feedbacks, but those 2 were pretty much unanimous. They have stated they are working on HTML5 this cycle (usually about 1 month) and have not yet announced the features for the next build. In short if you want HTML5 it will likely be fixed in the next two months so just go ahead and get your game put together and by the time you're ready they should be ready too. (However you'll likely have to spend another $19 when it's ready). UE4 reminders: 1. The 5% only counts after your first 3K per quarter per game. So very few people will ever pay it. 2. You can cancel at anytime and retain github/new launchers access until you hit your 30 days. I always signup, check I have access then immediately cancel and have had full access for the full 30 days. 3. After your sub. expires you lose github access a bit after, and you stop getting access to new Launcher builds. However the launcher will still work and you can still re download all the builds/content that was available when your subscription lapsed. As an example of this if you paid $19 at launch and then immediately cancelled and then opened the launcher now (roughly 6 months later) you'd still be able to get access to UE 4.0 or 4.1 if that came out in the first 30 days and the sample content that launched, but not the new content like vehicle game. As it happens the vehicle game wouldn't work with UE4.0 as it required engine changes in UE4.2. It's an insane deal and I wish I wasn't so bought into C# (10 years experience) to make it unappealing to use.
|
# ? Sep 11, 2014 06:58 |
|
Stick100 posted:1. Which editor are you using, Windows Mac or OSX? Stick100 posted:2. WebGL doesn't "require" 64 bit exactly. If you make a shipping build instead of a development build it will likely work on 32 bit browsers. There is also a setting you can change (it changes the heap size) that will make it much more likely to deploy to browsers. Thats good news. I haven't actually tried a web build yet, but it sounds like it is coming along faster than I thought it was. Stick100 posted:
I didn't know a about the first thing. That is pretty good news too. As far as it being unappealing to use because of your C# experience, have you messed with Blueprints? I am curious how a more experienced programmer feels about them, especially a C# programmer. I have a friend who is a professional programmer who does almost everything in Java or C#. He has mostly been playing in Unity 3d because of that and and hasn't tried Unreal Engine because he likes C# a lot more than C++. After I have been playing in UE4, I have been telling him about Blueprint and I think he would like it but I am curious about how another C# programmer feels about it.
|
# ? Sep 11, 2014 12:53 |
|
Lowen SoDium posted:curious about how another C# programmer feels about it. Me too! Because I don't "get" my OOP so well, blueprints let me keep trying things till they work. It would be a MUCH slower process with code. For a seasoned coder who made little errors I can't imagine much benefit in them. Currently my disappointment is with how temporal AA is not suitable for something I am trying to do (leaves artifacts) but it overall looks much nicer than FSAA which I'm left using. Welp
|
# ? Sep 11, 2014 19:59 |
|
Lowen SoDium posted:As far as it being unappealing to use because of your C# experience, have you messed with Blueprints? I am curious how a more experienced programmer feels about them, especially a C# programmer. I have a friend who is a professional programmer who does almost everything in Java or C#. He has mostly been playing in Unity 3d because of that and and hasn't tried Unreal Engine because he likes C# a lot more than C++. After I have been playing in UE4, I have been telling him about Blueprint and I think he would like it but I am curious about how another C# programmer feels about it. Blueprint is all right I've spent a couple hours with it and like it. It's REALLY well implemented and a much better visual programming language than anything else I've seen (like Scratch or Playmaker). If I was going to work full time on game development I think I could throw out my C# work and switch to blueprint + C++. If you are not a professional programmer and have no interest in being one then UE4/Blueprint is the best solution for making games currently. I'm still working days (and probably will for years to come) as a C# dev so I'm in C# already 40 hours a week using Visual Studio and Resharper and very comfortable with that tool chain. Now that Unity works with Visual Studio and Resharper out of the box I can't really choose any other tool set or my sanity and career. The experience transfers exactly between web dev and game dev, you're dealing with nearly the same issues so I can't really deal with another programming environment. As a programmer who has already had to deal with certain issues (like OAuth calls to Google, or updating large data sets quickly) in his day job, being able to use the exact same code is really useful. In the last 3 years I've done Java, C++, C#, VB.Net, Javascript, a couple flavors of SQL and by far what I've found most productive is when I can spend the most amount of time in C#.
|
# ? Sep 11, 2014 20:27 |
|
echinopsis posted:Me too! Because I don't "get" my OOP so well, blueprints let me keep trying things till they work. It would be a MUCH slower process with code. For a seasoned coder who made little errors I can't imagine much benefit in them. Blueprint (and all visual programming languages that emphasis state) are useful for a professional programmer but certain things (like doing a parallel foreach) will always be way easier in a non-visual programming language. This is why often times people will mix the two with certain developers using one much more heavily than another. Unity actually has a pretty good state based system in Mecanim Animation System. In the training video for Unite they explain and show that Mecanim can actually be used like Playmaker to create Finite State Machines completely unrelated to animation.
|
# ? Sep 11, 2014 20:32 |
|
The Gamer's Choice phase of SA GameDev is now live, if folks want to check that out. There are some... really, really weird games this year. It's awesome
|
# ? Sep 11, 2014 20:56 |
|
Speaking of Mecanim, how in the world can you properly check animation completion? We're working on a 2d game that uses converted spriter -> Mecanim animations, and we need events to fire at the end of certain clips. From what we've seen with some off and on googling, the last few weeks, this may be impossible to do reliably? That can't be correct, can it? We've tried checking the current clip normalized time, and that doesn't do the trick. Should we be looking at coroutines for this?
|
# ? Sep 12, 2014 13:43 |
|
|
# ? May 28, 2024 08:43 |
|
superh posted:Speaking of Mecanim, how in the world can you properly check animation completion? We're working on a 2d game that uses converted spriter -> Mecanim animations, and we need events to fire at the end of certain clips. From what we've seen with some off and on googling, the last few weeks, this may be impossible to do reliably? That can't be correct, can it? You can't. You can use an addon someone made called Mecanim Event Manager but even then the events don't fire sometimes if it's mid-transition and basically: Unity is hot garbage and I'm in pain every day until I finish the game I'm forced to work on in unity
|
# ? Sep 12, 2014 13:50 |