/v/ - Video Games

>>15325673

>Clothes less complicated than weapons

What?

>>15325716

Last time I made weapons they looked very awkward, even with references I could never make complex weapon meshes without also having horribly messy UV maps.

>>15326042

Develop that one thing you want to do first, save a template and move onto the next, then put the puzzle together.

>>15325673

What are your prices? I just need an anime blonde girl dressed in a maid outfit with a blue ribbon.

>>15326085

You need the girl with the dress or just the dress? rigging is usually the hardest part, since I am not putting in any complex textures I would go for around $100 to $200 USD but it really depends on what exactly you need, also 50% upfront (last time I got a commission from a developer from another channel I got a $500 USD commission with no upfront payment and the client has gone completely silent which basically means I wasted my time), you can contact me by e-mail: vanfanel.soma@gmail.com

>>15326042

It's worse when you start development, then have to do uni work then develop more ideas for even more games when you can't even work on your first idea.

I now have 4 game ideas in my head. It's torture

>>15326042

This happens to me on anything that's big. The trick is organization: make a spreadsheet and break down your project into it's components. Then break the components into smaller ones and so on. Then you'll end up with a list of small, achievable tasks and you can focus on one every week. This isn't just a productive strategy, complex projects are too big to fully conceptualize, that's why one gets overwhelmed. It's not the size, it's not knowing what you should do first.

>>15326254

even the logo is cute

>>15326226

Hey I got a reply, my shitpost paid off

How much would it most likely cost to hire someone to do sprite art? I'm doing it all myself right now, but I'm hoping to eventually get someone to replace all my art assets with better ones if I can afford it.

>>15326298

Depends on the level of detail I guess.

>>15326085

I have to assume you aren't interested or can't afford the average I posted but it would be nice to have a confirmation for that, silence is not the best way to convey a polite way of declining the offer, thanks.

>>15326298

It's different to want art that's above programmer sprites and art that's Metal Slug tier. It also depends a lot on how much art you need and how flexible do you want the artist to be (e.g. "let's work together" vs "make a thing for me I'll give you $xxx for it").

How do I mix a tactical shooter with a mediaeval dungeon crawler with total war with her set Radio future?

>>15326696

By axing three of those themes.

An introduction to an interesting series on audio production. Wonder when the first episode airs

>>15325378

Tore my Godot project apart and renamed a ton of stuff. Writing some code today, nothing specific. Not much to report, new video soonish.

>>15327782

Got a portfolio m8?

>>15326042

Give up for a while, change the hobby, I had the same problem, I'm only interested in making single mechanics, I've never finished a single game, I switched to learning music and it's more manageable, you can get results faster, it's still related to gamedev since I always wanted to 1MA a game, when you get something done you'll be more motivated to finish other things, so take a breathing and comeback to deving when you're feeling better about it.

>>15329669

original might be best for financial reasons. You can always add a vivian skin as unlockable.

>>15329827

I am building it, otherwise I can link you to GDI's channel on YT, my very early works are there.

>>15329897

>financial reasons

I am making a game for free anon, i might not even share it in a year or so.

>>15329669

From my experience developing things for other people will never bring lasting motivation, make the thing you love instead.

>>15329219

Because your code is optimized for your own case.

That is, Unity devs have to account for every use case, figure out the most common one, build around that, but still make it flexible and accessible for others to use

>>15330745

With C#/XNA, they had a few different awful ways of handling input. You'd have a state object for mouse, keyboard, and joysticks (and xbox controller gamepads that not even MS supported) but they all had different shit and even though they were all inputs, polling them was different for each kind. It was just awful

>>15330837

I'm using GLFW's input handling. It's pretty consistent and simple for mouse and keyboard, and I figured that out fairly quick, but joysticks are ridiculous. There's only a callback effect and that's for when the controller's been plugged in or out. The actual controller names are pointers, so they're a pain in the ass to use in any functions. There's also no way to check one particular input on a controller, instead it's all in an array, which is also a pointer. I just want some simple fucking booleans so I could output them and have to go on a journey to get there.

>>15330944

What I did for my case is to make a simple helper class to manage keyboard inputs … it also works for joysticks, but they're a little more difficult because they can have floats for different axes.

Basically I make a struct of however many bits in size (eg SFML defines up to 120 keyboard keys, so it's a 128-bit struct), with a number of utility methods to flip particular bits.

https://hastebin.com/ebanibaqid.cs

What are some comfy dev podcasts/streams to have on the side? I can only find faggots and tryhards. So far I only have Jonathan Blow and Solar Lune streams.

Give me feedback. Should I make this game?

>>15327108

>>15327108

>>15327108

>>15332280

Translation:

>Should I write this story and attach it to over ambitious generic description of game features?

Can C++ use namespaces or something so you can refer to properties of structs without naming the struct? C can't and it's getting on my nerves.

For example instead of having to type "something.position.x", I want to change the namespace of the position in the 'something' struct, so that I can just type "something.x". For example;

struct Vec2f {
    float x;
    float y;
};
struct Somestruct {
    using Vec2f position;
    int somethingElse;
};
Somestruct something;
something.x = 10; // valid

>>15332700

The feature you are referring to is called "anonymous struct", and is, ironically enough, available in C11, but not in any C++ variant. Unfortunately, it comes with the limitation that you can't use an existing struct declaration, the anonymous struct must define a new one.

However, you can use inheritance instead to get similar behavior, which is probably what you want.

struct Vec2f {
    float x;
    float y;
};
struct Somestruct : public Vec2f {
    int somethingElse;
};
Somestruct something;
something.x = 10; // valid

>>15332764

>>15332700

For clarification, this is an anonymous struct:

struct Somestruct {
    struct { // We can't specified a predetermined type, has to be a new one. (This is because of struct padding rules.)
        float x;
        float y;
    }; // No name, this is what makes it anonymous.
    int somethingElse;
};
struct Somestruct something;
something.x = 10; // valid

>>15332825

I'd like to know that as well

>>15332837

I did a quick check, and this feature makes great sense when dealing with unions. For example, if you had a CPU register type where you need to refer to each half of a 16-bit register as an 8-bit one, you could just have a.h a.l and a.x rather than say a.r.h a.r.l and a.x.

>>15332825

In that particular case, none. However, consider this case:

union Vec2f {
    struct {
        float x;
        float y;
    }
    float data[2];
};

When not used in combination with unions, anonymous structs have no functionality.

You could theoretically use them to try and improve readability, but if I'm going to use structs to group some struct members together then I think it's better to name the struct whilst I'm at it.

>>15332700

It's not quite what you want, but you can just add float pointers to Somestruct that point to the address of position.x and position.y, then access them with *something.x and *something.y

struct Vec2f{
	float x;
	float y;
};

struct Somestruct {
	using Vec2f position;
	float *x = &(position.x);
	float *y = &(position.y);
	int somethingElse;
};

void test()
{
	Vec2f somePos;
	somePos.x = 10;
	somePos.y = 20;

	Somestruct something;
	something.position = somePos; //now something.position.x = 10, something.position.y = 20
	*something.x = 30; //now something.position.x = 30
	*something.y = 40; //now something.position.y = 40
}

>>15332912

Sounds like a bad idea if you care about performance.

>>15332927

If you care about performance, you'd just fucking write something.position.x

If you're too lazy to write those extra characters you're too lazy to optimize

>>15332931

What has namespace related things got to do with performance of the program? It's purely for programmer convenience and code organization.

>>15332764

>>15332781

I use anonymous enums in a couple places, but yeah what I want is to use an unified struct across many structs. Kinda defeats the point if I have to change the values in every parent struct anyway.

>you can use inheritance instead to get similar behavior

I'll look into it, though it sounds a little uncomfortable. I basically just want a nicer way for referencing a property of the struct, not some fancy system that does more than I want.

>>15332912

I don't really want to add anything to the struct itself, but that's an interesting trick.

Anybody know anything about Unity Web? Is that discontinued now? Does it work on mobile?

Can you just code something in C# and compile it to work for unity web?

>>15333273

>I'll look into it, though it sounds a little uncomfortable. I basically just want a nicer way for referencing a property of the struct, not some fancy system that does more than I want.

The thing is, when you think about what exactly you want, and how C++ handles it, it's pretty much the way of going about it. There's some minor interfacing differences, but ultimately those would be superficial in the setup you want.

The issue is that if you want a struct's variables to be part of the struct you put it in, then you can only do that once. This is because otherwise you would have multiple variables of the same name. (E.g. you have a vec3 position and a vec3 normal. If both were "integrated" into your struct, then .x could reference to both the position.x and normal.x, which is obviously a no-go. If you want to be cheeky and say "well I want normal x/y to be named nx/ny", then you might as well call the normal member "n" and access it with .n.x/.n.y. Same behavior, better clarity.

On the other side you also have to consider the advantages C++'s inheritance offers. If your object is also a position, then you can simple toss the object into any function that takes a position. So if you create a generic function that returns the distance between two positions, you can toss two entities in there that inherited a position and it'll just werk.

I understand that you're trying to avoid C++'s features and want to take a C-like approach. But writing C-like C++ code is considered bad practice, and you'll have a harder time solving issues. Either because any documentation you find is C++-like, or because the people you ask to help aren't used to C-like C++ code. If you aren't even using classes, why use C++?

>>15333518

You mean "how the fuck do you use git".

Anyway a tl;dr that skips over many details is as follows:

>clone: creates a copy of the specified respository. (Normally you do this to copy the contents of a remote repository on a remote server to your own now-created local repository on a local device.)

>commit: create a new version with the changes you've made to your local repository (You can specify which changes you want to include, and leave a comment/description)

>push: push a version from your local repository to a remote repository

>pull: pull versions from a remote repository to your local respository

Do yourself a favor, install a GUI client, set up a free repository on one of the many sites, mess around with it and figure out how this stuff works.

>>15333566

>>15333518

Gotta be honest I personally tried to use Github's GUI client and a few third party ones for other gits and I still could not understand how it worked at all.

It's unnecessarily complex. Just set up a google drive and enable version control.

>>15333496

>The issue is that if you want a struct's variables to be part of the struct you put it in, then you can only do that once. This is because otherwise you would have multiple variables of the same name

Isn't that the same as having x in the parent struct and x in the position struct? It's just a name conflict, I don't see why you couldn't inherit more than 1 things if they don't use the same names.

>So if you create a generic function that returns the distance between two positions, you can toss two entities in there that inherited a position and it'll just werk.

You can already do that by sending a pointer to something.position, you don't need a class for that

>If you aren't even using classes, why use C++

I like C but I'm looking at alternatives precisely because C lacks very specific little conveniences like this. I'd also love to have better varargs for instance. I'll probably eventually move to Jai anyway but I want to do my best until then.

>>15333591

>Isn't that the same as having x in the parent struct and x in the position struct?

Yes.

>I don't see why you couldn't inherit more than 1 things if they don't use the same names.

You can. I was referring to inheriting the same object multiple times (a vector for position and a vector for normal), which doesn't work.

>You can already do that by sending a pointer to something.position, you don't need a class for that

Yeah but the whole discussion is about syntax ricing anyway. This allows you to omit the .position.

>I'd also love to have better varargs for instance.

What do you want out of varargs anyway?

>>15333734

Yes it is, I was inspired by Shadman's take on it and made a partial costume for anyone who wanted to crossdress their Genesi 3 figure in DAZ3D.

Finally got a few hours of dev in and added a weight system. Now you get encumbered if you carry too much shit. Ground breaking new features like this is what keeps my interest going.

Let's say I can make a simple clothing mesh of your choice just to build up portfolio, bring it on, give me something to work on, just don't expect me to make detailed textures or rigging, 3 people max, go.

>>15333935

I mean I don't need anything but can you make a wizard hat in the shape of a massive dong and a beard made of much smaller dicks?

>>15333952

That's quite the complex task, mostly the beard.

>>15333957

Right, you seem like an alright guy but like a lot of artists I don't think you understand how gamedev works, don't take this as a personal insult but more as an explanation as to why the way you're doing this won't work most of the time.

First of all, asking for 200$ for a single character is ludicrous, no indie dev has that kind of money to use on a single model. If I were working on my own and needed an artist, a single model would be completely useless to me because I would still need to make everything else. And since I'm not very good at modeling, it's likely that your character would not fit with the rest of the meshes I'd end up making.

And that's without considering if I need more additions to that model, working with a 3d modeler friend we've had to redo a lot of stuff, either because the first iteration didn't work, or because once we actually put it in the game it just didn't look good. If I payed you 200$ to make a model and realised that I needed a few changes down the line and you would not be able to provide them for me for free, that money would be completely wasted.

You can have all the talent in the world, but I seriously doubt a vast majority of people will agree to go along with this system you've set up in a long run, especially if you still need to make up a portfolio to show what you can do. It's just not going to work for the person paying you in the long run.

And you can see a clear example here in this very thread. Despite you offering your talents for free, nobody has asked for anything useful simply because they just don't need a single mesh.

If you're so desperate to expand your portfolio that you're offering to do things here for free here's two better options

>learn a minimum of programming and create small projects

Oskar Stalberg creates small interactive scripts to show off what he can do (http://oskarstalberg.com/game/planet/planet.html)

If you wanted to get some practice with clothing, why not create a small script that selects clothing randomly and have it visually show off the models you've created?

>find yourself someone that can program and work on a small project together

I think this is one of the best things you can do, sure you might make no money from it until the thing is released, but neither is the person you're working with. In the meantime you get to build knowledge about how models for games should be made and if the project gains some traction it'll put your name out there.

Either of these is better than waiting around for someone to ask commissions. There's thousands of people doing commissions for your same exact prices, it's crazy to think doing the exact same thing as all of them would be a success.

Sorry for the rant, but I keep seeing artists asking for commissions of hundreds of dollars from programmers that aren't making ANY money from their games and are just doing it to get their names out there and it's baffling to me that such few artists would be willing to also work for free just for the experience of working on something.

>>15325673

Those clothes would look good if you did a few things.

Physics simulation of the clothing to add wrinkles, and to give a lifelike appearance. This is very easy to do in marvelous designer.

Detailed fiber textures. This is pretty easy in zbrush or substance painter if you get a few free brushes, and after you can bake out a normal map.

Add different materials per piece of clothing. For example, you can add cloth material to to the clothing in substance painter, and add a metal shader to the metal pins. The materials are exportable to the usual engines/formats.

Look at clothing like that in nier:automata for what you should be shooting for.

Then it wouldn't look so stale, lifeless, and boring.

Also here's a reality check, do you seriously think work of THAT quality is worth $200? No, no it's not. I wouldn't even download it if it was free w/a CC license.

>>15334107

Yup.

>>15334297

>Also here's a reality check, do you seriously think work of THAT quality is worth $200?

NOPE, those were just cheap samples that would never go that high, my 200 price tag would be for way better stuff, I think most of you are stretching it, those are the things I can do for just a few bucks with not much detail added onto them.

>>15334107

>create small projects

There's no reason for an artist to work on projects for the sake of portfolio. You just need to have good looking art/models/whateveryouwanttoactuallydo in it.

For that matter that's why I am asking for some simple requests to build up some credibility, in this case I propose the following, make a request for clothing for a standard character in MakeHuman for example, I can pretty much add it to MH and it will rig it with little to no problem.

>>15334326

Yeah good luck standing out from the crowd with that.

There's no reason not to diversify what you know and good presentations (especially if interactive) can make all the difference.

>>15334313

The point is, your current posted portfolio doesn't lend weight to the claim, "my 200 price tag would be for way better stuff".

Yeah, better stuff, but what evidence do we have that you can do that?

You've also stated you don't want to be asked for detailed texture work. Like, what? This is basic shit.

You don't have your workflow sorted out, and don't have enough experience to say you want to start doing commissions imo.

Basically, the way you presented yourself didn't inspire confidence.

Also, pro-tip, don't show low-quality work on your portfolio, unless you show a progression from start -> finish to demonstrate your workflow w/the finished being high quality, otherwise low-q work is a no-no.

>>15334326

Well if you want commissions, and want to entice a wide range of potential clients you need to be slightly different as to catch their eye.

You can also just be really skilled in their wanted artstyle, with what they would deem reasonable prices, but it's better to pull in as large a crowd as possible.

This is especially true if you're an independent contractor.

>>15334339

You'll have a better chance to be hired as an artist if you spend your time making kickass art pieces instead of some half assed dangly doodad for your portfolio, trust me.

>>15334377

My last failed commission was a waste of time thanks to not asking for a % upfront, but whatever, that's why I was trying to grant small requests for free just recently, and even then I am met with this attitude, right now I just don't have the ideas.

>>15334394

what is the problem you have with the tail?

>>15334400

>and even then I am met with this attitude

lmao

Alright then

>>15333518

Save yourself the trouble and use gitgud. It basically drag and drop.

>>15334400

>even then I am met with this attitude

If you don't want honesty, go back to reddit.

We're giving u advice, and not just shitting on you - nothing is free.

>right now I just don't have the ideas

You lack creativity, and yet you want to be an artist? What?

If you really have no ideas, and want to practice. Then recreate some models from games you like, or anime. Find a scene you like in an anime/game, and re-create that scene. That is, in terms of models, textures/materials, and the ambiance.

>>15334409

>>15334440

I mean, disregard the pricetag right now, I was trying to prove myself on a basic level and people are still acting as if I was trying to force people onto paying me to do sub-par works, I cannot just work on a project right now for free, I can make small stuff maybe.

>>15334446

If you ask me you should consider using a sea mammal's skeleton for reference on the lower part, otherwise that looks like a sideways fish soldered to half a human skeleton, you need to blend those together.

I'm working on a turn-based game and using callback functions to handle the effects of status effects. For things that increase or decrease player stats this works fine, but I'm running into issues with health regen.

int HealthRegen = 0;

		if (GetHealthRegenEvent != null) {
			HealthRegen += GetHealthRegenEvent (this);
		}

In testing, I've got multiple health regen effects applied to the character. Each returns an int for how much health is supposed to be restored, and I know they're being called because they're logging a debug line I put in, but it's only setting the HealthRegen variable to the value of the last one that gets called rather than the total amount regenerated by all effects combined. What am I doing wrong?

>>15334394

Is that a mermaid?

Add tits

>>15334470

>Either just slap on what you have outlined, or just have it end in a tip and say "it's magic I ain't gotta explain shit"

maybe use some kind of ghostly glow to indicate the fluke?

>>15334500

no, add just the seashell bra

>>15334504

Then they should move according to the physics of the tits they're supposed to cover.

>>15334508

It depends if it's supposed to be a mammal tail or fish tail.

>>15334509

I can try something in Maya, got the OBJ file?

>>15334537

https://drive.google.com/open?id=10xVioa1zsF2wN1_liiBy-zTeZX9lh3BE

>>15334500

no, its a merman

>>15334612

Yea, and that was the best auto projection

I can use follow active quads to fix the worst branches, but it is a lot of manual labor

If you can't find a quick automatic solution don't bother, it will take too much of your time and if i have to make changes in the tree your work will be wasted

>>15334539

>>15334646

OK, for starters, looking at your model I think all branches could just be sharing the same UV map most of the time, did you make it by duplicating branches? doesn't seem like it does it the same way Maya would do it, in the future I would like to give you a tip, sculpt the first branch AND unwrap it, then just duplicate the object and merge them as needed, that way all branches will share the same projection regardless of shaping, to change shapes I would recommend using non-linear bend deformers instead.

I can't fully replicate your tree but I would have a different method of doing something similar.

>>15334663

i made the tree using curves and converted them into meshes after the tree shape was done, i don't think i can set the UV when they are curves

>>15334671

That's because you did all the shapes first, different methods, also, your main tree has almost no edge loops.

>>15334587

do male dolphins have penis bones?

>>15334694

Yeah I used different decimate methods on these because i made them first then merged, don't lose your time anon, i appreciate the attempt, but i will probably remake some of it

>>15334742

I don't know much about how Blender does these things but it shouldn't be much trouble when you merge objects like these, it would be an interesting experiment to try replicating something that looks similar.

Eh, sorry anon, I can't rescue your tree's UV maps from this mess, but you gave me an idea at least.

>>15334808

Since the tree was made using curves and then applying cylinders to it, It might be better making it all via script

On unity, since I know shit about python

Very low priority, however

>>15334835

I have a more "primitve" idea, I will make a straight branch mesh, then I will rig it like a tentacle, you can then add a "tentacle" control rig on blender (I assume it is possible, I can do it on Maya) to allow it to be controlled by an IK spline handle, then you can just twist it and move it as needed, duplicate it and deform it a little depending on your needs, then freeze all transforms on said thing, delete the joints, rinse and repeat until you got a bunch of branches, then merge all objects in the end, just one thing, do a merge but don't use booleans, I noticed a few cuts under your model.

>>15334867

Probably would work, but my bro is now thinking about making a script to generate the tree from a skeleton with small branches and leaf generation, but won't be something we will do for this next demoday

>>15334694

>bothering to do UVs on a clusterfuck mess that is a bundle of vines

Skip that shit my dude, focus on using materials that ignores UVs entirely.

>>15334902

Good point, I'm pretty sure these are the same guys who showed a webm of two green objects intersecting and their textures seamlessly merging. The same should work for this tree, shouldn't it?

>>15334943

Yes, I made a floating island recently and I Just ignored the UVs all entirely since it was sculpted and I didn't bother to waste my time doing it either.

the guy I'm working with who's more on the coding side decided to use a triplanar shader on the mesh which also includes noise map and some normalmaps for each randomized texture, looks about right.

>>15334339

recruiters (for modelling - environment and character models) want still renders that show off the model and your lighting and composition skills.

Animated shorts, interactive shit, all that shit goes in the trash and your portfolio with it. They don't have time for that, they don't care, they want to see your modelling, texturing, and materials work and that's it. As quickly as possible. Next please.

>>15334400

This place is about 70% underage gremlins asking the most basic shit about Unity and starting programming language battles. 20% amateur-tier trolls who are bored with /tech/. 10% legitimately crazy people.

Go do some mods for games. Honestly though, what you showed wasn't all that inspiring.

>>15334980

in whatever case, this has to do mostly with problem solving, when you have a shitload of mesh that is going to basically repeat itself, procedural or triplanar type might solve some pretty obvious issues, that said what I would usually do is I would focus on parts of the mesh that should have those details and actually apply uvs on them but ignore the other parts entirely.

>>15335037

Did you change the noise tiling? that could cause issues and expose seams, no idea why it does that.

>>15335138

Are the textures you're using seamless?

specifically the normalmap and heightmap.

>>15335037

>117k verts and 230k tris

>tri planar shader

Wew.

>>15335330

Blender usually has a different axis setup, nothing fancy, you can try using that branch for other things, obviously, not expecting you to remake the tree.

>>15335037

>Few potential solutions for you

You could try just hiding the seams.

Apply some vines, moss, and other little meshes that are properly unwrapped + textured.

There is a solution to texturing for this scenario but it may be a bit too performance heavy.

You can do solid texturing, which you basically have a 3D volume of noise which you apply in a 2D projection similar to planar projection, but the actual projection data originates from a 3D data set; so it's seamless.

You also may want to look at the blend strength for your tri-planar shader, and ease up on that parameter as to mitigate the sharp seams from blending each of the textures per projection direction (X/Y/Z dirs -> tri-planar).

From experience, I've got a triplanar shader that can do spheres and cylinders perfectly w/o seams; in addition to having AO parallax effects (which need no seams to look realistic).

>>15335281

>trying to bake CHAOS

HNNNNG

>>15334902

How can materials "bypass" UV mapping? Everything that gets pushed to the card needs vertex information. Or is it just a thing that automagically works?

t. non artfag

>>15336406

Less of an artfag thing and more of a graphics programmer thing.

You can do that by using world coordinates. Basically you pass the world coordinate from the vertex shader to the frag shader which is interpolated to the exact world coor position (pixel position) to shade that pixel anyway u want.

After that u can apply some texture projection to that pixel or other algorithmic magic to get cool effects.

Pretty much the same for UVs but in a UV coordinate context.

>>15335530

I wouldn't worry about it too much, as it seems easy to fix with a few manual touchups. Interesting goal for the tree. It'll look amazing when u guys get there.

>>15335769

Working on rotation tomorrow, working soon.

>tfw working on editor window shit is suffering

>need a GUI for assembling AI behaviors

>coding the AI framework was leagues easier than this shit

Having robust AI is nice, but creating the tools to make it usable is a pia.

I've been doing a side-project to make nice inspectors for all my custom components that aren't usually supported by unity OOTB to get some practice.

>>15337018

when I worked in C-Battlefront 2 mods I recall the AI being more or less workable by using hint nodes in the level editor and the rest was LUA variables using classes and class parents.

>>15336989

In Godot, how do I apply a texture to one side of a cube mesh? Someone mentioned I had to get the UVs of the texture and it is to my understanding that a mesh has to be put into Blender to set this up. I need to put test textures on some nodes so I can make a 3D tileset.

Guy who thought my game was a Fallout Tactics clone, I hope you're still out there somewhere.

>>15337057

just one side but said side is the one showing? you can use the method I listed without sewing the edges, that way all faces will share the same texture.

>>15336778

>Autumn Willow

Now that's a name I haven't seen in a while. You back at it?

>>15338121

>You can still headpat/kiss/etc while she is picking stuff up

>bend over just behind the duck

>she goes to pick it up

>kisses you herself

I am an evil mastermind.

>>15338463

I've been dicking around with engine trash the whole time and making little progress,as usual.

>>15339264

I like #4 the most. It looks the most barrier like

>>15339292

>>15339264

>>15339321

Thought I was looking at a high-score screen before reading your post. Looks good but I'm wondering if they're too computationally expensive for a single element.

>>15337033

Not sure what AI battlefront 2 uses, and can't really find anything online; so hard to compare it to my scenario.

For my case a GUI is so convenient it's difficult to pass up… as without it, the AI would be really difficult to organize.

I should also mention the AI algorithm was designed to have a visual representation (GUI) for all the data/logic/contextual behavior and so on.

Imo trading initial complexity for eventual simplicity (w/the perquisite to simplicity being to craft tools) is a fair trade.

On the note of designing the GUI, I've recently passed the initial learning curve on the editor window stuff; smooth sailing from here on.

>>15339555

>>15339478

all of them are equall to compute and only so ever slightly slower than drawing an animation, it's a very cheap effect. I'll make the resolution larger tomorrow so I can record some video. At full speed the colors blend buetifuly

>>15339660

https://sites.google.com/site/swbf2modtoolsdocumentation/getting-started

ODF sections and all other LUA documentation can give you an idea.

>>15340540

not sure what language it is, but if it's c++ you're making a terrible mistake by using all these new (and no delete btw.)

I'd just always draw 4 squares and only change their color depending on the input. Some pseudo code follows (and I'm on the phone so brace yourself)

void drawShit (bool up bool down bool left bool right )

{

drawBox (x, y, up==1 : green : red);

//repeat 3 times for the other boxes

}

>>15340655

for some reason I saw these squares as WASD. I see you have more, but same idea applies, although try avoid using more than 4 arguments for your functions, make it a struct or use one int and do some bit masking (seems like you already are doing that anyway since the input is represented by some hex number)

>>15340655

It's C#, which doesn't have delete/a way to force memory release. At best, I can release the references to those objects, but it's all held in the local function scope, and I force a GC collection every frame anyways.

The debug shit is just temporary anyways - the actual code that handles stuff is much more conservative and proper.

>>15340859

What approach are you using to handle your timestep? It's one of my things I need to tack down at some point.

Since you're tracking average frame time, I assume you're maintaining an average/rolling sum, say 300 values large. Is that enough to track if the game is running fast/slow? I keep trying to break shit into smaller components when it doesn't need it

When making ledge vaults and other 'animated actions', should I lock the character into it, or let the player cancel?

One of the issues of letting the player cancel is because he might cancel when the character collision is inside something (as character capsule and movement will not match during these).

If I don't let the player cancel, how do I handle the character being hit during these special actions?

>>15340872

you count your frames until one second passes. You measure the time between frames and divide it by the time scale (so if you measure in milliseconds you divide by 1000) and use that value in all your fps depended code like speed of your character or whatever (velocity * framwTime)

>>15340946

A lot of my game is tick-based, because it is a puzzle game and so it cannot have any random elements or floating-point error to it. I don't use a variable timestep for that reason.

Or rather, I advance a tick every 16.6ms that passes, but I haven't tried counting ticks yet. I'll see if that fixes the small issue I set aside

>>15340954

fps locking is lazy cop out and there's always a way to make any game, any mechanic fps independent. Give me one example where this "floating point error" would break your mechanic. I'm sure it can be fixed.

>>15340954

also your approach will not work if your game wipl be too slow for 60 fps (even if it's pixel shit or whatever, don't assume you will always have 60fps, computer can lag sometimes or whatever)

It is more work than doing fps consistent code, but it's not impossible, my 2d game doesn't have any bugs (going through walls etc.) even at 1fps

same anon different ip

>>15340992

>>15340993

The default speed for my game is moving from one tile to another is 30 ticks, or half a second. Whether that takes exactly half a second, or lags and takes two seconds doesn't matter - internally it's consistent as long as ticks are followed

>>15340993

sure, so my approach would not break this, since the frametime is global for the whole frame and all objects would move the same amount and arrive at the destination at the same frame. Anyway, you can keep your 16.6 ticks, but update your game every frame and summarize all the movements to a float. Then at the 16.6 Mark round that float into a move or not a move. Again, your objects won't go out of sync, since their value will grow by the same amount.

>>15341002

this is exactly what I was warning you about. Think about it, quake and half life can run on all kinds of frame rates but still keep a constant pace (i.e the gsme doesn't slow down on low fps but skips frames) and yet you can record a demo and replay it just fine. You aren't doing it right.

>>15341039

But I do have a replay system roughed out and it does work fine, independently of FPS.

While you're solving a puzzle, it accepts player input and sticks a timestamp on it. When you're replaying a puzzle, it resets to the initial map state, and then at the designated timesteps, it sets the input based on what the recording says. To the gamestate, playing and recording are exactly the same thing. If the player is 14 ticks into its movement, it's 46% of the way through its movement, whether it takes 0.5s or two minutes for the ticks to occur

>>15340568

Might be GOAP as there's references to goals, heuristic variables, and setting blackboard parameters.

>>15340908

I'd look at assassin's creed, some of the best free-running out there (best aspect of the series).

For assassin's creed ledge hops/vaults are quite fast animations, and cannot be canceled by user input, or damage. However, damage can introduce something like stumbling/faltering a little and losing a bit of speed after landing.

Certain actions where it's slower, such as ledge hanging, and a ledge pull up (from ledge hanging) can be canceled by damage or user input.

For massively damaging hits that add an offset velocity to the character just enable a ragdoll component.

Think of it like this, fast animations like a ledge vault cannot be canceled because the forward velocity of the character carries them through the motion despite any hiccups (unless there's an explosive force or a great sword hitting them).

You can have a "stumble" landing activated if damage passes a threshold, and a ragdoll if it goes beyond the massive damage threshold (between entering and exiting the state of the animation).

>>15340823

High quality work anon, nice. I especially like the bioluminescence touch.

>>15341160

From what I can remember while working on it the level editor had these things called hint nodes which basically worked alongside path finding and collisions to tell the AI where they can take cover, where they are supposed to snipe, if they would kneel, where to face, in which direction to jump, etc.

>>15341171

Sounds like they got inspiration from FEAR (game that made GOAP popular), as they used a similar tactic, but their level design really strengthened what is actually not very complex AI.

>>15341178

Precisely, if you made a flat map without paths and nodes the AI would just go straight ahead to the objectives, they would get stuck trying to get through walls and props, also, the AI could only use a 2D space for pathfinding, the objectives would be designated by variables in the game mode ODF and so the AI would know when they have to capture the flag or capture a zone, they would engage enemy units within the range of their weapons which had their own class parent ODF as well, as for capturing and the like you had to designate a cylinder or cube in the level editor, usually in the same area you would put the 3D object representing the objective, in general the game had tons of options delegated on different files which themselves had calling for class parents.

>>15341160

>I'd look at assassin's creed

I guess it's time to play AS1

Besides the 'skip gameplay grappling hook', nothing was added to the parkour in the series ever since the first one, right?

Does this look okay as a hash/checksum functions? "Anonymous" would yield a value of 40030 as a result.

public ushort HashFunction(byte[] values)
{
	int result = 0;
	{
		for(int i = 0; i < values.Length; i++)
		{
			result ^= values[i] * 17;
			result <<= 1;
			result %= 65536;
		}
	}
	return result;
}

What I wanted to do was let each player profile have a name associated with it. This name would be used to generate a hashed value (above) which acts as a fingerprint that is attached to replays. The replay files would have a checksum, the fingerprint, and then the replay data. This way, you couldn't easily cheat by downloading another player's replays - you'd have to complete the levels yourself, but you can still re-use them between devices if the profile names match. Or just change the fingerprint in the replay data and recalculate the checksum yourself

>>15341289

AC1 has the best free-running feel, and is my favorite of the series; definitely worth a play through.

>nothing was added to the parkour in the series ever since the first one

Nothing major.

AC3 added some tree-free running, and better IK/transitional effects for running but that's about it afaik.

Though, I haven't played any AC after 2, and the knowledge of AC3 is from developer presentations I watched online (GDC I think?).

>>15341211

Neat

Sounds like a bit of a pain in terms of user friendliness for assembling AI behavior, but man can you get some great behavior from GOAP if you can wrangle the behaviors, and use the tricks from FEAR.

>>15341332

Looks fine to me, you're multiplying by a prime, and doing some standard hashing logical operators.

Can't say for sure without plugging in a bunch of values, and testing for collisions though.

If you're doing it simply for a checksum (offline calc) u can just use sha256, md5, or something with a long for more variations than 65k.

.>>15341465

39.99 monkey bucks for a 10 years old game

gonna pirate it instead

>>15341465

Well, not sure about user friendliness considering is basically a split on how to do some complex modding, if anything it was very easy to setup when you just had to tweak some values here and there on your custom factions and weapons.

>>15341478

yup, better to pirate, as I wouldn't want to support ubisoft either.

Too bad the original ubisoft montreal went down the drain, they really had something w/AC1, but sadly they didn't innovate anymore after the initial bout of creativity.

>>15341489

Ah, my mistake, I meant in the context of the gamedevs not modders.

Modding is in the context of the developers already assembling everything, and giving you a bunch of behaviors which you just tweak a few values for.

I mean in the context of crafting all those behaviors and such from the gamedev perspective, and it being a bit of a pain.

>>15341504

Well, that is always the hard part, once you figure it out however is just a matter of using text.

>>15341520

>just a matter of using text

Sort of true, depending on your approach to

>figure it out

For perspective, their idea of making it easier was to make a LUA front-end for assembling said behaviors, and the original/base code would be in the C++ backend where all that logic is plugged into.

Which puts it into perspective of it being a PIA without some kind of wrapper to make functionality more accessible.

>>15341841

The first half is fine, but the second half is a bit too distorted. It cracks in the middle of itself unnaturally, almost like a robot would.

>>15341841

You can look at some of the videos by Marshall McGee, a professional sound designer who goes through his process of making sounds. There isn't anything for creature sounds but watching his workflow will probably help.

>>15342195

to be noted its not running at the intended 48 fps so the timing is funk/ blender doesn't like showing more than 60 and i like animating on 12's so 48fps is the best time for me to work with

>>15341465

AC1 is the only asscreed worth playing, and yes: the movement is what the game brings to the table (historic setting too).

>>15342195

Is it smart to have so many bones, anon?

>>15342195

The odd bending on the upper/lower legs, when they're transitioning from step to stretch, looks really weird tbh.

The frame when the leg is raised looks good, but otherwise it looks like the legs are supported by jelly not bones.

I just use usual bones and IK + constraints and it's offers plenty of control.

You may want to look at dedicated animation authoring software. Akeytsu is what I like.

>>15342415

bendy bones, its one bone with 11 segments and a curve drawn between them adding in the angle of the lower leg and the well an extra bone added in the upper leg that has no attached mesh

>>15342427

kind of intentional wanted to make the upper leg a bit more "meaty" but the weight painting is unfixable with the bendy bones auto weight painting for the upper leg

Shodanon, are you still around making Beelzebox? I've been waiting months to hear an update lad.

>>15342516

Lets try this again.

>>>/agdg/32100

>>15342548

Get dropped off at daycare, kid.

>>15342574

OP complete, go redeem your internet hitman points.

>>15342477

I get that you don't animate them all manually, my question is whether having so many bones affect performance. Do they? Or is it just a bit more memory used? Or is the impact insignificant?

>>15342673

I once downloaded a gamedev ebook torrent with 43.7 GB of material. I was so overwhelmed that more than two years later I haven't even began to sort through them.

>>15343059

GURPS or Avernum/Geneforge, or Elona or TES

>>15343073

I really want to do GURPS but I doubt I could get a licensing agreement that would let me release the game as FOSS. I want to sell the game, so I can't release it under the radar or anything.

I'll check out the others.

>>15343189

It could be as simple as a point buy system and you get points for doing certain thinga

>>15342195

You should keep those cubes visible and make a /clang/ game.

>UInode->fill.color.gradient.gradient.simple.color_start.rgb8

I feel like I went wrong somewhere

>>15342032

>you don't need expensive equipment, look I just used my phone mic!

Conveniently omitted all the software he's using for the compositing and effects.

>>15344443

Pirate it.

>>15344449

>adding status effects

>realise encumberance does not make sense as a status effect

huh I guess I never thought about it before

>>15344468

didn't mean to reply

>>15344539

Broken video.

>>15344548

seems to play fine on my pc.

>>15344548

>>15344561

vid is broken here too

>>15344565

>>15344548

https://twitter.com/ca3games/status/1032428937094279168

>>15344573

Doesn't look bad. I don't particularly understand the rules, though, is it just a simple number of blocks in one cluster to remove them? I thought it'd be a certain amount in one row or column, and would remove all the adjacent blocks too, but that doesn't seem to be the case.

Seems like you're introducing more colors to escalate difficulty. If I can make a suggestion, you should add the ability to rotate blocks. It might make the initial start really easy, but when more colors are introduced, it could add quite a bit of strategy to it. Might have to adapt the current game for that to work, though.

>>15344622

It's basically a single player tic tac toe with tetris pieces.

>>15344637

Well, not sure about the original anon, but remapping them when certain segments don't share the edgeloop make it harder, I ended up making my own version and for the most part it seems to work.

>>15344691

>Dancing wizards in geometric spaceships that exhaust rainbows

Sure, why not?

Anyone tried Nim before? I stumbled on to it the other day and it seems like a pretty interesting language.

>>15343879 (CONT)

I have no fucking clue what i'm doing but I got this from all the synthesizer.

>>15344870

JUST FYI, I ear raped one guy and I didn't know it was loud, so yeah, warning, it might be loud for some.

>>15336989

https://www.dropbox.com/s/pogz2xkbyo3qusd/ANUBITEK_game.exe?dl=0

If anyone wants to check out how the movement and rotation feels, feel free to download. Not much to this, but I guess this could be considered a demo.

oh fuck I finally made something another person can actually play

>>15344691

Beautiful.

>>15344691

dude, your third project I want to see completed.

I love that 3D trick you're using

what's a proper nice way to increase dificulty (level up)?

>>15345371

Depends on the game.

>>15345377

a tetris clone

>>15345382

Well, for those games you usually increase the speed, if you are planning on a vs AI mode allow it to plan bigger combos for a comeback.

>>15345390

yeah, but how do I determine when to increase the speed?

>>15345394

Tetris usually increases the level upon deleting a set amount of lines each time.

>>15345371

>>15345382

>>15345422

It depends on the long term gameplay of the Tetris clone in question. How do you generate the upcoming blocks? You could generate them in batches of, let's say, 10, and guarantee a certain amount of each type of block appears within each batch, and then feed them to the player one at a time. To a player, it would look like you're just feeding them blocks normally and in a well-balanced fashion. With every couple of sets/score targets/level changes, you could generate each set with slightly less convenient blocks. This would let you throw curve balls at a player without just letting RNG ruining the entire run whenever it feels like it.

Some versus Tetris games usually add dead blocks to your opponent's area to reward you for doing well. You know, where the game drops a row of blocks that don't contribute to actually clearing your row, but can be eliminated by clearing blocks adjacent to them. It might be weird in a single player game, though, but it might be worth experimenting with in some way. Maybe deploy dead tiles that, if not destroyed within a specific amount of time, slowly begin to expand, or something.

>>15345413

you don't even need to implement the line breaking feature

https://twitter.com/ca3games/status/1032760193757728768

>>15345443

>It might be weird in a single player game, though, but it might be worth experimenting with in some way

Instead of generating garbage blocks, it generates some kind of points (equal to the number of garbage blocks you would have created). Then, you can spend points on certain powerups, like bombs, remove entire row/column/color, etc

>>15339321

Cool effects. I like the 246,11 version best.

>>15340823

great stuff. wish I could model this well. love the 5th pink glowing version.

>>15346080

But will this work for non square uneven surfaces?

>>15346980

>uneven surfaces

yes

>non square

sorta

curved ledges and whatnot need to be manually added, besides that, it will work normally.

Might be possible to detect curves in a plane, but at this point i'd rather manually add them instead

>>15339536

its not mine

i got it from https://www.textures.com/

You can get the normal and height map there too

>>15347018

You may want to model your testing environment to reflevt your game. Your town doesnt look anything like your blocky orange map nor does it share block shapes very much.

>>15345413

i think it would be neat if there was a solution to a sequence of blocks Then one could call it the dark souls of puzzle games cant you see a club of people trying to figure out how the set pieces fit together and then trying to find someone to pull it off?

Anyone got a link for a decent youtube channel with tutorials for 3D modelling? There was one posted a little while back, it had some dude who mainly modeled anime-esque stuff but he actually structured his tutorials and they weren't 45 minutes of "and um then we" but I can't remember his channel for the life of me.

>>15348651

Depends on what you want to do.

>>15348675

Get a basic enough understanding of modeling so I can mess around with UE/blueprints.

>>15348703

Well, it still depends a little on your approach, a basic model that is not UE's dummy would suffice, something with pre-made animations maybe? otherwise you need to learn on the go, people recommend Blender but it heavily relies on keyboard shortcuts.

1. Is there any benefit to building specifically for a Linux platform?

2. How much trouble will I get in if I share paywalled studies/papers about 3D graphics?

3. Has anyone found any studies hidden behind a paywall they really wanted? I've recently been downloading whatever I can while on campus since uni students have access to a bunch of different studies aggregators

>>15348836

>Is there any benefit to building specifically for a Linux platform

You get to lift your nose up to people who make money with dev even though you make none.

>>15348790

That's sort of what I was hoping for in a tutorial. Someone who talks through the process without a break so their tuts can act like drills for the keyboard shortcuts.

>>15348836

>1. Is there any benefit to building specifically for a Linux platform?

No. If you even care about Linux, just try to be compatible with Valve's new fork of Wine. No one who values their time (has money) uses desktop, Pottering-Poz Linux.

>2. How much trouble will I get in if I share paywalled studies/papers about 3D graphics?

use volafile?

>>15348836

>2. How much trouble will I get in if I share paywalled studies/papers about 3D graphics?

Send them to https://scihub.org/ if they don't have them yet.

>>15349665

can he unhinge his jaws, because it looks like he can't bite shit with such long teeth.

>>15349689

he stabs and thrashes, no need to bite

>>15349787

to stab with

>>15349808

stabbing with your mouth is more commonly known as biting

>>15349808

I'm pretty sure what the one anon is trying to say is the skeleton doesn't look like he can open his jaw wide enough to bite anything.

>>15349205

In Blender's case it will be full of people doing things without even mentioning tutorials, sadly it doesn't have that many visual aids when going to menus, I'd rather pirate Maya and check those tutorials or even the online documentation.

>>15349863

>>15349880

it doesnt need to bite, the teeth bend outwards, it can jab at its target, stab them with its teeth then thrash around

>>15349992

So it kisses it's enemies to death?

>>15349924

The trick is to use the [space bar] search menu (just need to remember part of name), and to use the more advanced radial pie menu addons.

It just takes some practice to get the hotkeys.

>>15349205

Any blender tutorial with screencast keys is rapid firing shortcuts at you.

Just find one that's at your level, and work up from there.

>>15349479

>>15348836

You can just upload them here:

http://gen.lib.rus.ec/

Login: genesis Password: upload

Other paper resources:

https://genofond.org/viewtopic.php?p=9000/

Also, it's a good idea to make our scientific advancements, hypothesis/theories, and discoveries available to all academics/people as to allow for open access to information.

The (((paywall shit for academic papers))) is a crime against humanity, so good on ya m8.

>>15348997

I will consider this when baking the ledges

>>15347662

baby steps

>nothing worth going for

That's exactly what death would say were he trying to protect his treasure

>>15350763

its a merman

>>15350968

Is it supposed to be a skeleton?

>>15350991

yes, its the skeleton of a merman

>>15351078

OK, my only complaint is the fish tail, that is, the fins, a skeleton true and through so far seems to lack the fins at the end.

>>15351372

I don't think that's how it works when you specifically refer to a male.

>>15351400

I know, he is an idiot.

>>15349992

the teeth should be more horizontal, then

>>15351400

What other animal besides a human is part of people kind?

>>15351450

niggers and jews

>>15351450

Has nothing to do with things other than humans, just that using mankind as opposed to womankind is oppressive. Although I think dolphins are classified as people, at least in some places.

>>15351570

GLFW only grabs data from the joystick when you call glfwGetJoystickAxes(). Call it once a frame to update your float values of the joystick positions.

>>15351597

God damn it, I thought so, just wanted to make absolute sure. Damn niggers on Stack Overflow swearing otherwise and making me waste half an evening. Seems awfully wasteful, though.

anyone here used raylib? 2.0 version seems decent. I might try it for a quick prototype just to find out how good it is with 3d rendering.

>>15352319

That's a really neat touch.

>>15350379

>baby steps

I think you are taking the wrong approach. A square even cube testing world is not good for stress testing your world which has varying ledges, normals, and faces. You are setting up your algorithm to be fundamentally wrong because it is relying too much on eveness. Id like to see it work on uneven ledges with grooves and bumps.

>>15352639

the ledge detection and ledge handling systems are separate

the ledge handling will be able to handle anything

you will just have to set the ledges manually if they are too crazy

>>15353034

I'm using SFML in C#, which is built over top of a OpenGL context, but it's basically the same environment. It should have a render target that you draw to. Draw commands usually take a Vertex[], plus a starting index and count, eg:

>RenderTarget.Draw(Vertex[], int start, int count, PrimitiveType)

Vertex data is 3 floating point values with XYZ coordinates, plus texture coordinate and color data. (SFML restricts this to 2D position which is annoying, so I can't do any kind of 3D filtering or work). You can use transformation matrices to alter the vertex data, acting as a transform for scaling, rotation, etc. This is more intuitive in a 3D context with view frustums and such, but in a 2D setting, you can just as easily think of it as a camera pointing directly at a flat plane that can zoom, with the correct transform applied.

Good luck with your SS13 engine :^)

>>15353034

Also in context to your specific questions, not exactly.

>Is it just a bunch of 2D planes with textures on them with a camera sitting further away in 3D space?

Each RenderTarget is a drawable area which contains color/pixel data (actually depending on the open gl context settings each "pixel" is assigned a bit depth ranging from 8-32 bits in size, which is important for shaders and handling massive graphic data, but it's helpful to think of it in this way). This data can be used however you want to. They don't really exist on a plane and are just kind of there in graphic memory. What you do with them is up to you.

Typically, you load your image/texture data into memory at the start of the program, then run your game/draw loop continuously. The draw loop usually clears the screen, then tells it to draw some rectangular region from the source texture, onto the target texture (the render target - changing textures is expensive because of graphics card bus overhead). After you've done all your drawing, the back buffer is swapped with the screen and the process begins again with the next frame. This technique is called double buffering (having an active/working screen, and a displayed screen). Don't get fancy with this part - 99% of consumer hardware is designed around the "clear, draw, display, flip" process

>>15353034

OpenGL draws 2D points, lines and triangles. You don't need to mess with things like projection, w clipping, depth and all that nonsense if you use it just for 2D.

struct {
    float x, y;
    float uvx, uvy;
} vertices[4] = {
    .[0] = {0.f, 0.f, 0.f, 1.f},
    .[1] = {1.f, 0.f, 1.f, 1.f},
    .[2] = {1.f, 1.f, 1.f, 0.f},
    .[3] = {0.f, 1.f, 0.f, 0.f},
}; // Load this into a vertex buffer. Draw as GL_TRIANGLE_FAN using 4 vertices.

// Calculate transformation (We use mat3 since that's all you technically need, but you can use mat4's instead.)
mat3 screenScale = {.x.x = 1.f/screenWidth, .y.y = 1.f/screenHeight, .z.z = 1.f}; // Inverse screen resolution so we can use pixel coordinates instead.
mat3 quadScale = {.x.x = width, .y.y = width, .z.z = 1.f}; // Size of the quad in pixels.
mat3 position = {.z.x = xPos, .z.y = yPos, .z.z = 1.f}; // xPos and yPos is the position of the bottom left corner of the quad in pixels. 0,0 is the center.
mat3 transform = position * screenScale * quadScale;

// calc matrix

// Vertex shader code
in vec2 position;
in vec2 uv;

out vec2 UV;

uniform mat3 transform;

void main() {
    gl_Position = vec4(vec3(position, 1.f) * transform, 1.f); // We're tossing 1.f into z here, which might mess up if the depth buffer is on.
    UV = uv;
}
// If you can't imagine what the fragment shader code looks like there is no hope for you.

>>15354834

>>15343879

Can you fuck it?

>>15354834

Saw a mudfrog the other day. Horrible creatures

>>15336989

>>15344934

I'm a dipshit, the game had broken dependencies so here is an updated version. This is ANUBITEK as it stands right now, gotta code my camera/player mover node today so wew gonna learn some collision limits today. Gonna talk shit about Godot today, I'm sure.

https://www.dropbox.com/s/7u6ynto7wek0zl7/ANUBITEK%20%5B0.0.1%5D.zip?dl=0

>>15355478

controls:

arrows - move

a - jump

x - (hold) run

~ - console

>>15355532

Interesting suggestion, I'll think about it. Looks very ethereal, the glow effect in particular looks very cinematic.

I genuinely like my flat low poly style, but I will experiment with lighting and color palettes.

I need some advice: I'm planing a top-down shooter/action rpg, sort of twin stick, with assault rifles and guns.

The problem is that enemy hitscan isn't fun and invalidates player movement for combat gameplay, but slow bullets conflict with gunpowder weaponry. I started thinking of giving enemies melee weapons and making ammo rare for melee gameplay but it's starting to get a dark souls vibe that I don't want.

How can I have gunpowder automatic firing gameplay while making player movement relevant?

>>15354992

I think he means that the tail at the end doesn't have the webbed fin part, which he is correct about

>>15355700

just telegraph the attack then you can make the bullets as fast as you can

>>15355890

I recognize no hitscan on that video, but thanks for the suggestion. I guess enemy aim has to be retardedly slow.

Do you think it would be worth having demo day only 3 times a year instead of 4? Say Nov 11, Mar 3, and Jul 7?

>>15356382

I don't particularly see why the change is necessary. If you don't have a demo worth showing, don't contribute. Three months is a decent chunk of time, and having more of them means that there's probably fewer demos per demo day, so more of a chance that yours will get played if you do post one.

>>15355986

>How do I assign AnimationState variable through script?

If it''s state based use the StateMachineBehaviour base class (assignable to the state itself). It has stuff functions like: OnStateEnter, OnStateUpdate, etc.

If you want to access the state/clip info from an external script, you need to get the related instanced class: Animator.GetCurrentAnimatorClipInfo -> clip info (length in seconds, etc, Animator.GetCurrentAnimatorStateInfo -> (state info, active, etc).

There's also other tricks to get clip length from state info, but that's a non-runtime hack (editor only, i.e. not available in build).

See here LINK: https://answers.unity.com/questions/692593/get-animation-clip-length-using-animator.html

>Is there's some other way to check how long a certain animation clip is, without assigning it by hand in the inspector

Do you mean from the assets folder? If so, definitely, but you still need to know the name. Would require loading from resources folder, loading the clip into memory, and checking length.

Or do you mean when the animation is running? Explained above.

>but that seems really unreliable

It's unreliable if you gather the data in an incorrect order (i.e. gathering state info before animator executes, animator executes, get next frame, etc), and also on the wrong timestep (intrinsically unreliable then); animators can run on a normal update or fixed update, match the animation state info to that timestep.

>>15354834

You going to polish up combat, hitboxes, reactive animations (so attacks/getting hit and such have weight), and generally polish mechanics? Combat looks frustrating and airy, not fun.

I know it has parody game elements, but still.

>>15355700

Use some type of heuristics/behaviors for aiming.

Such as overcompensation of aiming direction (from the original aiming quaternion to target quaternion use an unclamped slerp), all actions have a reaction time (makes agents "appear" smarter), IK + velocity based recoil which throws off aim (i.e. re-adjust aiming when forward of barrel is off by X amt from correct aiming direction via DOT product), and not straight up hit-scanning; i.e. be creative.

You could also just do an imitative neural network, and train it based on different skill levels; so you can plug in difficulties (i.e. trained neural networks).

Bit overkill, but would be neat. It wouldn't be performance draining either if you don't train it during gameplay, and just let it run a pre-weighted NN during play.

>>15355532

Damn that went from looking good, to looking amazing. Nice

>>15333590

>Just set up a google drive and enable version control.

This is a fucking horrible idea, ignore this poster. It would literally be better to zip the root folder of your project and email it to yourself as version control, at least that way you have discrete snapshots of your project at a given point in time.

Also if you aren't using git / other version control systems, you can't have continuous integration and all sorts of other nice stuff.

>>15360987