/v/ - Video Games

Gonna enter crunch time and do mostly coding for the next while. In the meanwhile, I wrote this song. Quite happy with it.

https://files.catbox.moe/7zv5jt.mp3

Not sure when I'll have much worth showing from here.

Just a little treat - something I've been working on

>be me

>making minecraft clone

>have chunk class that generates mesh for chunk (16x16x16 grouping of blocks, quads only drawn on blocks adjacent to air to prevent overdraw)

>want to create association between textures and quads for my chunk

What's the redpilled way to do this?

Some stuff I've considered:

1: add another byte for each quad in my vertex array which corresponds to the texture that I should use, and just make one texture in my fragment shader for every texture that i could possibly need (lol), and adding logic in the shader to select the texture based on that value. I currently like this one the most, because it seems easiest.

2: Store quads for different textures separately, and give each its own vertex buffer object; in draw call, iterate through these data structures, binding the proper texture before each is drawn. Maybe bad for draw speed, because I'm making a very parallel operation more synchronous by iterating through my data structures before putting things through the rendering pipeline?

>>15280972

>and adding logic in the shader to select the texture based on that value. I currently like this one the most, because it seems easiest.

You're not allowed to dynamically select a texture in shader code. I've made the same mistake before, adding a flat int layer value to the vertex / fragment shader inputs, and using it to index into a sampler2D[]. The end result is graphical artifacts; fragments of one triangle will sometimes end up using the texture of fragments from another triangle, which can be a different texture than you intend. Note that I'm talking about "fragments" here, individual pixels will sometimes, depending on how the GPU batches it, use the wrong texture. Not even the entire triangle.

The correct solution is to accept your lord and savior GL_TEXTURE_2D_ARRAY. It doesn't have the downsides of a texture atlas (which is texture bleed), whilst being a performant solution.

>>15280970

Is this for SL?

>>15281182

I can try to help if you want. Most of the storytelling in my game isn't given directly, though, since it's a fucking platformer.

>>15280970

What is it going to be and what are you using to make ?

>>15281033

Okay, I think I see how to use texture arrays now, but how would I get the correct depth to shader for drawing the quad? Binding the layer number as a uniform before calling draw?

I'd have to do this for every quad though, isn't that slow?

>Game status: still cornflowerblue.jpg

>Interactivity: can press F1 to enable a flag or F2 to disable a flag

>58 fps

Its gonna be rad

>>15281395

You put the texture index as a vertex attribute.

>>15280972

How would you handle transparent blocks or stuff with altered geometry (eg thin column blocks)?

What about polling adjacent blocks between chunks?

>>15281266

Well, the concept is it's an FPS-RPG, and it's basically a Deus Ex ripoff, except with other elements of gameplay and story-telling to make it feel unique. Maybe it would be like Deus Ex x STALKER, linear progression but large and explorable areas. Or maybe something along the lines of Watch_Dogs done right.

It's a futuristic sci-fi game with hacking elements to it, and shooting. And the same immersive sim elements that made Deus Ex special.

Here's bits and pieces I have to the story so far:

>Large tech companies have gained monopolies on everything. There's one standardized operation system that most software is written for, and only big-brand services on the internet are in use. Most of the general public has become unaware of the alternatives, and access to such information is generally censored on these services

>massive censorship and surveillance going on. The big tech companies sell information to the government

>media featuring degeneracy and mind-numbing nonsense are promoted by these companies, intentionally giving the false illusion that it's popular

>general population starts acting like a bunch of degenerates and as a result lose their strength and independence. Resulting in dependence on the government

>doing certain, inappropriate actions in real life will result in temporary bans from proprietary services. Not really a game mechanic, but something to tie into the story for other conflicts

>game starts off in small town America. the censorship and authority enforced by technology are blatant even in this small town

>eventually the player travels to Silicon Valley in SanFran hellhole, where the issues are even more apparent and the city looks dystopian, even to modern and well accustomed people

That's all I have so far. I need some bigger underlying issue going on that triggers all of this, though. Something beyond just, "the internet is censoring people and big businesses are selling information to governments"

Maybe the player has a cyber security role for a local branch of a bigger company, and something happened at this branch that they're trying to censor. Hacking in this game will be a well developed mechanic and will rely on problem solving. So, the first few missions will involve counter hacking and improving your company's security.

I don't know, I don't want to rip off of Deus Ex too much.

>>15280970

Whoever this guy is, he's quite accomplished in his life

https://www.ejgold.com/

>>15281700

well, this sounds like a good solution, but for some reason glTexStorage3D is giving me a very hard time.

I've been using code from the wiki trying to get it to work on my machine (https://www.khronos.org/opengl/wiki/Array_Texture) but this particular call seems to segfault every time.

However, when I was reading about texture arrays, I came across texture atlases, which also sound like a very reasonable, easy to implement solution. I think I'll try to implement that.

>>15281726

I'm not sure how I'd handle blocks iwth altered geometry yet.

Preventing adjacent quads between chunks from being drawn is something I've given a bit of htought to, but not yet implemented.

>>15281979

Have you used https://www.khronos.org/opengl/wiki/Example/Texture_Array_Creation ? Anyway, with atlas textures, remember this: there's no (reasonable) way of doing texture wrap. So if you want to collapse quads with the same texture into one and repeat the texture across it all, you can't do that. Well actually, you can, but you need some extra data. For instance, you could have per-vertex texcoords in the UV plane as normal, which describe the coordinate within the texture (not the atlas) and a second set of textures that tell you where in the atlas the texture is. Then you just need to do the coordinate wrap manually. Depending on your precision requirements, you could squash this into just 4 bytes, but it's tight.

>>15281733

The premise seems alright and current enough to spark some interest. If the player character is not just some blank slate for you to take over and fuck around in the world, he must have strong motive that would make sense for him to delve into the conspiracy swamp and risk his life. Like for example: gov hides the truth about some nasty event and the protag is made into a scapegoat so he has to clear his name, but it's not so easy because he has whole system against him or something like that.

I think the world you described should be a background to the protag's personal struggle against whatever pushed him to act, unless he's meant to be a hero/revolutionary that wants to change a world itself, not just his own slice of it.

> I don't want to rip off of Deus Ex too much.

Just dew it. No point in trying to be original unless you come up with the better idea for something.

>>15282272

>I've gotten my texture atlas working at this point, so no need for me to use a texture array

Sure. Just keep in mind that supersampling & bilinear/trilinear texture filtering can cause texture bleed to happen.

Theme for Ludum Dare 42 is Running out of space

>>15281733

I'll just list some notes as I can come up with them.

>How far in the future is this? Monopolies are far from unrealistic, however, people being blissfully unaware that anything exists outside of them is quite dystopian. Dude Sex takes place in 2052, 52 years after release, so at the time of launch, going from the state of the world to where they portrayed the world being was actually possible.

>Why do tech companies need to sell the information to the government? Can the government not just demand it be handed over?

>exactly how is the degeneracy pushed onto the general population? Internet and media of course, but what else is being done? This should be showcased in the structure of the world itself. Was religion perverted beyond recognition, or straight up outlawed? If outlawed, would then there be underground cults, like Deus Ex?

>how is the population generally controlled? With violence, like 1984 or North Korea, or more like China and Tencent's promised system, where their social media scores people based on loyalty to the government and association with the right people? In that system, if you don't have a high enough score, you can't get loans, housing or work high-paying jobs. You did mention that ban from proprietary services, I suppose it's similar to all the shadow bans and stuff like the Unpersoning of Alex Jones. Could be an interesting subject to tackle, though I think that it should be pushed further, unless the internet is more important then than it is now. Perhaps there is no longer physical money and people could have their credit accounts frozen for disobedience.

>I know that even small towns are embracing technology, but how reliant are they? For this to work, to be without constant access to the internet must be the equivalent of being homeless in this world, or at least close to as significant. Refer to my suggestion in the previous point, but how far will it go? Perhaps a significant portion of work is now over the internet, or a significant portion of schooling and education.

>If there was a closer link between the social media monopolies and the government, then social media itself could be used as a government ID/Passport. This could address my previous point. Perhaps everyone's assigned an "account" at birth, to serve as a birth certificate on top of that. It could also explain how newer generations become so used to the state of affairs so quickly.

>Dystopian is easy to say, but how dystopian? With the implementation of technology everywhere, I'd expect a lot of screens and LED billboards everywhere. However, a big part of Deus Ex and 1984 was the disconnect between the rich and the poor, and how widespread homelessness was. It could be a flashy city on the surface, but dingy and abandoned in reality. But it's all good, they still have their smartphones!

The idea for cyber security isn't a bad one, actually. It allows to further highlight the discrepancy between the rich and poor, or the party and the people. If cyber security and hacking is a big focus, then I have two important points:

>what goals do the hackers have? are their numerous factions with different goals?

>what is actually possible with hacking?

Consider my idea of unifying government IDs and birth certificates under an account. You would be able to entirely change your identity to escape major crimes. You would also be able to pretend to be someone else, to get them arrested, and so on.

Also, the idea of linking banking to that as well, I think it's quite obvious what hacking that would result in. This gives me the idea of certain factions of hackers simply doing it to help the general state of affairs and giving the money to organizations who need it, whereas others do it solely for profit and gain, though that may be a bit on the nose. Perhaps the supposedly nice organizations are not as nice as they seem.

A big part of Deus Ex was having it's own story, which in the background predicted many real world happenings. You could take that angle, predicting where technology will go next. I've given you some examples, but some can just be part of the worldbuilding itself. How big is VR in the future? What utility does it have? Could VR be used for propaganda, or reeducation, of political dissenters? What about Smartphones? They will obviously be a big part of culture, but what are they like now? Did technology like Google Glass and such take off now?

I've written a bunch of general thoughts, but can't really give you an exact focus. I did suggest those angles on hacking though, and those could do you well, I think. I hope those notes help, either way. Oh, and don't worry about ripping of Deus Ex, it's been as good as dead as the first game.

>>15282333

Why do the ones I like never win, Mr. Trips?

>>15282440

I wish I knew.

If I'm moving my character with rigidBody.Moveposition() what's the best way to calculate his speed knowning that rigidbody.velocity doesn't work?

As of now I'm doing something like this

Vector3 position = rigidbody.position;

{function to apply movement}

Vector3 velocity = (position - rigidbody.position).sqrmagnitude

but my velocity is really small so it's pretty useless.

Any suggestions?

>>15282550

I'm a fuckin idiot I should just divide that by Time.deltatime

>>15282384

Thanks for the input anon. Really appreciate it. I like the idea of a Chinese style "social credit" system, and having government I.D. being linked to your online presence.

I don't want the game to take place too far into the future. I'm thinking maybe 2040, about 20 years into the future. I know that kind of conflicts with the idea of people forgetting about non-proprietary services, but I think that's already starting to become a reality. Most normalfags with a phone these days basically have their entire internet access pre-installed on their phone. They use Google search engine, YouTube app for watching videos, facebook app, and every other service that has been monopolized by Google or someone else. This problem is even worse if you're a normalfag with an Android. I feel like even if someone wanted to start more "ethical" alternatives to YouTube and Twitter, they'd be drowned out because they aren't as easy to access and most of the general public simply isn't going to hear about it.

So, saying people completely forgot about alternatives may be a bit of stretch, but during 2040, when these companies basically monopolize all services for communication and accessing media, I doubt the average 25 year old Joe is going to know that there are currently other operating Systems besides the ones that Current Big Tech Company is putting on their devices.

(1/2)

>>15282694

>how is the population generally controlled? With violence, like 1984 or…

I'm actually still thinking about it. Trying to weigh the pros and cons of either a brutal authoritarian dystopia, or a dystopia that was created from humanity's negligence and submission. I had originally thought about robot police roaming the streets, or people having a chip in the back of their neck with a paralysis switch that the government can enable at any time. But I think a brutal 1984 dystopia might be a bit too bizarre, especially considering it's only 20 years into the future.

My overall idea is that the government tries to create a certain illusion of freedom. You are free to work any job you want, and you can buy any computer you want. Read whatever book you want, and buy the house you want to live in. Our good government isn't going to punish you, but these big tech companies have the right to restrict access for any purpose and at any time. Maybe with the exception of government social media I.D.

On the other hand, they did take away a few rights. 2nd amendment is gone, and people still have freedom of speech, but their credit will be harmed by what they say. And, recently, people have been getting banned from services for doing things in the privacy of their own home. There was an outrage about this, but all conversation is being silenced.

>exactly how is the degeneracy pushed onto the general population? Internet and media of course, but what else is being done? This should be showcased in the structure of the world itself. Was religion perverted beyond recognition, or straight up outlawed? If outlawed, would then there be underground cults, like Deus Ex?

Actually a really good point. It's a problem in today's world, with the media glorifying weakness and shaming ambition and achievement. Teenage girls are acting like sluts because their favorite pop stars told them to, and 20 year old men are acting like they just got their balls chopped off, because they feel like that's the only way they can fit into their "subculture" of socially awkward nerds.

Obviously, this tactic is something that will only change subcultures and hive-minded countercultures. It won't change the general population. But, lately even the mainstream media has been trying to promote degeneracy to the average person. CNN published an article a few months ago encouraging men to let their wives fuck other men. Imagine a world where that was a common thing to see when turning on the news, and we had no means of publicly ridiculing them. I imagine that would actually begin to confuse people enough to mentally break them.

Not to mention the power shills have to hijack internet threads and create an illusion of common consensus. What if LITERALLY every time a man complained about his wife cheating on him, 10 or more people tell him that what his wife did was perfectly okay and that he should accept what happened? That's how degeneracy is pushed. By making an illusion of consensus and shaming people who don't participate.

>Lol, you want to get married and have kids? What a fucking loser, I'm gonna go have a 3-way with my open-relationship girlfriend and her new boyfriend Tyrone.

I think I should definitely expand on the concept of degeneracy being normalized. Have the consequences be portrayed in this world. You know, one of the bigger fears the older generation have about millennials is that they won't be adept enough in the workforce, and some people are even worried about what happens when the boomers die and Gen X retires. Will we have a workforce full of unskilled and unmotivated workers? I'll probably explore things like that.

I really appreciate the input anon. You made me talk myself into a lot of things. Maybe I'll focus on world-building for now and write up a document. I'll come back with that later.

(2/2)

>>15282694

>>15282696

Godspeed, anon, you have some great ideas there, just a matter of fleshing them out and good execution.

>>15282721

>Out of space

>A simple bus traffic sim

>>15282748

Very neat.

Hi there, soul-drained drunk pro-dev here.

Ask me anything I guess

>>15283217

wich 2hu wud u fug and why do you want to gas the kikes

>>15282748

Looks like someone skipped leg day.

>>15283227

ur mom twice lol

why do you want to exterminate cocroaches?

They shouldn't spread.

>>15282333

>>15282721

>tfw haven't participated in a Ludum Dare since 32

I would participate this time just to do it alongside some of you guys but I'm working on my current game and have too much to work on after that. Don't forget to post your entries for us to see.

>>15283588

I should point out that I drew those feet/legs/walk cycles based on the 2D sprites I did of the Xenogears character. This was in 3-4 hours at work, in between slow parts of my shift on Wednesday night. These came out pretty decent I think, especially the 3-4 at the center of the paper. Gonna keep at it, then upload those into Gimp so template editing might be easier in the future. This is not a focus, right now I'm planning, building my project, constructing base scenes in Godot, and putting together a portfolio binder.

>>15283752

MGS uses motion capture

>>15283752

maybe you should try to be more of a gamer. I figured that out for my game in about 10 minutes.

How to do I go about implementing accurate animations for players/characters moving over deformable terrain?

where they step on debris and lift themselves up with their legs or only move if their feet actually make a solid connection?

>>15284051

It's called Inverse Kinematics, start searching for your engine plus that term.

>>15284065

Thanks, idiot.

>>15284065

Is this why graphics libraries and engines have a matrix object that stores it's inverse?

>>15284051

You do it backwards. You don't make the legs move the character, you move the character and then make the legs appear as if they are doing the work.

If you want to do it the other way, where the movement of the legs drives the movement of the character as a whole, you are entering uncharted territory. I can't think of anything existing that would work for games. There's stuff like the IK rig and Motion Matching out of Ubisoft and some neural net animation controllers that all produce really nice animations, but those are all reactive, moving the legs to match the movement of the character as a whole rather than moving the character using the legs.

>>15284150

I've just played a game recently (Exanima) where the character leans to one side (as if the legs are just dangling from the body and the player is looks as if they are just controlled by a single position) when you move forward and turn at the same time and I want to avoid that because it looks like shit.

Found something regarding realistic character movement. Possibly the most impressive tech I've seen in years.

>>15284152

The overgrowth dev went over the way he did procedural animation, vid related. Probably your best bet as far as something that is easy to implement and looks nice. While it would be nice to have something like Motion Matching it's not really feasible for a one man no budget project.

>>15284220

The tech has advanced to four legged locomotion since then.

https://www.youtube.com/watch?v=uFJvRYtjQ4c

>>15284243

Mean for >>15284167 not >>15284152

I'm considering of taking a game dev course. Is it worth it? I'm ADHD and doing stuff by my own leads to a dead end.

>>15284923

>doing stuff by my own leads to a dead end

Gamedev will be a dead end for you.

>>15284936

Everything else is a deader end for me, I don't like them. I can actually work well with a team that watches over me while I work and tells me how to do stuff.

>>15284948

find another hobby

>>15284976

What if there's none.

>>15285011

Try and do things, but I wouldn't waste money on a course. And I definitely wouldn't expect to make money out of gamedev.

>>15285098

It rained here, 70-something deg. F. from 115 for a week or two. Going to a coffee shop to do more work today, might get rained on. wewlad

>>15282049

https://www.ejgold.com/cold-war-spy/

I don't frequent these threads so forgive the ignorance but is that guy who made the Gondola vaporwave thing still working on his game?

>>15285019

>wouldn't waste money on a course

>wouldn't expect to make money out of gamedev

Why so antisocial and pessimistic?

>>15287041

I view those things as a test. Don't let it deter you from your goal.

>>15286949

Remind me again what your engine does that modern engines don't? Where is it limited?

>>15285596

I believe it was taken out back like old yeller

Anyone here using blender know how to force it to only use defined keyframes? I made 3 keyframes I want to use, but as soon as I spaced them out it generated a fuckton of them that lerp between the ones I actually want.

>>15287390

Already messed around with all that.

>>15287390

>>15287400

I should be more specific. That killed it lerping from frames, but it still exports all the frames with the keyframes even if they don't do anything.

>>15287249

It's written to use the same idea as the quake engine, goldsrc engine, except it uses modern graphics API's and formats, while still being backwards compatible with older systems.

It doesn't really do things that diffrently beyond having access to better image compression, etc. by being made in 2018. I want to eventually make game engines that do things modern engines don't. But right now it doesn't really do anything that you couldn't do right now with something like idtech3.

It's at least faster than a modern engine on the merit of doing less things.

>>15287403

Blender is baking the animation, probably because you are using a proprietary format like FBX that blender doesn't have great support for.

https://docs.blender.org/manual/en/dev/addons/index.html#import-export

>>15287497

Doesn't have one. I'm going to resort to janky workarounds.

>>15287501

>FBX

No. I'm the anon wasting his life away at GLTF.

>Don't want to use default ISerializable shit in C# because it's "bloated", same with JSON. Instead my game is going to look like

<../data/user/someLevel.ma0

<../data/user/someLevel.ma1

<../data/user/someLevel.rep

Where .ma0 is the map metadata, .ma1 is the map tile data, and .rep is the replay file

>>15287512

>GLTF

Godspeed anon.

>Just want to make some vidya music.

>Spend the next 4 hours fucking with VSTs, FL Studio, and Kontakt.

Kill me. Why is Kontakt so fucking buggy?

How do I maintain a stable framerate?

Let's say my game is designed to run at 60 fps and "one second" of gameplay consists of 60 ticks, so that things essentially happen 1:1 to real time.

Right now, I'm having a rolling average of the last 300 frame times (5 seconds' worth). At startup, there's a big spike - therefore, for the next 5 seconds, the average frametime looks high, so it might update more than necessary. What do?

>>15287848

Look up delta time.

It ensures everything "works" no matter the update speed.

>>15287860

My game works in terms of ticks, though. Objects have to update in a deterministic way, so I can't afford to have floating point error, so I have to use a tick-based system.

Or would I just accumulate time each frame, and either call 0, 1, or 2 tick updates depending on how the math works out?

>>15287879

Number two is I think how you'd implement it in that sort of system.

It's a little less smooth but it *should* do the job.

>>15287879

>>15287909

So yeah, accumulate the time and then update as many times as you need

>>15287520

why not just separate your format into sections? My format does things like print:

LEAF_LIST

NODE_LIST

,etc to show what part of the data should be read.

>>15287848

write a FPS limiter to guarantee that a frame takes at least a certain amount of time. If this amount of time is higher than the maximum amount of time it takes to complete a frame without the FPS limiter, then it will be stable. The way you do this is record the frame time and if it doesn't meet this minimum, check time in a loop until it does. To limit CPU usage if you have a lot of time before you hit this minimum, call sleep to give up CPU to other processes. But if you are close to meeting it, don't call sleep and hog the CPU to make sure it's accurate. Also, never call sleep for more than the minimum.

>>15287919

>why not just separate your format into sections?

I've never really done something like that before, so I didn't try it. Stuff like variable-length fields or looking for termination characters seems scary to me so I avoided it.

>write a FPS limiter to guarantee that a frame takes at least a certain amount of time

My first attempt kept the average frame time, and then tracked a minimum and maximum threshold value, flagging the main loop as being either too fast or too slow, which would result in either pausing or calling two updates, but I kept noticing subtle timing errors with it.

What are acceptable values?

>>15287934

60 FPS is a good default. So basically you would have a variable for your timestep and a variable for your minimum timestep, which would be 1/60 seconds. And you would block by checking the time until you meet or exceed this minimum timestep. As long as you can always complete a frame faster than this framerate then it will be stable. You would have to lower it if you couldn't do that. Some pesudo code would look like:

time_update(time_t time){
    time.new_time = get_time();
    time.tstep += time.new_time - time.old_time;
    time.old_time = time.new_time;
}
...
while(1){
    time_update(time);
    if(time.tstep > time.min_tstep){
        update_game();
        time.tstep = 0;
    }
}

Of course it's more complicated than that, if you want to look at the actual code I use I can show it.

What I do is, since it's in the file the size of the buffers are known, so I just have a header that has a bunch of numbers that tells me how big each buffer is going to be.

>>15287993

Is there a particular reason why the inner part of the while loop uses if(>step) rather than while(>step)?

>>15288095

while is redundant because it is only true once, with the time.tstep = 0;. So, its misleading because it makes that look like a loop, when it's not.

>>15288130

You assume tstep will always be smaller than min_tstep, leading to less frames per second when your assumption is wrong.

Using while instead of if, and subtracting min_tstep from tstep instead of setting it to 0, would always attempt to update and draw every frame, even if one takes longer because of some irregular expensive computation.

There are arguments for both cases, just pointing out what c6a995 presumably meant.

>>15288144

That makes sense- I never thought of doing it that way. I was only thinking of this system in terms of each individual frame when I wrote it.

>>15288914

has to be started from a monobehaviour

also not sure if you could call them with a string

>>15288926

Wait so my only solution is making another function that is used exclusively to start a coroutine?

That's absolutely retarded

>>15288931

assuming that you only have one player, do this

have a static reference to the player monobehaviour

then from the object you showed do Player.refer.StartCoroutine(…);

if you REALLY want to do everything from within the object, you make a function in it that checks the progress and call that from a monobehaviour (like the player) every frame or whatever your timestep is

>>15288939

That worked but damn that's dumb

>>15289070

I wanna cum all over your spritework

>>15289334

And what does this have to do with /agdg/ in your imagination?

>>15289334

>no sound

Why do you keep living indeed.

>>15289070

Damn nigga, you hand drew all those sprite animations?

>>15289070

>Chest equip item is called a vest, and its sprite has no pauldrons.

>Overworld sprite gains pauldrons from equipping it.

That aside, excellent.

>Update and Draw steps should exist independently, one should not hamper the other

>Therefore, I need to check the total frame time, and both methods to ensure that one isn't excessively fast or slow

>But code duplication is a bad thing, which means each "group" of functions can be extracted into an object instance, so I now have a Draw instance and an Update instance

>Each contains the target and elapsed times, etc, as well as a rolling average time to make sure we're within valid bounds

>Now each component is measurable but I can't just have my Game root object behave differently either

>Game instance is also a "component" with the same pattern so it can now monitor the speed of any of its child components

This feels weird but it works and I think it's secretly a good design. It's basically a tree structure with no restrictions on structure

Can any C macro autist tell me if it's possible to make a macro for this:

>arr_pushMany(&octaves, 4, (f32[]){10.0, 1.0, 2.0, 0.2});

so that I can type it something similar to this:

>arr_pushMany(&octaves, 4, f32, (10.0, 1.0, 2.0, 0.2));

>>15290227

afaik C macros are just dumb substitution so no, but you can do that in Rust.

macro_rules! arr_pushMany {
    ($a:expr, $b:expr, $c:expr, ($($d:expr),*)) => {
        arr_pushMany($a, $b, &[$($d),*]);
    }
}

and it looks like

arr_pushMany!(&octaves, 4, f32, (10.0, 1.0, 2.0, 0.2));

>>15290227

This code seems pretty redundant. Is this what you're trying to do?

void push_many(void *dst, size_t ecount, size_t esize, void *src){
	memcpy(dst,src,ecount * esize);
}

int main(void){

	float *buf = calloc(1,sizeof(float) * 10);
	push_many(buf,4,sizeof(float),(float[]){ 10.0f, 1.0f, 2.0f, 0.2f });

	for(int i = 0; i < 10; i++){
		printf("%f\n",buf[i]);
	}
}

>>15290337

I'm curious how many indie devs are using Rust. I've never heard of it being used for gamedev.

>>15290398

That's more or less exactly what I'm doing. I'm just using it as a makeshift variadic arguments since C's stdargs are shit.

>>15290450

You don't need variadic arguments for a simple memcpy. The type system isn't that important since it's just blocks of memory, so you don't need to make it explicitly aware of the type of memory you are passing it.

>>15290465

But I need the type when creating the array into the function call anyway. And I have other similar functions where I want to loop through the items rather than just copying the data.

So for my game's battle system, I want enemies to randomly select one of up to three party members to attack, excluding any party member currently in the "KO" state. How can I do this in C#?

>>15290676

Add all non-KO party members to a temporary list, and then randomly select from that list.

>>15290894

Is it worth learning that language? Every time I try and pick up Godot, I hit a mental block trying to learn that awful Godot Script or whatever it's called. Getting flashbacks BYOND's Dream Maker.

>>15291732

If you really hate GDScript (which is really close to Python), Godot also supports C#.

>>15291732

is easy as fuck though.

>>15291975

then make your game, expect some bugs and deal with them if you give a fuck at that point.

>>15291975

What errors are you getting with godot (or does it just crash?)

Might be the new renderer, they're working on backporting it to the old opengles version so maybe that'll fix it fo you?

>>15291975

>love2D

>trash language made by thirld worlders.

meme spacing aside, you do realize that Love2D is a framework for Python and isn't itself a language?

>>15291988

>>15291990

>>15291993

I'm talking about my overall experience with game engines, and I'm very angry right now.

Will start to enginedev a pong game tomorrow.

>>15291975

On this topic,

What do you want to see in an engine? Be as verbose as possible.

>>15291998

>>15292016

>>15292017

I knew it was Lua, I don't know why I wrote Python.

I'm very tired and spent all day working on my engine timestep and still have no idea how to organize shit.

>>15284220

movement systems like this lead to every character action becoming animation-based and you end up with shitty, unresponsive gameplay and frustration.

>>15292808

But muh CINEMATIC FEEL™ is more important than gameplay.

Is there any place I can download 3d models of popular games? I want to see triangle counts of models in various games over time.

If anyone can just link a resource that tracks this stuff that would be better.

>>15293278

https://www.models-resource.com/

>>15293290

Nope

>>15293292

>get quite literally what you ask for

>n-nope

>>>/out/

>>15293304

Think of a model from a PC game that's popular.

Now try and find that. Regards.

>>15293316

It's the best place available, if it's not on there you're shit out of luck. They have a forum where they'll help you rip shit yourself, which you're going to have to do if you want shit that's not on there.

>>15283588

Changing the codename of the project to Sync now, can't think of a better one yet.

I have the HUD working better, resizing at any size with the game. Had to make sure my project's "shrink size" and "aspect ratio" window properties were set to 1 and ignore, respectively. Xolonium is a fun font, I hope I'm allowed to use it for my project. I think I can.

>>15293709

Watch vid related, it will give you an idea of what the settings mean and tell you what to turn down so it doesn't take 15 hours to bake.

>>15293844

It's just a checkbox, dude.

>>15293844

look up the action editor on blender and make every new animation as a diferent animation.

unity should load every animation as a new file.

>>15294629

The first steps take the longest.

>>15293316

https://www.models-resource.com/pc_computer/leagueoflegends/

Are you fucking stupid?

>>15294629

It takes three months because it needs one month for each dimension

Are there any resources for making your game maximally moddable? I'd like to have something where the player can just drop mods into a folder and they get loaded at startup, but I have no idea how mods even work.

>>15294676

You could do it like Bethesda: You should already have folders for all your assets in the game, just make a folder that has priority over the other folders and has the same asset folders, then whenever someone downloads a mod they just drop the folders into that folder and everything gets sorted.

You could also do the same thing with zip folders and just have the game access the zip folders inside the mod folder and replace everything.

>>15294686

But that only helps me with json-tier mods and texturepacks. I want to know how to load gameplay altering code files.

>>15294703

Same concept, just have the game load those code files instead of the ones you currently have.

>>15294703

i know c# lets you compile strings at runtime

>>15294709

You'd have to run a compiler with your game

>>15294657

That would be 1.5 months since he did 2 dimensions in 1.

>>15294709

That doesn't answer how you can load multiple mods without them immediately conflicting with one another.

>>15294722

Mod tools or establish some kind of order, like date modified or whatever.

>>15294716

You're basically just making an engine then. You might as well have the game compile every time you launch it at that point.

>>15294722

>mods conflicting with one another

Welcome to modding. In a perfect world the game would iterate through the mod folder and just use the last example of a file it came across.

>>15294730

Date modified for what? My game .exe? What are "mod tools" in practice and how do I add them for my game? This is why I asked for resources on the topic because I don't fundamentally understand how a code altering mod works.

>>15294755

What programming language are you using?

>>15294817

C

>>15294884

Then to my knowledge you don't have the ability to change code on the fly like that and you would have to recompile the game every time you launched it if you wanted to support mods that edited the code. If you want something like that without compiling every launch then >>15294713

>>15294958

I could probably include some iteration of mingw and call it from the launcher, but that still doesn't really answer how you preserve compatibility between mods. You can load tons of mods to Minecraft with them affecting the map data and item properties and stuff without them breaking one another, and I want to know how that all works.

>>15294985

Just have the game go through the mod folder and give everything in there precedence over their respective game files. It doesn't matter how many mods you have. If two mods replace the same thing then just have it use the final mod it iterates through's files.

>>15295026

What files? Am I supposed to have a separate file for every part of every feature?

Though now that I think of it, I wonder if I could somehow read the source code of the mod and mold it into the original code. For example if there's 2 definitions of a struct, I could combine all the properties. That way multiple mods could modify the same struct without breaking each other.

>>15295084

Yes, you are. In your Minecraft example most of that data could be stored in .xml files which the mods could replace.

>>15295133

make something like pong or breakout or flappy bird.

>>15294884

What you have to do, is you have to load the code for mods out of a dll or so. This can be done on windows and linux respectively:

https://msdn.microsoft.com/en-us/library/windows/desktop/ms684175%28v=vs.85%29.aspx

https://msdn.microsoft.com/en-us/library/windows/desktop/ms683212%28v=vs.85%29.aspx

https://linux.die.net/man/3/dlopen

Then document what functions you will try and read out for any mod. This way a modder can inject code into your program in an arbitrary way based on how much control you give them. For example if you are defining some items in your code, you would have an external function that queries the dynamic library for items to define, allowing a modder to create their own items.

>>15294733

This is something I don't want to do, but have it be moddable. This is why my simple fucking puzzle game is taking months to figure out because I don't know how to approach it.

Like I can create all sorts of states and flags for objects to use, and then rely on data to plug things together, but I don't know how best to check all these interactions and make specific behavior

>>15295133

Read the docs

>>15295133

>Don't know a single programming language

Do Code Academy or something

>>15295133

Learn literally any language, there's very good interactive tutorials for javascript/python and shit in codecademy. Once you know how to code in one language, it's not that hard to move onto another.

>>15295137

I think I recall handmade hero using those to load something, but I ignored it at the time. Seems like the right direction to look in, thanks

>>15295133

If you're just starting out you are going to have to start reading and watching beginner tutorials until you can understand the documentation. Once you can understand the documentation that will be your main source of knowledge on top of just googling your questions and reading whatever comes up.

>I just want to make a Digimon World 1 clone.

DON'T MAKE YOUR DREAM GAME FIRST. Or even second, or third, or forth. Make your dream game when you know how to make every individual part of it or when you have the money to hire people who can collectively make every part of it.

At this point make a clone of some Atari game such as Pong or Breakout. I made asteroids and lunar lander as my "learning games". Even when you have some games under your belt making simple learning games is a great way to figure out your workflow. For example if you get used to Godot but now you want to learn Unity, you make space invaders or something in Unity to learn how Unity works and figure out your workflow with that engine.

>>15295133

recreate Snake

>>15295376

Rather than having to control the curve by hand, which it seems like what you're doing, have you thought about doing something like this?

https://en.wikipedia.org/wiki/Catenary

>>15295401

well. I had no idea what was the name of the curve, one more thing to do tomorrow

>>15295384

Nokia Snake*

>>15295133

There are tutorials all over youtube.

>>15295208

If you want modders to be able to make game logic, like conditions and dynamic stuff, you're going to have to add scripting ability to your game. The plus of this is that much of your own game can be scripts, which makes changes easy.

>>15296167

Is occlusion automatic or do you have to lay down brushes to determine where occlusion happens as you do in valve games?

>>15296184

Occlusion is automatic. It's also automatic in valve games as well. It's just that valve games also include tools like hint brushes that let you do it manually, even though it's not a requirement. valves engines are using the same exact kind of automatic occlusion culling as my engine, in terms of the theory behind it.

>>15296222

it's not automatic in valve games, it's done during compiling, vvis being the tool which determines occlusion. Maps must be blocked out by designers so they are sealed, which helps with vvis. It is best to treat even outdoor areas as rooms with a floor, 4 walls and a ceiling when compiling for the final .bsp.

-t. used hammer for like 8 years

>>15296260

I have also used hammer editor, I can understand what you mean but maybe we are using the wrong words. What I mean by automatic is exactly that: such a tool like hlvis constructs visibility portals for each leaf and calculates a PVS, and it does this automatically, just like my engine, you don't have to plot out portals or anything like that. I thought you were talking about how you can manually place visibility portals with hint brushes.

The requirement to block out the maps isn't an absolute requirement by the algorithm but it's a very sensible requirement to impose on mappers. Basically you can use the visportals to check which leaves (and therefore, polygons) are never going to be visible to the player, and you can use this to prune the map and basically cut the amount of polygons you need to consider in half (about). My compiler actually doesn't do this optimization yet and so it's slower, it generates a more bloated BSP and PVS then the hlvis, hlbsp, etc would.

This map editor and engine actually work exactly the same way as hammer does, after the map is ready to compile it does a CSG merge (like hlcsg) on the brushes, it compiles the resulting polygon list into a BSP tree, then it creates a PVS out of it for occlusion culling.

In the future I want to make it do a two-pass BSP compilation, the first pass is to generate visibility portals and cut the polygon list down, so that it only needs to consider the polygons on the "interior" of the map, and then the next pass is to redo the BSP compilation on the reduced polygon set so that the BSP and PVS are smaller. I can also make it faster by keeping a list of the original polygons from before the CSG operations are done, and then eventually only storing polygons from that original list, which would stop all of the problems with an increased polycount from polygon splitting.

>>15296294

Yeah, I do want to add hint brushes because I don't want to make compromises on optimization, if I have extra cycles I am sure I can waste them on some other effect, and make the game look better. I haven't actually checked performance on a computer where it matters yet, I just got this to work. Right now i'm working on some debugging HUD so that I can look at how many polygons are being drawn and how many frames I am getting. Then I want to test it out on my windows XP machine which only supports OpenGL 1.3, it is the perfect computer to test it on since this actually will make a big difference in performance for it.

Also, right now it's missing frustum culling and back-face culling, so the polycounts can be reduced more, beyond the ways to optimize the BSP like I talked about earlier. Kind of a long post but those are my ideas for it.

>>15296359

>you don't have to plot out portals or anything like that

No, but you still have to do quite a bit of work, especially if your game has vertical gameplay like in tf2, you need to set the walls and ceilings quite high and account for surrounding geometry/visibility. The games handle it by using the data vvis generates, vvis generates it based on the brushes you mapped out for it. The less you do, the longer the compile time and the worse the optimization.

>>15296359

>Kind of a long post but those are my ideas for it.

Don't worry about it, I always enjoy reading your posts. If I didn't want to hear I wouldn't have asked.

>>15296388

I just mean, it can do it for you… not that it can do it very well, or that you don't have to be aware of how it works to get good results. I guess I just interpreted it to mean something more extreme than what you meant. You're completely right.

I only used hammer with goldsource so I am mostly thinking of making those kinds of maps in this engine. It should be able to do a good job at that.

>>15296393

Thanks, I'm glad you like it! I wrote that mostly because I was surprised about how long the post was. I'm going to test out my framerates on my weaker hardware since it's not going to make a difference yet with my main machine.

OK, so looking at it, on my T61, with occlusion culling, I get 190fps on the worst spot and 250fps (my cap) on almost all of the other spots. Without it it bounces from 75 to 90fps. This is on an Intel 963/965 integrated GPU. So this is the computer that really benefited the most from it.

On my T41, which is the computer that really needs this feature, or so I thought, just stayed at 50FPS the entire time, and went up to 100FPS if I raised the framerate cap. Kind of confusing. It didn't make a difference if I had occlusion culling on or not. When I set it to wireframe mode with occlusion culling off it totally fell apart and couldn't render fast at all. So it makes me think that the Z-buffer was enough to keep it rendering fast. This card does hardware occlusion culling with the HyperZ II unit inside it, which uses a hierarchical Z buffer to do occlusion culling in hardware, so it's pretty interesting to see how effective that is.

It's pretty impressive either way that it has such a better Z buffer. The card is an ATI Mobility Radeon 9000. Even though the intel GPU has a better shader, since it supports OpenGL 2.0 while the Radeon card only has OpenGL 1.3, i'm wondering how this card will end up performing in the end. It has a fragment shader which is using a platform specific shading language, not even ARB assembly shaders work on it.

>>15296167

Dude how did you learn to write a 3d engine in c? I'd love to try making my own but i'm such a tard with programming

>>15296611

You have to follow a lot of tutorials, these are the ones that I used mostly. The rendering engine that I wrote works the same way as the one described in the last link:

http://nehe.gamedev.net/

http://winprog.org/tutorial/

https://vulkan-tutorial.com/

https://software.intel.com/en-us/articles/api-without-secrets-introduction-to-vulkan-part-1

https://vulkan.lunarg.com/doc/sdk/1.0.57.0/windows/tutorial/html/index.html

http://www.alsprogrammingresource.com/bsptut.zip

https://www.cs.utah.edu/~jsnider/SeniorProj/BSP1/default.html

https://www.cs.utah.edu/~jsnider/SeniorProj/BSP/default.htm

Work your way up by writing simpler and slowly more complicated engines. Eventually you will get there. It's not about skill, it's just about writing more code… you will learn more than you realize by challenging yourself that way. At one point, I really didn't know very much about programming at all, but only spent my time working on engines.

if i bake a skinned mesh to a regular mesh is there any big pro/con?

i'm getting rid of the skeleton so less gameObjects, but i'm also using a new mesh, so more memory, right?

>>15296718

Thanks man

god i fucking hate steam reviewers fucking self entitled little shits

>>15291370

Store metadata along a model describing the attachment points. Obviously you need at least a position, in model space. Possibly an orientation, depending on what you're trying to do. Like, say you wanna have a turret hardpoint on a space ship. You probably want the turret to snap to the hull, so you'll need a vector or something describing what should be up for the turret. In addition you might also want a roll angle, so you can point the turret/cannon backwards, or whatever. This can be combined into one thing by using a quaternion. In engines that use a scene graph, this is trivial to implement. When loading the ship, create a node for the ship hull, then for each hardpoint create a node using the model of the turret that goes there, parent it to the ship's main node, then set it's local transform to the location-rotation of the hardpoint as specified in the metadata. You can then (probably) make it point at the enemy by simply doing turret.lookAt(targetPos), as (usually) those functions takes the nodes transform hierarchy into account. It's real easy doing it manually, too. You basically figure out what transform in global space needs to be applied to point it in the right direction, then transform that down through the hierarchy and you end up with the transform in turret local space needed to turn it the right way, then apply that.

To do something like a sword, you'll want to implement parent-child relationship between hardpoints and model bones. This is also easy to do in a scene graph, since the bones can easily be a part of the scene graph. You just do what you normally do, I.e. drive the animation in the scene graph and such and parent the sword to the hand bone. It'll stick like glue. If you're doing ragdoll physics, you could even make the sword an actual extra bone that's just added to your skeleton, like all the rest and have it collide and affect the pose. Same thing for anything else, like armor. Just parent it to the appropriate bone and make sure the animation traversal visits it and updates it's local transform.

As for 3d editor, you'll probably have to write your own plugin. Unless there's some kind of generic metadata export script. Though this isn't that difficult, especially in blender. Just find some good script that does something similar and use it as a reference. Use the console inside blender to explore the data structures. I'd personally use an Empty to represent a hardpoint. You can load a turret model into blender and use modifiers to snap it to the Empty, applying all relevant transforms in the process (location, rotation, even scale). Have the export script indicate what it's parented to in a way that you can find it in the engine, like the name of the parent bone. You could unify it by giving all models an implicit root bone, or letting parent-less nodes be implicitly parented to the model root (i.e. the root node of the model instance in your engine). You could write a small editor that shows the model graphically with all it's hardpoints rendered and presented in a list. Then you can set any additional metadata, like size restrictions, plasma only, max power available, etc. So no, you'll want an extra tool. Or export the metadata to a text format like json and hand-edit it. Actually, with blender you can make UI extensions that lets you turn an empty into a hardpoint and set it's properties right in the editor. You might even be able to create ghost objects to show what turret you've selected for that hardpoint. But it's a lot of work figuring out blender's api and some things, like having an animated preview of the turret firing would be exceedingly difficult. With an external tool, you have max control.

>>15295376

Anon I don't want to disappoint you but there's already a program integrated with unity and unreal that does this

>>15297707

yea but it's not free

i too have a generic spline thing to make fences, but that's about as much as i need, not gonna spend whatever houdini is charging

Does anybody know how Aurora 4x does the graphical representation of planets and ships? I wonder if it uses some external library or if it's just VB6 and .Net.

>>15282011

Impressive. Very nice.

>>15297707

cute, but no runtime generation.

We want players to be able to make the bridges wherever they want.

>>15297652

>end up at my random number generator function

>it's taking in a long as a seed instead of double like it's meant to

>somehow worked correctly until this point

One of these huh.

>>15298273

This is some top-tier writing anon.

>>15289373

It's an amateur (free) game in RPGMaker

>>15289376

The original has no sound :^(

>>15298976

These threads aren't to discuss free amateur games, they're for developers of amateur games to discuss what they're doing.

>>15291975

>I just wanna make games but every fucking retards keep telling me that raw Opengl and SDL are just impossible to use and that I shouldn't make my own engine and that is hard as fuck

First of all, quit meme spacing, we can tell you've been getting advice from Reddit because you format your post like a retard.

Secondly, you don't need to learn raw OpenGL. If you want to a make a 2D game, I recommend SFML. It handles input, graphics, audio, and network. You don't need 10 lines of code and autistic memory management to load a sprite. It's really easy to use, while letting you use the complexity offered by C++.

Also, most online OpenGL tutorials are pretty shit. I'll admit that. There's a lot to uncover, but most of them just teach you how to draw a 2D triangle and then say,

>"Yay, you've learned OpenGL, now fuck off and make a game with this."

So, if you want to make a 3D C++ game from scratch, I'll point you to a good resource.

His name is Benny. His voice is full of soy, and he always sounds like he's ready to give a warm, fuzzy hug. But, his videos are actually quite in-depth, and he actually takes the effort to explain the low-level work going into the code. I like him because he's not some meme-tier shit programmer teaching people awful practices. And honestly, his voice is rather refreshing compared to the typical "groaning bored teenager" tutorialist, or the Indian with the thick accent showing you how to build a meme-tier game engine

>off for 4 days in a row

>only work for 2 hours at most

Fuck

Got a C++ question.

A cool thing I learned from the SFML book is modern C++'s "smart pointer" feature, which I thought was really nice, because C++'s memory management is probably one of the most dangerous parts about the language to a less skilled programmer.

But, even today, in modern source files, I don't see any projects using smart pointers. They all use the classic new/delete functions. Is there some kind of con to smart pointers that I don't know about? You'd think that something so convenient would be in awfully wide use by now. I just want to make sure I'm not doing something absolutely retarded by using smart pointers for things like Scene Nodes.

>>15299852

Smart pointers have basically superceded classic ones

>>15300018

https://www.khronos.org/opengl/wiki/History_of_OpenGL

It's not that hard to make it compatible with 1.1 if you really want to go for it.

>>15298182

From where to where? Your town is small.

>>15299852

Many programmers will just write C code in C++ and ignore many of the modern features. This isn't because smart pointers are terribly slow or memory hungry, people are just lazy. They learned about pointers the C way (which is how people should learn it) but then they don't bother to learn it the C++ way. Their decision to use raw pointers is based on what they know rather than what is the best tool for the job at hand.

>But, even today, in modern source files, I don't see any projects using smart pointers.

I'm surprised you haven't found any modern source file that uses smart pointers. Here are three I found with minor searching.

https://github.com/meta-toolkit/meta/blob/master/src/analyzers/analyzer.cpp

https://github.com/grpc/grpc/blob/master/src/cpp/client/client_context.cc

https://github.com/godotengine/godot/blob/a8e845a474c01c4cde1416ed00e6ae456786c1fb/thirdparty/etc2comp/Etc.h

And here's some articles about the overhead and performance of smart pointers compared to raw pointers, in case you are interested. Long story short if you've decided to use smart pointers only use shared_ptr when you can't use unique_ptr.

http://www.modernescpp.com/index.php/memory-and-performance-overhead-of-smart-pointer

http://blog.davidecoppola.com/2016/10/performance-of-raw-pointers-vs-smart-pointers-in-cpp/

>>15300056

The island will be large, the player will be able to choose any area of the island to build their farm and will eventually be given or sold an artifact to teleport to and from the city

>>15300304

Some meshes do that. Especially landscape meshes like rocks, or anything involving round architecture.

But, character models contain a shit ton of polygons, and the result you get from directly using character models as a collision mesh isn't exactly worth the performance cost. So, the compromise is to put a low poly mesh on each body part, with geometry accurate enough that you really can't tell the different unless you autistically examined each pixel on your screen.

>>15298090

Good shit.

>>15300304

There isn't always a need for increased collision detail. Stairs, for instance, often use a slope collider for player collisions because it's smoother for first person cameras.

>>15300304

what is that from?

>>15300304

>Is it another case of things not improving because of devs not bothering to optimize their stuff to make use of resources

It's literally the opposite. Using a handful of simple 3D shapes for collision testing is significantly better performing than testing collisions against tens of thousands of triangles.

Also I'm pretty sure 3D models are transformed in the GPU during animation, so the triangles' in-world positions aren't readily available.

>>15300573

Unless it's a MGS game, in which case every stair is modeled and interacted with as they run up/down them, and the camera still behaves smoothly anyways

>>15303291

Of course it doesn't, otherwise GW does their usual legal carpet bombing.

>>15297198

I like how this is coming along. I think it's missing some of the "bulk" that the reference has which, in my opinion, makes the reference look better. Your helmet, for example, doesn't have the extra plate that goes over the top of his head. Yours has quite the slope from the neck to the shoulders, whereas the reference appears more stout and has a little more bulk to his chest plate. With him looking more lithe than normal, he looks more like an air caste dressed in a commando suit.

The last thing I would point out is that the antennae on the backpack seems superfluous. The helmet antennae would already handle things like communications. I would naturally assume that a special operations helmet would be better at whatever communications they perform. I'd go a step further and point out that Pathfinders don't use extra communications equipment, instead relying on their advanced helmets. I would suggest pulling the back antennae and refining the helmet a little bit.

What variations do you plan on making? Are you doing just normal Firewarriors, or are you doing loadouts for all of the various teams/pathfinders? I actually really like the default Tau color scheme, so I really hope they're released in a default scheme, or are at least easy enough to recolor.

>>15298090

So how's this work? You're drawing a bounding box around the starting radius/bounds and the projected region it will end up, then comparing against geometry?

>>15303783

A lot of it is just common sense though. Where's that one autist that wrote the book about WL4's level design which is like 600 pages long? That's apparently a good read.

What kind of game?

>>15303783

Most game design classes are useless, study other games and figure out what makes them work and what could be improved. Other helpful exercises include reading source code, studying neuropsychology and hypnosis, and developing an interest in architecture and landscaping.

>>15303820

It's probably some instancing error with the node/scene setup you've done

>>15303783

The autism behind perfect level geometry and psychologically leading the player along certain paths sounds compelling at first, but it's really overrated.

Just do what Nintendo does. Make your hard levels first, then after that, ask yourself, "what's the most non-intrusive way I can teach a player how to get used to this gameplay?"

In the hard levels, you don't really need to worry about such psychology. You only need to make sure that the enemies are in the right place and that the level geometry is relevant to the player's or enemies' objectives.

>>15303783

>>15303933

On second thought, play Super Metroid. In my opinion, Super Metroid is the epitome of good level design and non-intrusive tutorials. It basically does exactly what you're talking about, giving the player a huge world to explore without ever feeling lost because the game is designed to be figured out without any lines of dialogue and no walkthroughs.

>>15303750

The AABB is used to get the list of potential collision objects from the spatial hash table. If the entity has a velocity large enough that it would move further than it's radius in one tick then it does additional tests first and stores the how far along the movement vector those collisions would happen, then uses the earliest one to correct the target position. Then regular collision checks happen after that for more fine grained responses. The extra checks are just circle-line or line-line intersection tests with heights interpolated based on where along the movement potential collisions are tested.

>>15303783

You learn good game design by playing video games. You notice what works and what doesn't, what you like and what you don't like. Whatever you don't pick up through playing will be learned by listening to your playtesters. Game design classes are a scam.

Does anyone have a pdf of "the pushing points topology workbook"? I tried different torrent sites but can't find it, I checked the share thread but didn't find it. It's supposed to be really good for learning good 3d topology and quick to get through.

Does anyone have good topology books, sites or youtubers? I'm getting pretty good with blender hard topology modelling but I'm still learning.

>>15304649

Look at the wiki, on the gamedev resources page.

Has everything I'd suggest already compiled there.

>>15304649

>the pushing points topology workbook

It's on CGPeers. Accounts are open now.

I'm uploading it to volafile too but that will take longer than making an account and torrenting it: https://volafile.org/r/f8vk0tec

This is an interesting paper. It's basically path-tracing for sound with specific optimizations that focus the limited samples on areas that matter more, giving better results than the alternatives in the same amount of time. They used an Intel i7 3.5 GHz processor and 32 GB of ram. They generally got good results at around 16ms per frame on a single thread, meaning it's viable for video games.

http://gaps-zju.org/bst/

Here's a direct link to a video showing it in action and comparing it to other algorithms.

http://gaps-zju.org/bst/bst.mp4

And because I know someone is going to post it, vid related is the Aureal A3D 2.0 Rooms demo from 1998.

>>15306580

>vid related is the Aureal A3D 2.0 Rooms demo from 1998

I remember having that shit on our computer, and I was always confused because it looked like a videogame but I couldn't find any objective. I was still mesmerized by all the things making sounds though.

>>15305079

Yeah there's some nice things on topology there http://8agdg.wikidot.com/resources

But the guide I'm asking for was recommended by one of the main guys who makes blender hardops so it's going to be good. I'm pretty good with topology now, can get some nice smooth curved surfaces with detail but there are some scenarios I get myself into that I find I don't know how to get out of.

>>15305785

Nice! thanks anon!

>CGPeers

Registered

>>15305079

ignore my reply here:

>>15306902

wrote it first thing after waking up XD

polite sage

>>15306626

>second webm

Hair brushing when?

>>15303880

>It's probably some instancing error with the node/scene setup you've done

And how can I check if I screwed up on that part? Godot doesn't complain anything about Node instance being wrong.

>>15304092

>just naming the function _on_Tank_shoot won't do anything without it.

Not even when the Player tank is inheriting from the Tank scene? By that logic I would have to copy and paste the same script for every Tank out there that exist on the map, so if I would have for example 10 Tanks on the map I would need 10 scripts for it just to make the shooting signal to work.

If I use this for the map script then the shooting works now.

extends Node2D

func _ready():
	set_camera_limits()

func set_camera_limits():
	var map_limits = $Ground.get_used_rect()
	var map_cellsize = $Ground.cell_size
	$Player/Camera2D.limit_left = map_limits.position.x * map_cellsize.x
	$Player/Camera2D.limit_right = map_limits.end.x * map_cellsize.x
	$Player/Camera2D.limit_top = map_limits.position.y * map_cellsize.y
	$Player/Camera2D.limit_bottom = map_limits.end.y * map_cellsize.y

func _on_Tank_shoot(bullet, _position, _direction):
	var b = bullet.instance()
	add_child(b)
	b.start(_position, _direction)

func _on_Player_shoot(bullet, _position, _direction):
	var b = bullet.instance()
	add_child(b)
	b.start(_position, _direction)

>>15306626

That first webm is excellent

I have an idea for a mechanic and I would like feedback.

The player wears different cloths. Each item of clothing is assigned to a different body part or set of body parts, sometimes with "bleed over" into other body parts. For example a t-shirt covers the entire torso, none of the neck, and some of the arms and hips, so it bleeds over into the arms and hips. Sometimes you can pick where you wear an item. For example you could put your pants on your legs but not your hips to represent sagging your pants and suffer a movement penalty or you could put the pants on your head if you wanted to roleplay a character that's literally pants-on-head retarded.

The cloths are stored in a list, with the innermost cloths appearing later on the list. So if you were to tuck in your shirt, the shirt would appear after the pants. But having an un-tucked shirt lists the shirt before the pants.

You can also attach some wearables to some other wearables. For example you could pin a badge to your shirt. Then any items that goes over or under the shirt goes over or under the badge as well. Also a byproduct of this method is that you could pin the badge to the inside of the shirt if you really wanted to.

All of this together has mechanical significance with something like a concealed weapon.

For example you can attach the holster to your pants. Attaching it to the outside of your pants makes for a quicker draw, attaching it to the inside of your pants makes for a more concealed weapon. Then you can either have your shirt tucked in or not tucked in. Tucking in your shirt reveals the concealed weapon, but makes it faster to draw, while not tucking it in covers the weapon and better conceals it.

>>15309313

Are you me? I had a similar idea

>>15309313

TF2 and Dwarf Fortress do something similar.

In TF2's case, it's for hat flags to determine what cosmetics can be worn together; eg a helmet could cover the face, which includes the eyes, so you can't also wear sunglasses with it.

In DF's case, each worn item has a coverage value associated with it, a unitless value that determines how much it covers up the underlying clothing or how bulky it is. Basically, to prevent you from wearing 30 cloaks at once, but also lets you layer clothing in an appropriate manner.

Actually, most roguelikes have a stack-like implementation like you do for clothing. You put on your enchanted underwear, then put on some pants, then you wear that belt you found, only to realize you can't take anything off because the belt is cursed

>>15309453

Neat. That bulkiness and stack-based clothing looks like a good idea. I think I'll steal those. I suspected I wasn't the first to come up with something like this.

>>15309420

This seems to be a natural outcropping of figuring out cloths in a game. "How am I going to store the cloths the player is wearing? What about a list? But how will I order the list…" and so on until the mechanic is fully formed.

Just letting /agdg/ know that if they need someone to do music for free just let me know. I just put out an album and some singles and would appreciate the feedback about it and if you want work sometime. I'll probably be busy with things, but I'll make sure I respond as soon as possible. Here's my new single, Ridgin' On, it is basically 8-bit Ridge Racer and it is in STEREO so get your headphones or stereo set to listen to it.

>>15309607

Forgot to post the track

Fuck you, Microsoft. UPDATE OPENGL! I'M NOT USING DIRECT3D IT'S SHIT. I JUST WANT GL_UNSIGNED_INT_8_8_8_8

>>15309313

I was thinking about a physics-based-clothing system just the other day.

You can tether accessories anywhere on your person so you can stick a bunch of ammo pouches anywhere in different places much like Kerbal space program. Then, the tether point for the pouch could only be one place and low so in game the accessory jiggles around a lot.

>>15310327

Also, it would be cool to have a weapon accessory system where you can superglue attachments on wherever so you can have 20 flashlights on a gun or even an ammo pouch in a gun

>>15310334

>>15310327

Sounds interesting, fun, and difficult to implement.

>>15307622

>It still irritates me thou that why the map scene has to the handle shooting part which sounds illogical to me if anything the weapon itself should handle it so either the turret or the tank node itself.

Putting the on_Tank_shoot method in the Map node makes sense because it's purpose is to add a bullet to the map scene. If it were in the tank or turret node you'd need to get the tank's parent and then spawn the bullet as a child of that node. This is easy in a simple scene, but a more complex scene may have several nodes of depth between the tank and the map (for instance, a Navigation and NavigationPolygonInstance node).

Anyway, the method would be better named "spawn_bullet".

>>15310393

If it's not hard, it can't be fun.

>>15309694

Why the fuck don't you just use GL_UNSIGNED_BYTE instead?

>Fuck you, Microsoft.

You mean Amd/Intel/Nvidia, right?

Does anyone have those low poly models someone was posting a few days/weeks ago?

>>15311892

>bcs despite going to uni for compSci

>bcs despite [me currently attending]* uni for compSci

polite sage

>>15311438

>Why the fuck don't you just use GL_UNSIGNED_BYTE instead?

Because GL_UNSIGNED_BYTE doesn't support streams of unsigned integers. A pixel in my buffer is one 32-bit unsigned integer, so if I use GL_UNSIGNED_BYTE then I everything is in greyscale.

>You mean Amd/Intel/Nvidia, right?

AMD/Intel/Nvidia aren't the ones who locked away everything that's >OpenGL 1.1.

>>15311921

>GL_UNSIGNED_BYTE actually works

Oh damn I look like an idiot.

Well, I still needed to use GL_BGRA, so my previous post still applies.

>>15311487

Work on shitty ms paint adventure games in The Games Factory when I was like 12 and had no idea wtf I was doing -> many years later learn javascript from codecademy -> work on various browser game/software projects over years -> learn C mostly from "intro to C" and early episodes of handmade hero. I haven't learned a thing from a book in my life though I've tried. I just do what I want to do, and search for answers when there's something I don't know how to do.

>What was your first finished project?

For game projects I've only "finished" a small idle game where you gathered various resources, fed npcs and built buildings. I want to redo that idea some day. For non-game projects it's hard to say what counts as "finished".

>What was your first failed project?

>How many failures/unfinished works/projects did you have in your belt before feeling balanced?

Don't think I've ever completely abandoned anything because I rarely work on something I don't have a strong interest in, I just lose motivation/interest and put it on hold. I still plan to work on all the things I have on hold. I've only really worked on a couple projects, but I tend to redo them completely when I reach a certain point and realize the current architecture isn't that solid. I think I've redone my main game project 3 times, though that was, as far as I remember, the first game project I ever worked on not counting The Games Factory shittery.

>>15311932

>Well, I still needed to use GL_BGRA, so my previous post still applies.

That's because you're ignoring the data format.

What you seem to want is a uint32 to represent one pixel. This however introduces stupid little-endian nonsense, which is something you have to deal with when you're moving data between different devices or big-endian files.

If you don't want to deal with endian bullshit, just don't do that. Just store/load your image as uint8[pixels*channels], instead of uint32[pixels]. Clean up the format of your data, and the simple elegant code will follow.

>>15312026

Yes, that's what I used to do but, a single UINT32 seems to be faster than 4 UINT8s.

>>15304021

and it with 3. any number anded with 3 will be 0-3.

What's the logic that makes total war combat work? Like, you have a bunch of little battalions that march around but each unit in that battalion has their own path finding and combat SEEMS like each unit actually harms each other unit.

I can't figure out how this is accomplished .

>>15312479

the original shogun used a neural network for the individual pathfinding i think

or something of the sort, i watched these ages ago

>>15312479

To clarify, i imagine that there's a logical object for each battalion which is used for movement/path finding. Then, each unit is assigned a number (as part of an array for example) and that specific number corresponds to a position for a particular formation. That AI's position is adjusted based on the direction of the battalion object and it path finds to that location. When a unit is killed then that element of the array is popped out and each unit with a lesser number moves up one place. In battle, each unit attempts to hold formation but fighting takes precedence so it will attempt combat until it can escape and reform.

There's a lot of optimisation in the engine naturallu because you can reuse the same textures and models for each unit in a battalion. I think I could make a more efficient total war…

Great… This is the third game I've become obsessed with completing despite having no time to actually work on anything.

>>15312498

Il watch these

>>15312893

did you check if the models actually have their armature exported? reimport them into whatever you're modelling with

>>15312979

They do, if I open the animation and try to select a root node all the parts of the armature are there.

They just don't seem to show up in the file itself

>>15313030

post screenshots, i'm clearly not getting something

>>15311932

Remember that this isn't a feature every implementation has. Query GL_EXT_bgra and if it's not present reorder the bytes to RGBA.

https://www.khronos.org/registry/OpenGL/extensions/EXT/EXT_bgra.txt

Sharing a tune I made during my days fooling around with PXTone. If somebody want some music for their game, I'd like to help.

>>15309611

Nicely done, mate. I'd like to hear this in an action setting.

>>15280943

Sounds comfy at the beggining and dramatic at the end, keep doing stuff like this.

What's your setup for music making?

>>15313308

is that only the project view? hierarchy only goes as far as 1 set of children there

did you actually try putting it in the scene

>>15313348

try changing the root bone in the import settings

>>15313348

>>15313365

also make sure Preserve hierarchy is true

>>15313365

>>15313376

I have, but the model has already been imported and the hierarchy has already been broken at this point, so it's not going to fix itself

>>15313385

try exporting as an fbx instead of blend, even though that's generally supposed to be worse for blender

>>15313388

>>15313385

Shit nevermind, I reimported a few times and now it's working as intended

My mesh hierarchy is still a mess but now the animations do look like they're working

>>15313792

i think you're supposed to pick the highest+closest ones

is there still a market for a text-based/low sprite open world lewd rpgs like coc without 90% dickgirl npc saturation?

>>15313792

I think there's a paper written by a Valve employee about this. Give me a little and I'll find it.

>>15313829

right, well i guess the ones closest to your preferred height

if they're above it, they're not climbable because you're a manlet

i don't know tbh, i just put colliders for this thing

>>15313792

Nevermind it was a power point not a paper. Starts on slide 29 and it's for AI not for player characters, but it should still be helpful.

https://steamcdn-a.akamaihd.net/apps/valve/2009/ai_systems_of_l4d_mike_booth.pdf

>>15311892

how helpful is python for creating a game? Is it a good language suited to that purpose?

>>15313859

I guess I will try reading this

>>15313956

Oh actually I see, they're looking for a point that is lower than the previous ones, that's pretty smart

>>15313985

>>15313956

You were right the first time. It is trying to find the edge closest to the character. But you could use it to find a lower ledge.

>>15314016

Oh no I get that, but I think they're finding the edge closes to the character by casting down and moving back and comparing if the hit point is lower than previous ones checked

>>15314025

Ah I see, that makes sense.

>>15313289

I'm glad you enjoyed it. I've been doing chiptune for the past four years, it is honestly refreshing compared to where I started composing on GarageBand with some help from my dad six years ago. Learning the ins-and-outs of certain sound hardware have helped me find my style since a lot of modern chiptune composers seemed to emulate Tim Follin or Jack Kaufman which is not bad. FM Synthesis is incredibly difficult, so I staying with 2 operators for now until I have the time and mathematical knowledge to understand four operators.

Also, I liked how your composition really emphasizes the echo and simplicity as if the character is inside the cave. My composition only uses 12 monophonic channels at most(5 from the 2A03[percussion, bass, melody], 4 from N163[chords], and 3 from the Konami VRC7[Bass, melody due to how the chip handles with custom instruments I could only used 3 channels at most].

>>15313943

Your piece kind of reminds me of a theme song to a soap opera which is fine if you're aiming for that JRPG-esque style. Also, nice usage of your melodic motif. Hope you find your style in the meantime.

My music education spans from elementary school to early high school, played in many bands inside and outside of school, and I grew up learning the keyboard and doing some percussion at school. Composing for me was like a seed, started making up introductions to songs for my piano recitals then eventually here composing on 0CCFamitracker and Ignite which is really cost-effective for the kind of music you can create although I wished they added more selections for woodwinds, brass, and strings. Ignite is only 49.99USD, so if you are on a tight budget and have a Oxygen product that is not registered you can get it for free. Chiptune is also good, but there is a huge learning curve to optimize how your composition sounds on certain sound chips. Fun fact: Most of the samples I used on FamiTracker usually consists of some percussion samples I salvaged from my hard drive, samples ripped from NES games, recording my voice on my phone then converting it on the program, and using some stuff from Ignite. I believe good music comes solely from a good composer NOT good equipment.

I guess I should showcase what Ignite is capable of doing, so here's something I made for my English class that wasn't used. I was trying to do an audio interpretation on the novella, Heart Of Darkness.

There's one thing you have to consider when using Ignite, and that's the DAW doesn't support VSTs but you can get some expansions which I highly do not recommend since they are just simple sfxs and percussion loops.

>>15314054

>>15314025

>>15314016

>>15313985

they are actually just moving back until no more raycast hit, therefore that is the "correct enough" position of the ledge.

They don't even care about the ledge orientation

>>15312796

It looks so damn great, I hope you end up making a shitload of money.

>>15289090

As long as it's just pixels and not TILES

>>15311487

>How did you guys start your journey in game development?

When I was 13 (so about 15 years ago) I messed around with a shitty program called RPG Toolkit. It was pretty terrible in hindsight. Think RPG Maker, but done by a 1MA with 2002 tech. As bad as it was, it introduced me to scripting and basic programming structures (if, while, etc). Obviously I was too young to make anything worthwhile.

Later, in 2007, I joined the development team for Mordor XP, which was a community mod for an obscure Windows dungeon crawler. Fell apart in a year or two, but I picked up drawing in Photoshop around this time, and contributed in that way. I still wanted to work for videogames, so I went off to college.

My 2-year course ended up taking 4 years to complete, because I didn't want to move to (((Toronto))) and also had difficulties finding a suitable co-op placement (I also failed my final semester because of WoW and had to redo it. I later sold my account to a friend for $200). The education was pretty awful; they did not cover a C-family language, and we had two semesters introducing us to Java. They didn't even get as far as generic typing. So, in a way, I feel all my coding has been self taught.

One co-op was at Blackberry doing server tech support for their enterprise clients (Outlook emails pushed to mobile devices). At this point they were already sacking people left and right due to their layoffs so I wouldn't get rehired. The one client I can remember, they wanted a simple phone list for all their connected devices. Simple enough, right? I didn't know how to do it and the solution seemed to be to clone the server into a new environment and import the list and view it that way. Turns out this was a bad idea and I let them dictate how to do it, and brought down their email server for about 3 hours. US Department of Treasury and Mint, New York Division. So that was fun.

My other co-op was at a small local game studio. They used XNA framework, and I think that's why I stuck with C# and XNA for so long, before finally moving to SFML last year. I didn't contribute much, I almost feel like I wasted their time. They did make a game before and after I worked there, so that was cool. I mean, you can check out Gravonaut on the iStore, but it's ancient by now.

So basically, I messed around with a little bit of everything, learned algorithms and more compsci stuff on my own after graduation, but still haven't just like made game yet.

Oh, actually one other fun thing I did in high school (using Turing). I made a shitty "dodge the falling balls" game without OOP somehow, since instancing confused the fuck out of me. I put it in a shared class folder, but I also made it spam the user's personal drive with hundreds of empty folders every second. I only got one person with it, and I had to stay after class and delete them. I just sorted by date and it was no problem for me, could've gotten in a lot of trouble

>How much did you know before going in?

Nothing really. Just slowly accumulated knowledge, plus focused learning on whatever topic I needed at the time.

>When you first started what resource did you use to help you learn? Youtube tutorials, torrents, books, pdfs, or just reading the documentation?

This was before YouTube. Or at least, I never used it. RTFM is a fantastic skill to have and will carry you far. Discussion boards about how to approach certain problems, StackOverflow/QA sites are useful, too.

>What was your first failed project?

All of them, technically. I haven't finished a single playable demo in the 10+ years I've been hobbying at it.

>What was your first finished project?

I made a "game" where you can bump into a block from a certain direction to activate it, and then flip it into an adjacent tile, reversing the allowed directions. It was more or less a prototype and the drive I had it on seems ot be slow and failing, else I'd show it off, even though it's a non-game

>How many failures/unfinished works/projects did you have in your belt before feeling balanced?

I mean, I must have lost count of how many projects I've started and abandoned by now

I've been active in these threads since 2012 or so, back when "Motavio" was a thread mascot and we had Recap Mondays. I remmeber Catmouth Island was a popular game at the time.

>>15311487

>How did you guys start your journey in game development?

Making scripts in RPGMaker XP and attempting to recreate cool things I saw in FF7/8 and others RPGs of the same years. Of course, I never managed. But I did a script that enabled you to use the mouse to move/click menus (although the move part was taken from another pathfinder script)

>How much did you know before going in?

How to open a browser and yahoo things

>What was your first finished project?

Never heard of those

>How many failures/unfinished works/projects did you have in your belt before feeling balanced?

There is no such thing

Just like start some shit, even if it's too wild of a dream, you will certainly learn something along the way that will help you in your next crazy project.

I don't even care too much about my current project, but I know that if I don't finish it, I will never finish anything. This is the best chance I ever got on finishing something, and I will not let it go away for nothing in this world

>>15313876

>how helpful is python for creating a game?

Python is a great entry level language.

Because python is an interpreted language there's no need to wait for compilation, and it's suited to the purpose of learning the very basics of programming (using logic in a programming context).

Also, I have some good news, if you're a logical person you will have a natural affinity for programming logic, as you have a network of transferable knowledge, and it will just be a matter of your brain adjusting those neuronal pathways to a programming context. Thus, python is useful here bcs it's easy to use, and you can get more practice time in faster w/o having to mess with basic shit; as to solidify those logic pathways in your brain for a programming context.

After one solidifies the understanding of said logic, I'd recommend to learn a low-level C language (preferably C++ or C), and learn the low-level aspects of programming. This is extremely relevant to gamedev as games have a real-time performance requirement; which requires knowledge of smart memory/cache/cpu usage practices. Note: the importance of understanding these low-level aspects cannot be understated (pointers/memory addresses, heap/cache, and how this interacts with the CPU).

The logic of these steps are rooted in this idea: take the most efficient route to building transferable knowledge.

I.e. solidify an understanding of programming logic via python (transferable to all languages), and to then learn the low-level aspects for how a programming language interacts with the CPU (transferable to all languages).

These are the basic foundation stones to learning to code in a gamedev context.

However, there are additional building blocks of knowledge one needs (design patterns, high to low level abstract, design principles, language specific quirks, etc), but with those cornerstones of transferable knowledge any language will be quite accessible to you.

>Is it a good language suited to that purpose?

Python isn't suited to games, and with enough time produces bad-practices that one learns due to the intrinsic qualities of the language (very high level language, interpreted, etc).

It's really only useful for practice/learning, and shitting out some quick logic w/low effort + no concern for performance (that's one reason it's so popular in the scientific field).

>>15317174

>Not pictured

The rust boat refusing to move because the captain can't mathematically prove that there aren't rocks somewhere in the ocean that the pilot could potentially navigate over.

>Also not pictured

The two men with paddles on the inside of the ship actually pushing the damn thing around.

>Also not pictured

That the boat takes four days and nights to unmoor, but only the first time, hopefully.

>What is actually pictured

An unmerged PR against the develop branch which only compiles on nightly with a handful of opt-in unstable features using a very specific set of dependencies.

Come back in a year or so and you might actually have something developer ready which looks vaguely like that.

>>15317174

The one thing Rust succeeds at is being an EVEN HARDER language to write in than C++. To think that it also runs slower and takes 10x as much time to compile as C++, a language that already takes ridiculous amounts to fime to compile.

>>15317214

you could just check the bounds of every collider in the scene yourself, but that's about as tedious as using ontriggerenter

>>15317222

It's very fun, but a lot of things you take for granted in C are just not as stable or battle tested in rust. It causes a frustrating amount of friction, especially if your project is in an area where you're not comfortable enough to just dive into the internals of whatever libraries you're using and start fixing things yourself.

I don't "work" with rust, no one does. But I've written enough hobby rust to have a love/hate relationship with it.

>>15317255

So basically it's a toy language.

>>15317262

It's not racket levels of toy language, but it doesn't make me want to fucking kill myself like half the languages currently in production do, so it must be at least slightly a toy language.

>>15311487

>How did you guys start your journey in game development?

BASIC on the color computer 2. Later C and some assembly on the Amiga 500. Then Turbo C and Turbo Assembler for my 486dx2. So yeah, like 1986 - 1996 self-teaching. I finished a few little games along the way. Now I just mess with GPUs exclusively; OpenCL and GLSL for simulations and even a few simple games.

>>15317222

Rust isn't at a point that it can be taken seriously. Maybe someday. It has a very long way to go.

>>15317465

so, no one interested in some 3D modeled clothes right now? I think my latest "client" died or something, they don't respond.