/v/ - Video Games

Because of the steam sale I'm going to throw some shekels and good reviews towards any games active anons have on steam. Are there any other than Hellbreaker and Speebot? inb4 >steam. I found a gift card in a used book

>>15050055

Is that purple artifact Z-Fighting or the result of something with the portal?

>>15050099

It's Z-fighting, the portal isn't being properly clipped by the leaf behind it and it needs to be split again, that would delete the fragment intersecting with the polygon.

>>15050099

Keyreal has another game on Steam, a vertical shmup with anime girls, but other than that I think that's about it for 8/agdg/ projects on Steam.

Anyone got recommendations for a C/C++ library I can use to create a window and blit an array of RGBA values out into it? I used SFML last time, but it seems it might be a little heavy for that purpose alone

>>15050296

Anything you use will ultimately be a wrapper around the underlying window managers (win32/xlib). But if you don't want to go that far, SDL 1.2 looks good for this purpose.

>>15050334

hey buddy, I think you got the wrong door

>>15050356

Well fuck, I'm dyslexic today. Sorry.

>>15050305

Don't you tempt me sigmadev, but SDL does look straightforward, thanks

>>15050529

If you want to write OOP shittery, then why are you using C?

>>15050453

Just make sure you're using SDL 1.2 instead of SDL 2. 1.2 is much cooler, it has true software rendering and on Linux the fullscreen mode uses XF86VM to change the video modes for real (SDL2 fakes it). It also has way more rendering targets.

>>15050414

What is your first call? Where do you create your OpenGL context? It sounds like your context creation is failing.

>>15050529

You've got to free yourself from this mindset to enjoy writing in C. Try and restructure your code to stop using objects and see how that goes. It's better when you stop thinking about higher level structures and instead start thinking about the components that make these structures and how they interact with each other.

>>15050529

If you're not doing Data-Oriented Design then don't use C.

>>15050573

>It sounds like your context creation is failing

Problem was that my window creation function did some stuff which ended up in a roundabout way calling windowTick too early. In other words I'm retarded.

>>15050453

I used to have the mindset of wanting only cross platform libraries, but once I realized how quick it is to pick up the native apis and how much more you learn about the platforms I realized it really is the better way.

>>15050806

Nice, looks good.

OP is a sperg, he also made the OP with Bin Laden and mice.

>>15050930

Ein bisschen vermisse ich den Doktor ja schon.

>>15050529

there are more than 2 programming languages anon.

>>15051249

>implying they can compare to C/C++

>>15051249

for some reaosn in /v/ they really shill C++, i remember back in 2014 that people hated java whitout knowing all the frameworks and optimizations, nowdays java got even better.

>>15051269

you can because C++ is based on C.

>>15051287

You seem to be hard of comprehension. I said that all other programming languages cannot compare to C/C++.

>>15051287

There is no excuse for Java. If you must, go for C#. Java is the flash of programming languages.

>>15051269

>>15051302

I've heard good things about Go (mostly for the CSP-style concurrency, which you also get in C through Plan 9's libthread) but the garbage collection hurts its usefulness for most games.

>>15051345

Limbo exists too, you know.

>>15051451

I'm not touching (((Go))) or (((Rust))).

>>15051287

Java is good because I can immediately remind people to use C# instead

>>15051520

Javascript it is!

>>15051522

javascript > no game

>>15051525

Look at EvaXephons game and tell me that's not a work of absolute beauty

>>15051469

>(((Go)))

Come now, I'm pretty sure Rob Pike isn't Jewish.

>>15051287

Java is shit for games and always will be

>>15051330

C++ is shit for games and always will be. All the game devs who use it ignore almost all of the features is brings. Inheritance, exceptions, templates, you n ame it.

>>15051555

Name a good alternative

>>15051555

I've played some pretty good games made with Java, I don't know what you're talking about. You don't really have an argument there about C++ either, just some strange generalization about game devs that isn't true.

>>15051646

He says it's shit for games but his reasoning is that it's not used properly by devs. He's either shitposting or doesn't know what he's talking about.

The language war shit is one of the worst parts of this thread, lots of broad accusations tossed around and very little actual discussion.

>>15051688

If Java brought anything to the table, we could discuss that. Instead, all we can do is agree that C# is the best compiled, managed language currently available.

>>15050806

This is OP on the phone, yes it is about a journey to being a game developer, the backdrop is a map of Poland but this isn't a journey in the physical world, although it does take place in Poland because of how many people here who are polish

>>15050917

Like the other anon said you really should have a grasp on who the Arab dictators are, I mean you did learn your state capitals right? It's kind of like that .

>>15051688

>>15051799

[your favorite language] is objectively shit and you're shit a retarded nodev if you use it. All the respected pros and intelligent people use [my favorite language]

>>15051854

Forgot I was posting a vid

>>15051858

While you're at it I can post a similar video too. it's been posted before by me and others a lot but it's relevant here

>>15051879

Here's the mars rover one that guy mentions.

>>15051799

>c#

>compiled

>this is what the average c# """programmer""" believes

lmao

>>15051854

>>15051858

>muh caches

So your argument is that there might be cache misses and that means c++ is shit?

into the retard bin you go

>>15051330

i never said that C++ was bad, actually its better if your game damands too much graphic hardware performance or makes complex/too many CPU calculations (like supreme ruler were every unit in the world map makes various dice rolls every game hour/turn).

>Jesus this thread gets worse every time i take another look

now you know why.

>>15051688

you are right, it is full of programming laymans here, there is no much reason to use C and C++ if you are making a small side scroller game, you could use java for that, but if you make a game whit fluid simulations you should go for C++ since it interacts better whit the hardware (still java has already the framework and libraries to do the same). Anybody who says that there is a definitive programming language just doesnt has professional experience.

>>15051898

>So your argument is that there might be cache misses and that means c++ is shit?

Not might. WILL. For example: Inheritance has a cache miss to hit the look up table, and another cache miss if it has to load an overloaded function. Even Java handles inheritance better, because the JVM can optimize around useless inherited functions. Only solution is to avoid inheritance. This vid outlines that, and how hard cpu cache rapes performance. But clearly you're more interested in shit flinging than learning

>>15051916

Wrong vid. It's amazing how many of these there are though.

>>15051898

Yes? Raw C# code is compiled into an intermediary virtual machine language, which is abstracted from the coder so it just works across any hardware configuration. That's literally the point of virtual machines, portability.

>>15051990

It's interpreted not compiled because the resulting code is not native.

>>15051920

Since you seem to have archived all the c++ videos in the world why don't you find the ones that show you how to use inheritance propely and how to avoid common performance mistakes?

Do you also believe everything should be written in c?

>>15052027

If you want to be technical, we're both right. Even though the virtual machine code doesn't belong to a real, physical computer, the raw code is still compiled to that "native hardware", so your code ideally runs against the VM as quickly as possible.

>>15050573

Well I started with SDL2 but the whole "#define main SDL_main" thing is making problems in my environment and really turning me off. Is it that way in SDL1.2?

>>15050052

>zero cost abstractions

>zero cost

ahaha

>>15052285

Literally just used a fall through case statement today. Why would you intentionally limit yourself.

>>15052027

>It's interpreted not compiled because the resulting code is not native.

Compiling doesn't require something to become native assembly code. You're latching onto a partial definition and then insisting that everyone who uses the actual full definition is retarded. I swear half the people here are illiterate and use that as an excuse to act like a fucking idiot.

>torpedo

goddamn they just keep coming

>>15052108

I don't know about actually using SDL (as I use the underlying API's) but I have looked at the feature set and some of the code but only a tiny amount. I would just check it out since it's the window manager, ultimately not that much code for you but I don't know how the API actually is to use.

>>15051898

>muh caches

kys

>>15052284

>instead of warning you your code needs fixing, it just does random shit you can't predict in absolute stealth leaving you wondering why the fuck your function is returning arcane and impossible results according to your logic

Sounds about right. Why nobody has tried to enforce a better language to be a standard in web development is beyond me. Javascript is the worst thing mankind has produced, and that includes kikes.

>>15052486

>Sounds about right. Why nobody has tried to enforce a better language to be a standard in web development is beyond me.

WebASM is supported in most browsers, and has JS fallback for IE. Most people who doesn't want to use JS just compile it from some other language, and osme of them are just JS with less bullshit/features.

>>15052486

I've spent hours trying to debug code because number+undefined = NaN, and number+NaN = NaN, which creates a chain reaction from a single mistake and everything becomes NaN without giving any error or clue where it originated from.

>>15052759

The joys of programming.

>>15052759

To be fair, that is defined behavior

>>15052880

Technically javascript's autocoersion is defined behavior but it's still retarded.

>>15053096

Oh, even in C#, that's how floats work. As soon as you pass N/0 or a value with infinity around, other numbers become it as well.

>>15050529

Use Nim

So I'm still plugging away on my QuadTree implementation. If it creates a bunch of child nodes, then gets values removed from it, should it still keep the empty subtrees, or should it dispose of them?

I think it should still keep them, unless it is specifically trimmed away.

>>15051555

But you're objectively wrong. Running with Scissors used templates and inheritance for Postal 1.

If I have a 512x512 quadtree, and I tried to set a region to a particular value, then it should only take one call if it's the full size (the root gets set to that value). If I tried to set the region (255, 255, 2, 2) to a value, it should take 40 calls, correct? It would have to find the smallest tree that completely contains it, which 512 log 2 = 10, x4 (one for each quadrant) to find the 1x1 region in each

>>15054363

What, no, that's not at all how quadtrees work

>>15054416

The problem is OP's autistic normalfag C/C++ meme. I'll abstract your face at zero cost.

>>15054416

It's alright, my dear Lainposter. Conflict is the path to life.

>>15054430

>>15054416

>>15054416

You didn't cause it, you just made a shitty contribution to it

>want to make progress on game before I go back to wage slaving

>getting warning because a class conflicts with itself, in its own file, in its own namespace

I bet I can ignore this shit and get away with it for a while

SHIT

>>15054771

nope never its too hard

>>15054771

>beta after only half a year

You are either a speed demon or using an engine.

>>15054632

>>15054657

Okay, so I think my confusion stems from the different types of quad trees.

>Region quadtree: Every leaf node has one value that applies to the region, eg a pixel grid

>Point quadtree: Every leaf node splits at a point, the point is at a "corner" and then it gets divided to its children, ending when it only has one value per node

>Point-region quadtree: As region, but instead each leaf can have any number of points

Point qtrees also seem to be a KD tree, which is a generalized binary tree of N dimensions, so I shouldn't worry about it in my qtree implementation, then, right? But I'd still need to store the XY position and value per thing

>Point-region (PR) quadtrees[4][5] are very similar to region quadtrees. The difference is the type of information stored about the cells. In a region quadtree, a uniform value is stored that applies to the entire area of the cell of a leaf. The cells of a PR quadtree, however, store a list of points that exist within the cell of a leaf. As mentioned previously, for trees following this decomposition strategy the height depends on the spatial distribution of the points. Like the point quadtree, the PR quadtree may also have a linear height when given a "bad" set.

>>15054804

>or

>>15054805

I'm still no closer to figuring it out, but I just remembered why I'm doing this in the first place

>Trying to make sprite drawing easier, so made a vertex batch to manage my draw calls

>Still need to optimize it with draw batching

>Could make my own SpriteSheet class to manage stuff, which would rely on atlasing

>Sprite atlasing would be another optimization in itself, on top of sprite sorting, the result should be very fast and simple draw calls

>Atlasing would require stitching together multiple textures at runtime

>Best achieved through quad trees

Polite sage

>spend a few hours on a feature

>realize it's fucked

>I'd be better off starting over

>It's sunday night

well shit

Anyone have that image on why you shouldn't use unreal because of tencent?

>>15055746

Are these videos broken for anyone else? I switched to linux and they're not playing here despite working locally

>>15055750

They're broken

>>15055750

Same.

>>15055780

No idea what the fuck is going on here

>>15055785

works for me

>>15055857

>explaining pointers

explain this

int* p
int* p const
const int* p
const int* p const
int const* p
int const* p const

>>15055870

Why bother explaining esoteric language abuse that nobody will do in practice.

>>15055746

nice

>>15055870

Because you can do this

int *a const, b, *c;
const int* d, *e const, f;
int const* d, *e const, f; // identical to the line above
// a is a constant pointer to int
// b is an int
// c is a pointer to int
// d is a pointer to constant int
// e is a constant pointer to constant int
// f is a constant int

>>15055870

const int is the same as int const.

Always choosing the latter, i.e. putting the qualifier to the right of what it's referring to and putting the asterics close to the name makes it easier, imho.

Try reading it from right-to-left now.

int *p // pointer variable to integer
int *p const // immutable pointer variable to integer
int const *p // pointer variable to immutable integer
int const *p const // immutable pointer variable to immutable integer

Keep in mind, when we have int *p, we have 1. a pointer variable (*p) and 2. the type of the thing it's pointing to (int). So, we describe not just one but two things.

>>15055573

I was just hoping someone would have it on hand because I lost my copy somewhere in the thousands of folders I have.

I have reason to suspect it was made by a tencent shill as it only displays """""intellectual property theft""""""" instead of all the shady shit tencent does. Just a hunch I had thinking about it and I wanted to check and make sure.

>>15056023

>i hate meme retro shit

then why are you posting it

>>15055870

Read right to left.

>>15056103

this, but also I'm autistically angry that the pixels are square and not stretched to fit a 4:3 monitor like they would be on real hardware and a huge faggot for fucking up the spoiler and reposting

>>15055746

>cuck armor

<I'm a male feminist, I can't be a rapist!

>>15056023

What program?

>>15050529

C is a very good language I am making a roguepike in C very good I hope so too. Pythong is a good lakeuade too but maybe it if is slow and fps is impostant for a rogues.

>>15056866

It's one of those Fantasy Console things. It tries to recreate having an old BASIC computer with some helpful software to help make games, like a spritesheet editor, map editor, and sound/music creator. Instead of BASIC though, you use Lua and I know with Pico-8 you can send your games as playable-only image file. The downside being, for all of them besides LIKO-12, you have to actually purchase these fantasy consoles.

>>15056943

You look into your target hardware and learn how to make the best of your registers and cache, and when it makes sense to improve data locality or use a different data type. Maybe read some of these: http://agner.org/optimize/

On a more practical note, you profile, measure and if it looks like you're spending a lot of time in one place, look into making that particular spot suck less.

Don't bother with low-level optimizing until you need it, otherwise, you might end up shaving cycles off of insignificant areas or having to completely rework your optimizations when you think about restructuring your data.

Getting 10x the performance in 0.1% of your overall workload is likely only going to make things less readable; find the hotspots and measure both before and after to make sure your optimizations are both in the right spot and are actually having an effect.

>>15057264

Singletons are referenced with their name directly, like a global variable. You shouldn't be using get_node().

If it's a YouTube video just set the speed to 0.75x.

>>15057861

Forgot to mention it but I modified real-life firearm designs to make firearms for them and added the photos in for comparison's sake.

>>15056943

Now try to partially unroll that for loop to avoid checking for the condition after 1 addition for more speed.

I'm really curious on why would the change you made be faster, you can't just benchmark it once, you need to do it hundred of times and average the time, debug/release can have a major change on the result, I got tricked once when "slowest" version of my code was insanely optimized by the compiler after some flags it made me doubt the whole thing.

Seriously tho, why would allocating a new variable, adding to it then overwriting another variable be faster than just adding to a variable?

>>15057861

Make a shortstack female one that's a friendly NPC you can take along as a companion

>>15057861

Maybe give them pistols to carry as 2handed weapons instead of SMGs to emphasize how small and weak their are, at least at the beginning. I think objects that are immediately recognized as small compared to a human would be better to convey they're midgets and not just shrinked down humans (SMGs look like shrinked rifles). If you plan on making more variants of the goblins you could still use SMGs for elite goblins or something like that

Hard to tell whether they should wear gloves and boots or not if we don't know what their hands and feet would look like, but I think this design looks pretty nice. Though I'd like to see how a goblin looks like next to a human, because alone he just looks like a chibi stalker.

>>15057960

A little bit of the context is missing, the m_Total is somewhere else, so writing into it causes you to access that memory location on each iteration, which is a lot slower than writing into a value that's close in memory.

>>15057983

Quick sketch. I feel like if I did that I'd have a lot of work to do to make it not look like Shy Gal

>>15058031

In my sketchbook I have drawings of them where the gun is as large as they are. I kind of lost the plot as I was drawing these digital drawings. I was thinking about emphasizing how weak they were in their animations where they'd only be able to fire in bursts because the recoil almost knocks them over One of the animation ideas was a Goblin hovering over the ground from firing into the ground and recoiling upwards.

They're hands and feet are just simple all-black with little triangle feet and 4 (including thumb) pointy fingers. Here's a scale with humans and their Elite style counterpart included.

>>15058048

Somewhere else how? variables that are used repeatedly are cached unless your values[i] is so big the CPU can't keep both, it's not like your dealing with pointers, is this related to scoop? are local scoop variables closer to CPU than normal variables?

>>15058141

Yeah, fair enough. The way they look now is pretty unique. You should still definitely put in a friendly female one

>>15058165

A variable in a class (hence a member or m_ variable) is declared on the heap, whereas a variable declared in a function is declared on the stack. Stack access may be faster depending on CPU and its addressing mode.

That said, if you look up "is stack access faster than heap access" you get a whole lot of people claiming its either true or false, and it certainly shouldn't be 12 fucking times faster either way.

>>15058141

>shy gal

Nah, giving her irises and filter make her completely distinctive. I assume the gask mask won't be white, so the color is also in your favor. Also, that design with the gun has much better proportions than the previous one.

>writing program

>Have array of size 2,000,000 bytes

>Worried that it's going to saturate most of my RAM because 8GB = 8,000,000 bytes

>wait….

>Remember 1GB = 1,000,000,000 bytes

brainlet.jpg

>>15058579

If you find yourself needing a 2GB array, chances are you're doing something wrong.

>>15058654

It's 2 MB, guess you're brainlet.jpg as well

>>15057960

m_total is most likely a class variable, as that is standard C++ notation used by some (weirdos). That means the variable is not local in memory and is actually accessed using this->m_total, where, of course, this is the pointer to the class instance.

But, I believe the any modern compiler would take care of this. In fact, at -O3 and expensive_optimizations GCC will probably unroll that loop.

>>15058654

>If you find yourself needing a 2GB array, chances are you're doing something wrong.

No. Chances are you are working on very interesting and/or cutting-edge problems.

2GB array is really not all that impressive or unusual in scientific computing.

>>15058767

You might want to look up array-of-structs versus struct-of-arrays, or AoS vs SoA for some information on what you might be dealing with. The latter tends to be more performant in many cases, and the former might be easier to wrap your head around with less indirection involved.

I went with the SoA route, except I just have one big manually allocated buffer for each struct.

The way I have things set up, I have an unsigned integer as my entity. It's just an index.

Then a struct for each component, holding the amount of active components and some other information, each with some fields for data. Except those fields are all just pointers and their data is held in one contiguous buffer.

And related to each struct is something to map every entity to an index in each struct.

Each component has its own updating system that just hammers through all the data in the struct.

For example, all my graphics-related things are in one such struct. So every frame, everything in that struct gets processed and possibly drawn or skipped, depending on the data.

So I have an entity with a graphics component and I want to animate it. So I tie an animation component to my entity. The animation component, based on its own data and logic of what frame to play when, updates the data in the graphics struct corresponding to that entity. If the animation component ever gets destroyed, it'll still draw the graphic, but just hang on the last frame. If the graphics object gets destroyed, it'll animate a dummy graphic without any output and not horribly kill the program.

So far, I've really been liking it. It makes sense and it's fast. It also allows for some really quick spaghetti-ing and hacky code if you just need to see if something works.

>>15058813

>Chances are you are working on very interesting and/or cutting-edge problems

Such as? Which videogame feature requires 2 gigabytes in a single array?

>>15058767

organizing things in structs doesn't make it OOP anon, generally you only need a source rect, the destination rect is barely used unless you want to stretch a texture, even then you can just draw transform instead.

>>15058952

>not having at least 2GB of particles on screen in every frame

>>15059030

>>15058952

>he doesn't load a 2GB collection of reaction images into RAM for convenience

Plebs.

>>15059846

>if you only have one

I meant

>if you only have one, get_nodes_in_group("groupname")[0] will give you that specific node no matter where it is in the hierarchy

>>15059001

>organizing things in structs doesn't make it OOP anon

As a C# programmer, I am confused by this. It's said that C has no OOP features, but a criticism of C# (and other managed langauges) is that it treats everything as an object. How is a struct or a pointer/ref not an object? Or rather, how can you have these things and not consider them an "object"? In either language, they're first-class language features so they can be used very flexibly.

>>15060058

The criticism is not the idea of an "object" but the application of this idea to every single thing. In C structs can be used for the same things you would use objects for, but the reason that it's not seen in the same way is that it doesn't have to meet any of the other requirements of an object in OOP, in that structs are used as a data format and not as an agent in a system of objects.

>>15060058

What >>15058951 said, particularly struct-of-arrays bit I think explains some of the confusion. It's a pretty good explanation.

It is a different way of thinking, and if all you have ever done is OOP it might not even make sense or seem possible. The best way is to read other peoples' C code written in a functional and/or data-oriented design.

My main complaint about a functional style is that you end up passing around data and pointers everywhere. At least I do. Classes allow you to naturally associate data and functions when it makes sense. I end up using a mix of classes, C-style functions grouped into namespaces, and integer indexes into huge arrays of data (as >>15058951

mentions), which I think qualifies it as 100% spaghetti coding style.

>>15060270

Sure but if a struct is just a bunch of data, you can do the same thing with C# - a struct here is an object that represents a data/value and has pass-by-copy semantics applied to it. You can make functions that have a ref parameter too so it ends up being passed by reference, rather than creating a new local copy. You can even enter an unsafe/manually managed context to work with pointers.

I just don't understand the difference between the two, since you can pretty much do whatever C does, however you want it. The only thing that's very obviously different to me is automatic memory management, but even thing, you can suppress or call the GC manually when doing certain things, too.

>>15053784

>we delicious flat chest now

>>15060404

Ultimately you can program in different ways even if its not as fluid depending on what language you use but the distinction will come in the mindset that surrounds the design and use of the language.

An "Object" is more than just the data that an object would contain. It's the idea of an object as an agent that separates the two. In C#, the object can control how it's values are accessed via getters and setters, it has public and private interfaces and variables, etc, all these things establish it as an agent instead of as just data. This agent is also related to other agents, with inheritance, etc. The program is a system of agents that contain data and interact with each other.

In C, the program is just a system of procedures, when a data structure or some structs are created there is no regard for ownership or relationships, they are just passed in and out of procedures for the whole program. In that way the data does not have any agency and the agents in the system do not have their own data, just the data they are working on right now.

Yes of course you can program in both ways in both languages (file I/O in C/UNIX establishes a file object as an agent of the filesystem, just like C#) but the idea is, how the data is considered and thought about, not as much about whether you can do OOP in C or procedural programming in C# or not.

You could watch this for someone talking about how OOP works vs procedural to hear more about this viewpoint, it might make more sense after.

>>15060856

No I think I get the gist of it now. C# has more meta information with its things, thus an int value has type data, instance information, functions and methods attached to it, even though it's "just" a primitive number, whereas in C, it's literally just a chunk of data in the computer.

>>15060891

That's the difference in the implementation, but I think that the real separation between the two is where the agency is. In C a function has all of the agency on it's structs, while in C# the objects all have their own agency and interact with each other by imposing that agency on each other. So where in C you can write a pass_message() function to impose agency on your data, in C# you make a "messenger" agent with it's own data to pass messages between other agents that all have their own data too. So you can have code in C where some conditions are checked and then the operation is done, while in C# since these conditions are a part of the agents you are operating on, those checks would be done in a getter or setter, and then the pass message agent would do the actual message pass. So the code is separated out into multiple places since each object has its own agency. Meanwhile in C there is just one function that has absolute agency over the interaction between it's data.

Maybe this is dumb, I just wanted to ramble about this since I was thinking about it a lot…

>>15062088

What compiler, about how many LOC, and how many source files? It sounds like the biggest problem here is the amount of #include's you have, are you adding #pragma once to every file to ensure that everything is only included once? That should help reduce these issues a lot. You also might be using a compiler that is just slow, or your CPU is slow. What CPU/OS are you using? You might be able to switch to a faster compiler. It sounds like your build script is just building the whole project all over again instead of compiling just the changed files again… which it shouldn't be doing. You can set up your makefile to only do that- but I don't use makefiles so I don't know exactly how that works.

>>15062088

To greatly speed up C compilation, compile the .c files into .o, ignoring .c files that don't need changes, and then link the .o files into an executable.

By passing -c to a compiler with a file, it'll spit out a .o file. Then you can just toss all of the .o files into the compiler again to link.

cc -c main.c
cc -c game.c
ld main.o game.o

In addition, as >>15062139 mentions, use #ifndef guards (or #pragma's like a pleb) to prevent the same header file being parsed multiple times.

>>15062139

>What compiler

GCC/MinGW.

>how many LOC

754

>how many source files

8 .c files and 7 .h files

>are you adding #pragma once to every file to ensure that everything is only included once

Fuck, what page of B&R?

>It sounds like your build script is just building the whole project all over again instead of compiling just the changed files again… which it shouldn't be doing

Yes… I am just using a .bat file and it looks like this:

@echo off
cd src
gcc main.c -o ../obj/main.o -c
gcc graphics.c -o ../obj/graphics.o -c
gcc engine.c -o ../obj/engine.o -c
gcc fleet.c -o ../obj/fleet.o -c
gcc star.c -o ../obj/star.o -c
gcc playerInfo.c -o ../obj/playerInfo.o -c
gcc research.c -o ../obj/research.o -c
gcc spritefont.c -o ../obj/spritefont.o -c
cd..
cd obj
gcc -o ../a.exe main.o graphics.o engine.o fleet.o star.o playerInfo.o research.o spritefont.o -lgdi32
cd..
echo Build finished

My PC is fast, so I think it's my fault, unless Windows 8.1 has some kind of problem with MinGW.

>>15062156

>ignoring .c files that don't need changes

How can you do this in a .bat script?

>>15062173

>spoiler

I meant K&R

>>15062173

>How can you do this in a .bat script?

That is a really good question and I don't know. I don't think bat scripts are meant for that sort of thing.

Even in makefile, the way to do that is a bit fucky, which is by making this the last line of your file:

-include main.d game.d

You also have to pass something to your compiler to make it produce depency files. Apparently -M is the flag for that, and there's a bunch of variants. (I use -MMD. Cannot recall why, don't care, it werks.)

>>15062189

>Merging everything into a master header file

You sicken me.

>>15062088

Another thing you can do is minimize use of #include in header files, instead forward declare an empty version of what you want to include in the header and move the actual #include into the source file.

>why is this useful?

Imagine A.h includes B.h

If you change B.h, A.h has to be recompiled as well, since including literally just copies the included header into where you include it. Recompiling headers is slower than source files.

Downside to doing this: if you radically change B.h, you may have to manually change the forward declaration in every file that includes it.

https://brianbondy.com/blog/5/slow-compilation-time-in-cc

>>15062199

>You sicken me.

Yeah yeah, I'm sure the progress I make while you're writing header files sickens you.

>>15062210

Post some of that progress then.

>>15062088

>>15062173

#include all the other .c files into your main.c and compile the entire program with one gcc invocation

Something else is probably fucked in your code or system if you're less than <1kloc and getting such terrible compilation times, but it's solid advice regardless. The time the compiler will spend recompiling files that haven't changed is dwarfed by the time that would have been spent restarting gcc multiple times, opening and closing files repeatedly, serializing and deserializing data, etc with the standard approach

>>15060937

It's not dumb, your point about agency seems absolutely correct

>>15062173

Use a makefile or something, holy shit

>>15062242

It is 2018. Your computer can simulate physical interactions between millions of objects at 60FPS without breaking a sweat. Yet you are spending time mentally tracking the build dependencies of a project that could fit entirely and easily into the RAM of an utter piece of shit computer from 20 years ago.

>>15062338

What are you talking about? Why are you implying that his windows BAT file solution is somehow a marvel of modern computing? Are you sure you replied to the right post?

>>15062338

>it's literally the current year, why aren't you transitioning shitlord?

And do you have an alternative you'd like to suggest? He's trying to cobble a build system together out of batch scripts and your solution is insulting established build systems because "it's 2018"?

>>15062610

>Should i do x?

>>15062632

>I might not though

Then why fucking ask retard

And the answer should always be YES to fat fucking anime titddies

>>15062656

at the very least the scribbles aka designs will have big fat anime tiddies im just unsure about 3d

>>15062667

what are you making the models in??

>>15062173

time for makefiles

>>15062610

make a game before thinking about stupid shit like that

>>15062722

its been too long and it looks like i applied the subdivision to the model so i cant show how the topology was before it

>>15062722

Jiggle

>>15062753

>expect pixel-perfect quality on heavily compressed video

have you tried giving up

>>15062753

I use premiere pro with the webm plugin

I could never into ffmpeg

How do I become as good as this man?

>>15062763

Because lossy algorithms don't work like that. They're generally meant for normal cases where the individual pixels don't matter, like the kind of video you would record on your phone. You know, the kind for which JPG is a better file format than PNG?

>>15062767

Hard work and dedication.

So better just give up now.

>>15062767

Step 1: be finnish

>>15062750

oh shit nigga thats great

>>15062775

>the kind for which JPG is a better file format than PNG

If you don't like that my blurry 8000x6000 photo is in PNG, then you clearly need to get better internet faggot! Hard drives are so big it don't matter lol!

>>15062793

Now you have no excuse

>>15062830

yeah I do, I cant draw kek

>>15062860

why do you need to draw to make a model? Thats like saying you're bad at sculpting because you cant draw.

>>15062912

because I need a reference for the model??

>>15062927

I model without reference, just adjust it as I go. Though I did turn to Gelbooru and such for certain proportions, and models-resource to see how N64/PS1 era games modeled certain features, like noses.

>>15062610

yes, especially if she is small, but also what >>15062721 said.

Haw many OpenGL[tm] (GL for Graphics Library[tm]) shades do I need to make #? 4 I thinks but aslo I have no gpu so then!

>>15062927

Unless what you're modeling is relatively alien in design, you'll probably be better off taking a generic set of good-looking images from Google, modelling that, and then making changes. If you have any particular original features that you want specifically (tails, weapons, accessories, etc.) then you can draw those and simply add them to the model separately.

>>15063045

No one will ever know unless you outright tell them.

>>15062953

What do you mean? You need two at minimum to render

>>15063045

There's nothing shameful about references or even outright copying. They're both good tools to learn and improve

>>15056872

>>15062953

Anon whatever aerosol you're huffing, you need to cut back on it.

>>15063386

I want to fuck that loli.

>>15063045

Only if you try to sell it. Then people will find out, they will ostracize you and they will make sure you never succeed. In short, if you're copying something, try to change it enough so that it's not blatant theft.

>>15063426

https://youtu.be/UMV4mK5LGhc

Is it really copying/blatant theft if i'm just drawing over a mannequin though

>>15063439

Maybe not in that scenario. But you will probably get lots of ridicule if people found out you did so

Oops

>>15063569

My suggestion is to use defaults that people are already used to:

Movement: WASD

Jump: Space

Punch: Mouse1

Duck: LCTRL

I've never encountered camera centering buttons in a m+kb game before, are you sure it's needed?

>>15063569

Use T to recenter the camera, Left control for duck, space for jump and either LMB for punching or some other key like F.

>>15063569

Maybe you should implement keys for "intensity" together with WASD, to simulate analog controls, or woukd that be retarded?

>>15063969

The closest I could come to that would be a Mario 64 DS-style run button. I think crouch for slow walking would be much more convenient than that, since there are few times when you don't want to run at full speed in these games.

>>15063620

>>15063796

Very interesting suggestions. For some reason, LBM for punch never crossed my mind, but it would work out quite well, I think. I'm thinking of perhaps giving a few presets for KB+M controls, using mouse for camera only as some sort of alternate control scheme.

Really, I've just been ideaguying while making some food, I'm still a ways away from actually implementing this stuff.

>>15063569

Use the standardized KBM controls that >>15063620 suggests.

>I've also considered mouse controls, but they seem best left to camera control, using the two buttons to center the camera.

You don't really need that function with mouse look. You need it on controllers since analog sticks are more-so a function of time then they are of distance, but mouse look is purely a function of distance. A player can just move his mouse further in the instant they move it.

Also, why the heck do you need two buttons to center the camera's. Also, you forgot that most mice actually have three buttons, third one being pressing down the mouse wheel. Some console ports meme souls actually use that as a refocus/lock-on button.

Hey, I'd like to discuss some game design if anybody is interested, when does less become more? and when it's better to settle for less? let's take Dwarf Fortress vs Rimworld for example, the former has your typical roguelike feel, entities are represented by a single sprite and movement is snapped to a grid of a fixed size, while the later has animations for entities and seamless movement, both are approaches to the same thing "an entity existing and moving on a map", but they vary greatly in term of complexity and execution.

Scale is another subject of interest, again, in DF you can manage a big colony with tens of citizens, and the map spans multiple levels with different terrain type, while in Rimworld you can only have few citizens and there's only one Z level for the map, but even with the huge difference in execution, both games are loved and enjoyed by a lot of people.

I couldn't stop thinking about this when I was trying to port Dragonebone runtime, a fancy 2D skeletal animation with mesh deformation will certainly look sexy, but it is also a lot of work that will probably wouldn't matter in the whole scheme of things, pixel art and ascii art are still widely liked.

My original goal was to simply make a colony manager but a lot of parts can be done in multiple of ways with a huge difference in complexity and resources, the map size and existence of multiple Z levels would greatly affect the size of the colony and what you can do, digging and exploring the depths has certainly a charm to it but you can totally ignore that complexity by having one Z level and a "mine" building where miner just enter and spend some time before returning with resources that may or may not be infinite, the result is still the same "player dispatching miners to get resources to spend on the colony" and only few people would care about the difference after spending a couple of hours playing.

I'm also concerned about the presentation, I've been discovering a lot of isometric games of this genre that vary greatly in presentation like Shipwreck and Stormland, the former is really simple while the later has a complex isometric look.

*suddenly, out of the darkness, a hooded figure emerges*

*his eyes are glazed over; this horse is blind*

*sends out his forked horse tongue, tasting the airs*

>"Ah yes, this is the place…"

*opens his gaping maw and begins to speak:*

How much of computer graphics do I need to understand to actually make something in opengl?

I'm reading a book on graphics atm and it's very rough. I wouldn't be able to implement this stuff, but I can mostly follow along.

>>15064494

Do you want to spend more time dealing with animations and design or programming and logic?

Because Rimworld is the former while a dwarf fortress clone is the latter

Are you looking to sell the game or release it and get followers?

Because nobody wants to pick up another dwarf fortress clone and such simple graphics don't sell anymore.

>>15064494

>when does less become more

When you fail to integrate said "more" into your game properly or don't have the time or resources to do it. If dwarf fortress had the UI and graphics of Rimworld, I swear it would be among the top of indie games.

Games that have a lot of shit but seem to fail to gain a significant following pretty much always have an obtuse shit design that makes it very hard to get into. Just think about how interesting and cool the concept of Wurm is, but it's not that big because in order to play it the first thing you need to do is start reading a fucking wiki.

>>15064494

Ask yourself how many things a player has to keep in mind at any given moment, with the obvious follow up question: "Is this too much?".

Game mechanics that elegantly fit together and introduce complexities from interactions are a lot easier to keep track of than many mechanics that don't fit together and are complex by themselves, but don't introduce any meaningful complexities from the way they interact.

When it comes to sim games, the best way to do that is to make all resource manageable in a similar way, and introducing them over time by making some resources only relevant later.

In Dorf fort, for example, metal isn't really that big of a deal until you actually need to forge weapons, axes or pickaxes. And even then, you can still make due with your starting materials, buying shit off of caravans and wrestling everything to death. Instead, you focus on things in a specific order: Food first, followed by beds, followed by luxury to keep your dorfs happy, which is then balanced with overall productivity and military.

First you set up, resource X, then resource Y, then resource Z, and then expand resources as necessary. You only start with resource Y when you don't need to worry about X anytime soon, and you don't worry about Y until you have X set up.

>>15064572

OpenGL is simple. All it does is draw points, lines and triangles. If you're not a brainlet you should be able to easily pick it up and get something nice going in 2D in a matter of days.

3D, however, takes more effort.

Either way go here: https://open.gl/

>>15064580

>>15064705

Domes feel more "academic", give it a regular roof.

>>15064752

yeah, i guess the roof matches it better, also i could make the top room of the tower be the Innkeeper room, since she is also one of the leaders of the community council it would be convenient for her to just get off her room and cross the bridge to do meetings in the office that will be in the third floor of the community center

>>15064595

>Are you looking to sell the game or release it and get followers?

The later, I want to release something enjoyable and maybe gain support later, and that's a big maybe.

>>15064624

>If dwarf fortress had the UI and graphics of Rimworld, I swear it would be among the top of indie games.

I thought DFHack solved a lot of UI problems, and a tileset can make DF easy on the eye, but I agree that the UI can make or break a game.

>>15064630

>Ask yourself how many things a player has to keep in mind at any given moment, with the obvious follow up question: "Is this too much?".

But when does it become too few or too minimalistic?

As for the resources, do you think DF really benefit from the overly detailed minerals? there's a joke going about some dude restarting a fort just because he couldn't find Iron and he kept ignoring the rocks that give it just because he didn't know the actual name of said rocks, a lot of alloys and some metals can be seen as pure noise as they provide little to no value in the whole scheme of things yet they still have a place in the game.

Previously I asked if DF would still be enjoyable if it had only one Z level and some anon said that people argue it was good before it had multiple levels, I only played the recent version so I wouldn't know, while verticality allows buildings that weren't possible before but I find that having a horizontal fort can be as interesting and enjoyable

Verticality is an interesting subject since I couldn't even choose a satisfying depth, 256, 128, 64, or heck any non power of 2 number too, obviously the more depth you allow the more heavy the game will be on RAM and CPU, pathfinding through all those levels will only make those FPS deaths come sooner especially if the game is turn based instead of real time, I kinda got used to use multi-frame functions to make it easy on CPU which you can't really use when every entity need to be synced in a single frame.

Another question I find interesting, how much can you strip away from DF before it stops being the same game?

>>15064572

There are way too many variables in that question for you to get a decent answer. It depends entirely on what you want to implement, how you want to implement it, and what knowledge you already have.

>>15064923

>But when does it become too few or too minimalistic?

When the game becomes simple to the point of being un-engaging.

Simplicity vs complexity is a balance between boring and confusing. Engaging, which is where things are easy to understand and get into, yet continue to deliver interesting interactions throughout the game, is the sweetspot in between these two.

>As for the resources, do you think DF really benefit from the overly detailed minerals?

I think DF could do a better job at expressing what minerals are smeltable in what kind of metals. But I do think it benefits from it. The advantage of adding different kind of metals means that players cannot rely on a fixed baseline for all metal tools, it depends on what kind of mineral veins RNGesus blesses you with. The disadvantage, of course, is complexity. But I think the added complexity is low enough to justify the addition, in part because the main progress throughout the game (get metal -> produce tools) has not changed at all.

>Previously I asked if DF would still be enjoyable if it had only one Z level and some anon said that people argue it was good before it had multiple levels

I haven't played ancient versions of DF, but I can totally see the game being well enjoyable without the addition of multiple Z levels. Aside from being able to make more compact fortresses by expanding into a third dimension, I don't think there's significant improvement from adding multiple Z levels.

>how much can you strip away from DF before it stops being the same game?

DF has its reputation for being very complex, yet in intuitive ways. Simplify it too much, and it would eventually become just another simulator, losing its reputation in the process. As long as stupid shit like embed related happens, which is the result of a bug combined with many different mechanics, DF is DF.

Despite the complexity in DF, the main hurdle for getting started is its shitty UI and lack of proper tutorial. Give that an overhaul and you could probably ship that shit to even casual players.

I've been programming for five years now and don't know if I'm using OOP or not.

>>15065577

Quake

Quake 2

Quake 3

Unreal

Unreal Tournament

Team Fortress

Team Fortress 2

Counter-Strike

Counter-Strike Source

Counter-Strike: Global Offensive

Dota

Dota 2

Literally any game made in C/C++

Literally any game made not in (((C#))) or (((Java)))

>>15062767

Learn OpenGL. Don't even bother with above versions higher than 2. They require you to throw out the fixed function programming and do shaders.

OpenGL itself is already depricated since Vulkan is out, so if you care about learning the newest stuff, go straight to Vulkan.

>>15065603

>you've made

>>15065628

I just posted them :^)

>>15065603

Damn you're fucking ignorant

>>15065602

You are. Everyone uses OOP design principles, and if they claim they don't they're either a fucking idiot or a liar.

>>15065661

There are some languages that don't allow for it, I wouldn't make such broad claims if I were you

>>15065602

Then you're likely a bad programmer. Not because you do or don't use OOP, but because you don't know.

>>15065677

Probably. I've never taken a course on programming, have never had a job doing it, and I use GameMaker. A quick search says that GM uses a kind of OOP. Is there a good source to learn about the alternatives?

>>15065974

>props don't cast shadows

>>15065994

Working on it

>>15065661

Kill yourself Javanigger

>>15065698

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

In particular look into procedural programming.

>>15066091

I have never once written in Java

>>15065666

You can still use OOP principles in most languages not designed for them, even functional languages. Some of the principles are quite handy, especially on the higher levels where performance isn't key. If you use them absolutely everywhere, I find that code tends to lose some coherency as slightly related code tends to end up in a hundred different places and you have to navigate around to find what you're looking for, what's overloaded, what's inherited, what's a special case handled in some super class.

>>15067475

1.) You need to learn the differences. Go YouTube search some videos.

2.) Not an easy question and not an easy solution. If you are hardcore you'd dip your feet in Vulkan or you can just use OpenGL and DirectX. You can achieve what you want with either 3.

3.) You are the ambitious anon with the physics particle universe right? Take a step back and work on something a lot smaller. A LOT smaller. No answer either to this question. Is a couple years worth it for you?

>>15067475

>Microsoft™ DirectX©

>future of graphics

Do you even need to ask?

>Is it worth the time investment to development for Vulkan vs OpenGL

If you want to make game then no. If you have extra time in your hands and want to learn something that might be useful in some distant future point in time, then sure.

>>15067475

OpenGL will be more than enough for a long time and it runs on everything. Revisit Vulkan in 10 years.

Is there ever a time you would need more than 255 bones?

>>15067528

What engine are you dealing with? You should be more specific. Is it 255 per object or can you combine several objects? 255 is relatively low for AAA quality models. Facial gestures and physics objects can take up a shitload.

>>15067528

When you're an infant?

>>15067494

1. I have been looking up the differences and from what I can best tell: They can exist alongside one another, vulkan offers lower-level access to the GPU while openGL is more abstracted API, and development for vulkan will more complex in various areas

3. I'm enjoying this stuff so much I may make a career out of it; already going to graduate with degrees in IT and maths.

I wouldn't mind developing terminal ballistics simulation systems for the military

I'm primarily worried about lack of documentation and tutorials for someone who isn't educated in traditional graphics systems. The benefit though is that like with learning C, as opposed to like python or whatever, learning vulkan may force a more deep understanding of how the GPU works.

>>15067511

I think this is true in traditional game dev but from what I've seen, all the demos showing massive performance gains have been in systems that have to render and create lots of different objects. This is opposed to games that exist which have tried to optimize in every way to get around these bottlenecks. I'm not too concerned with shipping a final product as much as I am with the development process of this engine.

Also, if I go on to get a masters and have to write a thesis, vulkan may be more relevant at the time.

My Patreon was just approved. Here's just the intro video without any links or names so you guys don't think I'm shilling. Funding is pretty much only for sound, music, and voice assets. This is a hobby, never meant to replace my real job. Thanks for attending the wedding if you did back in September.

>>15067734

>prostituting your waifu

I expected better from you.

>>15067734

inb4 old hags freak out and get you banned for not liking old hags

>>15067734

I saw that thread you made on some board about 3D modeling your waifu. May I say that was some deeply autistic yet enjoyable shit to read.

>>15067734

>Programming and modeling is my way of relaxing

beautiful

>>15067734

Are the Japs aware of any of your stuff?

>>15068238

>>15067734

fugg, tweeting out my support and discovered your twitter account is suspended. RIP

>>15058366

>A variable in a class (hence a member or m_ variable) is declared on the heap, whereas a variable declared in a function is declared on the stack. Stack access may be faster depending on CPU and its addressing mode.

It might also just hang around in a register, especially in that tiny example.

>But, I believe the any modern compiler would take care of this.

In the toy example, probably. In real code you can be doing stuff inside that loop that makes the compiler doubt that such an optimization is valid.

>>15062173

754 is nothing, maybe you're programming on a toaster or including a ton of garbage.

#pragma is compiler specific IIRC the portable way to single-include is

#ifndef SOMETHING_H

#define SOMETHING_H

#include "something.h"

#endif//SOMETHING_H

>>15067734

Is it possible that I can compose music for your project, anon? I'll do it for free, because I do not want money to propel my artistic integrity. I do chiptune and some actual as a hobby. Do I need to post a link for you to hear it?

>>15068722

You just have to do it.

>>15068722

>How do you motivate yourself to go through menial drudgework that you can't get around?

By remembering that no one else will make the games i want.

>>15068722

I'd like a sauce

>>15069505

>added slowmo to make it easier to aim

Have you thought of adding hotkeys for those? Unless you intend to have a shit ton of turrets, at which point, have you thought of having control groups?

I think having to individually click turrets in order to aim them is going to get really annoying really fast.

>>15067734

You are one of the coolest turbo-autistic devs on here. If I weren't a poorfag with no money, I'd support you.

>>15069076

Lolis are meant for two things: bully and fugg.

>>15069537

Yes and it's already annoying with such small selection hitboxes. There will be option to assign selected turrets to a hotkey, so that when it's pressed the assigned turrets aim where the cursor is.

>>15068279

Oh I disabled it a few years ago because I stopped using it. I don't really do social media.

>>15068751

Well I'm mostly interested in sort of guitar themes. Like this song. It was used in JetMoto 3. Not so much modern music. But yeah post a link for your actual music.

>>15069603

I've never done "guitar" music before, but I actually appreciate that you're aiming for something different because I've been expanding the genres I've been doing for some time. I'm six years into composing and I mostly do ambient, techno, chip tune, jazz, rock compositions. I have more tracks that I do not post online but seems to fit what you're aiming for. I have been primarily playing keyboard and percussion, so I'm up for the challenge. Just give me a week, and I'll post it in your thread on /agdg/.

Mind you, I'm do not have high quality equipment or DAW, but I'll do the best to my ability.

https://soundcloud.com/muscises

>>15070048

Trust me until around the end of March, I got a set of Bose Speakers with a sub woofer and I just got a pair of Sony Headphones with Noise Cancelling. I know I have to get the mixing down.

Raining Meadows is like that because mixing in FamiTracker is quite difficult especially when some of the chips are naturally louder than others, hence why I use specific instruments on it and the distortion is not on mine end rather than the program itself.

Equanimity for the Irrationals was supposed to sound drowning-like since I was aiming for a progressive ambient tone, however, I noticed the percussion had some peaks but I didn't realized it until it was too late and the Bass was really loud. Again, when I made that composition I was using my old Turtle Beach Headset and they weren't made for Music Production.

Always Moving was taken from a more simplistic approach, little to no filters, instruments that had the most clear sound, and it was straightforward, which I guess why it has better mixing. Also, it might sound like Sega because the Bass was actually a FM Synthesis Bass.

Chiptunes are clear sounding because you can't put any low-end or band-pass filters, echos, reverb, etc. It was just that channel producing that audio signal, and I naturally adapt to that which could be the reason why some of the earlier actual compositions were weak on the mixing end.

Other than that, I honestly appreciate the criticisms because when I started in music school, I played in a band and mostly the guitars and bass overpower the keyboards, which I played. and the fact that I just started getting back into using a DAW once I had the money to buy a computer to run it since I did my music on a Windows 7 laptop then on a Windows Vista Desktop. You'll be surprised that the DAW I got cost $50, but I got it for free because I registered a broken MIDI keyboard I had.

Again, thanks for the constructive criticism.

>>15070158

Bose is a bad, overpriced meme and noise cancelling headphones use noise to cover up background noise. If you're going to mix audio you should either invest in studio monitors or a neutral pair of open ear headphones.

I finally talked Kowloon anon into replacing his crappy bass-boosted headphones with something much cleaner and the quality of his mixes skyrocketed.

>>15070217

What are some open-ear headphones you recommend?

Also, I got these mint-conditioned Bose Companion Set from a Secondhand shop for $30. I tried to adjust the bass to hear as less as I possibly could.

>>15070158

Any time.

I'm afraid I can't give you too much in the way of specific advice. It's been a while since I've dealt with any kind of mixing, and there's about as much subjectivity as there are concrete "rules", so to speak. You just gotta roll with it, combine your channels into mono every now and then, see what it sounds like quiet, what it sounds like loud, dick around with different sets of channels to find things that just don't work together and find the mud.

You can still mix and master in trackers, but that usually involves either manually adjusting the waveforms, the channels or the notes themselves. Sometimes that's not an option and you might have to make due with just cutting notes down in length or popping into another octave to serve as a different instrument.

Also, the "Sega arcade menu" isn't just the bass. It just has that specific jazzy… thing… that Sega likes. I can't pinpoint it to anything specific, but so many Sega titles had that vibe, usually in their menus, title screens or high score screens. Just needs a few more sparkly chimes and more active bass.

Those might not be the best choices in audio gear, but it's probably an improvement over what you had. Something open-backed might be better for mixing. You're looking for as neutral, possibly as analytical of a sound as possible; maybe hit up some more reputable headphone review people and see what they have to say.

I got a pair of Fidelio X2s a few years ago, and they're leaps and bounds above anything I've ever tried at any store or anything I've used prior. I wouldn't have imagined Philips doing so well with headphones. But apparently, they had a drop in quality at some point, and I lucked out with getting an earlier revision. Not sure what the state of the model is now, but it was pretty hit-or-miss for a long time.

>>15070306

I had my eyes sought on a pair of Sony MDR V6 headphones, because I heard they were the industry standard and they're cost efficient. Are those good?

Also, I think emulate a lot of what a Sega feels, because I've been listening to their soundtracks since I was 4 when I played Sonic the Hedgehog 1&2, then I got to their Arcade soundtracks. Say what you want about Daytona USA or Sega Rally Championship, but those soundtracks are funky as hell and like some of their Outrun, Outrunners, Thunder Blade, Galaxy Force II, Space Harrier. I could go on and on about Sega, Namco, Nintendo, etc. Video Game OSTs were my inspiration to get back into composing music.

>>15050806

Long time ago I wanted to make a boat racing game but it never went anywhere. I did however got almost accurate boyancy physics simulation which i never quite finished. Do you want it?

>>15070510

Why are you using SDL_Rect for this you imbecile?

>>15070528

SDL_Rects are how you RenderCopy() sprites onto the screen.

>>15050052

The thing about C++ is, you will have to forfeit all those nice features if you want to produce fast running code, for the simple reason being these features incurring runtime costs on top of your code's own computational cost. The pointers, the vtables, you name it. At that point you might as well be using C and even that's a stretch, realistically you could do with just Assembly - if it wasn't such gigantic pain in the ass to actually use. If your game doesn't need every last bit of CPU power available, then it's OK to use slower languages and/or slower language constructs. If you aim to create a game that absolutely needs maximum available performance, you use C.

>>15070538

And the reason you use sprite-on-screen-pixel-integer coordinates to store your characters' locations is because ???

>>15070510

You should have the position and velocity using floats and then use that position to set the rectangle's position, rounded to nearest integers. The sprite is just a representation of the thing, it shouldn't directly be tied to the position if you want pixel aligned drawing.

Anyone want to post their favorite resources on learning computer graphics/opengl?

Atm I'm looking to build a little virtual space in opengl, and then moving onto the creation of a game (maybe I'll use an engine instead at that point).

Have been going through CMU's lecture slides and some other resources, am interested in what others have found helpful.

>>15070578

Thanks, this gives me the exact results I needed. I thought everything had to be tied to the sprite's rect.

>>15070573

because I'm retarded.

>>15070563

You can mix multiple levels of programming in a game. When you need assembly you should write assembly, when you need c you use c, when you need a scripting language you use a scripting language. The important part is good design principles and choosing the right tool for the job. Bad code can still be written in C and good code can still be written in C++.

Also, templates are a modern feature than can result in fast running code. They aren't a silver bullet though, and shouldn't be treated like one.

>>15070662

You don't need a new paragraph for each sentence.

https://open.gl/

https://learnopengl.com/

>>15061143

Hey, it's pretty cool you are writing your own portal scene manager. Its interesting stuff, and please keep posting progress.

>>15062088

Only #include the files you need. Look into pre-compiled headers. Avoid templates. Templates are concentrated cancer in applications. Sometimes useful in library code. Also, static link in your base codebase as a library if it's fairly stable. Check out QtCreator and Qt in general. It uses a multi-threaded make. >>15062205 what this anon says about include strategy.

>>15062610

Bigger is better. Make them fucking huge and squishy.

>>15065619

Starting with any OpenGL before you know 3D math and linear algebra is a mistake. First, write your own affine texture mapping routine (floors and walls a la Doom). Then 3D perspective view transforms for a box using indexed vertices. Then learn OpenGL 4 using the Shader Cookbook. If you don't want to do this, use Godot or Unity.

>>15065661

Millennial detected. Object-Oriented means Objects first, procedures second. All you know is hammer. Don't hammer UNF #10-32 bolts and call everyone else retarded. You know what, keep doing it. It's actually pretty funny.

>>15065974

Pretty cool. What are you using?

>>15067475

Vulkan has a shitton of setup code and the off-the-shelf debuggers for it aren't that good yet. Also, easier to multithread, but another shitton of code to do so. Good for huge projects or fundamental library code for huge projects. For OpenGL, GLFW + GLEW and go.

>>15068137

No. But you shouldn't be reading it yet, just using it. STB code is godlike. STB is a C god. Use his code. Reap the rewards. Do not attempt to understand divinity until you are a master yourself. STB knows the way. He knows the evils of C++ yet his code may be used with C++. Do not question, you are not ready.

>>15069603

Pretty good song.

>>15070662

I'm going to recommend OpenGL 4 Shading Language Cookbook and the associated code on github again. Also OpenGL SuperBible 5th or later isn't complete shit, but not as useful.

>>15071067

It's the manager here: https://www.cs.utah.edu/~jsnider/SeniorProj/BSP/default.htm

But I am just retyping it all to learn how it works and understand the code. I don't have enough time to write it from scratch. (it took John Carmack years of full time work to make the quake engine after all) I'm still debugging it, I will show off more later tonight I hope. I don't mind the bugs too much since debugging it is very educational since it is forcing me to understand how the code works instead of just copying.

>>15072692

T man. Savior of the universe.

You could also try baggy cloths like in Miegakure or an over-sized weapon. iirc I saw the test cylinder firing huge bullets, so it would make sense for the character to carry a huge gun. I'm not a big fan of the Chibi style, but it's ultimately your game and your call.

It's nice to see the player's shadow is directly underneath the character, especially because you mentioned platforming. That's a nice touch.

>>15068388

I didn't say your second line of greentext?

>>15073778

The guy's some autistic christian who made his own brand of religion by combining judaic shit with christian shit and some other stuff. He's not an actual jew.

>>15073952

https://www.random.org/

>>15074195

what?

>>15074209

I guess I've mistaken you for that faggot that works as a vehicle artist at CD-Projekt and shits up threads with his reddit tier posts. Unless this is you in which case GTFO

>>15073219

>doom

>curves

am i in bizarro land or something?

>>15074267

Stop talking like a fag and everything will be fine

whatever happened to that dmc like game with the cyber waifu

you know the one that isnt 2b or haydee

>>15074337

We do this every thread.

>guy spergs out about 2B being a ripoff of his character

>gets bullied

>leaves

He still works on it but he doesn't post here anymore.

>>15074390

>He still works on it

If his organs weren't collected by the polish mafia that is.

>>15070493

I'd like to take a look at it, if you're offering. I can't promise I'll use it because I'm not sure if it wouldn't be a little over the top for what I'm going for, though. There's also a possibility I'll not be able to understand it.

>>15074797

Nice, I like it

>>15074228

GZDoom is a farcry from the actual Doom engine these days.

>>15074797

>Eyes are just floating flat disc planes inside those huge hollow cavities called eye sockets.

It's so fucking weird when you know that's how the eyes are done, but it works so well.

>>15075197

Avast was bought by AVG recently which has always been a huge false positive thrower. Might be half your problem. Try Bit Defender.

>>15075197

>>15076900

I've heard Vulkan is hard, since it is more of a GPU programming language instead of strictly a graphics API like openGL, meaning you probably have to make your own openGL-style high-level functions.

>>15071067

STB?

>>15077310

Vulcan is still strictly a graphics API like OpenGL, just with different goals that make more sense in the modern day. It's not a GPU programming language. You can't access the hardware directly. You still have to go through the driver. Only now there's less between you and that driver.

Vid related goes over the reasons for why Vulkan was introduced and how OpenGL and Vulkan differ from each other. The whole talk is good but if you don't want to spend an hour watching it the first 15 minutes give a rundown of the major differences.

>>15075157

might want to use some vector map/array instead of the uvs for the flow direction

>>15077964

>reverse engineer the API

Reading the documentation is not reverse engineering, even though it may feel like it at times.

I'm working through one C++ tutorial right now in pure C. It is taking a little longer but it's nice because I have to stop and think about exactly what I'm trying to accomplish at times. I don't even know C++, but I do know C# and other OOP concepts so I'm translating it all at a steady pace.

Or maybe just wait until I finish my engine and I'll license it to youfor a fee :^)

>>15078045

It was already hard enough learning OpenGL when people just tell you to call x y and z functions without properly explaining what it does, like the infamous glEnableVertexAttribArray that literally nobody including opengl official documentation has a proper explanation of, but now I need to learn OOP and C++ and Classes before I can even begin.

>>15078233

Vulkan, like OpenGL, is implemented in C. But yes, every resource uses C++ for some reason so it is necessary to learn fundamentals like collection syntax, classes, members, etc…

One of the reasons I picked up Vulkan instead of OpenGL is because it just looked like a hacked together mess that didn't teach my what was going on at the lower level. I don't think anybody truly knows what was going on or at least that's the vibe I get from listening to the guys who had to write the drivers to handle every GPU interacting with every program. It's pretty much why Vulkan was created - the guys writing the drivers had enough of the shit.

any way i can get meshes in unity to fade out with distance from the camera without using a custom shader?

the fog option in the lighting tab basically does this, but i need the alpha to go to 0, not just the color

just don't want meshes popping out of nowhere when they get loaded

New thread where

>>187333

Does Unity not encode color as RGBA values or does it do RGB as one thing and A as another?

>>15078396

the fog color has rgba, but the a doesn't do anything in this case

>>15078457

Well that's retarded

But also

NEW THREAD

>>15078478

>>15078478

>>15078478

>>15078478

>>15078478