Amateur Gamedev General ~ /agdg/ /vm/

I feel like the gun that permanently kills them has way too much ammo on the map. I'll write in a different kind of gun that is capable of killing them, but give it really scare ammo. I think that's the best medium for this enemy.

Those hooded enemies aren't a joke- they do the most damage in a single shot out of all the enemies in the game.

Just make sure you avoid the trap of making guns only good for one thing. They can be used to excel at one thing here or there, sure, but if you have a healthy supply of ammo fot X, you should be able to clear the map with it without too much difficulty barring one or two things that it doesnt do well.

For example maybe a high powered nailgun machine gun. Doesnt do much damage individually but it could do decent knockback even pinning the corpse graphic to a wall. Alternative uses could be to shoot at a wall yourself to create a ladder surface allowing for shortcuts or secrets

This is in blender, does anyone not know what im doing wrong?

post the blend and check that you don't have any of these locked , because unless you are confusing object mode with edit mode I don't believe you anything

Implemented a new gun for Space Pirate.
As with everything else, graphics are placeholder. I have no clue what a gun like this would look like anyway, to be honest.

It emits a field of push from you that knocks back any enemies in range–the closer they are, the further they get thrown back. Any incoming projectiles coming at you are knocked away as well, if not destroyed.

The point of the weapon is to give the player a little bit of breathing room in more intense scenarios, playing a little bit more defensively by blocking incoming attacks and then blasting away as needed. They sacrifice the heavy damage of other weapons with the ability to stay a little bit safer.
Ammo regenerates but is severely limited–each shot uses ten ammo and there's 100 ammo total, so it gets used up very quickly.

First of all, don't use standard C even if it is "just for practice"

Secondly, SFML is better.

gyazo.com/c9d9e8283334b19527fdfe0700fd3a26
Its my first blend ever so I dont really understand how to create it as well as I do on paper. I would also greatly appreciate people telling me in what areas I should reduce the amount of edges, or where maybe ther are to little. I understand the plane looks lilke shit, but I want to know how I can make it better

finally stopped being lazy.
Worked on some animations.

are you trying to move towards the red arrow? just disable clipping

If you look in the gif, im trying to move the green arrow towards the mouth on the right side
But it wont fucking move

first: sorry, I didn't know it was a gif
second: holy shit, I don't have a fucking clue why would it do that, reboot your computer and see if it solves itself lol

I'm thinking about getting back into GameMaker. After a half a year of not working on it and wondering if I should even try, I think it's time for me to try again. I learned my lesson before, my first project was too ambitious. I need to start off smaller, but not too small as to not challenge me.

Game Maker is pretty easy though.

Make a simple platformer or puzzle game.
GM is (at least last time I used it, many years ago) unfortunately bad for ambitious projects.

Did you also try to get the art and idea done before making the game too? I wanted to make vidya and barely knew any scripting so that basically fucked me over. I didnt realize art was the last thing you do, even when it came to basic pixel art.

That seems like one of those glitches in Blender where you just gotta restart the program or your computer to fix it

Not the person you replied too.
But I did that exact thing I did, in the end I did some gamejams.
Forces you to work on your own and make a short game.

Tomorrow, I'm going to format my hard drive and make a nice comfy dev profile

And then I have to work 8 days straight, fuck me

How about not using non-portable apis that tie you to a single OS? SDL/SFML are both great and will suit the needs of any game.

So I tried to make my own rope constraints in Unity (since its lacks some basic shit like that), only to rediscover how shit the physics in that engine is. It doesn't take much speed for objects to just phase through each other, the ConfigurableJoint class doesn't obey its own rules/parameters, and some other shit is just driving me up the fucking wall no matter how simple a project I want.

Fuck it, I just installed UE4, and I'm gunna see if their physics calcs are any better.

why not today?

Because yesterday he said tomorrow

But today is yesterday's tomorrow

Look man, just do it. I know you've been dreaming throughout the night, but dont let them just be dreams.

It's 3 am and I've been waking up at 5 pm lately. I have to work Tuesday and I don't have my backups sorted out yet

Hey user, i'm not sure the scope of your project, but I was just reading about the nvidia gameworks stuff they have integrated into branches of ue4, and one of the packages might be of use to you. check out Flex: geforce.com/whats-new/articles/killing-floor-2-gore-goes-next-gen-with-nvidia-gameworks-physx-flex

I'm actually on the other side of the fence. I really really like Unreal, and I enjoy how comfy it is in certain aspects, but for the life of me I can't get this voxel system to run at an acceptable speed. I've seen that other people have gotten voxel systems to work, but all those people seem to be really advanced and doing stuff with the lighting internally that is way above my programming level. UE4 is also kinda heavy with features, and I'm thinking I need something a little more lightweight for my game.

I saw a demo this guy made, and I really liked it. He did it in unity, it was a full voxel cave setup, and I said, fuck man, it might be way easier to just program this whole thing in C# in Unity, than having to become a C++ pro and learn Unreals lighting system inside and out.

I'm just not sure at this point. It's just tough because its all voxels, i need to run dynamic lights on everything, but that then causes the engine to run slow as fuck. Even with some of the apparent advances in 4.13 coming up, I don't think it will be enough to make my game viable without a shit ton of code that's beyond my skill level for years to come.

Can anyone make any recommendations on 3D engines that are lightweight, reliable, and are capable of working with voxels?

I don't care if the graphics are kinda shit, as long as it runs really fast. I want a permadeath option in my game, so it has to run fluid all of the time or it wont be enjoyable if people die from being lagged out.

If I really have to, I'll use Unity, but I'd like to see what my options are first.

How do I put Denuvo in my game when it releases on Steam?

Paying for a licensethat costs like 10,000

how the hell do i have actual collision with rigidbodies in godot?

What are you even doing here?

what are you willing to sacrifice?

I feel adventurous today

I'm gonna integrate Lisp with my game so I can define game variables with Lisp and easily add modding

you know lua exists

I want to fuck all of them after I compile the kernel of my engine

lua is gay

I want to make games, but I'm just an idea guy at the moment that can sorta sketch and make music, and some people say I'm a decent .wad maker. What's the best course of action for learning programming so I can start bringing my ideas to something concrete?

what are my chances of learning anything if i start using only visual studio and the included WSIWYG designer. tried using it and i can't understand shit.

i managed to code a buggy pong game using codeblocks 2 years ago, using SFML. but i gave up and forgot everything since then, thinking of learning c# instead

Suicide x 10

Suicide x 5

If you don't want to get autistic with programming, either learn some high level language like javascript (usable with Unity), or Unreal Engine with blueprints (a visual scripting "language").

IMO the autism you find around here with C/C++/C#/etc is very harmful to new devs, especially to those who aren't particularly enthusiastic about programming. The performance loss from using a high level language is so negligible that you really should be prioritizing getting shit done right now.

It's buggy and broken as fuck and slow. It's garbage.

If user really wanted to get something done, he should avoiding raw coding entirely and just use an engine; Unreal, Unity, Game Maker, etc.

If he does use Unity, C# > JavaScript

How is it buggy and broken?

You can't avoid coding if you want to make games that do more than roll balls and shit.

The problem you get with people who just stick to scripting is that they don't properly understand data types, memory, etc and make a lot of stupid mistakes because of it. They should at least read a programming book and understand the concepts before just jumping in like they usually do.

I'm talking about people who don't necessarily like programming. Doing things period is more important than making things efficient.

For the average dumbfuck nodev that's never made anything before, drag and drop with icons to introduce basic logic is plenty sufficient. Even then, Game Maker would let them eventually expand to learn about render targets and shit if they want to get fancy; it's a very capable engine for 2D games and beginners.

pixelstech.net/article/1354210754-10-design-flaws-of-JavaScript

This is just the first Google result. It's poorly thought out and designed, and was pretty much designed to be a web language. You CAN make games with it, but it's a bad idea.

There's nothing wrong with a new programmer mucking around with scripting to make something happen. Sure, their programs will be flawed from a pedantic level, but it gets the results they want. After they decide whether or not they like programming, that's when they should decide to educate themselves more formally.

There's no point is reading about memory management if they're struggling to understand the different kinds of iteration, for example.

this
newfags to programming will only shoot their own feet in the process if they go through using JS.

creating game while not knowing how to optimize code for maintenance will only cause them to get jaded in long term

>pixelstech.net/article/1354210754-10-design-flaws-of-JavaScript
Have you actually used it, or are you just copypasting memes you found in Holla Forums?

Of course, I don't mean to make people jump right into memory management but quite often I see people jumping in without even knowing anything about data types or structures. They will often just keep hacking shit together and won't ever decide to educate themselves more formally. It would benefit them to sit down and spend an hour or so reading the first few chapters of a book. Doesn't even have to be c or c++.

Total beginners have much bigger issues than how maintainable their code is.

Very first non-placeholder original gun sprite.

S'pretty big gun.

I made big progress on learning how to make a game playable online, of course i can't post it because how would i prove that it's working online?

you may think that maintaining code is like some sort of a 'next step' after learning how to program, but I'd still argue that maintainability is very essential to any project.

especially when doing a large project like a game, which would most likely not be the only thing you do all day.

code maintainability is so important, i see programmers jaded facing the shitty code they face every single day and get burnt out because of it. every new feature you have to add takes longer because you have shit code that you have to read every line again after not touching it for just a week just to know what and how to do it.

How do you develop discipline?

Every day do something you don't want to do, and don't do something you do want to do. Don't give in to any excuses. Keep those things up and before you know it you'll have an iron discipline.

Resolve to do something at a given point of time, then do it. Repeat. Don't fuck this up, don't listen to your brain if it wants to do something else, your brain is a dirty fucking traitor.

I don't disagree with what you're saying but I think it's much more likely that a beginner tries to hammer square pegs through round holes and gives up. You need to know what solution to use in order to make a maintainable version of that solution.

I can't speak for everyone, but to me it's always more important to get into the action as quickly and easily as possible, because otherwise I lose interest. I learn best by doing, and I can do things very independently from any help.

I can keep working on it for weeks then start over and make it better, but if I tried to learn some intricacies of hurdle jumping before I've learned to walk, I would most likely have given up.

I'm still a massive beginner to programming, but is it a good idea or a waste of time to add comments on almost every block of code, describing variables, what it leads to or describes?

Damn, my comment caused quite the discussion. I should also mention that I understand code in theory, what with operators, operands, if/then procedures and shit. What I really want is a firm footing that I can develop from. I'm willing to put the time in, but I just need to be pushed in the right place.

Alright, if I can get a vote, would any beginner nodevs be interested in the following:

A.) Insanely gory survival horror game modeled after an 80's b-movie slasher fic (Texas Chainsaw Massacre, Friday The Thirteenth, etc)

B.) A 2D turn-based tactical RPG set in a fantasy setting.

strawpoll.me/11106131

Adding comments is definitely a good habit to get to, but you don't need to describe it in huge detail. One sentence is usually enough. Also add comments in-between the code and inside loops and stuff to describe what happens next.

I've gone through college about 5 years ago, yes. It was mostly focused on web languages and the most in depth we went was about two semesters of Java proper. We didn't even cover generics. I'm paying off the last $4000 I owe next week

In any case its taught me that I absolutely hate website design and programming (and nowadays you can just get things like Wordpress to shit out a coherent sote without coding it yourself). In that domain, PHP trumps JacaScript

But back to my original point, JS is still crap because it was originally designed as a scripting language for a web environment. Just because you CAN use it to make a game doesn't mean you should. If you did want to use a scripting language exclusively, use Lua or some general purpose nonweb language.

A compiled, strongly-typed language is still king.

Turn-Based RPG isn't much to go by since there are a shitload of ways to do it, but still gonna vote for that cause it would be actually attainable instead of the super gory / detailed horror game option.

I think KhanAcademy had a decent programming section. A lot of examples were somewhat handholdy but it gives you an online editor/shell to mess around with and see immediate results.

I still recommend Game Maker just because it was designed for people with little or no experience and you can see results inmediately.

As far as programming well I would recommend you pick up a textbook of C++, C#, or even Java (again note that JavaScript is completely different and not what you want). If you're comfortable with basic programming logic like ifs and while and shit, a non beginner book that explains the language in detail would give you the most value (I recommend Jon Skeet's C# in depth).

Of course if youre still struggling with basic shit then pick up an introductory book to a language. It probably won't go into deep technical detail about why X is bad but only that you should avoid it. Places like the MSDN documentation or Stack Overflow ate very good resources if you want to be formal and pedantic about your code

What should I actually be doing after I learn the basics? My guess is I should be doing incrementally more difficult small projects that also introduce some new concepts slowly, but I have no idea where I should look for a resource like this. Should I just grab a bunch of those coding exercise things on Holla Forums then work through them using documentation for the language I'm using?

Basically my problem is like having learned the basics of mathematics then not having any sheets of problems to work through.

Looks gud.

Why are you doing game dev?

If there's a game you want to make, aim to learn the things that are required for that game.

Read programming books that walk you through deploying real projects and build them yourself as you read. It's probably the best way to get better at programming. Build actual projects, not fancy Holla Forums fizzbuzzes, and iterate on them as you learn more.

Any ideas on where to get legally free assets apart from Turbosquid and the UE4 forums?

You may want to look into something like Urho3D or github.com/taylor001/crown although I doubt they support voxels out of the box, most engines don't aside from shit like Voxlap.

You can't move anything unless you have the vertices selected. Press 'a' in edit mode, and then you can drag it along the Y axis. Alternatively, switch to object mode and drag it that way.

idea guy here again

let me know if you need a brain storm

Anyone here have any experience working with sdl? How is it?

As long as you don't consider it to be anything beyond a convenience abstraction for getting your program talking with the hardware, it's hard to be disappointed with it.

It has some features that go a little further, but essentially, that's what it is.

If you go in expecting something beyond that, you're going to have a bad time with SDL.

How do I do this, though

Make board games on your computer. Tic tac toe then connect 4 and so on.

Well does the api make sense?

Posting the dankest agdg oc

Is the gun supposed to be so powerful/character so weak that he needs to wave his other arm around to counter the recoil?

Just a small critique. I'm not like super /k/ or anything, but you should have your ejection port on the right side of the gun rather than the left. Real guns generally have their ports there so that you don't eject hot casings into your face.

I've never had any problems with it. Seems pretty straight-forward.
If you want something a bit more minimal, you could try GLFW.

Not sure how SFML stacks up.

I have a mind blowing idea for you user.

How about an insanely gory 2D turn-based tactical RPG survival horror game set in a fantasy 80's b-movie slasher flick?

Unity's JS compiles to the same managed Mono code as C#, and it's specifically designed to get around some of C#'s retardation. There are legit uses for it.

As long as it has shaders, texturizable quads/rects for 2D, "render this list of polygons at this position" (optional: functions for operating on those polygons, just so you don't have to go through implementing maths autism) for 3D and camera controls, you are pretty much able to do everything you can imagine. Access to FBO already makes a graphics library pretty much perfect.

If it doesn't have any of those things, forget about it or get some other library to do it for you.

Thanks for the suggestions user, I spent a little bit of time looking into both.

They actually seem to be sorta kinda what I'm looking for, I just worry about future support for smaller engines down the line. I figure my game might take a couple to a few years to make, and it would really fuck me up if support for them died while I was working on the game. I hate to say it, but at least with Unreal and Unity, I know they aren't going anywhere.

Also, while both crown and Urho3D seem to be capable, there is a lack of examples and documentation. It makes me a little weary of committing my project to it.

They are good suggestions otherwise though and I'm going to keep poking around to see what else I can find. If you think of any others, I'll check them them out for sure.

>keep buying them and playing with them
I've lost control of my life

First thing I thought of was this.

Make a tabletop game, then.
I've made for my little brother before, it's fun!

I'm creatively bankrupt

I really don't see how that's a problem, though.
There's literally nothing wrong with clones or derivative works, so long as they're fun

I said Starcraft-like not c&c-like
rts general thread would hate my game

Thats exactly what I'm doing atm with these toys
Got a bunch of d10s just now too

Just have fun with it, user.
Maybe someday you'll want to turn it into a computer game too, maybe.

It's more that it had absolutely no coherency between the different rule groups, and things were kind of forced.

It was like Warmachines/Hordes in that you used 2d6+POW+STR to determine the effect, but you could also play cards on a 7x7 (or was it 9x9?) board in different orientations, like some unholy mash of Magic and Yugioh and Chess

Quake C
cataboligne.org/extra/qcmanual.html
github.com/id-Software/Quake
quakeone.com/
Do it

So, my friend took graphic design in college, and I want to learn coding so we can work on something together.
I'm a total helpless faggot, but I saw that nehe.gamedev.net/ link, so I bookmarked that.
Can you guys give me any other resources for beginner reading and practice?
Thanks in advance!

...

So I'm reading through the thread, and I see I need to be more specific.
I wanna start with something really small. He says he'd like to work up to something like Salt and Sanctuary, so what I'm really looking for is the beginnings of sidescrollers, I guess.

Go learn Python to see if you really like programming or not.
Learn Python The Hard Way(Read this book)

dear god no

Thanks very much, user.

Once you learn Python, you can start using the engine "Godot", it uses Gdscript, which is pretty fucking similar to Python.
It's also a quite robust engine, probably one of the best ones for 2D games.

Use the move command for kinematic bodies

Anyone have any thoughts on this

?

I see this book recommended to every beginner to programming, and I just can't understand why. The book does very little to teach any useful programming concepts. Most of the time the author just dumps a load of code for people to copypaste and sporadically comments on a few things about python's syntax. Take the chapters on objects for example. Zed doesn't even explain what OOP is, let alone defines what an object is. He doesn't say much of anything regarding classes, objects, or the design patterns, other than that "They are hard to understand and are like modules but not really."

I don't know, maybe I'm being harsh. After all, the intention of the book, in the author's words, is to give the reader enough insight to learn programming on their own. However, I just don't see how this book accomplishes that. I read it when I first started programming years ago, and I didn't really take anything away from it. It was a slog honestly; a very linear hand-holding brand of instruction that I couldn't apply to other concepts or languages because it was just too simple and limited in scope.

TL;DR I don't know what I'd recommend to novice programmers, but it certainly wouldn't be Learn Python the Hard Way.

Look, the book is just meant to be a introduction to programming through Python, it assumes that the reader will look shit up on his own. The author himself states many times that the book's only useful because it will teach you how to look stuff up on your own.

What you can reccomend to a novice programmer is: Get programming, and this is what LPTHW does. And it's way better than the codeacademy crap that's usually reccomended too. After he finishes that, he can just branch out to whatever he likes since there are a thousand of other books out there.

Zed Shaw cancelled his Learn C the Hard Way book because someone criticized it and told everyone to learn Rust instead because the community wasn't full of meanies.

Nope!
learncodethehardway.org/c/

Zed only took down the page where he shit talked K&R being reccomended for begginers, since there was a huge backlash about that

pure ass cancer.

GLFW or SDL2.x + OpenGL/GLSL 4.x core profile only, and C++14. Use GLM for your 3D math functions. Use C++ standard containers, streams, threads.

It is called Learn Python the Hard way

What are you trying to say?

I'm going to make a fighting game using 2D sprites in a 3D environment.
According to the engine documentation, sphere or rounded cylinder shapes are better optimzed for collision hitboxes, but more 2D fightans use rectangular hitboxes.
What should I use?

He must have changed his mind since I last looked at it, pretty sure he wasn't charging for the book back then either.

OpenGL 3/3.3 is a more reasonable starting point since 4+ is restricted to D3D11-level hardware.
ogldev.atspace.co.uk/
open.gl/
opengl-tutorial.org/
web.archive.org/web/20150311211412/http://www.arcsynthesis.org/gltut/

Man you're right, it isn't up for free anymore like his other books.
Well, fug him
ptgmedia.pearsoncmg.com/images/9780321884923/samplepages/9780321884923.pdf

Note:
I don't actually reccomend using Zed's book for learning C

Give me a concrete example of where JS would be a better, more easily readable choice to implement some behavior over C# in Unity.

You can't

What the fuck are you trying to collide?

I find myself questioning why you would do this, but it doesn't really matter I guess.
If the characters are restricted to a 2D plane then there's no point to a 3D hitbox.

No, if you look closer, the other arm is catching the bullet casings release from the chamber after firing.

If it's 2D, then there's not really a point to 3D hitboxes…unless you want them to interact with the 3D environment some way.

If you want to be a rebel, use pixel perfect hit detection. Note, it's harder to program and doesn't work great unless you actually take into account the detection when creating your animations.

I get that it's an introduction to programming, but to me it just comes off as a cheap, linear book designed to make anyone who reads it feel like a "programmer". In my eyes, it's no better than those codeacademy courses, or the countless webdev tutorials you can find floating online. I feel like it's books like LPTHW that make pajeet-tier programmers, because it removes a lot of the logical and mathematical aspects from the whole endeavor. I guess what I'm trying to say is that there are programmers, and then there are computer scientists.

I'm assuming you're referring to this.

I don't want to get in a Zed Shaw debate but it's hard for me to recommend a book (LCTHW) by a guy who runs away and pouts when a community of programmers with significantly more collective experience disagrees with him, relatively gently.
Can't comment on his Python book, don't know Python. It's probably more his speed, though.

How does an adult play with dolls?

...

Now when the player fires the magazine dry the slide locks back instead of the way I had it before where the player cocked the gun every time they reloaded.

Not all guns lock open on an empty mag, and some may lock open but then drop the slide when you eject the empty magazine.

Having it lock open on empty is generally a good idea for vidya purposes, as it functions as a visual cue.

The question is, can you load a new mag and keep the extra round in the chamber? Bonus point for avoiding this with open bolt weapons

like how artfags do with their tiny wooden mannequins
and like how fa/tg/uy does with his GW miniatures

Alright for practice I'm going to build a bunch of board games with Lua and Love2D, seem like a good plan?

As long as you're enjoying it, it's probably good.

Yes. On top of this,
If the player reloads the shotgun when it is completely empty the character loads the tube magazine to maximum capacity then pumps the shotgun to put a shell in the chamber. But that means that the tube isn't filled to maximum capacity any more, so if the player presses reload again they will put that missing shell into the tube.
When reloading the revolver, the player loads the bullets in one at a time, and when the player unloads the cylinder if there are unspent bullets the character will cover the section of the cylinder with the unspent bullets so that only the spent shells fall out and reloading to full capacity doesn't take as long. Also the S&W style revolver rotates clockwise from the players perspective I did this for the soul purpose of fucking with the especially autistic /k/omrads.
If there is a round in the chamber then the player can fire when reloading. Although I have a bug with this that will get fixed eventually where if the player does fire when reloading the player cancels the reloading animation and gets the next magazine instantly.

If I'm going to have real looking guns in my game they should at least behave like them on some level.

just use paper, pencil/pen, scissors, and dice like any sane individual does

Yeah, the point of the game is that you only have one bullet.
So the bullet will bounce around and you catch the bullet and load it back into the gun.
The catching the casing is just for a look I guess.

Read above :)

That sounds pretty neat.

Currently working on a 2D platformer in Godot. The project isn't intended for release but I want to get it to at least decent state before shelving it. Specifically I set out project goals in my design document which I will achieve before moving on to 3D which I have real plans for.

You mind sharing the source code? I wouldn't mind taking a look

Nah mate, it's in very early stages and there likely isn't anything of interest for you and I'm far too autismal to let anyone look at my work before it's complete

git gud

Yeah thats true, I just like doing things in inconvenient ways if it means I can learn how they work.

I've finally hammered out enough of my design document to start coding some basic core systems for my game. It's weird having your whole game laid out on a document; it's like taking a herculean task almost manageable.

*making
Fuck, I wish it were possible to edit your own posts. I always end up making some dumbass error.

I'm still working on mine. Got any tips for it? I've been looking through the Saint's Row one that was posted in a previous thread.

For me it was really helpful to describe everything in terms of the most important mechanic. For example, my game's main goal is to have a very flexible and dynamic magic system that allows the player to "code" spells that can affect any object in the game. Once I figured out how a system like that would work, I brainstormed a few premises for the game. Now all I have to do is fill the game with balanced content, and make sure that any content/systems I add don't detract from the main one. It might not work for every game, but that's what I'm doing.

Would any of you guys consider to stop working on your game if you got a job offer?

I'm just saying, if I happen to come into some money I'd like to get a proof of concept to get some more funding.
Anyone familiar with UE or Source?

So, guys. I'm an ideafag/drawfag who wants to jump into finally making vidya.
Would it be advisable to jump in and start using Visual Studio and monogame or should I use something like GameMaker instead?

Where do I start my journey into being a dev?

The last time I touched Visual Studio was back in High School.

just like make game

I've got a job offer now, actually.
I just plan on doing indie stuff during off hours.

Balancing dev time with new job hours is going to be FFFUUUNNN.

I mean industry job offer(dev/coder/modeler). Or are you saying you got that?

Right now my game's in the infancy stage, still working out the mechanics and how the game works. Still programing up the graphics engine in OpenGL.

It would go back on the back burner. I'm already busy doing other shit so it's not going to be much different.
I mod a UE game in my off time begrudgingly.

Install emacs

Not happy with UE?

From the Grimm Eclipse sucks thread.

This is the same stupid user from the 2000 dollarydoos edition thread. So it seems finding westerners who can model good anime-style characters are indeed rare.

But I want to know if it's possible to get in contact with some of these god tier MMD modellers not just TDA and pay them some license or right to use their base mesh at the very least.

Of course it'll have to be re-rigged and stuff, but the core modelling and texturing would already look good.

Reaching that hard and being that autistic

typical weeaboo

I wouldn't say he's "reaching". He just spent a LOT of time looking to be able to see how shit RWBY is.

I didn't know I needed this.

Finally wrote my last exam today. Just checked my calendar. I spent the last 7 weeks doing nothing but learning. Se-ven fucking weeks. I don't even remember the last thing I did for my game.

this.

Nope, just an ordinary job. Sorry, I misunderstood.

They use the same API for physics user, heh, they use physX; so anything created in either could be done in either engine (depending on the support they have for the extensions to the physx API, as there's a bunch out there…).

You're also probably looking for soft bodies if you want a super realistic rope simulation (probably need to use a wrapper for bullet physics if you want this in Unity, or Unreal's flex branch), or you could do cloth simulation (slightly more overhead, but realistic), or joints but it would be a more "approximate not realistic + performance" simulation of rope.
Or u know… u could use the Unity community wiki to see an example of it being done… wiki.unity3d.com/index.php/3D_Physics_Based_Rope

What is your current approach to voxels?
I've been working on a voxel framework in Unity for awhile now, and have a wide range of experiences in this topic; so feel free to ask more specific questions.

The main issue I've seen with Unreal is the implicit overhead of their engine; i.e. requiring nice hardware to run even a basic game (as the main overhead isn't YOUR game's script itself, which it should be, it's the rest of the backend of the engine).
As in Unity, with nice physical based shaders, hundreds of thousands of triangles, particles systems everywhere, and the like; the backend takes up to 5-10% of the ms (w/o this it's less than 1%); while the rest is expendable budget for my own code… which is how it should be.

Lets see that link, and I can most likely tell you how they did it, and what their main voxel related algorithm was.

It doesn't really matter which engine you use.
As long as you're competent in the language supported, and the engine has support for what you need (i.e. the API required, say to do GPGPU, or some physics API like flex for Unreal, etc etc), and not to mention if the engine isn't horribly optimized…
Also, it's best to say "shaders", and not "lighting" system; as this is the universal term.
As the shaders are about the same across all engines, and just Unreal's visual programming system is unique.
As they all use a form of HLSL/GLSL or CG for the actual shader programming; which is what you should be doing all intermediate and above shaders in anyways, and will prepair you for actually knowing wtf you're doing in a visual programmer for shaders.

There must be some other overhead occuring in-engine. As in Unity I can do the same, and still have great performance (even with the physical shader model, which is complex shader bound algorithms to calculate lighting).

Just look at the shading model supported.
If it can run shader model 4+ (up to 5), and it has support for PBR (physically based rendering), you can make it look good with nice textures/normal maps/etc.

You're talking about networking here right?
Well, I'd start thinking about building you game around networking as this is a major distinction to make here.

Implemented new gun sprite. Feels good to finally have a non-placeholder sprite for it.

Making the muzzleflash randomized was surprisingly a pain in the ass, but it looks pretty good in action.
Still need to do that color-to-alpha thing so that it doesn't look like a big blob, though.

You see, this is a major grudge with UE4 for me too, the overhead. I thought that a literally empty scene would run fast as all hell on my rig, but it runs at 40 FPS at 720p on low settings. Granted, I have a 6670, but still, I have almost nothing being rendered.

I suppose there aren't any solutions to this problems except digging my teeth deep into the source code and tearing unnecessary shit out, or using another engine, no?

windows.h sorts this out for you, with preprocessor definitions. Here is some of the code in their header:
#ifdef UNICODE#define CreateWindowEx CreateWindowExW#else#define CreateWindowEx CreateWindowExA#endif // !UNICODE

I don't have too much of a need to use the API, so I just use it to create the window and get input. It could be better, but I would rather do it myself than using someone else's library to do something for me that I can just learn.

Have you actually tweaked anything, or did you just use the default settings? Unreal's defaults are set to up to look really good already. I'll do some benchmarks either later or tomorrow (getting quite late here). The laptop I want to test on has an i3 mobile CPU and a GeForce 410M, neither of which are fast by any stretch of the imagination.

All graphics settings (postprocessing, foliage, antialiasing, shadows) were set to low, and resolution percentage was set to 100%. Apart from that, nothing else was tweaked, no console commands or anything.

Wew. Let's see how it goes with my shitty laptop.

Looking good.

I know you posted this ages ago, but w/e.

First off, unless there are non-obvious tutorials buried somewhere the stuff that site has is ancient. If an opengl tutorial contains glProjection, glMatrixMode or in particular glBegin it should not be used. It teaches a 20 year old api that is catastrophically slow on modern implementations. You will quickly learn to hate the matrix stack, the state juggling and the unflexibility of the fixed function pipeline. It sucks nigger cock. Do yourself a favour and go straight for modern opengl. 3.2 if your toaster is from this decade, 2.1 if not. Use shaders, buffer objects, generic vertex attributes and a math library you get along with.

Also, please don't use windows-specific apis. They do nothing multiplatform libraries like glfw or sdl don't do far better.

Thank you for your help, I appreciate your answers and advice.

Honestly, I'm doing things a little different than most people, one of the things is that I'm not going to have the world editable through normal voxel placement. Instead, the world will be generated at the start of the game, and then most of it will be static geometry.

For how I have it set up right now, it runs a few steps.

First it generates the voxel data by creating a series of 8x8 x 64 chunks.

Next it takes that data then does a sort of serialization on it, so that it reads through all of the cells and copies the information into another structure to be saved. During this process, it removes all blank cells and then saves the blocks ID, location and chunk index number. It saves this data to the HD.

next step is the loading step; it loads the data from the HD and then makes instanced static mesh for each block.

Heres the thing though, theres no code running after I've generated and then loaded all of the instances. All of the frame drops I'm getting are purely from unreal doing the calculating on the lighting. After it loads the blocks the first time, I'm not having it generate any more while you are playing.

Also, I'm not running any sort of shader other than the default one from a new project. Because I'm using Instanced Static Meshes, I can't use static lighting at all, all of the lighting I use has to be set to dynamic or it wont render.

I haven't yet added invisible block occlusion yet, nor am I at the level yet where I'm just drawing each chunk as triangles, but even with that amount of optimization, I donno if I can get it to run fast enough with fully dynamic lighting. I'm not drawing nearly a big enough scene to have these things make that much of a difference at this point.

I should have been more specific when I said "lagged out". Nope, single player. I don't want people to constantly have rendering lag or their computers to seize up from having to generate chunks. For now, I'm programming this myself. If the game is successful, I'll get a team going and make the sequel of it multiplayer. But for now, SP only.

For example though, there's this guy who made a minecraft'ish clone in Unreal: youtube.com/watch?v=EjCandEgtHo

I'm not sure how he got it to run as fast as he did, but that said though, the lighting is shit if you try to go underground. So ultimately his approach wouldn't work for me either.

red-aurora.com/?cat=5

Here's that game. It's not an exact fit for the kind of graphics I'm aiming for, but its a step in the direction I want to go.

In terms of questions for you specifically. Wheres a good place to get started with voxels in Unity?

What is everyone's opinion on CryEngine vs Unity or UE4?

Good tech but small to non-existent community compared to UE4 and Unity. You can modify the actual code of the engine, something that developers have been able to do with UE4 before Cryengine, and something you can't do with Unity unless you pay. Cryengine isn't really for people who don't know what they are doing, and with UE4 blueprints and C# with Unity, along with what of heard of people who have messed around in Cryengine, development times are going to be comparatively longer. Cryengine probably has the best payment plan, as it is pay what you want with a minimum value of zero, and they don't take royalties of commercially released games, whereas Unreal Engine takes 5% after you pass 3,000 and Unity holds back features (such as access to the source code and better built in multiplayer support) unless you pay.

In short, if you really know what you are doing and don't mind development times being longer than UE4 and Unity I'm sure it will serve you well. But honestly you should just clone a really simple game using each engine and see which you like more.

I'll keep in mind. Now in terms of asset creation, how about 3DS Max, Zbrush, and/or Maya?

I use blender so you're going to have to talk to someone else, but in terms of workflow I'm sure that all three of those are better than Blender.

Well, the HDD is overhead best avoided… unless of course you don't mind loading screens.
Directly generating, and then rendering is the best way to implement procGen; as to get around the slowness of the HDD (main bottleneck for this type of stuff is always the HDD when it's used).
As you want the generation algorithms to be deterministic, as to allow this dynamic procGen of content at runtime, and to completely avoid the overhead of the HDD.
That is, even without editing of geometry this is warranted, and saving on the HDD adds so many implicit limiting factors that it would take too much time to go over.

I assume this is done at runtime… well here is already a huge limiting factor of using the HDD; as there's implicit delays for de-serialization from the HDD storage format, and not to mention the slowness of the HDD itself even with a SSD (don't expect everyone to have one of these)… so compare it to loading from RAM or the CPU cache, i.e. it's always way slower.

So, it iterates over every block, as to determine if it's empty or !empty.
Here, you could get around the slowness by a bruteforce approach of using multithreading (only around slowness in one of your steps… as loading from the HDD is intrinsically serial… i.e. one at a time).
Although, it would be even better if you use an octree or some 3D spatial data structure to completely skip large chunks of empty data (as it subdivides each chunk into !empty or empty sections, which can be completely skipped, or to only go over which sections actually contain at least 1 !empty block); this spatial data structure would be build when generating.
You could also then use multithreading to go over each of the !empty cells of the spatial data structure to build the faces (triangles).
Also, skip the HDD step completely, and use deterministic noise/procGen algorithms (same input == same output).

May also be due to the in-engine automatic culling algorithm, if UE has that… that is.

I'd determine with a profiler where your main delays are coming from… as it shouldn't be coming from the graphical section if you're not rendering much; although I don't use UE so it certainly could be afaik.

Ah, I c, my bad, and understandable.
You'd benefit a lot from the optimization methods I recommended above then, and determining your current bottleneck (HDD, iterating over empty blocks, cache misses, etc).

Looks like he just used the same approach as Notch, but with the nice post-processing graphical bits from UE.
Initial generation is the largest overhead (noise algorithm w/input based on the seed + position of chunks), and as you can see he waits for that to load w/a loading screen.
Rendering of his stuff is extremely simple, as it's only rendering the visible faces in close proximity.
When generating each chunk's face data he more likely determines if each face is visible (no other face is touching this face) before rendering it, and only the chunks visible to the camera (occlusion culling entire chunks that are not visible, hence no need to build faces, etc).
So he's barely rendering any triangles in those scenes, and any data for generating new faces is more than likely loaded from RAM or the CPU cache (i.e. generated right then, and used in bite sized pieces); so it should be quite fast.
For actually building the triangle faces, that approach is extremely simplistic w/barely any overhead, and requires a search to figure out in full detail (say, compared to more complex contouring algorithms, for say smooth terrain contouring).

Generating only the chunk data required, building only the visible + not occluded faces/chunks, accessing data from RAM/CPU cache (not HDD), and more than likely simplistic noise methods used for actual chunk data generation; not to mention the "contouring" (triangle building) algorithm is extremely simplistic.

cont. in next post (body too long)

cont.

So, do u want blocks, or do u want smooth surfaces? Both are possible (lmk if u want to know the smooth voxel algs).

Also, same things as that earlier Unreal one, they're using extremely simplistic contouring algorithm, and simplistic noise generation w/the above steps.
Also, they do use octrees for splitting up their data spatially (per their trello), as to only generate noise + faces needed (visible, contains surface, etc); thus eliminating a brunt of the empty blocks/chunks that waste CPU ticks.
They probably also use multithreading per set of cells (spatial data structure) to generate only visible faces (!touching other face) that contain surface data (!empty… cell of 4x4 blocks, etc).

Well, you can use UE, as your previous video demonstrated… but that's your choice.

For Unity specifically, I'd take a look at this megathread forum.unity3d.com/threads/after-playing-minecraft.63149/
There's also a tutorial series for coding a mine-craft esque voxel generation scripts in Unity: alexstv.com/index.php/category/voxel-tutorial

That is if you want to generate blocky voxel worlds, and not smooth worlds; although even if you decide to stick with blocky voxel world then keep my above optimization steps in mind.
Although, even if you decide to go for smooth voxel worlds, they do go over the basis of noise generation which is worthwhile, and how to create triangles/etc from this noise data.

For an in-depth review of figuring out, what is noise, check this blog out: redblobgames.com/articles/noise/introduction.html

Thanks for the response.

Let me explain my process in a little more depth, I don't think I really explained it fully. I'm doing a few things differently than other voxel games, so I think its leading to a mix up.

Unlike other voxel games like Minecraft or whatever, I'm not trying to do an infinite streaming world. I'm actually trying to do something closer to Dwarf Fortress for my procgen in that it will creates a fixed size world, then iterates over multiple times in order to generate itself. So I can't use deterministicly generated maps that generate in pieces, the whole thing needs to be created at the start of the game. I'd say maybe like think of it like generating an RTS map rather than a normal Minecraft map, because I want to be able to control what goes on in places that the character hasn't been to, for example.

I also am willing to break up the "infinite" world into smaller levels, which is what the HDD loading here is for. I'm okay with loading screens and stuff, but the idea is that once that map is loaded from the HD when the player enters the zone, that's it, there won't be any more loading from the HD (at least map wise) until the player goes into another zone.

The idea though is, generate the whole world at the start of the game with a long loading/generating screen. The player will know that it will take a while to do, thats fine.

The reason for serializing the data into just the pieces that have information, is to actually reduce the size of the data that needs to be saved and loaded, which also speeds up the loading times tremendously. Because you are doing all of the calculations before the player ever enters the world, you don't need to do any sort of octtrees or other optimizations like that. Before I had to save the whole chunk (even with the empty pieces), then upon loading it would have to iterate over each piece to see if there was information or not, but now since it only saves blocks that have been placed it doesn't even have to check if a cell is empty or not at run time, it doesn't have to run a massively nested loop each time it loads the level, it just loads each piece and its location and places it, that's it. And because I don't need to edit the voxels during playtime, I can leave out all the redundant data. Theres no need for octtrees and stuff since its only loading data that you know for sure is going to be displayed.

(cont)

shoot, I accidentally lost the other half of the text that I was going to paste here, let me write it again, 1 sec

What I dont understand is how the guy in that one video I posted got his running okay. I looked up the technique he was using "Instanced Static Mesh" and I used the same thing. So he wasn't just drawing faces that are visible, he was actually drawing the whole world a cube at a time. He also has a little bit of overdraw, but I don't have any idea how he got his lighting that fast.

So while I havent added in culling yet, I know its not the ultimate issue because the scene I'm drawing is only 15x20 chunks big (8x8 x 64 x 15 x 20 = max blocks), and only about 20% of those are filled. Even if I was able to cull half of them it doesn't help me much because the scene I want to draw is probably at least 100 times that size.

Consider that screenshot that I originally posted, its not super complicated, but it makes the PC I just built (i7, gtx1080) crawl and drop frames constantly unless I turn down the dynamic lighting to its most basic level, and thats even after I've compiled it.

I know theres a way to write it directly with triangles, but the component for that in UE is still considered "experimental" and crashed on me a few times when I was testing it out. At this point, I'm not going to wait around for them to finish that part of the code before relying on it for my entire game.

While his thing is a nice tech demo, I don't really think Unreal can cut it here. I haven't seen a single voxel game from unreal that has enemies, full lighting, underground areas, etc. I did see one voxel game that had some of it, but the guy developing it used to work for Epic for 10 years, so he really knew what he was doing, and his was also at a tech demo level, no mobs, npcs, etc.

Thanks for the links, I'll check them out.

working on a new level, this time i've got a metal grate that makes for a really great windowpane.

This level is called "radio tower" and will take place up in a skyscraper I think

cool man, nice to see a visual upgrade, looking good so far

You sound pretty set in what you're doing, and I'm not sure what you want help with… so best of luck.

There's more to it than that.

A calls take c strings.
W calls take wide strings. Only Windows uses wide strings for unicode.

I know what I want to make, I just need to find the right engine and toolset to make it. From that demo that guy made, I thought that I could get it done in Unreal, but doing a fully dynamically light scene doesn't seem to be something that the engine does well.

I'm not trying to be a jerk either, I've spent a few weeks working on this researching all of the different methods that people are trying to use to get it to work in unreal, but it always comes back to, learn C++ and change the engine somehow.

So I'm kind of back to my original point in that, if I need to go through all of that effort to get Unreal to do what I want it to do, why not just learn C# and do the whole thing in Unity instead?

If you could tell me what you think of that one cave game over at red-aurora.com/?cat=5
that would be cool. I'd be interested in how they built their voxel system.

I guess I haven't run into this because I only use ASCII. But I really don't think that you need more than ASCII just for making the game window and maybe a few dialog boxes.

I'm glad that you think it looks better, Its a simple trick but it really goes a long way.

more work on this

Hey guys.

Experienced newfag modeler here. I'm just trying to make a simple rock. Well, unfortunately all of the models I make on Blender that I port to Unity end up having really weird, blocky shading on them.

Each face is either lit or unlit. I tried setting the shader to doffuse, but that didn't work. I also tried the "smooth shading" option in Blender but the lighting just makes it look fuck ugly in-game.

Any suggestions?

Well, you're not being a jerk, it's just you're not hearing me in what I'm saying; so I figured it was pointless to continue the discussion as I had better things to do.
Let me clarify, having terrain with a few passes is fine, it's just teleological terrain generation, and I do something similar. However, you seem quite keen on not listening to my suggestions on how you can keep that, but make it faster; which will work for teleological terrain gen or the traditional ontogenetic gen.

The only reason I can see is that Unity is a whole lot easier, and has less implicit overhead as far as I can tell; in addition to the point that it has GPGPU support which is great for this stuff… also Unity isn't functionally built for unreal touney style of games, and instead provides a template to build your own.

Already explained it in my earlier posts… also that tutorial link I provided goes over the basics, but not the octree system they're using.

duplicate normals or verts cause that.
Merge verts, and regenerate normals in unity.

Post screenshots. I have an idea what it might be, but I'd like to be sure.

Got basic inventory stuff going on. The player first acquires items, which then he can add to a loadout. This loadout is then added when spawning.

That's why people use normal maps, because that ugly lighting is actually normal. Apart from that, you could try enabling the EdgeSplit modifier in Blender, that generates smoothing groups automatically.

i already failed one exam, passed one today, and ive got the next one in two days (and the entire thing is both the practical and written part in one day, so i've got to be doing fucking exams from 9 am to 6 pm)
and after that they finally give me a week before my last calculus exam
i can already see myself failing the algebra one, since all i can do there is "find the matrix X if A*X=B"
and then i've still got to try and make some vidya demo for mobile to put in my portfolio
the only light at the end of the tunnel for me is that i get new yugioh cards after the algebra exam

into the trash it goes. Do you have any idea what it actually does to the model? It fucks shit up.

I have no fucking idea what the fuck those Blender developers were thinking this is a good solution, no it is not, messing up with the Mesh just so that it is properly smoothed is a STUPID solution.

Wings3D has a auto smooth function and it leaves the model intact, it's changing ONLY the SMOOTHING and NOTHING else.

Shit's retarded, your putting in more clicks which makes it annoying

Just make a list you can scroll through with all the items or even better a Steam Inventory-like interface

Maybe even allow to select more than one and then add them all in the same click

it's the only way you can alter the smoothing, though. It isn't something magical. You can either have a flat or a smooth shaded model, and that's it. One workaround are smoothing groups, which are done in Blender like this. You're just not doing it right. You're not supposed to apply the edge split modifier right away. If you have been using Blender long enough, you might have realised that there's an option for applying modifiers when exporting to another format. That's the moment where you have to apply it. Then, when you see it in your game engine, your shading is not fucked, and skeletal meshes will bend and move appropiately without any fuckery.

It was my most readily available option, I have to leave my rig soon, so I'm working on programming and making everything work as expected, graphical design will be dealt with later. I had planned for such inventories, but I don't have the time to make one.

You aren't supposed to apply it right away. Just keep it on the modifier stack. Blender applies it when you export. You can either let it automatically split your edges based on angle or do it manually by defining sharp edges (select edge, crtl+e).

"The C Programming Language" is the place to start

Why are you too retarded to use Blender's auto smooth? Fuck that modifier. You don't need it. Ever.

Also, as is clear from your rant you don't understand what smooth and soft edges even are. Go read a tutorial on the basics of 3D modeling before you post anything ever again.

Is auto smooth supposed to increase the number of vertices once you export a model to UE4?

Been working on UI lately. I added panels that show character details on them in Fire Emblem style. I'm probably going to set it up so the character acting is shown on the left and whoever you've got the cursor over is shown on the right.

Also working on menus, but they seem a bit cluttered. Currently it's just two menus max. One for generic actions and if you select something more specific like a special attack or item usage, it opens a second menu. I'm considering switching the menus to a more Final Fantasy Tactics style… The first menu is move/act/pass turn, then under act is basic attack/special attacks/items, and then special attacks or items would open a third menu for that. Really not sure the best way to streamline things.

Long time no see, guys! I've been busy. Here's the progress on Speebot so far.

The latest addition: NPC chit chat, currently working on NPC models and animations.

This looks cool as fuck

I meant smooth and hard edges.
Auto smooth does the same as the Edge Split modifier. You can select which edges of your model will be rendered smooth or sharp. Your game engine will render sharp edges as double edges which increases the vertex count. That's how 3D works. Keep that in mind and only mark edges as sharp when they're really needed. Like when modeling the edge of a knife or a crystal shard. Other than that, keep your faces smooth.

Looks really cool man

Looks awesome. What did you use for this? It looks like Unity but I can't be sure.

Thank you!

Thanks, this is my custom engine written from scratch in Haxe.

So like these? I'm getting 5000+ vertices on the AK-12 here, and I'm about done with it.

Anyone got demos? I'd love to give feedback on any work-in-progress bits any of you have to show off.

Remember, a good demo is the key to a successful low-profile game!

Looks about right.

Something about the presentation and the sound effects is giving me serious doujin-soft vibes. Not just Japanese, but Japanese indie vibes. The OUT is, of course, helping.

It's baffling, as most people who have try to emulate that feel come off as insincere, but it just kind of works.

Stylize away or enlarge some of the finer details, replace the text with moonrunes, throw in a MIDI soundtrack and it's there.

Are you a nip?

I still want to help with this bad as fuck.

Haha, thanks, but I'm not trying to emulate any specific style. Just trying to make things look fun and fitting. And nope, I'm not Japanese.

Thanks for the offer again, but I'm afraid I don't need any help.

I just did my tests. Specs:

I tried running the C++ first person test on that laptop. Sadly I had the default settings set to high (which the laptop sure didn't appreciate) and couldn't be bothered with writing my own graphics settings. However, I tweaked the settings at runtime using the console. Simple put, this laptop isn't made for gaming, least of all modern gaming. You totally can run a UE4 build on that thing, but it's not a nice experience. Both 30 and even 60 fps were possible, depending on how fugly you want the game to be.
Sure makes you appreciate the guys over at CryTek. I ran the old CryEngine 3 Free SDK on the same laptop. Logically it adds to the requirements of the game itself, but the sample scene (that beautiful forest one) ran at medium settings with 30 fps, if memory serves. The engine may be a bitch to work with, but it sure is blazing fast. From what I've heard, the new one is even faster.

I thought you'd never come back with results. Can you post your config?

Sorry it took so long, but I had to figure out how to make UE4 crossbuild to Linux. That took a bit of time with my slow ass Internet connection (had to download the engine 3 times before I got everything right).

I have none. I literally took the sample project and hit package as soon as it opened. The only thing I wanted to find out is whether that thing even runs on that laptop, which it does. It defaulted to high or very high, because motion blur and AA were enabled by default. Here are the commands I used:
>sg. 0
Then I experimented with the resolution (r.setres 1280x720w for example) and sg.resolutionquality. If there is something like a default config, tell me where to find it and I'll post it.
I did all of this because I need to set myself a realistic goal for minimal system requirements. As it stands, I'm probably going to build a rig using old used parts.

I haven't delved into nitpicking shit around the engine to make scenes run properly in my rig, as I said before, so I'm not too sure if there is a default one. There should be info on the documentation though. I can't believe my 6670 is this bad. It could run BF4 just find at medium settings and 720p.

I tried adding my github account to my unreal profile but could never access the source on github. Did you have to do anything special or just add your name?

Nope, just put the name in there, then log into github, navigate to their site and accept their automatic invitation.

But his design clearly points out that he is suited to function in long smooth planes(roads basically).And his name literally has the word speed as a root, The fact that the generic npc-bots are bipedal shows us that wheels aren't the standard locomotion for robots in this palce either.
I suppose underground tunnels or a some kind of sky platform would be possible but Underground doesn't seem like it would be good operational space for a robot on wheels,And a sky platform of anykind would be quite impractical for a being whose specific thing is going fast as flying off because you failed to take a turn becomes a high possibilty in such an enviroment.

Welp,I feel entirely too autistic after typing all that out.The game looks great

Huh. What did you have sg.resolutionquality set to, 25?

Looks rad as fuck, though given that you're supposed to be speedy, NPC conversations might slow things down.

I hope to one day make shit half as cool as you, senpai.

I tried 80, 75, 60, 50 and (for shits and giggles) 25 and 15.

And you got 60 FPS by setting it to 15, no?

Nope. user, I'll get the laptop again to test 4U. I'll post some exact results here.

Thanks.

So comfy

(checked)
Haha thanks, I don't want to spoil too much but there's a reason for that design and dialogue. Might reveal a bit more in the future webms.

Thank you, the game is indeed supposed to be speedy and dynamic, but also free roaming in a way (not open world, the levels are in a sequence) - there are no time limits or anything like that because that would put unnecessary pressure on the player. I just want them to have fun and make the game challenging with level design rather than time limits.

The characters are there to make things interesting and give clues and also tell about the world a bit, and there's no rush for the player to beat the level (there will be a separate challenge mode for that). This particular level is the first level of the game so the NPCs also act as tutorial guides, hence the frequency of the dialogue. I'm going to try to keep things dynamic when I design the rest of the levels.

This laptop has a 1440x900 px display.

8.9 FPS
9 FPS
11.75 FPS
19.45 FPS
20.75 FPS
22.22 FPS

Using this as a base, here's what happens, if I play with the resolution quality.
32.2 FPS
36 FPS
52.4 FPS
capped at 60 FPS.

After that, I set it back to 100.
32.6 FPS

Sadly fullscreen is bugged for me (or on UE4 Linux in general) and I can't set it to fullscreen without it also running at the native resolution. Otherwise I would have made some tests with that too.

That's comprehensive enough, great. I get 50 FPS on 720p and 50% of resolution quality, plus everything set to low. With quite a bit more stuff than the default project, regarding code. As for graphics, not much is going on. My map looks like this, but without planar reflections.

I think it's time to profile.

I'm running this in the editor, by the way, so when the game is packaged, it might run better.

Some progress, I'm gonna swap out the current animation I got here in favor of a SoTN style Crissaegrim/AoS Valmanway effect once I figure out how to pull it off.

How do you have the water rendering set up in Unreal? That may be why things are running so slow. I know that the Source engine has animation, bump maps, reflection, refraction, and fog in their water when using the "expensive" option, which is three times more expensive than the "cheap" water option.

Maybe try a map with no water and see how well it runs?

Note that I said no planar reflections. Planar reflections are very expensive. There are almost no reflections going on, actually. Only cubemap shit. The water I have going on is the transparent water plane you can get from the water plane example. I'll probably try out no water tomorrow. I don't think it will change anything, though.

...

Best Unity 5 C# course?

The fucking manual.

...

This will hurt your ability to get meaningful and important feedback. tune your autism to want perfection, and realize that you can't create that without getting fresh eyes on it from time to time.

I make music with Famitracker.

Care to give any feedback? Pretty proud of this one, but it's not my best. I try to maintain a melodic focus and steady pace, unlike some other Famicom composers who overuse arpeggios and abandon any kind of structure.

I like it, but I don't have great standards compared to an enthusiast maybe. I like how the music sounds casual and flows around, but I wouldnt want to listen to it for too long on its own because its pretty repetitive. Sounds like a good soundtrack for a driving game when you arent paying a whole lot of attention to the music and just going with the flow.

Is python a good language? I've been dicking around in it and it's pretty fun. Are the python bindings for OpenGL good?

preddy gud, but could use a bit more interesting riffs in there.

maybe it sounds harsh, because I do think that its a good song, but you have this repetitive beat or chord or something in there and it makes it so that It feels repetitive, like the bass line and beat doesn't change, so I feel like the song is repeating itself in the back of my head even though I know the melody is changing. I think thats why it seems off. I like it but I didnt think good feedback would be useful for making it better.

hnnnnghh

Keep up the good work, user.

no arps… 0/10
you didn't even try to be retro

I'm not a musician, but I really like it. A strong lead melody should definitely be the first and foremost part of any song, especially tracker music.

this is neat as fuck. How did you learned to use famitracker user?

It's the drums. I thought it might have this problem. Basically, the song goes through a bunch of different phases, and the bass changes a lot too, but since the drums stay the same through the whole song, it never sounds like it really "transitions". Thanks for the feedback, I wouldn't have figured that out.

you missed it, there was one, right at the end. Still not DUDE RETRO LMAO enough, I know

There's really nothing to it but time, I've been at it for just over 2 years. It's not as hard as you think, you just need to actually do it and improve your technique. Danooct1 has a good [if a bit dated] tutorial on youtube.
youtube.com/watch?v=bwNElW5IEo0

Should I post another?

Please do.

Haha, the cursor dancing.

Great song man. I enjoyed it.

A good language for what? Making games?
Godot uses python so I guess

IIRC Godot uses it's own language, but this language is pretty similar to Python.

Gdscript is literally python man, all the differences can be picked up in less than 5 minutes.

yes

Glad I was able to help. I'm especially autistic about that kind of thing so its hard for me to listen to a lot of chiptunes.

if you are having trouble with the arppegiation effect you could just do a fast pattern for and alternate volumes 20 40… not shure for famitracker though. =/

I just looked up your graphics card. Dude, that thing is utter trash. Performance wise it's apparently only marginally faster than my mobile GPU. Unless you're really hamstrung with your power supply, consider upgrading. For ~60$ on ebay or so, you should be able to get a used card no less than twice as fast.

What's your CPU btw?

An Intel i5 3570k, stock clock speed of 3.4 GHz.

I'm genuinely surprised. I expected you to have a full-on toaster.

I ran out of money a bit too early. I was surprised myself that I could run Battlefield 4 with this.

That's some crazy shit. I want to make sure my game is fairly well optimized, but that card is even way beneath my quite low min. system requirements.

You and I both, famalamborghini. I thought that "well optimized" meant that I could run my game on this card decently, but I didn't think that UE4 would have this much overhead or that the 6670 was this bad. I wanted this game to be tolerable on toasters, but it looks like that's not going to happen.

If you have a halfway decent Internet connection, would you mind checking out the CryEngine? I'm just really curious to see how it runs on that graphics card. If memory serves, it shouldn't have any problems (unless you crank up the visual fidelity). From what I've gathered, it's supposed to be even faster now, than it was when I last tried it out.

I have fiber optic installed, but I'm still connected to my router wirelessly. I'll get it now.

I just downloaded and installed Cryengine, the empty example map with the first person project runs pretty nicely, but it's very bare, so I wouldn't expect anything else. The first person template character is rather shit, so I'll get a proper project later.

There still has to be that forest example somewhere.
I googled a bit:

Hey guys,
New to game design but I've been coding for a few years now albeit mostly scripts and web / back end dbs.

Any thoughts on this Udemy Course:
don't use link shorteners.com/jlu49db

Any alternate good free tuts I could use instead?

absoloute solid track right there

do you make your own kits for famitracker, or download them?
i was using it a few days ago and was wondering how to make kits, it confused the fuck out of me

Getting closer to making this satisfactory

I have to input my personal data to download the project.

Can't you just use a throwaway account and fake data? I doubt CryTek has the means to do a background check for every new account on there.

I can, but I have no idea what they can do to me if they find out. Plus, I used my usual e-mail for this. Mistake, I know.

They won't do shit, if all you do is download that project. You're not going to make money off of it.
You can always create another account, log out with this one and try with the other first. Should you want to switch to CE for whatever reason, you could log back into your main account again. Once you start selling a game, you have to give them your personal data anyway (for good reasons).

Okay, I'll try to do that.

...

I need to know more

Music is from Turrican 3 on C64, but I want similar soundtrack. What else do you want to know ?

So long, eardrums.

ONE JOB

You need to find secret part for that :^)

What kinda game is this gonna be?

Top down shooter with mechanics usually found in FPS games such as quick saves, many monsters at the same time, many weapons, health and armour and ammo pickups, keykards etc.

The levels can be very big and load very fast, most likely the whole game will not have any level ending screen, just seamless loading (like in Half-Life or Quake2) It uses 256 colours only (32 in 8 different shades) and will be also playable on PENTIUM I systems (with MMX)

I love it.

Be still, my throbbing testicles.

I'd show you the "tittle screen" but it's not made by Salami INC and is just a placeholder to work with transitions…

There aren't that many Hülsbecks out there, and people trying to imitate his style almost invariably lack the sheer drive and intensity his punchier works tend to exhibit.

So good luck. I hope you manage to do it/find someone who can pull it off.

Yeah I know, there's tons of good musicians on C64 though, people make demos to this day. I doubt any of them would want to coop with a lamer like me though hehe. I have a friend who does MX music, but that's not really what I want. We'll see….

As I said this music is from game called TURRICAN 3. Here's a webm of me playing this game with music, it's a different track though. Enjoy

Oh. I thought you meant specifically like Hülsbeck, not generally something for the SID chip.

He's got a very distinctive sound to him.

I'm not sure if the guys who ported Turrican 3 to the C64 just rearranged music from the Amiga games or if there was some original stuff there. If it's original, it's one of the few decent takes I've heard.

This one does not in the style of which I'm talking.

The music in Turrican 3 is somewhat random, the guy who made the game (yes, it was one guy who coded it and made all graphics) just took bunch of random sids without even asking. All musicians said they don't mind and are happy with the game though hehe.

Which engine would handle something like randomized 3d maze-like levels best? Unreal or Unity? Or is there a better choice?

Depends on the scale and how complex you want it to be. But unless you're going for some crazy cutting-edge shit, either will be fine.

Nothing super insane. Likely low poly, or fairly low poly.

Levels will be self contained, certainly not large in size, but big enough it would take a bit to find your way through.

Is this a wad or are you just using graphics from doom?

Oh fuck me man.

I just realised that after rigging and 40 animations done with my character mesh, her hands are too small. I was bothered because when holding a rifle, the trigger finger would barely reach the trigger. I thought I hadn't posed the hand correctly, but then I bothered to compare my hand with the mesh's.

Either will do. Here is the short version:
Unity:

Unreal:

Unless you want to target toasters, go Unreal.

It's a wad.

As for the music I like… I have hard time explaining what I like in music, but I know one for sure… The sawtooths bzzzzz It cuts into my brain !

It's looking nice

Thank you, I've been trying to fix this and make everything better. Someone recently mentioned they didn't like the slash effects and I'm inclined to agree with him, but fucked if I know where I'm gonna find nicer ones

I already saw the "tittle screen", it's exactly what it says in the scrolling text on your title screen.

One more

Awesome, that's what I was thinking, but I wanted a 2nd opinion.

Thanks.

Sounds like the usual acid trance kind of sound you find in demos. A lot of that comes from just being a very suitable and convenient style to portray on the hardware. I think that carrying over to the later Amiga demos is more tradition than anything else.

Maybe you could hit up some ex- or current scene folks once you're getting closer to finishing the game. The general feel among trackers not part of the demoscene seems a lot smoother and harmonious, as they're not written to be part of a boasty visual display. More jazzy and standalone than the dissonant high-octane buzzing backdrops of old.

Any game with a chunky low-res font like that has to be good. Maybe look at some later DOS titles if you feel like tweaking it. Some of those fuckers are gorgeous.
I hope you intend to do something about that nasty scaling, though.

Yeah I guess so, i'll take your word for it. It's funny, because I don't listen to electronic music at all, I'm all metal, but C64 and Amiga sound has something special.
C64 is simply mean, mean techno machine

Last one… this time from different composer.

New font. Still not 100% on it, may replace it again down the line.

Monospacing would make typesetting a looooot easier, but monospacing also looks like arse in regards to things like I, L, and most punctuation.

Nice font. I'm using monospaced ones too since it's way easier to code for it, but I decided to code something with normal font too.

my little balls are quaking

fags who say mega man was awesome haven't played turrican

I've been trying to get a voxel system going in Unreal, it hasn't went well. I've come to learn that anything dynamically created (like voxels or potentially your maze gen) needs to be light with dynamic lighting. Unreal has a huge overhead for this, and if you are planning on creating a lot of geometry and using a lot of lights, it might not be an option.

I have to respectfully disagree with the user here:
Simply on the grounds of how these two engines handle this dynamic lighting situation.
I posted about it some here:

I believe Unity handles dynamically generated scenes much better. Unless you wan't to dig into some deep C++, Unreal is better suited for prebuilt scenes with baked lighting.

There might be a way to get it running fast, but if you go with Unreal, get the maze gen done FIRST, since that will be your biggest obstacle.

So how much processing do most engines handle by themselves?
After doing part of the tutorial of Godot, I assume the engine handles positioning, movement, physics, and most things rendering related.
Can anyone confirm/deny?

How does Godot compare to more standard game engines such as Unreal Engine or Unity?

I love it.

looking great dude

You have to stabilise them.

Well, to be fair, most of Holla Forums and Holla Forums in general is from USA / Canada, so grown ups had PCs and kids had their NES. Here in Europe it was mostly home computers.

this is my newest picture of the "radio tower" level that I am working on. it doesnt have gameplay yet because I want to put in a new enemy for this stage.

The game also has widescreen support now.

Do you have a form of vertical movement yet?

NPCs now turn to face Speebot when chatting.

yes, thats been in the game even from before I started posting pictures of it here…

I mean upward, not the ability to fall. You dropped down in an earlier video showcasing you making a level, and didn't seem to have a method of getting back up.

oh, I see- you can jump, and there are stairs in the game that work- so you could get back up if I put those in.

looking nice, shouldn't the NPCs get back to their original position when Speebot gets out of their sight?