/v/ - Video Games

Important, complete this poll so we know roughly how many people to expect for demo day/game day. Hopefully we can have more than 1-3 games show up this time.

http://poal.me/a63qxw

http://poal.me/a63qxw

http://poal.me/a63qxw

>TEN DAYS UNTIL DEMO DAY WHERE IS YOUR GAME?

I gave up.

I rather won't work on it or any other game ever again.

Game developement is too large pain in ass tbh.

>>14692913

Apparently if you run your Lua code through LuaJIT, it optimizes the VM it runs on, or some shit. I don't know.

>>14692913

Regular lua is slow as shit, it's luajit that's fast.

Demo day coincides exactly with my finals. So I seriously doubt I will have anything to show.

>>14692875

No demo, still pissed that Godot 3 sucks for 3D math. WHAT FUCKING 3D COORDINATE SYSTEM ASSIGNS (0,0,0) FOR A NODE THAT IS INSTANCED AT (0,2,0)?!?

>>14692955

>WHAT FUCKING 3D COORDINATE SYSTEM ASSIGNS (0,0,0) FOR A NODE THAT IS INSTANCED AT (0,2,0)

Sounds like it's doing a local vs world coordinate system. Eg, in SFML, which is strictly 2D, a lot of sprites/transforms can get a position based on it's local transform, or tree of transforms (world space)

>>14692913

I don't think lua being fast has ever been a real thought in anyone's head, but with LuaJIT and FFI you can offload all the important parts to C if you're really hitting bottlenecks.

>>14692959

What is the advantage of this? It makes no sense to me to have up to 3 fucking coord systems at a time (local, global, 2D).

>>14692920

>>14692921

>>14692961

Is there a way to group things together in luajit for cache's sake?

>>14693010

I don't understand the question but as long as you're not fucking with metatables constantly LuaJIT is optimized around regular lua usage. LuaJIT likes numeric tables more than keyed tables but that's a minor thing. The only other way to "group" something would be with FFI structs.

>>14692910

Are you doing some Jade Cocoon rpg ?

>>14693044

I mean. If I put variables or objects of the same type in a linear array. Then run through them 1,2,3,4, etc. Will it work as expected, or will it do something weird like make the values pointers to objects.

>>14693015

move_and_slide() returns a Vector2 that represents "the movement that remained when the body stopped."

Not sure if that's the same thing.

>>14693104

>>14693015

you can use move_and_collide which will stop when it hits a collision and return a collision object

https://godot.readthedocs.io/en/3.0/classes/class_kinematicbody2d.html

>>14692875

how do I properly submit my demo?

>>14693210

>run out of magic, fall unconscious

Make sense, but would it make more intuitive sense if you, perhaps, called it "energy" instead?

>overload magic, fall unconscious

Missed opportunity to not have them explode and then fall unconscious, if this is meant to be kid-friendly.

So with your gameplay mechanics, would buffs/debuffs behave in a straightforward manner, or would they lead to some interesting high-risk-high-reward gameplay?

>>14693417

>Make sense, but would it make more intuitive sense if you, perhaps, called it "energy" instead?

Probably. Other option is to figure a term out for it tied more specifically to this verse. In code, it's just called magic though.

>Missed opportunity to not have them explode and then fall unconscious, if this is meant to be kid-friendly.

Overloading was gonna have different effects from just depletion. Definitely a little more "boom"ey so it's clearly different. But yeah, either way, they fall over and zonk out.

>So with your gameplay mechanics, would buffs/debuffs behave in a straightforward manner, or would they lead to some interesting high-risk-high-reward gameplay?

Depends what you mean by straightforward. Risk management is definitely an aspect. Buffs and debuffs, or rather "conditions" if you're thinking things like statuses, can be good or bad depending on the context. If you're trying to raise your ally's magic up fast with a certain element, a condition that increases fire magic gain from "infusions", or attacks that give the target magic, then it's beneficial. If you're already too high on it though, then an enemy with a high fire infusion skill and stats can try to throw you over capacity and knock you out. That's a fairly low level example of how that management comes into play. More complex systems are in the works, more counters and setups for those sorts of strategies too. We're hoping to have that presentable with a third species by demo day. I'm not sure if that answers your question? I'm still working on presenting this stuff well.

I should also mention, this does the Pokemon sort of thing where a creature may have a lot of skills to choose from. Skills a creature can learn can come from their species as well as their currently attuned discipline, which is job system stuff that I've got in the backlog to implement. The intention is to not necessarily restrict learning skills from levelup, but also from exploring the world and quests and the like. Guild Wars 1, if you've seen it. Either way, there's a lot of possible skills to take, but only a limited number, haven't decided how much, can be chosen before going out of safe areas like towns. This makes the player really have to think of their party composition, strategy, and all that jazz. Things are still being worked out, but that's the gyst.

>>14693393

Properly submit isn't really a thing, as far as I recall. You can just post it during the official Demo Day topic. I'd suggest setting up a post ready so you can just throw it up and leave.

http://8agdg.wikidot.com/demo-days

The Winter one was kind of disorganized and last minute, despite everyone was rushing for it. It'd be nice if some talk was going on for what to do with the unfinished parts of this wiki page.

>>14693104

>>14693186

I'm already using move_and_collide(). Sorry, I forgot to mention that the NPC is following a path. Here's what my physics_process code looks like (in C#).

Vector2 motion = new Vector2();
        float OldOffset = follow.GetOffset();
        float StandardMovement = motionSpeed * delta;
        follow.SetOffset(follow.GetOffset() + (motionSpeed * delta));
        motion += follow.Position;
        MoveAndCollide(motion - Position);

        //This is supposed to stop the NPC, but it doesn't work since motion.Length() isn't returning the correct values, and the right side of the if statement should probably be reworked
        //if (motion.Length() < IDLE_SPEED * StandardMovement)
        //{
        //    follow.SetOffset(OldOffset);
        //    MoveAndCollide(follow.Position - Position);
            
        //}

And here's what it's supposed to be doing.

if(get_tree().is_paused()):
		return
	var old_offset = follow.get_offset()

	var standard_travel = walk_speed * delta

	follow.set_offset(follow.get_offset() + (walk_speed * delta))
	move_to(follow.get_pos())

	var travel = get_travel()
	if(travel.length() < movement_threshold * standard_travel):
		print(travel.length())
		follow.set_offset(old_offset)
		set_pos(follow.get_pos())
		travel = Vector2(0,0)

	update_travel(travel)

Right now all that happens is that the NPC can be stopped for a few seconds, but once follow (the PathFollow2D node on the Path2D) circles around the path to a certain position, the NPC snaps back to it. I'm not sure if I'm explaining this well, I'm tired and have a bit of a buzz.

>>14694010

I go ham with backups. I have my main dev machine, then an old secondary computer to the side that I use as a home network server for git. I want to get it to communicate online when I open the server so my artist can contribute through it, but my I can't even get onto my router's config page anymore for some reason. I don't want to host on an external git repository, but I probably don't have a choice eventually. On top of that, I also have a TB sized external HDD and I also carry everything on a flash drive. tl;dr, I'm not losing my data anytime soon.

For the record, I highly recommend using git in general, and home server if you have an old computer to host it. There's some comfort in pushing my work to it at 1AM and heading to bed. You can roll back when things break, local or server, anyways, so it's nice to have as a safety net if everything just decides to break during a major rewrite or feature addition.

>>14694010

Right now, 8ch is being shitty and slow. Pages are laggy and often freeze/crash the browser and eat up 30% CPU before doing so. I think it is related to the trending boards at the top. I disabled JS for the site and pages load almost instantly, it's insane. Replying is a pain, though, I had to paste your post number, I couldn't just click and have it go

Anyways, in a bit of good news for me, I thought my PC was dying and stopped using it like 7 months ago. I was getting horrible horizontal tearing on my screen, and I tried two different graphics cards, swapped my dual monitors around, different cables, assumed it was a mobo issue since that was the only thing I didn't change. I found out today that it was actually the monitors; I hooked up my Wii U again and had the screen tearing within minutes.

Don't mind me, just having nested classes 3 classes deep with generic types and a dozen virtual empty functions

https://hastebin.com/lodavotebe.cs

>>14694242

why though

>>14693982

i don't think you are explaining it well either so you should try to sleep on it if you are not already and come back later

one thing that might be useful is test_move() which returns true or false if it is going to collide with something,

another is to declare a variable as

new Vector2(MoveAndCollide(motion - Position));

not sure how c# handles variables as functions

and subtract it from motion

also have you tried checking the offset.length() instead?

>>14694269

and because 8chan add spaces around the code tags now i look like i reddit spaced the whole post

>>14694267

If I make a derived class, then I would have to typecast to access the new variables. Eg MyGame derives from Game and has a Content object that holds all my resources. If I made a state or controller object, it would see only a "Game" object and I'd have to cast down. But using generics in this way lets me avoid that

I realized that the state controller probably shouldn't have resource access, and I can't guarantee what would appear on the state object itself, so I was actually able to refactor it so that the state controller didn't directly rely on anything

Guys, I found a way to generate true random()

Who here

STRESSED

about the fucking demo day?

>>14694010

>>14694090

>external HDD

>dropbox

>google drive

Use proper VCS you double niggers.

>>14694559

I don't understand the low level of progress that has been made either. It's been three years and it really could look a lot better than it does now. Although maybe it makes sense if he only works an hour and a half each day while explaining everything he is doing to the camera. It looks like he's doing his stuff in a really slow and through way since he does a lot of things that aren't necessary in the short term, but make it better in the long term.

I think that he knows what he's talking about, but I would rather read it in online programming tutorials and books, which have the same information, just presented in another way. I guess the real value is how he's putting it all together if you aren't sure how, or any insights he has that might be unique to his experience.

>>14694628

In the latest video he was like "I'm not trying to make a well designed game, I'm making an educational resource" which leads me to believe he's not following any plan at all and the finishing line doesn't actually exist. And if he does "finish" its obviously not going to be good in any way.

>>14694580

who here wants to

GAME JAM 4 DEMO DAY

>>14694639

>GAME JAM 4 DEMO DAY

oh god no

Unrelated to game dev. Anyone have a clean AGDG logo handy, or a clean variation? I'd like to make you fags a webm when finals are done.

>>14694712

Also FYI we do keep logos on the wiki

http://8agdg.wikidot.com/wiki:thread-template

>>14694756

Its a single C# script.

>>14693713

cool thanks, will do

Do people implement dialogues as state machines?

>>14695135

Use a tree structure.

Just discovered this, which one of you lads is behind this game and when can I play it?

>Play as a curious kid who starts out cautiously lurking in the enigmatic and dangerous worlds of Infinity Land. Mysterious creatures await in a treacherous landscape you must explore to find Diana. She has been taken captive in another dimension by the Pig Merchant.

>Merchant

>Infinity Land

>Infinity Next billboard

>I Am Growing Stronger

>Red cap

>Fucking Spurdo in the fast food joint

>>14695135

they do

>>14692885

>have a demo but haven't made any progress since last demo day

I have failed you, /agdg/

>>14695202

Good stuff. Hopefully get this thing running this evening.

>>14694658

I don't know about using mesh combiner.

What we do is making the clothes use the target bones to skin, rather than combine meshes or bothering with sync animation.

so long your clothes deforms like your body, it works flawlessly

https://pastebin.com/b2qM2JGx (rather long to just paste here in [code/]) The method you want to look at is Stitch

ObiSolver is from here: https://assetstore.unity.com/packages/tools/physics/obi-cloth-81333

Delete all references to it if you are not using this package, it is not required

>>14695230

No, anon You have failed yourself.

I don't have a demo either, but I might post a small program asking people to tell me their performance and CPU, since I'm optimizing a critical feature and wanna know how it scales hardware-wise.

>university wants me to post a bunch of blogposts explaining the work I have done for the current year so they can make sure I've actually done work

>have to write about 55 weeks worth of progress updates in two days

I wanna kms

This is super fast

			fixed4 c1 = tex2D(_LeavesS, IN.uv_MainTex);
			fixed4 c2 = tex2D(_LeavesA, IN.uv_MainTex);
			fixed4 d = tex2D(_MainTex, IN.uv_MainTex);
			fixed4 e = tex2D(_Green, IN.uv_MainTex);
			//fixed4 f = lerp(d, e, saturate(e.a - _FLeaves - _Season));// e.a > _FLeaves ? e : d;
			//fixed4 c = lerp(c1, c2, _Season);
			//fixed4 color = c.a > _FLeaves ? c : f;
			o.Albedo = c1.rgb + c2.rgb + d.rgb + e.rgb;// color.rgb;
			o.Alpha = c1.a;// _DEBUG_VISIBLE > 0.0 ? color.a : 0;

This is slow fast

			fixed4 c1 = tex2D(_LeavesS, IN.uv_MainTex);
			fixed4 c2 = tex2D(_LeavesA, IN.uv_MainTex);
			fixed4 d = tex2D(_MainTex, IN.uv_MainTex);
			fixed4 e = tex2D(_Green, IN.uv_MainTex);
			fixed4 f = lerp(d, e, saturate(e.a - _FLeaves - _Season));// e.a > _FLeaves ? e : d;
			fixed4 c = lerp(c1, c2, _Season);
			fixed4 color = c.a > _FLeaves ? c : f;
			o.Albedo = color.rgb;
			o.Alpha = color.a;

Why? I've assumed that the tex2D was the slowest thing in the shader

>>14695757

>science bois

what the fuck is wrong with you?

>>14695765

He's retarded and has been raised on a diet of black twatter and Facebook "memes."

>>14695757

Just decades of material science research. Have fun idiot.

>>14695739

What is slow is based on a whole bunch of factors. Simplifying it to "x is slow" means you're an idiot.

Doing multiple texture accesses when your hardware can't do that many at once is slow.

Doing it when your hardware can is fast.

Add the fun nonsense that is caches, tile based rendering, etc. and suddenly it's a lot more complex.

>>14695757

>density and mass

The shape of the object matters, dude. And how concentrated your force is. 1 newton will break through anything if it's applied on an atomic scale.

>>14695774

real talk though ironic facebook memes are actually quite good

>>14695739

>>14695778

I've isolated the issue

o.Alpha = color.a;

is the whole reason for the slowdown

changing it to either (c1,c2, d or e).a will not slow down the whole thing

>>14695739

The only thing that stands out to me is the ternary in the slow one. Aren't conditionals supposed to be slow in shaders?

>trying to make snowboarding game

>physics are a problem

>search "programming snowboarding game", expecting nothing but unity forums pajeets

>someone wrote a university paper on it

wew

>>14695757

Here's a semi-realistic solution made by some GURPS nerd.

http://panoptesv.com/RPGs/Equipment/Weapons/Projectile_physics.php

>>14695788

Conditionals aren't the problem. Branches in code are.

GPU's can't really into branching. The reason for this is because many cores are actually combined into a single compute unit, with the compute unit handling giving all the individual cores orders. However, the compute unit doesn't actually know what branches the individual cores take, so it just has to do all of them.

As a result, branches with a lot of code in it are really slow. But branches with minimal code in them are a non-issue.

A simple ternary operator is probably not the issue, as many processors actually have instructions that allow them to pick one of two variables based on another variable/register/state without any sort of branch instructions.

>>14694773

yeah maybe you can bully us into actually doing work

>>14694773

Gonna be more like emo day when half of them are sobbing over being functionally nodevs.

>>14694101

Yeah, 8ch without js runs great, but not having the qr box, backlinks, or autoupdater really sucks. I've been thinking about combing through 8ch's js and try to make an 8ch js-lite or something. Or I could work on my game.

>>14695391

That's what I did. I didn't deal with syncing animations. The output is 1 combined bone rig and 1 combined skinned mesh renderer.

Tell me why I shouldn't spend 500 bucks on getting this gear.

>>14696143

Because your tools are only worth as much as your talent.

>>14696042

You can disable some of the scripts, like the fucking emoji one. The main.js is apparently a fuckton of scripts concatenated together, so it's probably possible to rip out the ones you don't want, block 8ch's real script, and add your own as a userscript.

>>14695739

Problem is that you are doing 2 conditionals to calculate your alpha. But I'm wandering how you profiled this.This is for your growing leaf shader right? Why are you using 4 texture samples for that? You can do it with 1 texture sample that have 2 channels and with no branching conditionals. Use the texture's red channel for just the raw base color. Green for heightmap. Don't use alpha channel for heightmap because typically alpha channel has less precision than RGB unless otherwise specified in settings. Multiply base color with a precalculated uniform float3 that was already lerped in C# instead of in GPU. Then use step() to clamp output alpha to your heightmap. Pseudocode:

fixed2 data = tex2d( texture, uv ).rg;

fixed3 color = fixed3(data.r,data.r,data.r)*_myColorUniform;

return fixed4( color, step( data.g, _myLeafGrowSize ) );

>>14696173

the motion leap controller is 75 bucks.

the breath controller 150.

a midi keyboard is like 170 bucks.

So, all the gear is like 450 bucks.

>>14696210

That looks much better than I expected tbh

Congratulations anon

>>14692913

>Where did the "Lua is Fast" meme come from?

Use this instead: http://www.squirrel-lang.org

>>14696210

Why does your character have scratches?

Please do not abuse the vidya characters.

>>14696310

Why not just check if the player is colliding with the npc with Godot's collision tests?

>>14696446

>Why not just check if the player is colliding with the npc with Godot's collision tests?

They changed all the collision tests for Godot 3 and it's pretty confusing now. KinematicBody2D used to have a simple is_colliding() function, but that's gone for whatever reason.

How does much time does it take to model, rig and make one animation with a single character in 3D with the necessary experience?

>>14696562

What level of detail? Give pics of the model similar to what you have in mind. It could either take from 30 minutes to hours or days.

>>14696659

>Zyklon BFG

Not bad, good name

>>14696686

The BFG name change already happened on a much older version of Moonman mod.

>>14694139

>>14694535

Alright, I have played your demo, and I think it has potential. Why? Because its simple. Plus the Quake-like gameplay sells it for me. Yet there are a few things I'm gonna sperg-out over.

First off, the weapons have no "weight" to them. Shooting the shotgun or the revolver(?) or the rifle feels dull. This weightlessness increases more so once you realize you have to reload. If you didn't have to reload, like in Quake, then the weight issue wouldn't be as bad. But I'm not sure that's what your going for. Are you the same guy that as making the Anime Stalker thing? Or is this something else? Either way, the guns need to feel like your shooting a gun.

Secondly, the art style is kinda jarring. Now, I know your probably going with a low-poly look, but said polygons have low resolution textures. But the environment doesn't. Coupled with the random Chinese everywhere and the lighting and all that it makes the demo look like some advertisement from the mid 90's. Like "Check out the incredible world of 3 DIMENSIONAL GRAPHICS!", throw in a bunch of screenshots in different locations along with some work-in-progress stuff and boom, a time capsule of horrific surrealism. Now I can understand the desire to make the players look distinct from the environment since a lot of shooters out there get to the point where you can't tell if your shooting somebody or if your shooting up a rock. But at this point, its weird. Like I'm trawling through some Quake version of Worlds.

I don't really know what your going for, exactly but it seems fun. Also I like the idea

of a "Waifus with Raifus" game. Although I don't know which direction your going for, so I can't really say which one you should go for.

>>14694139

Is that the gorilla warfare copypasta in moon runes?

>>14697560

hot shit, it was made over a weekend and we cut the core game play, with you playing as 5 girls over 30 seconds clearing a room swapping between characters in a sorta sudo turn based/supper hot loop. i didn't really get it but the guy had a full page written up already for the game play. and the style was a byproduct of compromise

idk there was talk about keeping going on it but i'm not sure the programmers want to revisit their spaghetti code and i'm burnt out on that ass.

shit forgot the >

>>14694756

The names are just because I needed names to distinguish from allies and was lazy. We're hoping for a playable demo, but we do have a backup plan if I can't make it in time. Art, we should be good on possibly. Programming, it can be hit but I might look at my priorities right now and see what needs to be thrown in fast and what can wait. I'd like to not do a trailer, but it's on the table if it has to be.

kind of feel like just ripping off cortex command since it's the most dissapointed I've ever been by a game other than like spore

Anyone ever use teachyourselfcs.com? How far can you guys go? Are you a SCIENTIST?? Is it missing anything?

>>14698219

CS guy here. Haven't used that particular website, but the sections of studies and the resources it lists are good material, SICP and the dragon book especially (a bit surprised Sipser's Theory of Computation isn't listed). Anyway, if you're disciplined enough for self-study this site should help you a lot.

>>14698299

Is there loss in that second pic?

>>14698330

MOTHER FUCKER

>>14698461

i also don't art well, so i would need someone who can. i can music ok, but would like to know what everyone else uses to churn out catch beats and tunes you know what i mean fam

>>14698461

>i figure i could start out like every other indie dev and make pixelshit

So you don't want to aim for quality from the get-go? I'd say you're on the wrong place then.

Maybe there's a reddit guide to weaseling your way into the industry so that's where I'd direct you to.

>>14698472

For music, I assume the standard is still a midi keyboard and FL Studio

It's not that big of a jump. Really, you'll have game specific patterns and tricks that completely blow away whatever .Net you know (eg making things like large mutable structs for performance), but for the most part, it's pretty similar. Obviously you've played games, so just apply your programming knowledge to make a game work

I want to make a game. How smart do you have to be to get into scripting? I'm at about average intelligence.

>>14698461

Take your poo to the loo.

>>14698520

>I'm at about average intelligence.

Lol. Your asking the wrong questions. Try watching behind the scenes of video games on Youtube first. You sound like you've never touched a piece of code on your life. Familiarize yourself with what you want to do first.

>>14698520

>I'm at about average intelligence.

What did he mean by this?

>>14698461

>>14698472

This is bait.

>>14698584

just use UE4 or Unity. It's the way to go right now if you want to actually make a game instead of fucking with making your own engine for a few years.

>>14698594

>UE4

I don't want my game to be owned by the Chinese.

>Unity

Ditto.

>>14698584

XNA is dead and buried, but FNA is an open source replacement, and MonoGame is an open source extension of it.

>>14698584

XNA is dead and buried, but FNA is an open source replacement, and MonoGame is an open source extension of it.

>>14698597

>>14698598

>You can double post and bypass the spam limit if you disable JS on your browser

INTERESTING

>>14698597

Monogame looks good, but I don't want to work with the pajeet trainwreck that is .NET/C#

>>14696548

am I missing something from your other explanations? Move_and_Collide returns a collision object if it collides, all you have to do is check that object to tell what its colliding with

>>14698607

>pajeet trainwreck that is .NET/C#

.NET/C# dev here. You only have to deal with poojeets if you're building enterprise solutions. The industry is absolutely infested with them.

I suspect that you will have nowhere near the same experience using the same framework to write games.

>>14698626

>pajeets

I have to deal with these while wageslaving and they're even more annoying to deal with than mexicans

>>14698644

Well like I said, I have no idea what I'm doing for 3D and thought I could get away with a cute little cheat like that. Like I thought I could just shrink the two right vertices inward, like a shear transform, and it wouldn't be a problem. Hmm.

>>14698584

Godot is promising.

>>14698652

>Your game is not owned by them, you just have to kiss their ass

Yeah, nah.

I'm not so much looking to create the best game, I just want a good programming practice. Enginedev is something I want to look into and I need resources. Good resources. Not shit from 2002 that has only one tutorial on how to do a Hello Triangle. I need tutorials on at least how to do lighting, model loading, and physics.

>>14698652

>>14698596

>>14698713

the unity team is a bunch of cucks but at least they are not chinks, that being said they basically have built their business in that fucking asset store they run. trying to shill themselves as the go to engine to shitheads who "spend money to make money" betting on the libshit, gamejammer, nodev to spend more money in the store than they ever will in royalties, on what ever pile of hot garbage they shit out. but shit if you make your own crap and don't buy shit it will beat the 5% off the top that unreal will take.

and godot is looking better every day

>>14698725

>the unity team is a bunch of cucks

>cucks

I really don't even know what "cuck" means anymore.

>>14698731

>I really don't even know what "cuck" means anymore.

weakling.

>>14698750

>in the civilized world it's called a deal

Then no thank you. I'd rather study and begin work on a game engine that I can call my own then use Jewish shit.

>>14698713

> Enginedev is something I want to look into and I need resources.

Get the Godot 3 source.

>>14698764

So you're going to use Jewish shit after you work on your own game engine?

>>14698713

>tutorials

If you want to get into enginedev, read the source code to some id Tech engines and find some papers written by developers or studios you like. I personally recommend the Bitsquid/Autodesk Stringray blog.

>>14698731

it means someone who would put other people before their own. be it other children before their own children, or a foreign people before their own people. and the people that run unity have made it clear that they support diversity before hiring the best staff. that they are willing to hold that party line of cultural suicide, by the deracination of traditional western culture.

>>14698767

the only people i've ever talked to online that latch onto a typo instead of trying to make a counterargument are SJWs leftards and other types of kikes

>>14698330

Alright thanks.

>>14698925

Radical. Is it all going to be desert boarding or are you planning on other biomes as well?

>>14699123

More biomes are planned. The rough story outline is that you live in a fishing village on the southern tip go the continent. Your village is invades and/o destroyed and you have to travel north through desert, then jungle, and finally tundra to get some kind of mcguffin to save your family or something, standard hero's journey stuff.

>>14698925

Nigga you need to lerp that shit

good stuff though

>>14699512

>that spelling

I really shouldn't phonepost drunk.

>>14699541

It is lerped, the rate is just really fast. I should be sampling the four points of the board's corners and then averaging the normal, but I'm primarily focused on fixing the camera and making the level longer.

anyone else notice that the quick reply box can't be resized to a smaller size after resized to be bigger?

>>14699512

is there a yeti

>>14699567

Personally, when I hold a gun or rocket launcher, I hold it positioned at my crotch. It hurts when I fire it, but looks much better from my own visual perspective.

>>14699567

It doesn't really matter but just make sure you actually understand perspective and don't make it point way too far to the left compared to where you're shooting and look retarded like in the new prey game

>>14699575

IDK, I haven't made any enemies yet whatsoever, so that's all up in the air. All I know right now is that each biome/region will have at least one village with a unique group of people. I really want to emphasize the importance of small, tight-knit communities, something that I feel modern society has failed to preserve.

>>14699567

I personally think hip-fire looks cool, even if it's unrealistic. It doesn't really matter though, the primary concerns for FPS games should be:

>guns feel good to shoot

>killing enemies is satisfying/cool

Quake did centered guns exceptionally well. The animations in Quake are rather bland, but the sounds enhance their feeling. The fact that enemies explode into gibs on death, however, is what really gives the gameplay "meat". FPS games without gratuitous gibs are trash.

>>14699729

Are these the same type of file?

In unity you can set different import settings depending on the type of file, so you might have one of those imports set to a much higher transform.

>>14699729

Check the import settings in unity, and make sure the scale is the same as you're using for the other model (use file scale [true or false], and scale factor).

Looks like u have all the other bases covered.

Also, how did u do that hair? Mind pointing me towards the tut or a similar method? Or did u just sculpt it?

It looks really good

>>14699729

>how do I fix this?

The way I'd do it: set the scaling to your liking (I guess all 10s will look ok) and save it as a gameobject prefab

-OR-

change the import options on the model

>>14699729

also try exporting both models in the same FBX, and importing in Unity; it should fix any unity-side issues (use CTRL+D to separate them in unity inspector).

Otherwise, if u still have issues, it's blender side.

>>14699784

They are the same .fbx. I even tried exporting both of them at the same time and each mesh is bigger. That said if I spawn them as the fbx container assets they have the correct scale. But this doesnt make sense. the internal Unity scale should be consistent. My torso, arms, legs, etc. have the same internal scale except for the hair.

>>14699804

The only time I've ever had that issue is when I didn't apply my scale in blender.

>>14699851

nice, glad u solved it.

I had no idea u had a thread on agdg, nice, i'll read up now. thanks

>Delta Scale is broken as of v2.79c

guess i'm sticking with 2.78 then

>>14699870

It's not that bad just copy the delta scale to your model scale and then set your delta scale to 1. Then just apply scale and it gets rid of it.

>>14700128

Modern games have anything from 20k to 150k triangles on the LOD0 of characters. Even stylized games like Paladins have 20k.

>>14700128

That's about right. Here's a video that goes through all of the Counter Strike SAS models. The current gen models are just under 16.5k polygons.

Let's go through the polycounts of other hero characters in other games.

Dead Space 2: Isaac Clarke - 21,595

Call of Duty: Black Ops: Frank Woods (full gear) - 19,777

Bayonetta: Bayonetta - 23,000

Gears of War: Marcus - 15,000

The Legend of Zelda: Twilight Princess: Link - 6900

Mass Effect 3: Shepard (male, N7 Defender Armor) - 25,440

15k for your main character isn't to many polygons unless you're making a game for the original Xbox. But the question you should be asking yourself isn't "is 15k too many triangles" the question should be "do I have enough detail to justify 15k polygons"?

>>14700243

I think it looks better than what you had before. What engine are you using?

>>14700243

Those cookies look delicious.

>>14698461

>how much of a learning curve is video game development if you're already an experience software engineer?

Depends which area you're talking about. I feel like there's a lot more iteration/freedom to shift goals around than for most software, games sometimes change genres entirely over the course of their development. You also have more freedom to create your own file formats and such instead of having to be compatible with obscure shit. On the other hand, performance concerns tends to seep in or at least influence everything, since you can't just throw more servers at it when something is sluggish.

I think game dev is also more intertwined with the art side of things, e.g. when you want to make effects look good.

>i've heard good and bad about unity, but what else is there that compares that allows me to use the .net framework?

SFML has bindings IIRC, so if you want to start by crapping out a pixel platformer you could start there.

>>14700268

15k tris isn't worth sweating. especially for a main character model.

alternatively you could lower the polycount and bake those details to textures instead

>>14700291

That's essentially what I was thinking.

I got worried about the tries because, believe it or not, the Unity docs suggest 4k tris for each model.

I can't imagine how outdated that is

>>14700268

15k is nothing, stop worrying about it. Modern PCs can crunch triangles by the millions, what makes a bigger difference is the material applied to those triangles (which, unless you're doing some really crazy shit in it, is negligible).

>>14700268

>I don't know much about modeling but we could easily cut off a third of the polygons without losing much detail at all.

If you're worried about it just use LOD's and do what >>14700291 says. But really you shouldn't be worried about it.

>Unity.

Can Unity do vid related? Have a plane underneath the surface with a hump that pokes out above the surface. The texture on that plane pans the UV's as the character moves. Put the rocks and dirt texture into a displacement map and have that pan along with the rest of the texture. The direction the panning goes can be affected by the normal of the surface because I know your character needs to move up and down slopes or along walls. Panning could also just be affected by the players velocity if it turns out translating world coordinates to UV coordinates doesn't work for you.

Though if you're worried about 15k polygons in the main character model then the idea of displacement maps and tessellation must be the stuff of nightmares.

>>14700322

It can and that was the original plan. We just kind of discovered that it looked really cheap and old and decided to go with something else.

I waited until now to work on particles because the unity particle system is just really confusing

>>14700332

Fair enough. The only thing I would change with the particle system as you have it now is to increase the number of particles when the player moves to avoid the gaps between particles.

>>14699016

Just flip the coordinates when you render.

>>14700590

also this runs much smoother than it looks, recorder software is pretty choppy today

>>14700590

looks rally good

Some suggestions:

1. Perhaps the rocks should slide into the surface or shrink instead of disappearing (velocity, size over lifetime)

2. If that level will maintain this colour (looks like desert rock), I'd make the particle effect the same colour. This shade of brown would nicely fit a fertile ground, though (and if there was grass, I'd add blades of grass jumping up as well)

>>14700636

>1. Perhaps the rocks should slide into the surface or shrink instead of disappearing (velocity, size over lifetime)

I'm changing the size for them appearing (they start flat and increse in size with Y. I'm not too sure about the ones disappearing doing that though, I tried at first but it looked kind of off.

Current plan is to eventually add an effect for when they disappear, because apparently that's a thing you can do. So eventually instead of disappearing they'll explode or break.

>2. If that level will maintain this colour (looks like desert rock), I'd make the particle effect the same colour. This shade of brown would nicely fit a fertile ground, though (and if there was grass, I'd add blades of grass jumping up as well)

The level is not going to be that color, I'm using orange cause it's better than just having everything in white.

However the areas where DB can burrow are going to have a dirt-colored texture so they're probably going to fit better there.

>>14700590

I want to play this one day.

>>14700654

>eventually instead of disappearing they'll explode or break

sounds great

>>14695194

>October 2017

>No updates after a couple of days in the sites listed by the dev

It's dead, someone shut it down.

>>14700657

Thank you anon that means a lot

Should I catch up with math first, or keep learning programming? I want to make my own gameplay physics and Unity's character controller kind of sucks.

>>14700677

Math isn't really that important to make a character controller. Trigonometry is where it's at since you basically just have to figure out angles, rotations and vectors by using other vectors.

Honestly depends how much programming you know, trying to build a controller is an excellent way to learn if you already know a bit of programming and trigonometry.

>>14700693

>Math isn't really that important

>Trigonometry is where it's at

>Trigonometry isn't math

>>14700699

It's geometry without numbers.

I guess it's a derivative of math? Never really considered it that way.

>>14700693

Well, that still requires math, but not as much. I wanted to catch up with college level math because from what little I've read, you need calculus for physics, and I wanted to replace Unity's physics later down the line because they're not deterministic. Got any good resources for making my own character controller?

>>14700712

>Got any good resources for making my own character controller?

I just kind of took it piece by piece figuring out how to make stuff work one by one.

Some of it I learned by googling, and after a while it just kind of stuck and I started being able to write code without needing to check the forums every two seconds.

Start from physics and movement and add things as you need them.

>>14700243

When you're moving sufficiently fast like right when you make the turn the player's center of mass is hard to see. The particles come in patchy which makes it hard to know where the player is exactly because the particle trail is too far behind.

>>14700721

I'll try moving the particle's center forward depending on the speed, that should probably fix it

>>14700725

If that doesn't cut it then a slightly larger clump that rolls or something might serve as a landmark.

>>14699567

Originally, these games started off with blank slate, so they were mimicking real life to get started. Which is why there were center guns - that's what you see when you aim. If you lower the gun to a hip, it goes out of your field of vision mostly, in real life, and on the monitor it would be well below the bottom edge. So these side-guns are faggy cartoonish representations of the hipfire position. I.e. it's just for fashion. At that point you might as well not display the gun altogether in favor of cleaner view. If you want to go for realistic though, your hipfire should have gun low enough that it's barely if at all visible, and you need the ironsight aim mechanics.

>>14700590

This looks pretty great.

>>14700590

Needs to kick up dust that hangs in world space and fades out.

>>14700590

neat

The entry/exit seems rough. I think the particles need to linger a while after db has jumped out, although that might be easier said than done depending on how you've set it up. Perhaps a separate particle effect for dirt flying upon exit.

Also, I think it might be more effective if the particles start under the ground and translate up the closer they are to the origin (db), that would solve the disappearing as they would sink after db has gone past and would really look like dirts being displaced rather than being spawned. I might just be speaking out of my ass though and I don't know how hard that would be to implement.

>>14700890

>The entry/exit seems rough.

Yeah, it's not quite there yet.

>I think the particles need to linger a while after db has jumped out

Earlier I tried using a loop to have them disappear and appear but it looked extremely weird so I kind of gave up. I have yet to try that with this particular particle effect so I might just make it so that when he burrows the particle effect renderer is set to active and the Y scale is quickly lerped to 1 and then invert it for when he jumps out.

More particle effects for jumping in and out are planned, but again I've got no fucking idea how to use the particle system.

3D particles are really easy but getting 2D particles to look good? No idea where to start.

>I think it might be more effective if the particles start under the ground and translate up the closer they are to the origin (db)

This is happening when they appear, but not when they disappear.

The particle system is weird in a way where I can't actually set the size of the particles myself.

Basically I can set the size they have at the start and at the end, then to actually set it I need to change how far into their lifetime they are (a total of how long they last from creation to destruction they are)

This probably makes absolutely no sense but that's because the sytem itself seems really wonky.

I do get what you mean though and I'll see how it looks

>>14701521

Neat

>>14700357

Maybe?

I do want the top left of my level to be (0,0). Going "down" needs to increase the Y coordinate. It's the labels/API that's backwards, but if I flip that, then it's no longer universal

>>14701521

looks sweet, what are you working on anon?

>>14701594

>>14701610

He's got a point, anon. Toriis are generic as fuck these days. They come of as lazy attempts to make something look deeper than it is. Unless you're making an entrance to an actual Shinto shrine you're misusing them and it will show. Make something european or create your own design.

>>14701610

>>14701629

but this is the only part of the city that will have japanese (dracari on my lore) architecture, i want this part to contrast with the other city districts

>>14701521

Anon that looks fucking amazing

I'm getting a Tearaway/LBP fake cardboard vibe, love that shit

thank you for not going for the reflective lowpoly every single fucking indie game uses

>>14701547

>>14701599

>>14701642

Thanks anons, but I bought the asset of the asset store (I can't do art). I'm working on a maths game for kids. All my art is off the asset store, I just do the programming.

>>14701671

>i can`t do art

So what?

That didn`t stop me

Using only store assets will greatly limit what you can do

>>14701637

Forgive me, I assumed too much from the current visual style. It'll work if it's integrated with other japanese things in the scene.

>>14701695

>So what?

>That didn`t stop me

Well, learning to 1MA is years worth of practice in each discipline to reach the level of whatever you can buy off the asset store.

>Using only store assets will greatly limit what you can do

For sure. I'm fortunate enough to be able to call on help from mates if there's smaller stuff I need like GUI elements, simple models or simple animations. I'd spend a lot more time on creating unique assets had this not been a game for really young kids. It might be cynical, but I don't consider the asset store look a problem for such a young audience.

>>14701695

>>14701728

I'd suggest using placeholders or store assets for now. Once you game is playable you can use it to get artists interested in helping you out.

As someone who started not long ago:

It's always fun when you come back to old code/projects and you see spaguetti-code begging to be optimized and you laugh at your past self for that mess.

It keeps you motivated by showing you how much you have progressed.

>>14699016

>>14700357

So I realized that since my Position object is "game specific" and not part of my engine, that I can flip the helper function around and it works perfectly. Feels a bit dirty, but it's consistent now.

>>14702005

Myself. There aren't many games worth playing anymore.

>>14702052

Learning the academic way is fucking awful, reading docs and mashing shit together is the way to go

>>14702078

Yeah I wouldn't really recommend it unless you're a brainiac since it does involve a lot of headbashing otherwise, but hopefully having done it I can find interesting uses for it. I'm definitely not engine dev material though, I'd rather come up with fun wacky things to mess with than fixate over how my little dude moves around on the screen. I really like demoscene, amiga and ps1.

>>14702005

Anyone who wants a comfy game with deep mechanics, tons of lore and unique ideas

>>14702287

What is your game?

I've only seen blender screenshots so far

>>14702339

>first expansion

Kek

>>14702341

I had to cut something for later, if i had combat, alchemy and exploration (larger maps) on the first release it will probably take 5 instead of 3 years

Im already having feature creep problems within the planned features, so i need to put a hard limit on what i want to have on the 1.0 version and add more later

>>14702353

Most early access games like that don't ever get finished, we have no reason to believe yours will, especially if it takes another 2 years minimum

>>14702353

how's your bro doing?

Any people who compose music here?

Can someone tell me what this clashy instrument is at the 1:15 mark?

Does anybody here use Unity in linux? How is it?

Is it even officially supported yet?

>>14698299

>>14698345

You missed the Hotwheels lives.

>>14702586

Im shinobudev on Linux. Works fine so far. Using Unity 2018. I have to switch to my windows for photoshop sometimes.

>>14702133

To me, it's always good to plan shit just so you don't have to crawl through tons of commented code to figure out how something works. But other than that, yeah, the academic way is fucking shit.

>>14702713

but he was always alive

>>14702827

I will say that I probably know how to do more and better things that it would be hard to force myself to understand otherwise, or know about. Probably would become the if-else-hell meme rather than using data structures, OOP and language features like a good boi.

>>14702005

Myself, anon. For the most part, anyway, The genre of cursed runes has never been properly conquered and I wanna be the one to slay the dragon.

Anyone interested in doing a Kanye + Trump + USA gamejam?

https://8ch.net/v/res/14703201.html

>>14703213

I think the majority of anons have become disenchanted with Trump once it became clear he's just more of the same shit except also funny.

>>14702400

>early access

who said i will taint my game with that cancer? I will release when it has the first row of features

>we have no reason to believe yours will

Ok, so?

>>14702454

he is still contractually bound to do a few things for his old workplace for a few more weeks, but he is working on the game, he is making the character controller

>>14703213

was this a gamergate idea?

>>14703483

Jesus Christ that's fucking cancerous, and all the underage b& posting on it too.

But what does this have to do with game making?

>>14703626

Check out OGRE.

>>14703633

Problem with OGRE is that the documentation is ass, and they force you to use a "template" which, again- obscures a lot of code from you, making it hard to understand what's happening.

>>14703604

Just reminding everyone why we bullet cuckchan and reddit

>>14703626

Godot being open source means you can ignore the baby stuff and use the deeper framework.

>>14703671

No documentation for GDnative/C++ yet.

>>14703483

>>14703604

>WTF

>LMAOOO

>smh tbh fam

>"Can't post, and /b/ is blank"

The amount of normalfaggotry on halfchan is immense; in fact it's much worse than I thought. Did get a good chuckle out of that

/ck/ comment though.

>>14702615

I am charmed by your game's aesthetic and style (it's based on Ultima?)

Please keep working on it and doing your best anon

>>14703823

>>14703826

>Sleeping at all

>Wasting the first 30m "waking up"

>>14703846

I used 30 minutes to make myself some bacon and eggs and a nice coffee

Gotta get dem brain nutrients

>>14703858

>not feeding your mind and body with the energy of neetdom

It's like you're not even trying to make it.

>>14703879

>>made after dev lost all progress on an alien metallurgy sim

Hearing this always makes me rage incoherently.

HOW FUCKING HARD IS IT TO LEARN GIT YOU FAGS?

>>14704558

Very hard actually

Could be because I'm on windows but I cannot fucking understand how to work it

>>14704599

I know the real programmers will shit on be but there are GUI git clients. I use Sourcetree, for example

>>14704601

And I used them, couldn't update shit

I don't know what's wrong with me

>>14704483

A game with that style of combat should NOT have enemies that flip around 180 degrees instantly. It looks like it's set up for dodge>backstab but as soon as you go for the attack it's looking at you again.

>>14704603

try (((github))) desktop, is run by sjw but is simple enough to get a hang of git

>>14704605

>It looks like it's set up for dodge>backstab

Dark Souls was a mistake.

>>14704612

Yep, also tried it.

I think it has to do with some kind of hidden autism for badly designed interfaces. I see one and I just cannot handle it at all and just give up

>>14704599

>Very hard actually

Honestly, if you work alone, all you need to learn is:

>git add -A

>git commit -m "comment a shit"

>git push

Done, congratulations, you now know git. Literally no excuse is valid.

>>14704601

>GUI git clients

IT'S TIME TO STOP

>>14704558

>>14704599

>>14704633

>git commit -m "comment a shit"

Don't do that, use your editor like a real pro.

You should probably look into branching/tagging (tagging is very useful) along with merging, actually learning the point of version control.

>>14704658

>Don't do that, use your editor like a real pro.

What does that even mean? If your commit is as small as it should be (since we're talking about working like a pro), anything above 80 characters is a waste and I will punch you in the throat for forcing me to read all your stupid shit that nobody cares about.

>You should probably look into branching/tagging (tagging is very useful) along with merging, actually learning the point of version control.

This is why fags like fa3b8b never learns git, because people like you always throw them down the deep end. Learn bash, get the basics (i.e adding, committing, pushing, fetching), then you move on to branching, rebasing, stashing, you name it.

>>14704676

That's why you have your generic 'Add a new gay rainbow function to my gay ass fucking player' and then you make a new line, you fuggen retard.

Branching is also the fuggen basics.

>>14704684

>then you make a new line

No you fucking don't. If you need more than one line to explain what the hell your commit is doing then it's too big, split that shit or fuck off. Branching is not basics for a retard like fa3b8b who is unable to even get started at all.

>>14704694

>not having an autistic description about what you committed in case you forget

Branching is the basics, it's right after all that gay shit you just said.

>>14704599

The thing that helped me understand git was realising that it's just tags on dags. A dag being a directed acyclic graph, where each node is a commit, edges go from parent to child commit, and each commit can have one or two parents, depending if they are a regular commit or a merge. I think that commits can technically have three parents, but let's ignore that for now.

The tags, or 'refs' in git lingo, are pointers to a commit/node. This includes git tags as well as branches, and special refs, like HEAD. The only difference between a tag and a branch if you check out a branch, that branch becomes the 'current' branch, and if you make a commit, then the current branch ref will be updated to point to the new commit instead of the old one.

Now, git is a distributed VCS, so that means you have a dag of commits on your computer, but there are also other related dags on other computers, called 'remotes'. Typically, you only care about one remote, usually called 'origin'. What a fetch does, is download the bits of the dag that origin has that you don't, and update your dag so that your dag matches origin's, but it doesn't update any refs.

What merge does is create a new commit that is a descendant of two other commits that you select, typically your current branch and another branch. If one commit is already a descendant of the other, it can do a fast-forward merge, and just treat the descendant commit as the new commit. Since you are creating a new commit, this will also update your current branch to point to the new commit.

A pull is just a fetch, followed by a merge of your current branch, and origin's version of that same branch.

A push is the reverse of a pull. It uploads bits of your dag to origin, and creates a merge on origin between its version of the branch and yours. Usually, pushes are only allowed to do fast-forward merges.

A rebase is where you change the parent of the current commit, or one of its ancestors. You should only do this to commits that only you have a copy of. This is because you are editing the dag itself, not just the refs. If you rebase a commit that you have pushed, your dag will become incompatible with origin's and there will be wailing and gnashing of teeth.

HEAD is the pregnant commit. It is the commit that will become the parent of the next commit you make. 'Detached HEAD' is when the HEAD commit does not also have a branch pointing to it. This means that if you create a new commit, it will be very hard to find again, as it will have no refs pointing to it, and no descendants with refs.

Hope that helps.

>>14704572

just use gachimuchi sounds for everything

>>14704582

And yes, it was Frauki.

>>14704756

https://preece.itch.io/frauki

he's still at it

or was until December

https://preecedev.tumblr.com/

>>14704726

third webm, same cyberpunk dude I spoke of earlier

https://stomygame.itch.io/ctesiphon-demo

http://stomygame.tumblr.com/

I really need to make wiki pages for them, even if they don't post here anymore

>>14704813

Mario World romhacks, and after that, screwing around with the Hello Mario engine in Game Maker 5 or 6.

I've switched engines a lot throughout the years, and I never believed in the whole "start out with Pong" thing that most anons promote. Usually, I learned by necessity. I need a feature? I plan it out and research how to implement it in this engine/language specifically.

>>14704813

RPG systems on spreadsheets

>>14704813

Fucking about in Multimedia Fusion 2. If you're just starting out, start small. Pong, tetris, top-down shooter, platformer, some kind of progression of complexity that. The wiki may have more resources (someone should make a newbies page).

>>14704813

very first game project I ever made to completion was some really shoddy Flash arcade shooter with a single screen and enemies appearing to shoot.

>>14704813

I didn't really have a first project. Lots of dicking around in editors and stuff. I think my first real exposure was some program called RPG Toolkit v2, back from pre-2004 (before I was even in high school). I made a few dummy projects in GameMaker 5/6 for a while too, but never finished.

>>14706059

Evens the next demoday?

>>14706036

>>14706059

>this is happening 3 days before my final presentation for uni

>need to finish a level, add contextual button prompts so people know how to play and fix all my bugs before then

I don't think I'm going to make it unfortunately

fuck

>>14706176

Where does that come from?

>>14706306

Literally the first post in the thread.

>>14704813

Modding Warcraft 3, afterwards an RPG Maker XP Final Fantasy rip off, then a 2d topdown shooter flashgame, then a platformer in Unity.

>>14706059

Thanks, updated on the wiki.

>>14706368

what's this from?

>>14706368

>>14706440

>>14706447

Looks stylish af.

How do I convert the movement of a platform into a velocity to then apply to my rigidbody?

I'm feeling like an absolute retard right now

>>14706817

Do you not just take the difference in positions to get a movement delta?

>>14706850

That's what I'm doing but it doesn't seem to work great

>>14706852

I assume it's because you're adding the platforms velocity to the character, and then when the character updates, they assume (P+C) to be its new base velocity, then P.vel gets added to it again next frame and it compounds the error.

Dunno, haven't worked with physics libraries or Unity before, could be anything.

>>14706970

It looks like Popeye shrinks when he leans back to wind up a punch.

>>14707126

yeah, I dunno why.

>>14706970

mechanics look okay, visuals are dev art as fuck. Head proportions shrink during the punch windup, and that's why he looks like he shrinks. Fix his head size.

Oh right, found this nifty little site for 16x16 pixelshit fonts

http://www.pentacom.jp/pentacom/bitfontmaker2/gallery/

>teachyourselfcs.com

<please put in OP

>>14707618

We have a wiki for posting links to useful resources. Here's the page for programming. You are free to edit it and add that link.

http://8agdg.wikidot.com/programming

>>14707433

>that webm

It warms my heart that someone made that effort.

>>14707573

Nice! Most of those are public domain too.

>>14706029

im a tad drunk and phone posting but I’ll do my best

>1)

>low-resolution information as to where the intersection is happening

For MC, that’s exactly right.

For DC, it’s just the starting step; you refine it into a more exact intersection point per cube (single vertex per cube, using intersections per edge to get a single intersection point).

Cosidering MC, that’s one of the down sides (pre-computer tri-table for edge case of cube), and is what DC was partly developed to resolve.

Think of it like this, MC has a whole slew of triangles (thus vertices per intersection) you can get per cube depending upon the intersections, but in DC you get a single refined vertex (intersection point) per cube from all those edges/corners; so it’s much higher res.

Remember, it’s a “dual” approach, so we use adjacent cubes to form the triangles (quad per contour) from the refined vertex per cube.

Also, think of each of the corners defining an edge as: is this corner inside or outside the surface. If one is out and one is in = valid edge (isosurface passes through this edge).

>2)

>compute and store normal

Correct, store it per voxel.

Compute them via tri-linear interpolation (look at the ogre dual marching cubes implementation, it’s the matching cubes page).

Generally I use either a density field (also called level set), or a height map; allowing for easy tri-linear interpolation.

In my approach I store the vertex, and normal per entry in my separate “voxel data” buffer (utilized by my triangle induce buffer).

For MC, you’d calculate each normal per the intersection point vertex (using t-interp), and you do the same for DC.

However, in DC after computing per intersection normal, this Is where it gets a bit funky.

The normal per intersection point in DC is used in a step of DC to find the final vertex (or, points by which we can find the final refined intersection).

Basically though, it’s ((sum normals)/intersectionCount = final normal); this is also called a hermite normal (or a gradient, in form of: (normX, normY, normZ, density)).

Reason why it gets funky is due to using QEF, or particle decomposition to get the final intersection point/vertex; which treats normals as a “gradient”, as to find the direction (normal XYZ) and distance (density) to the isosurface.

>>14707696

/agdg/ is like a beautiful oasis of optimism in an ocean of despair.

>>14707573

Fucking excellent site

If I have a platform that is moving via animations, but my character is moving via rigidbody, is there any way I can get the movement of the platform to be updated during FixedUpdate rather than Update?

While my character moves with it while standing on it, it's really stuttery and I know for a fact that's a rigidbody update problem.

Any suggestions?

>>14704721

what made me understand git was viewing it from a save scummer viewpoint

imagine the ultimate save manager.

it lets you decide which changes to save (staging), which to ignore (add to gitignore) and even to revert parts or your whole save to a previous point (not sure about the cli command, i use a gui for git).

it even has a savestate method built in (commit). the ability to create multiple quick save slots (create branch) and to change from one savestate to the another (checkout) synchronize your data from one saveslot to another (pull) or even combine it in one save (merge)

all while keeping track of all the previous savestates and individual saves.

there are also some online functions like uploading your data (push), downloading other peoples savestates either for yourself (fork) or to cooperate (clone) (you use clone to cooperate with other people even if the other person is just the online version of yourself)), the last useful one is fetch which is as useful as "refresh this page" in your browser