/agdg/ + /vm/ ~ Amateur Gamedev General

Why are their mouths so fucking large? They appear to be twins, so we can assume that it's genetic. I bet they'd feel good.

Are you still here user?

I forgot to ask, but I'd also need some input on the bar to the right in my pic. Does it look out of place?

Hey kids! Sport shoes ready?

Got the idea from someone posting on a recent Haydee thread I made, kind of ironic, it gave me an idea for handwaving the concept if I can go further, I am mostly an artist but I want to make this character have different sets of pieces and all, basically the "human-like" head is my only concern since I am unable to animate that much, I would count on blendshapes provided by Fuse.

If I end up working in more gameplay I want it to be an action platformer, but the premise is still about the character being a modified gynoid.

I wonder if making rpg maker game with already existing assets is cancerous. I don't want to draw or program shit.

Absolutely. Most people aren't into /clang/, and would want the most human elements possible, to the point that they'd be willing to pay for them, even just for a sex bot. Much of life's identity is in its movement. Even if you don't get lip movement, facial expressions and body language, at the very least, are some of the core components of what would make a robot believably human. If they don't have these things, then they don't look real. If a robot doesn't look sufficiently real, then it has to look stylized, which humans will accept because stylization emphasizes traits that we understand as belonging to things which are alive. But even such a stylized robot has to emote in a way that we can relate to, which makes us accept it.

WALL-E and EVE are great examples of this. This is one of the main things the old sandbag animation challenge is for.

If you're gonna do that then for the love of god read the tutorials.

Facial expressions are added to Fuse models I make with the software with no problem.

2 questions if an user would be so kind.
Should I start with Unity since I assume it's simpler or would it be more beneficial to jump right to Unreal engine 4, if I did go with Unreal do I need a beefy graphics card or will the one I have be enough I have pic related.

As far as I have heard UE4 is a little harder to get used to and that's it, I guess Unity is just convenient since most shitty games from Greenlight are made in Unity.

So I guess the Graphics card won't be such a big deal unless I were to try and Jam tons of effects and crap into the game.

sounds like loads of work with little payoff.

I am pretty sure the only problem with Unity is finding a way to optimize your game at some point since nobody does (even Ubisoft)

That card is plenty fine. My "rig" is just a mini-atx with old-gen i5 and GTX750. The SSD helps a lot too.

My UE4 game is pic related, I haven't posted progress updates here in a while. At lowest settings it even runs on the cheapest Celeron laptop you can buy. It's entirely in Blueprints. All the models (except for the higher quality marketplace / startercontent / content example textures) are simply box-modelled, most of them in Blender, with enemies that need PhysX in Maya modelled by the other user here. My original goal was to get a extremely budgeted PSP / PS2 quality game in UE4 but with physics, and I'm really enjoying game development, definitely better investment of my time/money compared to getting a Hyundai Accent / Nissan Versa / shittiest/cheapset car you can buy etc.

I previously was a developer of shitty western visual novels, so I'm used to what the PC market wants (extensive campaign arcs broken into chapters, etc.). Reading about Life of Black Tiger on the PSN was one of the best news I've read, it means that the standard is now so low I can potentially get my game up there and not just be a yes pc dev but a yes console dev as well.

You better be a god tier musician then you lazy sack of shit

your topology is shit, what happened? I could have sworn that when you first talked about your Haydee clone you said something about knowing how to 3deezy.

Yeah, that's cancerous, and you sound kind of cancerous too, let me guess you have a neato story that must be told where the twist is that you are the zombies, cool man.

You need to fix that dress. The way it looks like legs instead of her real legs it hideous. My god who told you to make it that way?

Good to see you here, I know my models are not quite optimal but I keep learning.

Sometimes I don't get how topology is still shit.

Yes I am.

That's pretty good, only recommendation I have for you would be maybe to disable her saying "Lets do this" when retrying.
When you first start a level it would be fine but if you doing a really hard part, hearing the same audio clip over and over again can get annoying.

It's going to be really awkward if indie devs reintroduce things that haven't been seen in games for years and plebs think they're brand new game ideas.

I imagine those are people who say stuff on camera for money.
They are unsettling to look at not just because of their alligator smiles, but because they squint while they do it.

I don't normally post in agdv threads, but I do like to play armchair creative director or what have you where I think about a game I feel missed the mark and then write down what I believe would have been a better take on it.

Currently thinking about what my take on Mass Effect would be and really it wouldn't be a cover shooter, the space age setting doesn't lend itself well for that to be the core gameplay.
It would be more about offering a nice virtual choose your own adventure novel.
A multitude of endings and several morality choices that aren't labeled good or evil.
A story you could beat on your own but can hire 16 different characters that have their own character specific stories you can do with paths or events they can open up for you.
Incentivizing multiple playthoughs.
Romancing options but some would require more than persistent correct dialog navigation.
I wouldn't have a character creator as I prefer characters with more definition, you would play a male character with a defined backstory but I'd have hair, eye and skin color selection as well as hair style and voice pitch options.

Ideally the goal would be to not have a tone as sterile and robotic as Mass Effect.

Just aim to become better, good topology helps naturally form the features in the body and the face and it makes articulating the body much easier. I shudder to think what the rest of your body looks like just considering your neck and your whole mouth area.

They are sisters. They look like copies of each other. One probably dyed her hair blonde just so they could look different.

also hail satan

Actually the only thing I sculpted personally is the helmet, the body was made with Mixamo Fuse and some sliders it has, can't complain much except for the way the boobies are made in that software, those need to be refined for sure.

What's that whiteboard thing from?

What a mess.

I know using physics is considered the cheap / lazy way to go about things (especially here where the standard is you roll your own engine), but I really like it and have been able to experiment with some weird aesthetics, new vid related. Not shown is constructing all the artillery pieces tower-defense style prior / during the countdown. Each campaign chapter has slightly different gameplay style.

I'm also going to steal Kowloon user's gameplay idea, not sure if I've seen him for several threads, but basically it's close-quarters combat versus a mech that can punch through walls. I'm going to have a horror level as well featuring the beast you saw in the last thread. So at this point I'm pretty much not following my original design document and just chucking in all my favorite things into this game, you got yuri meidos, gunbattles, Germans, soon I'll be putting in the Luftwaffe as well.

I'm just following the character model sheet, and anime dresses are kinda pointy / floaty. I guess the lack of contrasting colors doesn't work in its favor.

I was gonna shoot you an email saying "Wazzup?" but I guess no need then.

Thanks, noted, the voice acting isn't yet complete, there are several expressions available, so I can pick them at random, and also only activate once per certain period of time and not all the time. Same with weapon expressions.

fiverr.com/theadtwins/create-a-promotional-video-featuring-twins

We call you idea guys.

You are certainly not helpful at all, but I suspect you just try to be a dick to anyone.

Here is the original character design.

While I do aim to improve where I can, my main goal is just to keep moving forward, so there's a lot of "good enough" decisions which I make. I do watch a lot of Jim Sterling's videos and Gaming Garbage and study a lot of asset flips and recently, Mass Effect Andromeda, they've been very useful in identifying common mistakes by bottom of the barrel devs, which I'm not ashamed to be part of (except the blatant scammers), just a hobbyist after all.

The only real advantages Unity has over UE4 is the huge cross platform support. There are other pros and cons, but they so dependent on the context of the project and the people working on it that really don't matter to the lone indie dev.

Its looking good man. Keep up the good work

time to do video games

To anyone's knowledge, as anyone ever done submarine mechs with jet turbine tits?

I don't think jet turbines work under water?

It's land, air and sea.

If it serves as a motivation or at least a reference, there was this rpg maker forum that I participated and like most of forums I know people really liked to show off, they would made beautiful levels using custom assets, abuse of scripts to make you wonder if it was even made in rpg maker, use and abuse custom midis from vgm to appeal to nostalgia, etc.

But there was this one guy, who's main project was creating a dating Sim in rpg maker 2003 entirely event based (there is no scripts for rm2003), in other words he knew his shit . One day he woke up, downloaded rpg maker vx, and thought "well, I see all these heavily modified stuff on the boards, but I never saw a game trying to be good using vanilla only resources, might as well try it." and then he proceeded to challenge himself.
Result:
A fully fledged 50 hour worth free jrpg using only rpg maker native assets and code. The guy won massive recognition from all forums at the time and it's probably an example to anyone fussing with rm to this day.

So, yes, it's possible to do it, and can be good, but don't expect people here to like it tho, in the moment they recognize it's rpg maker and that you are using default assets, you will get bullied. /agdg/eeks aren't exactly bad people, they give good advice and lots of help, but they are massively visual, they will give more attention and praise for a well done pixel art than to an interesting, but ugly alpha build, I mean, hell, one of the reasons I quitted posting progress was that people literally start criticizing my placeholder art instead of giving feedback to my game.

post an update so i can judge your game and placeholder art

It's quite old and I didn't work on it for a while, I'm focusing on making a "party" game before greenlight dies, and it's quite boring to show progress yet because I didn't code bots yet
also, I'm away from pc in the moment, but if you insist I'll post it later

Nah, I just haven't got anything nice to say to anyone right now.

wiki.openmpt.org/Tutorial:_Getting_Started

Trackers aren't too hard to learn so this should be fine. If you have trouble learning trackers with openmpt try playing with famitracker to get used to the basics.

I worked on a "game" for ludum dare.

dropbox.com/sh/fuoxtictnp4oihy/AACNBSeB2rNGmCOg-xL-QUiva?dl=0

It's barely even a prototype though. I misjudged the scope of the project and what I could complete in 72 hours. The "bugfix" file is me fixing some of the more glaring bugs after the fact, but it's still not exactly polished. Really I just want to move on back to my main project.

Eventually I'll see if I can get the Source files uploaded.

What a fucking piece of shit it is.

why must life be so hard

They're probably jews.

Holy shit so I was supposed to use it that way:
TNT1 A 0 A_Jumpif(GetPlayerInput(1)==BT_FORWARD,"Accelerate")
That is very noice, it is a bit pity that this functionality is for GZDoom 2.4.0 only hmm but at least it works now I guess.

I've been wanting to make my own 80's inspired fantasy cards for a future /dgdg/ project. Probably once I finish my exams I'll start doing this.

How do I decide how big my numbers should be for an RPG? For example, JRPGs usually give level 1 characters a hundred, sometimes several hundred HP to start with, and regularly end up with tens of thousands at the end.

Meanwhile, something like D&D gives you 10 HP to start if you're lucky, and only the godliest of beasts break 300 hp by level 20. Fire Emblem goes even lower than that: 50hp is an absolute beast. And, while not REALLY an RPG (but still applicable to my point), getting anything with 10 power or toughness in Magic: the Gathering is quite the task.

Why exactly did JRPGs land on "super huge numbers" and stuff like Diablo land on "reasonably high numbers?" I can't imagine it's just some kind of preference.

I think it has to do with the way dmg and other stats are calculated in videogames compared to rolling a few dice.

It's a sliding scale of increasing granularity vs. increasing meaninglessness. Pick where you feel comfortable.

Bigger numbers give more leeway, especially in the thousands. With JRPGs the story and exploration take precedent, so with padding numbers you barely notice until there's a difficulty spike (e.g. optional hard boss).

With smaller numbers you are literally micromanaging each hit point, so its no surprise they're paired with the more dice-rolley type of games.

The bigger the numbers, the more powerful your attacks will seem, at least on the surface.
Simple example, the Mario and Luigi games, it always seems like you're hitting for more than you are in Paper Mario, despite both games being of similar difficulty.

Determine who your target audience is and then adjust your number scope accordingly.

MMOs and grindy as shit JRPGs use bigger numbers as a progression mechanic. There's no real difference between 100 or 1000 or 1000000 if all the other numbers are in the same ratio. Bigger numbers are just a way to cater to people who want to see huge numbers flying at them over the screen and for people who want their skinnerbox fix. If you've just spent six hours killing rats to level up then seeing your strength go from 1000 to 2000 feels more rewarding than seeing it go from 1 to 2.

If you're not interested in catering to these types of players and are instead looking for SRPG players who might want to be running mechanics calculations in their head, then keeping numbers for your system small can be useful to them. But at the end of the day it all comes down to looking at your genre, looking at your audience and then determining what systems you need mechanically to suit them.

Funny since Diablo 3 has builds that hit for hundreds of millions apparently. The power creep in that game is crazy.

Thanks! This will give me a good place to start, at least.

Start out with the character having about 100-150 HP, focus the game's hitpoints and dmg around that…

End game would likely see the character having about 8-10k HP and doing around 9-10k DPS depending on how big the numbers you want to be seeing on the screen…

I never understood why JRPGs started players with like 1000HP, those fat numbers don't really mean shit unless they actually hit hard. A mob with millions of hp won't care much if someone's doing like 189k DPS on them… Make the numbers look reasonable yet endgame should be hitting like 3 or 4 digits to make it look like they're endgame or something…

TL;DR, JRPGs have a fetish for fat numbers that don't mean anything… Heck, you could do a DND style hitpoint system like Baulder's Gate if you want to and it'll feel more like a RPG.

Speaking of numbers, user, nice fucking digits.

Could be funny I guess.

The RPG Maker part can be fine, but to not be cancerous it's still going to take programming or at least customizing scripts made by others. Using pre-existing assets is always cancerous.

thanks user, I'll check this out as soon as I have some free time.

its retards like you that make it look bad
kill yourself, do not become a game dev

I'm streaming again
youtube.com/channel/UCQ2IcqxHhxir7wuaBzPHc5A

holy fuck user again?

...

fuck I didn't even notice. Also nice dubs you got there

I need more inspiration pics. I need to turn my rage into a game.

Damn, lost my folder somewhere.
Do have this from Resistjam though.

Yeah, they'll how to make a good game, unlike Undertale.

honestly been part of my development process

Undertale did some things right though. For one it was Kickstartered and didnt flop like every other project. I got what I paid for after playing the demo years ago

The soundtrack was really good and complimented the gameplay, and the ending ramped up appropriately and had good feelings behind it. I did cry I'll admit. It also had a good twist on otherwise boring RPG combat, I hope to see later games by other people extend such a system without plagiarizing it.

Really the only bad thing is how linear it is, lack of replay value, and the AWFUL community

The problem with leaving anything implicit these days is that certain individuals will consider it explicitly undefined and significant, as long as it can somehow have politics shoehorned into it. Once the projectors are fired up, that the framing is very hard to forget or ignore completely.
Somewhere at the back of your skull, you'll still recognize some choices or the absence of certain details as tumblr-bait.

So just throw in a singular detail that makes it problematic unpalatable to those types and you'll be golden.

Sorry but you have committed an unforgiveable sin and must now be burned on a cross as per the Holla Forums way.

So here's todays attempt at a memepoly character. It's got tons of nonplanar quads and the shoulder area is all kinds of fucked up. Pic 3 is my topolgy reference that I just copied and tried to model over pic 4 as anatomy reference. Not too easy seeing as the first looks like a HL1 scientist while the other is built like a space marine. I also have no idea if it will deform well, but it actually seems to shade well enough with phong shading.

I don't know if you are new or something, but in the first 3 weeks or close, Undertale was somewhat praised around here, and there was lots of discussion, problem is furfags kept making threads after the discussion was over for avatarfaggotry, waifufaggotry and disgusting furry porn, mods permitted it, dunno, because they thought there was actual discussion I guess, the op gave the false impression of being a serious thread (links and faq), so you take this cancer and add the fact that the game was severely overrated by tumblr, 4cucks and normalfags, and there you have it, 8/v/ disgust explained.

So was it only me and slimebro that took part in the Ludum Dare? Or are you guys stalling/shy?

I took part in it, I posted a link higher up in the thread.

yeah, I notoced, sorry.

let's be feedback friends

Tomorrow I'll try to analyze how the teleporting in your game works. It's a puzzle, right?
or is it random

i tried a game jam once and I felt like it was a waste of time.

Yeah it's a puzzle, it's not random.

Do you have a link to your game so I can play it?

Here's my game:
ldjam.com/events/ludum-dare/38/$26141

and here's slimebro's:
ldjam.com/events/ludum-dare/38/slimeworld

My (and my pal's) game is supposed to be an RTS but unfortunately the actual gameplay, AI and winstates were the last things we did, just before the deadline, which was 3AM here. So while I'm happy with how the game looks, the actual "game" part is clearly lacking, which is unfortunate because I felt like the concept could have actually gotten high marks. Doesn't seem so from the comments I get. Well, maybe we'll re-make this project once we're done with Beelzebox and/or my beloved dream Deus Ex clone…

Slimebro's game is a fun timewaster clicker thing. I think that was more of an afterthought, added to justify the amazing bouncy physics

I definitely will have to dig deeper into your game. I loved Rudy Rucker's musings on the gnarly fourth dimention.

sorry to hear that. I find the Ludum Dare community pretty cool although some political uindertones make me wanna puke. It's no /agdg/ but the feedback is pretty good. And the things you learn and come up with while making your game do seem worthwhile. Beelzebox was a product of a shitty jam after all.

Pretty much. I had a vague idea of a different more citybuilder-like game for it, but both time constraints and the idea being too vague mad that fail. I'll probably still make a rebalanced version next week, with upgrades, world increasing actually costing slimes, maybe more slimes and having them unlock as you go. Plus saving.
First I'm going to be helping out at a festival until Monday morning though, so I won't be around until then. No idea how long LD's rating period will last, hope I can vote on your projects before that's over.

They usually are, but sometimes they're not. And the main part about them is that they're only a waste of 48-72 hours of your time, which in the grand scheme of things isn't that much. People tend to get an urge to create new projects, jams can be an outlet for that. And occasionally you stumble upon a good idea that you can work out further.

Thanks. I'll check those out.

I had to work and there's no way I could make a game in that short of time. I'm stuck in enginedev mode and would fuss over pointless things and get stuck as soon as it started.

I'm glad

To add on to other anons

RPGs started because of D&D (not to be confused with older roleplaying tabletop games). You had FF1 basically plagiarize a ton of shit from it that were never used again, such as Mind Flayers and Beholders. There was an emphasis on the 4 core classes, Fighter, Rogue, Wizard, Cleric, that a lot of RPGs focused on (which would later become the holy trinity of tank/dps/healing used in MMOs)

Numbers in D&D are low because they used dice and were done by hand. JRPGs do in fact have inflated numbers and so do ARPGs. However, the numbers are strictly cosmetic. What happens if you took a JRPG where you have 4000 life, and allowed two decimal places for all your HP and damage displays? "40" life looks a lot lower. Thus, it creates a sense of scale. Most games don't use decimals though, so keep to whole numbers.

What really matters is how many hits it takes to kill the player or monster at any given point based on stats and progression. You kill trash encounters in 1-2 hits? You're doing fine. If basic enemies give you a lot of trouble and you kill them in 5 or more hits? You're underleveled and need to grind.

The other key thing is keeping in mind how much damage is being dealt, not as flat numbers but as a percentage of the total health pool.

...

Big chance they're jews with that fucked up facial structure. But to answer your question, dick sucking lips.

god I fucking hate the idea of grinding, RPGs need to fuck off

You should also consider how difficulty scaling would be affected by this.
For example, if you had 100hp at level 1, and gained 10hp every level, an enemy that hits you for 20hp will still be a threat at level 2. However, if you gained 100hp and had 200hp at level 2, the same enemy won't be nearly as threatening. How quickly you scale up can be used to change how long an enemy from an earlier area remains a threat.

I've been watching a lot of old game postmortems lately and I'm consistently surprised at how often D&D shows up in those talks. Doom, Deus Ex, Myst, all inspired at least in some part by D&D.

Don't forget players won't always use basic attacks, but the "1-2 hit" rule works.

An enemy weak to fire should take 1-2 from fire, but not other attacks to encourage players to use fire magic.
A tanky enemy should take 1-2 powerful special attacks.

IMO, if the player has to grind to carry on the game- either they are playing badly or the game is not balanced.
Design it assuming the player fights every (random?) encounter in new areas, and completes 75% of side-quests that can be completed at that time (but always assume they have the best gear).

Personally I hate it when you can grind and use certain tactics on loop to beat the final boss (like Bravely Default), but some people love being able to break the game like that. Again, target audience.

en.wikipedia.org/wiki/Pathfinding#Sample_algorithm

So I notice they used a queue and actually started from the endpoint. My implementation used a min heap and started from the origin. Does that matter? I notice that as a consequence, they didn't have to reverse their output at the end to assemble a correct path, but otherwise it looks identical.

Another thing I noticed is that they frequently add/remove points to the visited list.

Mine just doesn't add them in the first place, but it feels a bit shakier

For one, starting from the endpoint or starting from the origin is merely semantics. Unless your game has some weird one-way-only paths, the quickest route from A to B is also the quickest route from B to A. So that doesn't mean jack shit, don't worry about it.

About the queue vs. heap data structure…I mean, in the end, a queue will be probably be easier to work with. The heap probably has the benefit of lower memory usage and higher speed, but unless we're talking about some fucking huge goddamned maps and long, long, long movement, the difference in memory and speed will be so negligible that I'd recommend you go to queue or a list. Unless you personally just really prefer to work with heaps, that is.

Last, I don't see any appreciable difference between (1) checking each neighbor to see if its valid before adding to the "visited" list and (2) adding all neighbors to the "visited" list and then looping through to remove invalid items. Both end up with the same number of checks. I think the algorithm was written that way purely to make it easier to understand, not because its the better of the two options. Remember, pseudocode is rarely optimized. It basically goes against the entire principle of pseudocode for it to be optimized; it's meant to be a high-level, easily understood, language non-specific example.

I'm shy.

That's reasonable

Sort of… once you find the optimal path, you basically traverse the list of visited nodes and go to the parent nodes until you find the root and then that's your path. But if you go from A->B, the traceback will end up being B..A, so it makes sense to start from B, so that the trace ends up being A..B (which is more useful to the entity travelling it). I think that it would matter if you have many paths, or do it frequently.

It'll be useful to keep that mind if I make a puzzle game or something that has unusual pathing graphs

On a proper graph that isn't grid-based, it shouldn't even matter, unless the number of nodes is in the tens of thousands. Even then, better approaches like subdividing it into regions would make more of a difference than the data structure used. A heap should be faster, because it pulls the most promising values first… though it does use the weight/priority to order it, the cost of that is negligible.

There is very minor overhead for adding and then removing a value… it's cheaper to do an if check than it is to call a function or two. It might also end up creating more garbage overall. (Specifically, the memory allocation for the function's heap might grow; the default queue/stack/list collections in C# are implemented with an array that doubles in size as needed, and pushing needless values might make them stick around longer, since the GC is unpredictable with its cleanup; though, that is more of a micro optimization I think)

Today was a productive day for me. I managed to implement a basic character creation menu for my game.

You're not going anywhere son.

Imo maximo is subpar. What it does well is the face customization, but that's it.
The body proportions (especially chest/boobs) are mostly disproportionate messes that are impossible to fix without a 3rd party program.
I'd recommend that you use the alternative software makehuman, which is leagues better, and is less rigid compared to maximo; then stick the mesh into zBrush for final details.
The only downside to makehumman is you have to do custom re-toplolgy if you want reasonable poly counts; while maintaining high detail for say the normal map n' such.

My problem with MakeHuman is the face customization, I think it has none.

time to whip out the adderall and read shit
also if you liked this blog like comment subscribe and pls don8 to my patreon kthx heres all the donators goodbye

AGGY DAGGY
this makes me angry, but I am not surprised. also that white dude is plugged into the lifestream or some shit.

I downloaded polycode but haven't bothered to tamper with it, how is it?

What? If it can determine the optimal cost from A->C costs 1 ( [AB = +3, BC = -2], rather than AC = +4), then isn't that evaluating it and finding the optimal path? Negative numbers here feel semantic, since the ideal path should be identical if it's comparing numbers like 10 and 6, versus comparing 2 and -2, the overall difference is the same. The scale might differ, but I still see no overall change, if the relative change is preserved

This is only my first attempt at making a game, and I don't have a lot of time to code but you're right. Hopefully I'll get faster as I get more experience.

I haven't really done any contact with Polycode, I simply know that Polycode had taken literally years just to put out a single binary; the sheer length of time displacing any remaining hype pretty much became a small-scale meme.

Needs flavor text.

At least he's trying.

Try Godot again, it's getting a big update soon™. godotengine.org/article/godot-30-progress-report-6
Do try going through the tutorial, I find the node-scene setup quite easy to use: godot.readthedocs.io/en/stable/learning/step_by_step/index.html

Just checked the Polycode source, it was last updated September 2015.

there's plenty of options, but final details are of course best performed in zBrush.
Honestly it has the same range of face customization options compared to maximo, but it's just a less intuitive interface.

I never understood zBrush, can I use Maya instead?

i don't use maya, but the sculpting in blender is an alternative method.
i bet maya has an equivalent, so I bet u could.

Let's just say I suck at Blender and I am more used to Maya.

Added to my list, thanks. I don't know when I'll get the chance to add it tho.

This is honestly the best advice. Small numbers are better if you want your players to calculate down to the last damage point, penalizing them for overkilling enemies. Big numbers are better for the brainless, grindy RPGs.

If you want your game to be rogue-like hard, I'd suggest smaller numbers.

I like to consider challenges to have 4 main elements which define their type of difficulty.

Preparation: the work the player needs to do before even beginning the challenge (Things like grinding, etc).

Comprehension: The probing, detection, and comprehending of underlying mechanics of the challenge that the player needs to do while undertaking it. (Things that can typically be trivialized with a FAQ or strategy guide)

Skill: The fast reflexes, good timing, quick wit, good judgement, and clever strategizing required during the challenge. (Things that require practice and experience)

Luck: The favor of RNG Gods needed during the challenge.

3 of these 4 pillars of difficulty can be controlled by the player. Of the 3, only Preparation (or grinding) is not dependent on the player's physical or mental ability. Essentially, grinding should be thought of as a last resort method; allowing players who lack the physical or mental ability (and aren't favored by the RNG Gods) to win against an otherwise unbeatable challenge to them.

I've been messing around with godot recently. Got the pong example to work and I've been fucking with it. I want to get some sort of hit detection going so I got the _on_body_enter/exit signals working but they seem to be imprecise as fuck. Pics related

Is there a way to get that shit more precise?

...

Is that a commission or did you draw it yourself?

try using _fixed_process for physics calculations

personally i use area2d only for triggers and physic bodies for everything else

then use is_colliding() and get_colliding_bodies() instead of signals

it's commission, ask him to show you his 20k abomination

It's been a long time since my algorithms class but we did go over this. I'll see if I can explain what I remember.

Djikstra's fucks up with negative edge weights for a very simple reason: it returns the first complete path it finds. When you look at Djikstra's algorithm, what you see is that it's not actually concerned with finding the shortest path from A to B. That's the end result, but it's not what it's actually finding. It's concerned with finding the shortest path from start node A to every node it visits until it finds end node B. There's a nuance there that even a lawyer would find pedantic, but it's a very important nuance when we're referring to negative edge weights. Logically, if at every step you've minimized the distance between where you are and where you started, then when you've reached the goal you have the shortest path. Therefore, when you've reached the goal, you obviously have the shortest path, because at each and every step you've ensured that you've minimized the distance between the start node and that step. So it just returns that path immediately. Does that make sense? Djikstra's is a little greedy because under normal circumstances that greediness is not only acceptable, it's desirable.

But if there's a negative edge weight anywhere on the graph then Djikstra's has a possibility of fucking up. It won't necessarily fuck up, and it'll always return at the very least a locally shortest path, but it won't always return the globally shortest path.

Look at this graph and find the shortest path from S to T. The answer is S->U->V->W->T for a combined total of -2. But Djikstra's will return S->T for a combined total of 1. Djikstra's expands S to see that the cost to get to T is 1 and the cost to get to U is 2. Depending on your implementation, it can go to T and see that getting to W is another 1, but regardless, the quickest path it's found so far is 1. And look at that, we're at the goal. Since we've minimized the distance between the start node and every node we've visited, the path we've found obviously must be the shortest path. Right? You and I can see that it's obviously not the shortest path, but Djikstra's can't because based on its algorithm it returned the first shortest path its found.

tl;dr Djikstra's has a possibility of fucking up when negative edge weights are present because it makes the reasonable assumption in its algorithm that the first complete path it finds will be the shortest path. When all edge weights are positive, this will always hold true. When some edge weights are negative, this is no longer necessarily true and thus Djikstra's will not always return the globally optimal path.

But it never had a chance to evaluate the other nodes because it was adjacent to the goal; the negative value here never came into play, it was just a matter of the algorithm returning the first complete path it found.

If ST was +10, then it would take path SUVWT to find the goal (since S gets expanded and it can do ST+10, or SU+2); and even then, the negative weight would push it to be considered the lower value anyways.

Also, if I generate a bunch of random points between XY[0..1], can I basically just use those for a Voronoi graph?

It seems like I'd generate them, then iterate a 256x256 array or something; each value in the array would either copy or point to the nearest neighbour'd Voronoi seed?

…yes? You've stated the exact issue Djikstra's algorithm has with negative edge weights, but it's like you're attempting to use it as a defense of the algorithm. It's precisely because Djikstra's returns the first complete path found that it does not always find the shortest path when there are negative weights.

Of course we can change numbers on the graph and have it so that Djikstra's returns the shortest path, but at that point it's no longer the same graph, now is it? The point is proven: precisely because of how Djikstra's works there are graphs with negative edge weights where it is not going to return the shortest path. And because there exists even one graph with negative edge weights where Djikstra's fails to find the shortest graph, we cannot say that Djikstra's is guaranteed to work on graphs with negative edge weights.

If you want further proof, then try this: in between S and , make 3 more nodes X, Y, and Z, and make all weights between them 1. Then Djikstra's would still fail, despite no longer being adjacent to the goal node, because it would see the 4 unit cost S->X->Y->Z->T before it sees the 5 unit cost of S->U->V.

Djikstra's is a perfectly acceptable pathfinding algorithm in the majority of cases, since most graphs don't have negative edge weights (it's analogous to a toll road that pays you for using it; it's not a normal case). But when you add in negative edge weights, it has a chance of fucking up.

I get what you're saying but

>For a given source node in the graph, the algorithm finds the shortest path between that node and every other.[3]:196–206 It can also be used for finding the shortest paths from a single node to a single destination node by stopping the algorithm once the shortest path to the destination node has been determined.

With that graph you've provided, you can expand S, see that ST is the shortest path and arrive at the goal. Based on the above from wikipedia, you can either return that path (the first path found), or save it as the shortest path and expand all the nodes until you've found the shortest path. It will be guaranteed to find -2 via SUVWT.

Whether VW is -6 or +6 should make no difference to the performance or completeness of the algorithm. Hell, make every node increase by 6, and it should still be the same. I don't see how it could possibly fail.

Okay, something like this

And even if VW was still -6, it would then be the shortest path

I'm sorry, I guess I'm not understanding what you're getting at. That image shows that Djikstra's works perfectly when there are no negative edge weights. Which is correct. Djikstra's is guaranteed to work in that set of conditions.

If any edge weight marked 0 was changed to a negative, it would show that Djikstra's fails when there are negative edge weights. Hell, here's an example of Djikstra's following a path with an initial negative weight, and then still failing to find the overall ideal path. I don't know why you're so adamant about this. Djikstra's is simply not the optimal pathfinding algorithm when there are negative edge weights.

And yes, we can modify Djikstra's algorithm to deal with negative edge weights specifically, or we can modify Djikstra's algorithm to work through each and every path and find the shortest globally every time…but in the former you might as well just use a different algorithm (there are several that are built specifically to deal with negative edge weights) and for the former you're doing potentially hundreds of times the work for no point whatsoever (because, again, if you're going to be using graphs with negative edge weights, you should just use one of the algorithms that was designed specifically to work with negative edge weights).

you guys doing math in here or what

Considering learning SDL and making my own 2d engine for a P&P RPG inspired game. Is it worth it? Any comments from anons that have done the same? Seems like a lot of effort to make it all from scratch when my time is already pretty limited but I also like the idea of learning from it.

I'm proficient in c++ as it is. Are there any engine already there that use c++ for this? I know love2d is based on SDL but that's lua.

But it's not going to fail to find the optimal path in your image. That's what I'm getting at.

And that's the goal, to find the lowest cost, not the fewest nodes.

I don't understand how it "fails". It works in this case, still.

Alright. I picked them up and played them to completion.

Slime World

You pretty much summed it up in your post, it's a fun timewaster clicker thing that comes across like it's there to justify the well done bouncy physics. There isn't much more to say. I did have a hard time figuring out what exactly to click on and spammed the center big guy thinking it would be like but eventually figured it out after like 30 seconds I didn't realize the squares with numbers inside them were actual buttons. Anyhow it's alright. Not quite as addictive as cookie clicker but that's mainly because cookie clicker just straight has more content, something I know you can't exactly be expected to add in a game jam.

Tumblin' Down

I'm going to be honest, this is the first RTS I have ever played. I've watched others play and talk about Starcraft and Warcraft, but the genre has never really interested me so I never picked one up. I installed your game without reading the info page and went into it after only watching the trailer and after reading your post.
After finishing the game and reading some of the feedback people gave I have to say I do not understand the people saying the game is too difficult to understand. I found the whole thing extremely intuitive. Capture the points, kill the red guys, control the center. That's all I needed to know. Later when I read your information page I realized the points helped me get food. I thought that those were to help maintain manna or something. But I knew enough that I could actually play the game well enough to win. Which isn't saying much because the game is too easy. Those initial moments where I was fumbling around trying to figure out what I was doing weren't punished, the enemy should have been able to gain the upper hand when I didn't know what the hell was going on, but instead once I figured out what I was doing I steamrolled everything. I could just build a ball of death and move around killing any reds, capturing points and leaving some guys to defend, finding enemy barracks and leaving some guys to kill any reds that come out, and keeping some guys in the center to keep red's from gaining control. I didn't really think much about the strategy that's just sort of the first thing I tried and it worked.
In spite of all of that I did find the game fun. The sound of the crumbling world, waterfalls, ambient music and sounds, all added a lot of flavor. I feel like if I were to do some self imposed challenges to make the game harder the game would become more fun. Stuff like trying to win as fast as possible or only letting myself build a certain number of units. If the AI was better I think it would be a fine little RTS, but once again this is the only RTS I have ever played and I have no measuring stick to put it up against, so take that with a gallon of salt.

I'm trying to make a short game in RPG Maker just to learn about balancing levels for a bigger project I wanna do, but I simply cannot come up with a premise that I like. There was one that I thought might be good, but it was too embarrassing to even write any notes on it.

I need some help.

I'm trying to get my character to look in the direction it's walking in, and it works, except as soon as he stops moving the rotation goes back to what it was initially.

I have no fucking clue why that is
pastebin.com/02TdMaLG

I'm sorry, I really am, because there is obviously some disconnect between us about this. Djikstra's does not work like you seem to be thinking it works. Maybe it'll help if I explain it like this?

When Djikstra's algorithm comes across a node, it counts as a visited node. Djikstra's algorithm, when used in an environment for which it was developed (i.e. no negative edge costs), has found the shortest path to a node as soon as it has visited it. That's why it works. That's how it works. It has found the shortest path to a node as soon as it has reached it. The ENTIRETY of the algorithm ABSOLUTELY DEPENDS ON THIS. When you tell it to find the path from S to T and it reaches T, it does not visit any other nodes. It has found the goal, and according to its algorithm, it has found the shortest path.

What you are referencing in is not Djikstra's algorithm. It will not continue once it has found the goal. You can use Djikstra's to create a shortest path tree, which is sort of kind of similar to what you're suggesting, but you're missing a very important part of the algorithm in that it assumes it has found the shortest path to a node once it has been visited. It returns the shortest path to a node as it reaches it. If the algorithm can continue on after it has found the goal node and it is able to find a shorter path, then it has failed at its purpose. The algorithm is designed to find the shortest path first.

When you add negative costs into the graph, you're intentionally setting it up to fail. Djikstra's, as written, will say this about this graph

Shortest path from S to T: S->T = 1
Shortest path from S to U: S->U = 2
Shortest path from S to V: S->U->V: 5
Shortest path from S to W: S->T->W: 2

If you run it as a shortest path tree creator starting from S, it will say this:
S->T = 1
S->U = 2
S->V = 5
S->W = 2
because it assumes it has found the shortest path to a node as soon as it reaches it.

I think I've bolded that like 3 times in this because it's the crux of the entire algorithm. The way it is designed is precisely to find the shortest path to a node as soon as it has reached it. In order for that to be guaranteed as the case, there cannot be any negative edge costs.

If I still haven't been able to explain it well enough, then I'm just going to have to ask you to trust me when I tell you that it is not guaranteed to work on graphs with negative edge weights. If you have such a graph, use the Bellman-Ford Algorithm, as it was specifically designed to work with negative edge weight graphs (as long as there are no cycles with negative paths, but at that point there's never going to be a shortest path since the shortest path would just be to cycle through the negative path over and over and over again).

is correct. Dijkstra's algorithm does not work with negative edge weights or negative circles. Sure, you can construct some graphs where it works, by carefully placing the certain edges with certain negative costs in certain places, but that doesn't change the fact that it doesn't work in general. And that disqualifies Dijkstra's algorithm in any real world application, because you generally can't predict whether your graph will break the algorithm or not. Again, you can think up contrived cases where you can guarantee that every produced graph will happen to not break the algorithm, but these will be the exception rather than the rule.
You might want to look into the Floyd-Warshall and Bellman-Ford algorithms. The former can reliably handle negative edge weights, while the latter can handle negative edge weights and detect negative circles.

… where negative edge costs can occur.

Hey, thanks for playing it all the way through! It seems you actually managed to win way earlier than I ever did. The win condition is kind wonky.

Cool, I popped your RTS cherry. My first RTS also was some super niche thing. that one was relentlessly hard though

Well, we thought the same thing. Distill the RTS experience so that it's ez to make in a short time ane ez to understand. Hell, we even added autoattack to make it easier. Then again, we are missing many things people take for granted and it might be hard to adjust to such a minimalist approach.

We had different ideas for different shrines, we didn't have time to implement them. I'm glad there's at least one person who got the rules on the fly.

Well, my pal did his best to make AI quickly. It amazes me that it works at all, actually.

My friend actually made the AI do exectly that, but it seems it wasn't enough.

Thanks, although in retrospect we might have spent too much time on the audiovisual presentation side of things. Then again, this is what jams and postmortems are for. Learning from happy little mistakes

With almost 3000 games in the jam to try out, it's hard for me to imagine anyone would do that. It's best to just keep this idea in our minds and perhaps try to further work on it in the future.

The RTS gameplay can be worked on and perfected. I tried it once with a project called tic tac wars (decided not to use codebase from it thogh), tetrachromedev made an amazing little rts demo. It's the "world becoming smaller" part that is supposed to be the innovation.

Don't be. I'll leave you a nice comment on the LD page.

Anyway, it annoys me to hell that over 2000 games still have no comments given to them. Everyone wants to be a famous hipster gamedev but doesn't want to interact with others not that I haven't neglected leaving feedback on /agdg/ to concentrate on shilling and jams

Just look:
feedback.ld.intricati.com

Glad I could help. You and your friend(s) did a good job with what you had. I remember you posting tick tack wars here a while ago and I wondered when I played Tumblin' Down if you were the same person. Glad to see you're still around.
I'm actually really interested in that feedback site. I'll be using that to give the people who didn't get much attention some attention.

You need to check if your horizontal and vertical componentd are equal to zero. If they are, don't run Move():
float moveHorizontal = Input.GetAxisRaw ("Horizontal"); float moveVertical = Input.GetAxisRaw ("Vertical");if (moveVertical != 0 and moveHorizontal != 0) { Move (moveHorizontal, moveVertical);}

~~I get the logic where it can assume it has found the shortest path because it took the shortest route every time, and that's fine. But Wikipedia says it doesn't have to end as soon as a path is found; it can keep traversing the graph to guarantee that it has found the shortest path.

Like, look at pic related. If it always takes the shortest path at each node, then ABD will end up costing 1001. If it finds that path, but keeps continuing like its supposed to, it will eventually find ACD with a cost of 2.

Separately, you could replace the numbers with negative numbers and it would still function exactly the same. As long as it traverses the entire graph, it will find the shortest path. That's what I'm saying~~

Nevermind I missed the part where it evals the smallest point each time.

How can you figure out where "1" and "2" are? It's obviously a line segment perpendicular to AB, but how can you figure out where the endpoints are?

That depends on what the conditions for the end point are, could you elaborate?

Assuming you know the coordinates of the points of each sector, you can easily find out which points belong to both sectors. Is that what you mean?

Yo, this shit was finally added? Good stuff, and about fucking time.
Time to make use of it.
You tried any of ZScript yet, user?

Its voronoi diagram shit. Some algorithms are like >generate a bunch of points >now construct a voronoi diagram from them

Doesnt explain how. I can see the edges and they make sense but i see nl other way than picking random points and hoping its on the line, or else pollong every neighbour region which has to be expensive

And right now I have a bunch of other nasty bugs that prevents the vehicle class for it's proper function, such as when any weapons are being fired it will become impossible to move and for the movement function only 1 of them can become active so it is not possible to move forward and turning left or right at the same time. Here is the decorate stuff if you are interested: ghostbin.com/paste/gu3rs

However a decorate only based menu might still be feasible but this is something I don't need right now since I am more interested on having a proper vehicle physic in Skullcrap 3.0 multiplayer part since the code I used works fine in singleplayer but it bugs out in the multiplayer mode, meh I guess I will maintain Moonman mod instead for a while.

Nope I haven't learned programing at all.

Fortune's algorithm.

On a scale of 1 to diploma.jpg, how stupid should I feel that it took me years to realize I actually can draw simple primitives (lines, squares, etc) in XNA with the 3D rendering stuff but not having to change from the default 2D sprite screen projection?

I just never realized it goes from -1 to +1 on the XY plane, now I can draw shitty lines and not have to rely on spritebatch calls, which is cool I guess.

NEAT

if it keeps going after reaching the destination it's not dijkstra it's some kind of search algorithm

What's this physics model called again? Spring and dampner? It's used infrequently in some physics models?

Spring Joint in most cases, however i can differ between physics libraries.

Is there anything to work on that's fairly mindless? 3d modeling was pretty comfy for a while, but then when it comes to texturing I just stop at UV unmapping because it's tedious. I can't do code atm because between workshit and coding I get burnt out.

Is there a way how to purposely bottleneck my own system in order to test how my game would run a worse computer?

Alter your frame/update speed so instead of running at 60 fps, it takes X seconds to run one frame … just do the math and scale it by the number of FLOPs between your graphics card / processor and a worse one

I think there are benchmark programs, that test how well it runs on your system, and then compares it to other systems through a database. You could also do it manually, by comparing your hardware to older hardware and deduce the FPS drop based on that.

Use overclocking programs to underclock your CPU/GPU?

It's from a standup comedian. His name is Atsugiri Jason, he's somewhat famous in Japan.
He mocks the Japanese language, mostly things relating to kanji. That video is him showing how ridiculously difficult some kanji are to write.

Someone posted this in the FTL thread, thought it was interesting.

I would have thought UV unwrapping fits right into the category of mindless and comfy.
Maybe you don't have a right set up? I didn't like it too until I created proper key binds so I don't have to press weird key combinations over and over.

I will never understand what you guys do to make it super hard
just mark the seams and unwrap

You could rest.

In my case I spend the first half of a day placing those shitty seams and the other half of the day sorting those UV-maps again because Blender likes to leave a few large gaps here and here.

Not to mention this texture type makes it impossible to have about 95-100% usage efficiency because of all those little gaps that are everywhere which is impossible to fill in, tell me again why this shit is tolerated in the first place.

Holy shit I fucking hate anime, but I think that girl you posted might have broken me.
She's so fucking cute!!!

that is what i don't get, making seams should take like twenty minutes at most for humans and then the default unwrap should be enough unless you are doing something retarded like trying to use another's model textures; for machinery and static objects unwrap following active quads or from view should be enough; for plants hair and feathers pre unwrap them to the whole image then put them all sone corner and fix the ones that look wrong

if you care about that use unwrap as lightmap, it works like smart uv but arranges each face individually rather than in groups

Thank you.

I guess I haven't gone spam happy with seams. I only divide where there's a loop or a sharp change in textures (clothing, etc) because making a continuous texture for 2 parts is a pain in the ass. Like if I do a head I'll divide down the middle so I can loop it over itself. Until recently where I got a jumbled up mess of madness.

I never do humans modelling since that stuff is too complex for me and I haven't made a average looking one either, so I mostly stick to buildings, weapons, vehicles and tanks modelling.
I am wondering this do I even need to reorganize the unwraps in the first place when I am done placing seams? Since as I wrote before I do that all the time especially when Blender goes full retard like in the image.

Wouldn't that one require in general high resolution texture so in order to hide the seams better? At this point I rather prefer a .3do styled per-face texturing instead but Blender doesn't support that at all.

fucking Mate DE almost doxed me since on XFCE I used scrot instead of the Mate screenshot thingy

instead of making a game I've beaten gta 3 vc and sa 100% feels bad

Spring damper, which utilizes hooke's law for the math portion, and is used in games for softbodies n' such like cloth.

I see a few things that worry me about the screenshot you posted (namely that you don't seem to bake from higher detail models, and that you are not using materials) so I'm going to post my general workflow with hopes you can adapt it to your own

generally my workflow includes three models.
First one is my base which can be one or various objects and to which i will be adding detail until i'm satisfied or run out of time, each of these objects is separated depending on materials, physical location and repeatability, so you will end with a shitload of them
Second one is my shell, this will be used to bake textures to, this model is one single object and has no materials, is usually taken from the first model when there is enough but not too much detail to include in game. then i weld its pieces into a single mesh and uv unwrap it, it doesn't matter how is unwraped* most of the time, but still dedicate time to accentuate the details.

Finally the third model is the preview, a copy of the second model but ready for export, it has a single material that holds all the textures generated from baking

* it doesn't matter how it is unwraped because the third model (being a copy) will have the same uv maps, is worth mentioning that model three needs to be remade every time you update model two

Does anybody happen to have a camera control that works with a controller that I can quickly use?

I'm not gonna use it in the final game but I need one to check my movement and I don't want to code something I'm gonna throw away

Third person camera controller*

What engine?

Unity

Use QEMU to emulate some shitty CPU and virgl for graphics.

hmm is this really a needed for high polygon baking base model? I always model my meshes as 1 object and later when it is done then I separate it or not depending on the game engine.

So in this case my workflow should be this then right?:
1. make a high-polygon model
2. then make a lower polygon model
3. use the lower polygon model for unwrapping and texturing
4. once the lower polygon model is unwrapped then get the details baked on it
5. then do other steps necessary depending on the game and the model type

Could you elaborate on this part? If I understand you correctly, each object that the first model consists will have one or more materials where each material holds exactly one UV island, which spans the entire UV space. Then you "bake" this to another model? As in, all the models write their UVs and mapped texels to the other models vertices and texture space? How is the texture spaces of the parts of the first model mapped to the shared space of the second model? Manually? By the modeling suite? Are you doing this in blender? What buttons do these things?

...

That's just the movement for the character itself, not for the camera as far as I know

Tutorial series goes over third person controller with camera.

What if the camera was the character?

weird

Got to the camera part, I'll have a look, thanks

is that old shia laboeuf?

New game idea: you play as the third person camera and must film the MC as he completes the game. He can't see what you can't see.

dude skeletons lmao.

This is one of the 2 FPS rigged hand I could find on opengameart, the other one was much simpler but it wouldn't let me control the other fingers for finer control.

This is possibly the worst tutorial I've ever seen
the dude just rambles on using already written scripts, all zoomed out, without actually getting anywhere

could be a pretty fun little coop game where one controls the camera and the other controls the movement and they have to work together to do some 3D platforming.

I had no problems following it.

He even forgets to link his github with a finished version of the script anywhere
At least I found that and can figure it out myself

umlautllama.com/Audio/theory/theory.html
the zen of tracking

Okay so now I have this really good camera that just kinda follows the player more smoothly, the problem is that my movement is not relative to it.

How do I fix that? Anybody have any suggestions?

What's your movement code look like?

Man I have no idea how to explain it.

pastebin.com/Ppr7ieaK

Basically moveDirection is always opposite to where the camera is facing, which I suppose should have something to do with controlling the character from the camera's perspective.

But fuck if I know how to make that work.

Might just take a break, draw stuff down and see if that gives me anything

You should be using FixedUpdate () for anything that needs to happen independently of the framerate, including things like moving the player. Update () is for things that change every frame, like the cameras position/direction.

To have player movement relative to the camera you need to do some trigonometry to find the actual moveHorizontal and moveVertical for line 56. If you suck at math a lazy workaround might be to make the player a child of the camera but I'm not sure how well it would work.

answers.unity3d.com/questions/199744/how-would-i-move-a-rigidbody-towards-another-objec.html
This might be something you're looking for

Oh right, I'm retarded.
I thought I had already changed that.

Alright so I guess it's just about finding how to get the numbers.

My direction is just an angle though, not a Vector3. I have no way of figuring out how to make it like they're doing

glm::vec3 direction(cos(vertical) * sin(horizontal), sin(vertical), cos(horizontal) * cos(vertical));

KILL ME PETE

Take control of the input, user.

Get some good headphones.

Fuck, it just dawned on me

Since blind people are not spending all their time awake connected to onaholes listening to anime girls moaning I think it's safe to assume that VR won't be the end of western civilization.

That was a strangely autistic and outright random observation, user. I'd expect nothing less.

I'm not claiming to presume any such effects of the use of VR, but being blind is distinctly different from VR. Blind people do not receive any visual stimulation to such a degree it leads to their brains being wired differently than those with sight.

The whole "people will be completely absorbed into VR and forget about the outside world" thing is ludicrous, because it assumes that we'll be able to simulate all the intricate parts of life. Even with unlimited computing power, the amount of design and art work needed to make something believable at the same scale as the real world is impossible.

True but it doesn't need to be completely believable. Look at how absorbed people were in shit like WoW and Farmville.

There's finally peace. Now I can get back to getting nothing productive done.

arent those two sister who fucked each other

is not necessary but will make your life easier, also the shitload of models was more for static models, for humans and the like you can get away with by just making the edges instead of whole cloth pieces

the quick and dirty about baking is that you are going to use the second's model faces as cameras. here is a more in depth tutorial about baking in blender. it uses blender render instead of cycles but the principles are the same
katsbits.com/tutorials/blender/bake-normal-maps.php
yes and no, various objects can use the same material and various materials will not use uv maps at all but rather be generated with the object information
they don't write any data into the second model they all write it to themselves then the second model grabs this information and applies it to itself
they don't, the second model is a different object and will generate it's own data depending on what it can see, it will use it faces act as a camera, and each camera renders only to the texture space designated by the uvmap
"U" -> "unwrap" if you already have seams
"U" -> "smat UV project" if you are feeling lazy, it will not use the space as well but it works for testing
"U" -> "lightmap pack" if you want to use the maximum space, the disadvantage is that you will not know what face does what
all blender
by default only U is mapped, but i have bake to f12 mark seam to "," and unmark seam to shift + ","

it looks like you are looking for excuses to not to work

Mate, if you have that many seems, you do something wrong in the first place. Just divide the mesh "not literally but mentally" into few parts, maybe a dozen, and then use just one seam to "cut" a part off, and one seam to "cut" it open, like it's a a piece of cardboard that you want to flatten.
Most of the time in edit mode, use edge select, and you can just press "alt+rightclick" on an edge to instatly select an entire line/ring.
It's a pain in the ass at first, but after a few models you should get the drill. Also Illegally download some meshes from official games that are similar to yours, and compare how they UV unwrapped. Unwrapping becomes much easier that way.

And if a few gaps in the UV pisses you off, get yourself a Blender Addon like Quad Unwrap or UV Squares. Gernerally, if you take too long to make something in 3D, it either means you do it wrong or don't know about necessary tools to make it faster. Half of 3D is learning how to do shit more efficiently, and you can nearly always assume that there is someone who already solved the very problem you have.

Neat, I'll check this out too.

got any good explanations for IT related stuff?
our computer architecture lector explained branch prediction via trains

what exactly do you want to do
get the player to move in the relative directions of the camera rather than the player?

is a cache miss like the train hitting the twin towers

A cache miss is when you're loading the train that needs to head out tomorrow. The supervisor says, "Okay, next we need to load the dildos," and you go to find the pallet with the dildos, but it's not there. So you go to your supervisor and you say, "I can't find the fucking dildos here," so the supervisor goes to the gigantic 10 story warehouse behind the depot. Two days later he comes out with a forklift of dildos, a forklift of lube, a forklift of vanilla porn, and a forklift for everything else slightly related to dildos and lays them all out behind the train for you to start loading. After you load the dildos he says, "Okay, they say after the dildos we need to load the Mountain Dew," and you say, "There ain't no fucking Mountain Dew here, just a bunch of lube," so he goes back into the warehouse and he takes all of the lube and porn with him. Two days later he brings out pallets of Mountain Dew, Doritos, and Call of Duty games. You load the Mountain Dew and the supervisor says, "Next they want the Call of Duty games," and luckily enough they're right here in the last batch of shit he brought out so now the train's only 3 days late and not 5 days late.

thank you, i am forever grateful for this analogy

Working for the first time with someone that can model, kind of new ground for me.

If a model is separated in smaller pieces, can you use a single texture for the whole thing?

yeh

Hey faggots, looks like >>>/agdg/28915 is up for a new board admin

i'll take it for a dollar a week!
you can't beat that offer

I'll do it for 98 cents a week.

Dammit, I would take it if I didn't know I'd be wage slaving

Do you have a less triggering example?

Speebot is a long way from home…

Can't you ask to goy modeling for you? The answer is why did you think this could be a no, you just assign the same texture to the separate models. Unless you are thinking about seams in which case it's really an issue of normals.

He knows how to model, he doesn't know much about unity.
Also I tried assigning the texture to each part and of course it doesn't work.

I'm sure I've made multiple objects use the same texture in Unity before, try creating a duplicate texture and assigning that to the model to see if that is the issue which it shouldn't be.

I've already changed the model tbh, it doesn't matter

I will dance on your grave for that.

Any idea how do I check which joystick inputs unity is finding?
Jump binding was working yesterday, and now it's not, but just for a controller.

I TOLD YOU NIGGERS IT WAS A SHIT ENGINE, BUT DOES ANYONE LISTEN?

How are these two things even connected

Maybe this might help docs.unity3d.com/Manual/ConventionalGameInput.html

Its not terribly hard to build your own serializable class… it probably doesnt handle it natively because how would it know what you want done with a null subarray?

I did see that but for some reason the input inspector only allows you to check axes?
Not sure what's up with that

MoM user here, with a pretty late Update. Today I will be showcasing the Knight Class, his exclusive weapon, and so and so!

That being said, I may also upload another video tomorrow due to being a bit sidetracked, but who knows!

I am happy to answer any questions are per usual!

I'M GONNA DO IT
or i'll give up in a few days because i've got more important things to do

Impeccable Tastes!

Spoopy

No questions but your constant progress is really inspiring, keep up the good work!

I'm accepting ideas for new fun game mechanics in world 2. Any suggestions?

Icy floor.

water level rising if ice breaks
water freezing so you can walk on it
snowmen
santa hats
snow

ice that you can walk on but if you land on it from a jump it breaks

magnets that pull your robot to them, but can be used to extend a jump when placed over the robot

maybe with a mechanic to active and disable them

Thank you user, I am happy that I am able to Inspire other anons such as yourself!

What game are you working on?

The obvious choice is icy floors, but that feels overdone and it's not fun to limit a players movement. Here's some others:

delet this

A necromancy game. Got inspired by the skeleton thread the other day.

Lower the saturation of the water. It doesnt look cold

Would a Castlevania clone be frowned upon as an indieshit platformer?

I really dig that gothic monster slayer atmosphere

Just make the game you want. I'd play it if it's fun.

way ahead of you, nerd

only if you make the art pixelshit

multi-dimensional arrays are just single dimensional arrays that repeat themselves when they reach the next row.

so would it have killed unity to make them serializable? it's just them being lazy

I've just gotten into the habit of using single dimensional arrays for everything, even when something is multi-dimensional (which never really comes up).

How the fuck do animations work?
If I trigger my run animation the thing doesn't repeat and the visualization on Unity's Animator makes no sense whatsoever

How can I set up hitboxes for a sprite based game?

In particular, my character would be assembled out of several smaller sprites and I dont know a good method to assign them.

you have to set the animation to loop from the animation assets settings
select the animation asset, choose the animation tab that shows up in the inspector and then i think you had to tick loop and fiddle with where the loop starts and ends until you get a green dot (meaning that start/end frames are similar enough) - this is only if you're using a humanoid animation/skeleton, if you just made an animation asset in unity you don't get anything that fancy

is it fixed

first check that the animation that is currently playing is not the same as the one you are trying to apply, otherwise you will only play the first frame

trying to make a 3d model of a game demo anyone know if i have to follow to same topology for the torso even if it has clothes on or do i have to do something else for the clothes, because i am stuck right now and cannot continue until i know (i dont want to redo the model because of that).

Man that's dumb

Also the Animator panel keeps breaking on me. Stuff disappears and breaks things in the engine. Did anybody ever test this? No wonder most unity games have shit animations

I'm not sure what you're asking here?
You want to know how to make clothes?

what are you talking about
are you talking about the animator (animator controller) panel or the animation panel (where you create animations within unity via keyframes)

I implemented fighting game inputs for special moves, and a special move to test it with. Also set up a (slightly out of date) movelist, by a request of a friend.

There still isn't anything else to the game apart from the increasingly autistic juggle simulator part.

The way I'm doing it/would be doing it is have a superobject that has those multiple sprites, and separate hitbox data in form of manually fitted rectangles or triangles or something.

Didn't know there was an animation controller.
I'm talking about the animator

I'm really impressed by the framework you're building. Despite all the placeholders it looks really solid

this explains nothing to me
take some screenshots, you fucking mongoloid

Well I mean like, say I have an animation of the character making an overhead swing for their sword. Think Terraria, I guess

Obviously their clothes might change, but I'd need to draw the character out of their component parts, and figure out where the weapon sprite is for each frame of the animation. I don't know how I can do that, as placeholders might seem off. Also, if I do that, I'd have to basically step through hundreds of game states and set up several hitboxes (attack, block, etc) including players and monsters.

I understand what to do, I just don't know how to do it in a way that isn't diploma.jpg

No I mean I don't need help.
It can only be fixed by restarting Unity. I already fixed the looping problem.

I thought they were the same person and they do two takes and put them side by side, cause they never touch or cross paths and they look too identical if you know what I mean, even identical twins look somewhat different due to real life and health affecting the way they look etc

Either you do it like Terraria where you have sword hitboxes that you offset and rotate depending on where the sword is and how it's drawn right now, or you basically animate moves frame-by-frame, which is the way I do it & which involves sucking it up and painting manual hitboxes for every part of the sword swing or whatever you're doing at the time. Nothing wrong with that method, automatic hitbox generation has a hard time replacing human touch as far as things feeling good go.

Dude, just use unreal memegine and this plugin for unreal memegine called allrightrig that will let you animate memes inside the memegine itself.

alexallright.com/allrightrig/beta/

...

woops, let me fix that real quick
memes://meme.meme/memes/meme/
there

yes, i just want to know if clothes have to follow the same topology as the part it is on, like does a shirt need to have the same topology as the torso, or not?

it helps if your clothing has the exact same bones as your regular body, but ultimately it all depends on how you're attaching the clothing to the character

The animation system is a state machine, so brush up on what that is… then it'll make sense.

If you know it's a reproducible bug (else it's most likely user error), then report it.

related thread on unity forums:
answers.unity3d.com/questions/953674/how-to-have-swappable-character-clothing-3d-model.html

depends on how you do it, but in the generic case - like the provided example - you do not need a 100% matching mesh topology as you can just layer the skinned meshes so any overlap is culled.
Although, the rig should match, as explained in the link.

as an example in unity
both meshes share the same skeleton. in my case i'm attaching the clothing to the body by parenting every bone of the clothing to the identical bone in the main body, that way the clothing mesh still follows all of it's own bones, but those bones follow the real ones
i'm sure there was an easier way to do this but i never bothered to figure out how, too much work

Are you making a wrong door simulator?

i did, and every project so far is ded

tell me about your game user

http:// imgur.com/gallery/19nDKUO

looks pretty rad
shame it will never run on my toaster because unreal
where are the fucking skellys

Yup pretty cancerous.

You have some skill, you should finish a project, please.

no
finishing projects is for yesdevs

If he had made it in unity it would have the same shitty gameplay and terrible graphics, this is a good example of how using unreal makes everything look beautiful.

I need some advice.
I'm a amateur programmer.
I was always told that I never should make my own engine.
But since there's not good 2D fighting game engine, I tried to code my own.
So far I made one editor that gathers all the pictures and generates a list of names.
Also made another diferent editor that creates the hitboxes.
All were done in libgdx, every editor took nothing more than one or two days tops.
So far I was really happy with it, and it was easier than expected.
So, the goal of the thread is the following:
How easy and convenient is to make a diferent project for every small part of the game, as in one editor for the GUI, another for the physics, another for the intros, another for the cutscenes, another for the story mode.
Following the logic I could reuse every editor when I needed for other projects.
So, my idea is to have every editor export a file and then a single player binary will load every file and play the game.
So, if I need to change the game, I will simply make some new small editor and change some logic at the player, maybe making a player binary that takes files as a plugin system.
Do my ideas correlates with KISS (unix) way of coding?
It's this a good idea to make my own engine without being a retard?

what is your problem?
just improve any of the already existing engines.

What part of no good fighting game engine exist you didn't get?

you don't need a specific engine for a genre

I've made a basic figting game demo in unity, is shit.
Also gamemaker has not dedicated input manager, online or 3D shit and hitbox editors for what I want.

Also, unity doesn't allow non rigidbody physics which are needed, since unity physics are garbage for what I want.

I need something better than unity.

well tough luck
just code in some simple box collisions, you don't have that many in a fightan game

I've found the mechanim shit really convoluted and garbage for what I need.

so use regular animations instead of a controller

godot has all of those, except maybe online, a puny online plug in should be nothing for someone willing to make a whole engine, hell you can even key function calls in the animator

too convoluted and shit for my tastes, also the splash screen and 50MB binary size with empty scene.

Literally the definition of bloat garbage.

well you wanted an engine that works

What's wrong with making my own shit?

nothing makes me more suicidal than css

well why the fuck are you posting if you're gonna keep making your own engine

Learn to read, I'm asking if my idea is a good design for an engine.

what is wrong with improving already made shit?

if it was good you wouldn't be asking

well unity utilizes PhysX which is a rigid body physics engine; although you can use the kinematic setting so it's not effected via physical forces (i.e. only registering collisions, no forces like gravity/collider forces/etc).
What type of model of games physics are you going for if not rigidbody based?

answer my question.

it's not wrong but it is the inferior choice among its alternatives. creating something that only you can benefit or witness is not something to be proud or even strive to make when the alternative is to help yourself and others with less effort

that is a great learning experience, but from a pragmatic standpoint it's a larger time sink to make everything from scratch (or patching together, more like); in addition to lacking foresight.
if you have the time, and the will to do so; while in the end more than likely taking more time to create an engine than a game… then go right on ahead. It will be a valuable lesson.

Personally, I'd rather be spending that time refining, extending, and specializing an already made engine; which results in an order of magnitude more getting done.
The reason being, is that it's not just me working on the engine. It's entire teams dedicated to improving it, fixing bugs I couldn't be bothered to fix, adding features, updating various APIs to the latest version (then fixing those bugs that are introduced), etc etc. all the while I'm tinkering away at what I want specialized, or extended.

Though, do what you want, and if after reading all of the above you still feel that it would be what you'd enjoy most; then I say go for it user. Remember, it's a hobby, so treat it as such; do what you want.

Zun, pixel and Toady made their own engines.

stay mad plebian scum, greatness is not for you.

Unfortunately it appears that you're blinded by anger (or discontentment in a lack of consensus for your wanted response, more like).
Anyways, I genuinely wish you the best of luck, but I'd be best if you return to halfchan (as you reek of it).

...

nobody gave me a proper answer and everyone gave me the same advice I disregarded in my own post.
faggets.

Are you worried about triggering weeaboo's autism with the girl able to shoot that rifle like it's a fucking full auto weapon with a massive magazine multiple drums?

engine making is not something to be taken lightly, zun spend years iterating on the same engine chugging game after game after game, pixel took something close to seven years to finish one game, and toady has many games under his belt already.
from our own board sigmadev and speebot user are already yesdev

???

understandable

so seperate systems for each, it's a good approach, I'd take a gander at ECS ("entity component system" design pattern, commonly used for game engines as it's specialized for this type of software application).

so totally data driven, correct?

Humm, so are we talking like, construction set .esm files; thus being totally data driven, and you being able to layer logic, data, and other such things?
Sounds doable, and neat.

Yo user, your heart is set on it, do it man; it sounds like a fun project tbh.

thanks for a proper answer, and not meme advice.

of course user, and I'd definitely be interested in hearing your progress later on.

I am back with my horrible topology, time to unwrap and paint, but since it is for a gynoid and it has to look creepy-sexy, any suggestions?

Rate my sword.
It's meant to look like something manufactured by a modern gun company for military use.

It's a neat idea, but it looks like the handle is too short. You would only be able to grip and hold the sword by the lowermost part, after the handles "gun like" parts. It seems unlikely that someone could wrap their hand comfortably around that thing.

The hilt is way too short, especially in the 3rd pic.

I agree, I am not expert in swords but the handle is too small, not even a Chinese sword has a handle that small compared to the blade, a small rule that someone gave me back then was that a handle is more or less 1/3rd the size of the blade.

Actually on further research I found this. It could be a one-handed sword but then the blade is a little long.

Make the separation between the helmet and the "head" more clear.

that's still part of the helmet though.

if anything I think it needs some smoothing.

That depends on how easily an M1911 fit's in your hand.

It is meant to be a one handed sword, so I can extend the handle and shorten/thin the blade if that makes it fit that roll better. For reference the whole thing is now a little over 80 cm / 2.7 feet long.

You don't hold a blade the same way you hold a gun.

Doesn't Maya have "sharp edge" function or something like that?
It would look much clearer if we could distinguish between helmet and face.

You mean harden/soften edge, yes, the problem with that thing as you pointed me out was more of a sculpting problem, I was reluctant to add more polygons and sometimes I have to manually cut stuff, working on it right now.

...

That one looks pretty good.

Just one thing, make sure the handle is not perfectly round, for practical purposes anything with a blade needs a slightly flattened handle so you have a sense of "orientation" with the blade.

...

Sorry, practicality over fancyness, you can still work it around in a different way, use the hilt for the "gun-like" bumps and keep the semi-rectangular pistol part for the grip, just make it a little longer.

the problem comes with how the back of the gun would be to uncomfortable to wield, hell, if you turn it into a saber and let the trigger guard in you have one interesting design, the only real issue is the angle, if you want gun stuff on your grip pretend the gun is made in a straight angle.

I think elongating the handle and making it a little pear shaped would make your pistol-handle sword look more balanced without sacrificing the pistol look too much

No point in trying to be 100% realistic if you didn't plan to be from the begging.

Alright, so I lengthened it more and rotated the top parts of the handle to try and facilitate a more sword-like grip.
I'm not sure about including the trigger guard, though.

As for making the blade more Sabre like, I was going more for a backsword type design like in this last pic. The curved blade of a Sabre is a bit too elegant for what I'm going for.

The sabre thing was a suggestion for the trigger guard, it just goes for something like a single edge sword and the like.

Ah, I see.

I refined the helmet… so, any nice suggestions for painting it to make it look like one of those sexy bots you want to bang but you find creepy because there are no facial expressions?

I think you gotta make the lips really sexy. The whole gynoid effect is acomplished by the juxtaposition of secondary sexual indicators, in this case full, pouty lips, against the alien expressionless helmet.

Well, I can do some more sculpting on that, but what about the helmet's paint job in general?

Oh, I'd go with something dark, but it should coordinate with the rest of the character design

OK, I'll see what I can come up with.

I think this looks better.

Does anyone have a link to a good Blender animation tutorial? I'm clearly doing something wrong with the NLA editor, and I can't figure out what it is. Animations seem to prevent each other from running properly, even when they're not overlapped.

I can't even get the grasp of Maya's animation most of the time, stupid iks prevent me from rotating joints when I set them up.

...

Think total war but with TRPG combat instead of real time.

I had a similar idea. but it was going to be a turn based roguelike like C:DDA where you could perform necromancy and give orders to your minions.

Suffering is drastically preferable to comfortable apathy by virtue of what it yields.

Daily reminder only plebs use engines rather than learning the math and CS shit to make it on raw opengl.

i'm already failing university because of algebra and geometry, and i'll never finish a game even if i'm using an engine, and you're telling me to waste even more time on something that i don't need?

nigga come on, are u a brainlet?

go back and start from aritmetic bro.

can you name all the rules for a ring by heart?
if not then you just failed the exam

...

user, you clearly haven't done the same exam i've done, don't lecture me on shit you don't know anything about
there is no "bare minimum" here, you either fail or pass, if and only if you know every single theorem in the lectures by heart.
practical exam is easy, theoretical is the one where they fail you
source: about half of my group is doing the exam for the 4th time now

are u studying pure math?

IT systems
algebra is 1st semester of the first year, geometry is 2nd semester of the 1st year
i'm in my 2nd semester of the 2nd year, if i can't pass both by the end of the year i get expelled and have to reapply

how high is your IQ?

also, look up the TTC math courses.

Daily reminder that Enginefagging is not Gamedeving.

i wouldn't know, but i don't feel like taking a test that's too long

IQ tests were originally developed to find children who were under-performing in school in order to give them extra help. IQ tests to this day are only useful to determine one's chances to do well in school.

The free tests don't accurately judge it anyway.

you could at least link the new thread bunch of fags

C'mon, son.

Lies. IQ correlates with capability to succeed. Most of 3rd world countries, as south africa, have a very low average of IQ, while more succeeding countries, like norway, have higher IQ average.

IQ can be used to track more than just chances to go well in school.