Amateur Gamedev General ~ /agdg/ + /vm/

No embarrassment this time edition

Resources
>>>/agdg/
>>>/vm/

Links
>Wiki: 8agdg.wikidot.com/
>Beginner's guide: >>>/agdg/29080

QUARTERLY DEMO DAY SCHEDULED FOR MAY 5TH
Polite reminder that the wiki exists, you are encouraged to contribute to it if you can (even if it's just your game page)

Other urls found in this thread:

itch.io/jam/monogamejam
mega.nz/#F!UBp3xDrK!MRcdOY9QFHx1XtA4aXIOCw!pIZGFYyA
nintendo.wikia.com/wiki/Hiroaki_Suga
archive.is/elG5I
archive.is/E36GI
blogs.unity3d.com/2016/08/12/unity-oga-partner-to-increase-speaker-diversity/
github.com/OmniSharp/omnisharp-vim
stackoverflow.com/questions/538060/
godotengine.org/article/godot-2-1-rc1-out
themmnetwork.com/blog/2011/05/30/the-unseen-worlds-of-mega-man-legends
textures.com/
poliigon.com/
hastebin.com/carinamehu.cs
github.com/ostoru/vcsona
blogs.unity3d.com/2018/02/13/introducing-2d-game-kit-learn-unity-with-drag-and-drop/
twitter.com/SFWRedditGifs

finally implemented reflections correctly, weird bug in which diffuse material shows after one reflection, looks kind of interesting.

i want a smoothie now

Fuck.

Anyway, I expected depth-sorting to take ages, but it only took me an hour after i realized that i set the scale of the image to 0
To show it off, here's my attempt at a Löve-tan (or maybe it's just a cute girl who loves the engine a bit too much, who knows, I've never made this sort of thing before)

Some user asked how he would implement a list of requirements before starting a quest. I haven't slept in 30 hours.
using System;namespace QueryManager{ class Program { static void Main(string[] args) { // Functionally, we just have an array of prerequisites before the player can start a quest // If they've met them, they can start the quest (or whatever it's being used for) // Require the player to kill four NPCs var killRoute = new Criteria[] { new Criteria(QueryType.NPCKilled, 120), new Criteria(QueryType.NPCKilled, 15), new Criteria(QueryType.NPCKilled, 5), new Criteria(QueryType.NPCKilled, 200) }; // Alternatively, the player could just bribe a specific NPC var bribeRoute = new Criteria[] { new Criteria(QueryType.NPCBribed, 200) }; // And here's how we'd finally test it var manager = new CriteriaManager(); bool killed = manager.Completed(killRoute); bool bribed = manager.Completed(bribeRoute); if (killed) Console.WriteLine("Killing NPCs is fine"); if (bribed) Console.WriteLine("Bribing NPCs is fine too"); if (killed && bribed) Console.WriteLine("But bribing them AND killing them?"); Console.ReadLine(); } }}public enum QueryType{ NPCKilled, QuestCompleted, NPCBribed, Unimplemented}public class CriteriaManager{ // This is the function that actually checks if a quest can be started public bool Completed(Criteria[] quest) { for (int i = 0; i < quest.Length; i++) { if (!CriteriaMet(quest[i])) return false; } return true; } // Check if a single criteria has been completed private bool CriteriaMet(Criteria c) { switch (c.queryType) { case QueryType.NPCBribed: return NPCBribed(c.queryData); case QueryType.NPCKilled: return NPCKilled(c.queryData); case QueryType.QuestCompleted: return QuestCompleted(c.queryData); default: return false; // Not yet implemented } } // Check data stored in m_data and return true or false to each specific condition // These are all specific to your game, so in this example they always return true private bool NPCBribed(int id) { return true; } private bool NPCKilled(int id) { return true; } private bool QuestCompleted(int id) { return true; }}// I chose to use struct to keep it fast and simplepublic struct Criteria{ public QueryType queryType; public int queryData; public Criteria(QueryType queryType, int queryData) { this.queryType = queryType; this.queryData = queryData; }}

let me take this opportunity to make a public service announcement

How to webm for retards
ffmpeg -i input_file.mp4 -ss [time-in] -to [time-out] -c:v libvpx-vp9 -crf [quality rate] -b:v [video bitrate] -threads [duh] -c:a libopus -b:a [audio bitrate] output_file.webm

notes:
you only need the times if you're going to cut that part from the original video,
constant quality rate should be used, values between 30~40 are enough, the smaller the number, the higher the quality
the video bitrate is just a constraint, you should use 0 for pure quality control
libvpx doesn't multi-thread by default, it may or may not improve performance
the audio bitrate is needed as well, values of 48k~64k are sane, opus compress is extremely well
obviously only use the audio parameters if you actually have an audio input
you can also pipe input directly from your screen, for live recording, google-chan will help you there

i don't use vp(8/9) because of the glacial encode times and lower (as far as i've seen) quality for the same bitrate

I just use XMedia Recode, because it has a bitrate calculator, so I can just set it to 16 mb and it adjusts the quality as needed.

I remember that, will check better later on,
but what I thought is, for example, make a class solely for the triggers with three main values:

for example, what could be experience, some flag, and whatever,
how, is the operation to use, lesser/equal/greater, or possibly something more complex depending on need
and the value, obviously is what to check it against,

so whatever other class took this condition, would know how to read it and check it,
as for what happens next, I thought of coupling the condition with which ever other data it triggers, or maybe using forks:
when you reach A it can fork into B or C or D depending on what conditions check

and btw, using dark board style here, how to make code tags fucking more legible to read?

i just use any video converter and set it to the lowest possible quality
i'm a retard

copy the code into the editor of your choice

Why no 2-pass encoding? Isn't it better?

Hey /agdg/, I got a question.

Not sure if this is the right thread for this.
I started programming this year in school (c++) and it's a lot of fun, we are now supposed to make a game in class, but only simple shit like hangman or tic tac toe etc. which I think is lame.

I want to do something more advanced, I would absolutely love to make a short demo of a Visual Novel completely in c++.
Can anyone give me some advice on where to start?

That's not what I want, I want to learn how to code something like that.

Don't you know you're not supposed to show your power level? Come on now.

Usually when they want you to make hangman or tic-tac-toe they mean you make it using just the console, and otherwise they would have given you pointers on some graphics framework like SDL or SFML and you wouldn't have to ask us how to start since VNs are extremely simple.
If that's the case, a more advanced game with the console would be a roguelike. A VN in console mode would be shit.

how to know when to quit

Follow this handy chart

Spent the last few days tackling improved rendering and pretty happy with how things are going.
End result is something like 50-70% reduction of draw calls and CPU use almost halved. I could probably push them lower but it would probably make drawing sprites a lot more complicated. Next need to change to triangle strip format to reduce vertex count, re-add entities and particles, then time to start on map editor!

that is completely false, the only thing that even compares to vp9 is HEVC
and we're talking about shit 1 min max videos here, it doesn't take that long


that is incredibly inefficient, even if you had a 10 second clip with a still image it would still take 16 mb


exactly why I posted that, now you can pretend you aren't a retard


it depends on the case, with most modern codecs it reaps very little reward in terms of compression/quality for the time spent

and as a general rule of thumb, you can pretty forget about bitrate control with any of the decent codecs, just set a constant quality and let the variable bitrate fall wherever it may

Social media is some bullshit.

My class is chill as fuck its fine.

Yes that's correct, I don't want to do it in console mode though.

SFML, GLFW+GLEW+GLM+GLU+OpenAL(if you hate yourself a little), Allegro, etc. There's tons of options for getting stuff on the screen(and to the ears) in C++.

I've finally got the time to sit down and work a little bit, of course while I'm sick, for the third time this year fuck my life.

I've been trying to figure out a way of adding a native method of spawning a block upon block harvest, and I can do it easily enough if I just copy/paste a bit of code from blockIce to other block constructors, but I need to do away with individual block constructors at some point in favor of a universal constructor class.
I thought I would just do what I did for simplifying items dropped by blocks, and make a new map of key:value pairs with blockID and block to spawn when harvested and iterate through that in my universal constructor class. This doesn't work, I guess because I have to call harvestBlock from the super constructor Block class and that interferes with the other function that checks what needs to drop from harvests.
How can I go about tackling this? Should I use a single hashmap setup like:
and just iterate over those two until either a block matches a key and if not just return the default? I really don't know how to iterate over a hashmap with key{key:value} pairs.

Those are some nice reflections. How badly does it tank the framerate?

the framerate afaik isn't much higher without reflections


i tried reencoding a hard h264 video that i posted earlier, maybe it's bias but vp9 seems to leave different more annoying artifacts, whereas h264 seems to fringe edges and it's less noticeable.

for comparison

Alright will do some research on that thx.

I'm still having trouble getting my online play to feel right, even with basic stuff. Not that I'll get any useful advice on that here.

Reflections aren't all that costly, especially right now since they have a maximum depth of 1. Sounds to me like you're manually doing some kind of ray upon reflection by the way, you should just return the result of a ray spawned the usual way except at that location instead of the camera, which means it will recurse as long as it keeps hitting reflective materials. Don't forget to add a cap to how far it can go and return the diffuse when you hit that cap.
Refractions are a bit costlier, and glass even more so since you do both a refraction and reflection then calculate how much of each result to use based on angle.

Could never figure that shit out.

no recursion in opencl
current reflection code:
if(reflect) { afactor *= 0.8; float3 refd = normalize(r->direction - 2.0f*dot(r->direction, lnormal)*lnormal); r->origin = hp + refd*0.001f; r->direction = refd; spherelowt = 1024; cast(r, scene, &lnormal, &spherelowt, &color2, &reflect); if(spherelowt > 1023) return color * afactor; }


it's basically gl 1.1 with a shader doing all the rendering

muffled Tristram theme plays in the distance

Finally getting the thatching right was a breakthrough and a relief

I imported the old hat, its the only thing i saved from the old Adam
And made the shirt, it got done in record time, it has a shape key to make it look tucked into the pants if pants are present

Stay a while and listen

2deep

Oh yeah, that's true I guess. Something like this should work though, I think:
if (reflect) { while(reflect && depth > 5) { //set depth to 0 at the first ray depth++; afactor *= 0.8; float3 refd = normalize(r->direction - 2.0f*dot(r->direction, lnormal)*lnormal); r->origin = hp + refd*0.001f; r->direction = refd; spherelowt = 1024; cast(r, scene, &lnormal, &spherelowt, &color2, &reflect); } if(spherelowt > 1023) return color * afactor;}
You might want to turn reflect from a boolean type into an enum type called mattype, so you don't keep having to add more booleans for to check if it's reflective, refractive, textured, blinn, etc.

The walls and window frames are too blocky but that lighting and roof are fucking spot on.

Whelp nevermind, you already figured it out yourself.


That's a comfy looking house, that light looks inviting.

why didn't i think of this
thank you!
(to implement the 2 deep reflections, i just unrolled it again, which is kind of ugh)

You can also solve the "no recursion" problem by implementing a ray stack manually, should you ever need to implement shit like refraction and diffraction.

raytracing in unity is easy, just create a gameobject for every pixel

You just made me physically sick.

"ray stack"?
i think i have an idea of what you're talking about, but i'm still a tiny bit confused

user I'm sorry if just not initiated but what the fuck are you trying to accomplish?

Basically, if you were allowed to do recursion, you would write something like this if you wanted both refraction and reflection:
result trace(vec3 origin, vec3 dir) { result res = /* intersect shit and get diffuse color here */; if (res.need_reflection) mix(&res, trace(res.intersection, reflect(dir, res.normal))); if (res.need_refraction) mix(&res, trace(res.intersection, refract(dir, res.normal))); return res;}
But if you can't do actual recursion, you can simulate it by manually managing a call stack of sorts, like here:
typedef struct { vec3 origin; vec3 dir } trace_args;trace_args ray_stack[MAX_RAYS];int stack_len = 0;result trace(vec3 origin, vec3 dir) { result res = /* intersect shit and get diffuse color here */; return res;}result trace_stack(void) { result res = NULL_RESULT; while (stack_len > 0) { trace_args args = ray_stack[stack_len]; stack_len--; result tmp = trace(res.origin, res.dir); if (tmp.need_reflection) ray_stack[stack_len++] = (trace_args) { res.intersection, reflect(dir, res.normal) }; if (tmp.need_refraction) ray_stack[stack_len++] = (trace_args) { res.intersection, refract(dir, res.normal) }; mix(&res, tmp); } return res;}
Of course this is just an example to explain what I mean, you also gotta keep track of various coefficients and shit.

The blockyness befits such chunky characters.

Shit, this should be

there is a small bug here but like it looks cool so have a picture

A good start for a comfy village


thanks, I've been trying to get that roof right for days. When that was done, everything else went smoothly!

yeah I'll try to recreate those wall bonds next
Maybe I'll add some smoothing to the edges too

I'll keep the window frames for now, they work well enough.
compared to real architecture, they are likely too big but the lighting effect wouldn't work as well

and remembered that I still need a chimney!

I wouldn't worry about realistic proportions, seeing as the tictacmen don't even try to be realistic unless you want to use some other non-cartoony models later on and especially not about realistic lighting. This isn't autism simulator so you'd be only tying your hands, when it comes to creating a proper mood in locations.

Fug, meant for

Reducing minecraft's codebase into less classes that can reused logically by similar objects so that at some point a proper API can be built because forge is fucking garbage.
Once I've whittled down these excessive classes into widely used ones, changing basic game rules and mechanics should not just be easy but accessible to external configs. There is a LOT of work to be done and I have to learn this shit language as I go.

Looks like the sky reflection stopped working for some reason, do you handle sky by having a maximum ray length and returning sky if it doesn't hit anything?
You should probably add a skybox while you're at it, you'll probably want an HDR one ASAP if you do wind up implementing pathtracing. That's rather important for accurate results there if I remember correctly.


You do realize that's exactly what they're already doing for 1.13 or whatever, right?

The terrain rendering engine guy is still posting progress, this time he's trying to do dynamic lighting.

I haven't heard of this, but I am highly doubtful, the only evidence I've seen of an API in the works is command blocks with twitch integration.

nice vrchat avatar, faggot.

All recipes have been moved over to .json files, all world generation is .json files now, and they're most likely planning to make block definition entirely based on the resource packs as well although that is unconfirmed.

Creating those simple games help. They're small in scope, you learn a lot from them, and you're able to experiment with it.

VNs take a lot of work and they don't teach you much about game logic.

yeah i fixed that, now the issue is that the entire scene is present in the middle of reflections
if (reflect) { int depth = 0; while(reflect && depth < 5) { //set depth to 0 at the first ray if(dot(lnormal, -r->direction) < 0) { lnormal = -lnormal; } depth++; afactor *= 0.8; float3 refd = normalize(r->direction - 2.0f*dot(r->direction, lnormal)*lnormal); r->origin = hp + refd*0.001f; r->direction = refd; spherelowt = 1024; cast(r, scene, &lnormal, &spherelowt, &color2, &reflect); hp = r->origin + spherelowt*r->direction; } if((spherelowt > 1023) || reflect) return color * afactor; }

java or win10 edition?

Java


Okay, now that I have zero idea of for what could be causing it.

figured it out!
if (reflect) { int depth = 0; while(reflect && depth < 5) { //set depth to 0 at the first ray if(dot(lnormal, -r->direction) < 0) { lnormal = -lnormal; } depth++; afactor *= 0.8; float3 refd = normalize(r->direction - 2.0f*dot(r->direction, lnormal)*lnormal); r->origin = hp + refd*0.001f; r->direction = refd; spherelowt = 1024; cast(r, scene, &lnormal, &spherelowt, &color2, &reflect); if(spherelowt > 1023) break; hp = r->origin + spherelowt*r->direction; } if((spherelowt > 1023) || reflect) return color * afactor; }
everything opencl is stuttering for some reason now, not sure why

I just checked the gamepedia info on 1.13, and I really don't know what to think at this point. It's like they saw all my kvetching about what should have been done years ago and finally decided to do it.

I still expect this to be terribly lacking and it's still going to be with their bloated newer version so I may as well keep doing what I'm doing.

So anyone know if there's a better method of doing what I said here
?

If you're posting to faceberg, remember that it's an advertising platform first and foremost, and a social media platform second. Think of every post you see on that website as being a piece of native advertising, even if it's something that came from your little old aunt, parent or whatever. You either have the natural draw required for the system to push your post into other people's feeds or you need to pay for the privilege. The feature that allows you to pay the Zuck to get your posts to appear in other people's feeds is there for a reason.

I guess both posts are about the same thing. It does indeed look like a place blobby bois would live in. I'll keep it like that but I will keep adding extra elements and other buildings to populate the town.


are the bots treating you well? I lost half of the usual retweets after Twitter increased the character limit. got any tips other than spamming #gamedev #indiedev #madewithunity #unitydev #indiegame?


sorry I didn't leave feedback on the demo day thread, bro. I missed the chance to connect to your server, I guess. I love the fluidity of the game overall, though. needs fullscreen tho, I want to post scanlines


ummm, may I ask where the fuck is it asking for a friend


Why didn't you post House of Spaghetti? it looks real good or did I miss it?


wew

I think I might have forgotten with the holidays and all that. and I stayed up all bloody night on New Years making that, till like 10 am or something.

Besides, it was six weeks ago at this point.
I'll try to remember to link the new stuff here from now on. And that should be more interesting since I'm porting everything from GMS to Love2D, and will document everything from the beginning this time.
g-glad ya liked it, user.

Pic related, it's a feature in Xmedia Recode that lets you provide a specific file-size, calculates the bitrate of the video accordingly, and compresses them to that level. You can also configure stuff on your own, for example, if you need the audio quality to be better, or like in my case here, if you don't need it at all as I don't have SFX in these webms.

I don't even bother with Faceberg because you have to pay for advertising if you hit a certain threshhold. And I very much doubt a lot of people look there for niche game dev shit.


I get a couple of bots with every post. And I don't, sorry. My last few posts didn't even break 100 notes, but the ones before that all got 300-700. I don't know what I did differently then.

fugg, I was happy with 6

Really fits the feeling with the colours, I'm liking it.


No worries, you got to avoid the accidentally fucked netcode. The next session should be pretty soon, so I might try and schedule a day and time to see if I can get more people on at once.

How do you think a PS1 game like Silent Hill would have handled dozens of snowflakes that individually collide with the ground? Maybe I'm just overselling how computationally expensive something like that would be, but it seems that, given the severe hardware limitations, it wouldn't be feasible. Come to think of it, most of the terrain in that game is pretty flat. Maybe there's a hardcoded y coordinate at which the snowflakes stop moving. I dunno, I just like particle effects.

Trying out an alternative camera and attacking feel. Kind of gives a more action vibe than that MMO vibe. Anyone think this looks more attention grabbing? I kind of like how close to the action it is. Auto mouselook for attacking instead of previews rooted to the ground, too. I might try setting up a controls demo if people might be interested in trying this out if there's some interest.


Thanks for this. 2MB is easier to deal with than 20MB.

Whoops, forgot pic. and looking at my post, it's worded in a disjointed and autistic fashion. I just wanted to spark conversation about these types of effects and how they can be optimized

It's not that computationally expensive. I doubt there is real collision detection for the snowflakes beyond checking that they reach a certain Y coordinate. With PS1's low resolution coupled with CRTs of the time it would be hard to notice that.
The PS1's standard graphics library included such primitives as sprites, which are textured quads, and tiles, which are solid-colored quads. You could form draw/transform queues out of those and also out of more complex 3D objects, which were then handled by the GPU/GS automatically, much like draw lists and vertex arrays in modern OpenGL. I doubt that a couple dozen additional solid quads (or even pixels; there's even a primitive for a pixel-sized tile) really made any difference, since the draw queues could include up to several thousands of those primitives.

can't really tell where (the freaking glitches didn't help either), and if you're getting a larger file size you're doing something wrong

niggah, that is hellah wasteful, so much that you should be embarrassed

Do you plan of adding 3d bridges/room-over-room to your engine in the future?

Oh fuck, I just realized that my polygon splitting is going to be screwed once I try it because my polygons are formatted as a list of triangles instead of an actual polygon you can draw with GL_POLYGONS. Now I have to figure out a way to feed better data to my BSP compiler…

(the glitches are what makes that video hard to encode)
i paused on frames where the center green twotris are visible and looked at the edge to the sky.
h264 seems to overcompensate for the edge, vp9 leaves more spotty artifacts around the edge than h264 does

Anyone on MonoGame? They just started a week-long jam about 15 minutes ago.

Theme: "Change is good"
itch.io/jam/monogamejam

Here's a mega link that contains a bunch of game-dev and game design related books. Don't know if they're any good myself, but there's a lot of material.
mega.nz/#F!UBp3xDrK!MRcdOY9QFHx1XtA4aXIOCw!pIZGFYyA

RPGdev guy, spoilered jump animations. Wanting to finish these tonight but I might run out of time, gotta work in like an hour. Gotta do legs and arms for two and arms alone for one.

Also here's something I've been thinking about for a little bit: so Unreal and UT99 have Fire(float value) and AltFire(float value), which can take any value as they are arbitrary. So I'm considering whether I want to give all pawns I'm gonna use a weapon and pass values between 0-24 OR if I want to just give 1 weapon to the player (attached to a floating camera) and 1 weapon to an AI controller class that is passed information in the platformer mode or combat mode and just reacts using a tick(float deltatime) function. The reason for this is everyone will have 1 weapon that will give them about 25 abilities OR it will target a pawn and make them do certain things. Not sure if you can call fire() and have that somehow trigger a pawn's dojump() function, once or twice or whatever. All of the abilities come down to firing different projectiles or are movement abilities.

Speaking of abilities, if I can figure out how to do it, I might add wall jumping for a certain subclass as a passive in the platforming mode.

Also learned some better brush techniques for maps, apparently it is a better idea to build hallways/rooms out of subtraction brushes then add in semi-solids instead of solid brushes for things like planters, pillars, stairs, etc. It cuts the BSP of a room less and makes for less BSP errors and 'hall of mirrors' issues. The new map will have these, gonna do a Deus Ex training room sort of thing, but also using plain blue textures for a 'digital training room' sort of vibe.

Can that be possible in raytracers?

raycasting

Thanks!
just mentioning vr chat fills my mind with clicky sounds and tiny deformed knuckles HELP

Aren't they rewriting Minecraft in C++ and opening it up for modding?

They are still updating the java edition, because they've been working with it longer. The c++ is the mobile version ported to windows 10 and only has a twitch API so streamers can have their viewers send command block signals to their game via chat. So minecraft isn't being rewritten, but apparently they've started work on actually making the static game rules accessible via json files.

Man, Minecraft is such a fucking anomaly. Say what you want, but it's a game that succeeded despite it's flaws. Written in Java with piss poor performance, vapid updates and lackluster pseudo-lore that failed to build up to anything meaningful, and yet it's got a heart of fucking gold.

I hope Minecraft 2 won't be a disappointment. You know they're going to make it. Actually, nevermind, I can fucking see it already. It's gonna be pseudopixel shit like picrelated, it will have poor performance and disjointed visuals, and it will fail with everybody and be forgotten.

How is that pixelshit, when it's a more heartfelt rendition of what Minecraft Steve could have been?

Steve was adorable programmer art. It was genuine simplicity to represent something in a laconic fashion. That "heartfelt rendition" forces the game into a particular style, while the way Minecraft ended up looking is much more neutral and lends itself to much more variety.

Minecraft steve looks like fucking SHIT

You're so full of shit.

no you


fucking duh. It's programmer art

It was genuine laziness.

It was made by one ultra-autistic programmer until the game actually got popular. The assets weren't even modeled, he just smashed cubes together and put textures on em. Simple and effective.

The term being sought here is "programmer art".

Anyone know a good resource on OpenGL + SDL stuff?

...

It's pretty similar to other games like doom, and PUBG off the top of my head in that, none of those games invented the genre's that they are known for, but they managed to so totally eclipse what were previously niche genre's in gaming by creating a game that was able to show what the genre was actually capable of. Maybe that is why those games succeeded. They don't have to be amazing games, for their genere, but they just have to be the first to prove what can be done in that space. At least, that is what the pattern seems to be.

I think this is the right answer. I remember as a kid I always wanted a game with totally destructible environments. When Minecraft came out, it felt like catching a glimpse of my dream game. I think you hit the nail on the head.

Soory user, I miss posts easily when there aren't images nearby.


It's this basically. If your game covers new ground in a thought out sub-genre and is at least presents value (aka isn't complete garbage) then you basically become a leader instantly to that uncontested market space. If you're even a moderate success, you'll have cloners and followers, but it'll always go back to you as the founder.
Of course, you have to be making a product that has some semblence of QA, but if it scratches the itch and is marketted well, you win.

Should I invest any effort into OpenGL 3+ or should I stick with OpenGL 2?

There is nothing adorable about programmer art. Cats, kirby characters, and lolis are adorable. Programmer art is dogshit.

Kirby's design was programmer art.

You can't tell me that this ain't adorable, nigger.

Emphasis on "design" not "final spritework".

Kirby was literal placeholder art by Sakurai, the lead programmer.

Hey guys I was wondering, if you could hang out with your loli AI friend in VR what hand gestures would you do to her? I'm talking things like touch her face, bully, play patty cake or something. I'm asking because I'm building a sim and am brainstorming ways to build a fun Tamagochi-esque game.

I "knew" that, but googled it to double check. Which one is this guy?

Nevermind, I'm retarded.
nintendo.wikia.com/wiki/Hiroaki_Suga

Sakurai was not the lead programmer dipshit

bully her by punching her in the face to death

Rape, Paper, Scissors

Really makes you think.

Thumb War.

...

...

What do you need OpenGL 3+ for that you can't do with OpenGL 2? If there are things you can't do that you want to do, then it makes sense to switch.

Okay, why is MonoGame so retarded?

fuck me

I think to death is a little morbid but I wouldn't mind a little bit of punching. I'll think about it.


Not really feasible in VR.


A board game would be fun but Monopoly would be a game in itself. Maybe something simpler.


Good idea for bullying! I want to do lots of hand gesture driven reactions so that's a very good idea.

Rock, Paper, Scissors!

Yeah I also thought about that at first as well but it's not really fun to play. Maybe if there were things at stake it would make it far more interesting.

Headpats.

I am slowly adding support for my new map format with the textures in the engine, this is buggy on purpose since I just wanted to see it.

Is it feasible to have her respond to this?

ok, here it is with the textures actually working properly. The issue is that right now both the engine and the map editor have become unstable, so I have to find all of the bugs. Right now the engine crashes in Vulkan mode and the editor crashes when trying to export the other textured map I am working on, so once I fix those bugs I can move on to new features.

...

...

what the fuck

UNITY YES
why are you torturing yourself?

this as an easter egg

I use 2017.2 because 2017.3 was being a fucking bitch with certain things, particularly IDEs. I gave up because even with both versions installed, the former works and the latter just doesn't.

because I'm retired, user, and this is the only way I'll ever finish a game

I've constantly had so many issues with Unity that I've given up on ever fixing it.
Today is a good day, it only crashed once while I was trying to run my game

Is it safe and stable to develop with Unity on Linux?


Im more worried about being able to do that with the simple controllers. Maybe when the knuckles VR controller comes out.


Already in the works.

I'm just trying to export to linux so I can host my server for $4 cheaper

Fairly sure his raytracer is portal based, so yes it's possible. You can create bridges, rooms over rooms, or even non-euclidean geometry.

It's mostly obnoxious when it crashes and you have to go delete the lockfile so it'll let you open the project again. Doubly so when it keeps crashing and I have to run my DropCaches.sh script to flush it out. I would love to make my game in C++ or even Godot, but most my time and energy is in Unity and while I understand general programming concepts, the game I'm making isn't trivial and can't just be slapped together in anything while learning a new environment. Juggling classes makes this enough of a pain. Still, it's been getting the job done for me. If I had a jillion time and cash I'd probably make a studio do custom engine stuff while I focus on design and my artist does art.

the worst part is, unity does everything I want it to, just very poorly

...

Each unity particle system uses 3500 bytes just to initialize. I wonder what they're doing with that data?

Last I checked, their C# support is in alpha, and I have no interest in using another shitty engine-specific scripting language

this

why are their legs painted black?

Long socks.

are bows a good idea for weapons in a game that's neither fps nor targetting?
i know people generally disliked them in dork souls

People disliked them in dork souls because their implementation is crap since they're basically just shit spells.

Then maybe you should check again, because you're wrong

...

Well, absolute length borders (as opposed of proportional distance to center) are """"done""".
There are these issues to fix, the bad news is that I don't quite get the math behind these problems, even if I logically know why they happen.
The algorithm was:
for each point:
for each convex polygon it is part of:
get the last and next points (either from the convex, or the whole polygon)
get the line between this point and next/last points
extends the line in the direction of the center by 'edge'
get the center of the smallest line segment between these two lines, this is the 'edge' vertex

First webm is the result from 'last and next' being form the whole polygon
Second webm is the result from 'last and next' being from the convex polygon currently being tested

It would be quite fortunate if I can just mix those two 'where the last/next' comes from and the issues just fix themselves, but I doubt that will happen

Wew. Best part is I'm currently moving to another physical location so my ID is gonna change anyway. So long faggot.

...

Holy shit, I thought I was the only man using that shit on the entire planet.

Quick video testing Hexen texture atlas and some sound effects I made a few days ago messing around with Audacity. The music isn't mine though. Also redid my depth fog and colour clamping.


I wasn't planning on it but you got me thinking about it now. The way I do sectors should work with some modifications, it would mostly be modifying walls to have multiple sectors per face and an arbitrary number of sections to them. I'll look into it.


Yeah, I use portals now.


Here's to hoping you don't suffer too much.

name one thing sexier than a spline mesh

ooooh yeah this pleases me to no end
wish I had time to replay the old Raven games


post video of walking on it and pushing peasants off


I approve of this

holy filtering

What language is that?

C# ?

we did it reddit

I can't find any info on this anymore, maybe one of you lads remember. Are the Unity devs vocal SJWs or am I misremembering?
The new Godot version seems pretty promising but I'm not sure if it justifies a switch just yet. The devs of it seem a lot more tolerable, so it got me thinking.

they did a few ads about female pajeets making games a year or two ago, but that's about it
the real crime is that they keep shilling shit that steers towards walking simulators

They're no worse than the game industry. But, yes, they do set money aside to hire college graduates to write shitty blog posts.

Fair enough. Thanks.

Do You Want to Make Your Studio More Diverse?
archive.is/elG5I
Oculus & Unity Launch Pad Workshops: Promoting Diversity of Thought in VR [AKA Women in VR]
archive.is/E36GI
Unity & OGA Partner to Increase Speaker Diversity
blogs.unity3d.com/2016/08/12/unity-oga-partner-to-increase-speaker-diversity/

I knew there was a reason Unity was shit

HURRRGH

...

...

Yeah, that's C#.

it just needs to hold 8 bytes of data with absolutely no functions

So it looks like I'd have to upgrade my version of VS from 2010 just to have it fucking load the project right. Fucking Microsoft. It creates the project fine, but fails to reload it because it creates the initial dll hooks against the GAC which somehow works, but it's a behavior on load that fails. Fuck them.

So I'm not doing the jam after all, but how do these look?

It doesn't need to be that small, though. Honestly, it would make more sense to just make it as a (possibly static) method that takes two ints as parameters, and if you're so concerned about performance, pass them by ref.

It's at an awkward place where he's optimizing something that will never bottleneck, and at the same time, not optimizing it as much as he should. Eg, QueryType enum is still backed by an int, and could be reduced to a byte, so you'd only pass 5 bytes worth of data.

Use MonoDevelop.

I would, except I don't see why I have to jump through all these hoops just to get it to run

I didn't realize that and I thought that MonoGame and MonoDevelop was already a bitch to install on Linux. This is just painful. If you don't mind using Vim, you can use OmniSharp.
github.com/OmniSharp/omnisharp-vim

It's not too bad to be a struct if you know what you're doing. Keeping data on the stack is fast and you won't need to make as many allocations, not to mention accessing the data is fast since you don't have to look up the address in RAM if it's not cached. All I'd suggest is turning the enum to a byte and seeing if the int could be cut down to a short/ushort or even byte/sbyte. Obviously this is all splitting hairs, though.

how can automatic memory management be faster than manual memory management

I'll give Unity credit here, I'm glad they're attempting to implement Data Oriented Design in the form of their job system. It looks really interesting and start an anti-pajeet paradigm shift as one of the biggest products to be pushing it. I want to try it, but I think it's not released yet and unfortunately it will mean entire rewrites for projects that are partially rewritten. Maybe for a new small game when I get a chance, oh well.

Just guessing here, but doing all of it in one go might be faster than whenever it is needed, much like the principle behind DOP instead of OOP. The problem is that while it's faster, as an over exaggerated example it'd be a full 0.5 seconds stutter instead of 550 times 0.001 second. It'd be 0.05 seconds faster, sure, but also a lot more noticeable.

Well since Godot foss there is the possibility that they eventually squander their budget on womyn and tranny outreach.


They probably have/had someone developing that feature full time, so of course it will be better than the stuff someone just slaps together. They say generally so its not wrong when someone implements their own solution with their requirements in mind, while also knowing what theyre doing.

Unless Godot adopts the Contributor Covenant Code of Conduct.

I'd assume the same way that the compiler can optimize things. Ideally you'd never invoke GC.Collect() manually, except when testing things, and never rely on it for actual release code.

I refer to this as the canonical way to work with GC in C#, but just don't make a fuckton of objects in your main loop or update ticks and you'll probably be fine
stackoverflow.com/questions/538060/

Yeah I'm not gonna hedge any bets that Linietsky is a cuck
godotengine.org/article/godot-2-1-rc1-out

Funny you mention that, the same page of documentation suggests calling it every 30 frames on mobile.

I've heard automatic GC can be faster under highly specific circumstances, but other than that it's bullshit.

What is the file IO equivalent of SFML?

I want a cross platform library/framework for managing files.

I will just throw out any duplicate vertices. That should fix everything. I will probably organize the polygons into triangle fans after compilation since that makes more sense for walls and floors.


I can use mipmaps but I haven't turned them on yet. So, it's just plain nearest neighbor. Here it is with mipmapping and bilinear filtering on.

the roof still bothers me but that's a lot better

Sorry, there was actually a bug in my code where it was only filtering the first mip level, this is it with maximum bilinear filtering. I still need to add more flags to get the right balance between the grainy filtering and blurry filtering.

Is there a name for this art style that you see in anime all the fucking time?

What kind of artstyle?

The generic red eyes, and ancient ruins thing

anisotrophic?

I have turned on the feature in my code, and it doesn't seem to be making any visual difference here. I'm not sure why.

That's more of a motif than an art style.

You need to go Project > [project name] Properties… > Application > Target Framework

I haven't seen that style since Mega Man Legends. Best guess is that it's a combination of design elements from the style of late-80s to early-90s scifi and cyberpunk anime that were popular at the time, and Laputa's town, ruin, and robot aesthetics.
themmnetwork.com/blog/2011/05/30/the-unseen-worlds-of-mega-man-legends

...

Oh don't you worry, I've stumbled through this at length. Turns out that VS is keyed to certain released of the .NET framework. Eg, if you run VS2010, you have compatibility up to .NET 4.0. Basically I would HAVE to upgrade VS to 2012 for it to see the installed frameworks I have. It refuses to load the project without bumping it down to 4.0, even though it ran just fucking fine when I created it.


I guess it is fair to say that Megaman pioneered that style … the first game that comes to mind was yeah, MML and MMZ series. Code Geass also had a bit of a nod to that with Suzaku's Knight of Zero outfit (eg it looked pretty similar to Dr. Weil's stuff), and a few other JRPGs have that going on, too.

Why are secret society stories always the most interesting. Pure fantasy is always boring as fuck to me unless I just sit around reading all the lore. Even something like Thief 2 or Stalker, which are somewhat grounded, never ropes me in as much as VTMB. The closest reason I've seen is vid related, the guy mentions "Most of it should be grounded while the rest should be crazy". Although, there's something off about that.

Can I make a game using no baked lighting? Or is tech still not there?

I believe they have the renown mike acton working on implementing DoD.
He's also a no BS type of guy, as far as I can tell from his talks (especially the Q&A sections u can tell), and will defintely see to having an effect of implement better practices engine wide (or so I'm hoping).


There are a select few members of the dev teams that are vocal SJWs, but almost all of them seem to be normal people into engine dev and vidya.
There's also the current CEO who is someone to keep an eye on, as they had previously been involved with EA as their CEO, and since he got instantiated as CEO there's been a few "diversity" articles which is suspicious.

Its possible in lower poly enviroments using traditional rendering methods, there might be some stuff in academic papers that show rendering methods that may let you do it easier, but it's generally not done just because unless every single light you have is dynamic, it's a little wasteful.

In unity I've been using dynamic lighting almost exclusively.
Which is realtime GI and general lighting.
Though, it's only really feasible for consoles and PCs with dedicated video cards.

if your game is like 3 spheres and 4 triangles hell yeah

Just for reference I have scenes with 500k+ polys and multiple dynamic lights; with almost no overhead for lighting.
Not sure what everyone else is doing wrong. However, I do do aggressive custom culling (frustum, occlusion), and custom LODs.
Also I can't do baked lighting due to destructible everything.

What's your game?

Improving code quality is comfy.

I just dived straight into the world of CG, and to my surprise this is way easier than I thought, at least it's easier than drawing. I have a problem tho, how do I into public domain textures and maybe normal maps and similars? Is there such a thing? High quality stuff are behind paywalls or at least need login which is kinda fishy to my tastes.

Hey don't bully I just started.

Wow, I actually finished my jump animations. Gonna start plugging time back into coding immediately, also the trip laser may have to wait, I have to get the code working to have jump animations working properly.

Spitball: so I can do a trace when a player collides with a wall, if they are close enough then start a 1 second timer that tracks input. If the player presses jump again, wall jump. I can throw this in playertick(), which is the tick function that is tracking player input but then a timer isn't necessary so this might be what I go with. With these jumping animations though, if I grab like the 2nd or 4th frame it'll look kind goofy but it could work. Luckily I wouldn't have to worry about flipping animations, the sprite animation manager handles what sprites should be displayed.

Goals:

shitfuck sorry, thought I had spoilered these.

textures.com/
poliigon.com/
These two have some OK free shit, I don't know any others since I am no artist. Pic related is made of stuff on textures.com.

hello sigma dev
Yeah, I was talking about those two when I said about fishy logins and paywalls, I just wanted to know if there was any other good free alternatives before joining these with my fake emails. But, hey, thanks for the intention.

Ok, I just lurked the wiki and found out about this texturemaker program, going to fiddle with it and see what happens.

I fixed the bug that was preventing anisotropic filtering from taking effect in the OpenGL version, so now my walls are being filtered properly.


Only alternative that I was able to find was an old mirror of the free textures on graphtallica.com, I had to rip them out of a doom .wad. Not really useful to you though since it looks like you want to do PBR and textures bigger than 256x256.

I had to review my whole sprite system, I forgot what did what. lol

So the code I have iterates through all .int files and looks for this:
"MetaClass=RPG_Game_dev.SpriteParser"
then clears out any space characters from the first object line's Name= and Desc=, checking to see if the first Name could be put into a .IsA(actor(Name) check, which would basically be asking if the Name matches the name of the actor, then it sets up an array for each direction the actor should have for sprites (either 1 or 8) so it can assign sprites into specific spots.

It puts these together from Stand0_ as Stand0_0, Stand0_1, etc. then Stand1_x, Walkx_x, so on so forth.

Enjoy a screenshot from in-game import, working on these right now.

What I do is have a method called CreateFrames() that takes a handful of parameters, then generates a bunch of texture coordinate rectangles.
hastebin.com/carinamehu.cs

Then, I store it in a Dict and a Dict and wrap those in a helper class that lets me index it by ID or name. Eg, I can do Content.Frames["player_left"] to get the rectangle collection, then pass it through GetFrame to get the actual frame I'm on. I store my data for it in pic related

...

Just use GL_NEAREST or something.

Nigga that line's Unityscript code, you're using C#. Replace it with
Vector3[] vertices = new Vector3[4];

When you declare a variable, you do it in the form of [access modifiers] [type] [name] = [declaration]

"var" is a shorthand thing that the compiler will adjust to the proper type at compile time. Pick one:

thanks

While it's true that GL_NEAREST makes anisotropic filtering unnecessary, I have actually added options for all of the kinds of texture filtering that OpenGL offers. So, you can choose between all of them using the in-game console.

What's a texture coordinate rectangle? That'd help me understand a couple of things.

A texture mapped polygon will contain the vertices of the polygon, as well as texture coordinates that tell the GPU how to map the texture onto the polygon.

Most lower level frameworks (eg OpenGL or SFML) let you define which texture to draw, in addition to a subregion of that texture, expressed as a rectangular area.

Drawing from a spritesheet is going to be much faster than from individual textures, so people usually set up all their animations in a series, contained in one texture. This is because sending data to or from the graphics card is very slow, so if you can do it in one operation, it's much faster. Then, since the texture never changes, you just change which part you draw, again and again.

If you're using an engine, it might take care of this for you, but it's never a good idea to split an animation across 50 files

This is why I might be going to Godot, but I have other priorities first. Gotta buy a car before I do that, because I'll need a whole new computer. This laptop is dicks.

HELP

Isn't that what sprite.setTextureRect is for?
Why not? Just don't draw any sprite that's more than (spritewidth + screenwidth) * 0.5 away from the camera center.

Not really, no. That just defines what region of the source texture to use. I'm talking about clipping my draw call to only appear on a certain part of the target.

Eg, imagine a text box and I don't want to have it flood off into the rest of the screen

I'm working on a mechanic framework so it's more like a massive pile of black triangles at this point than a game.
Once I finish making editor tools for each system I'll start showing it more, and will be able to show off mechanics without hand coding each variant utilizing the various systems.
Yup

...

Yes

www.open.gl

how do i level/story design
i've more or less got a framework for everything in my game but now i have no clue what to do with it
where the idea guys at

Maybe read some literature from such a time period that your game depicts and maybe make a game that tells one of those stories or borrows some elements. Like the King Arthur and the knights of the round table, but if you made it into a video game? I can't think of old stories that depict such a place other than the knights of the round table, robin hood, and william tell, I don't know those stories seem to be set in a world that has the same aesthetic as your game

I was also having trouble level designing, mostly because I cannot use 3D modeling programs or draw for shit so I didn't know how to put my ideas down and give them to my artist.

I ended up buying a bunch of clay and making quick prototypes of levels until I got something that worked.

I used Godot on a shitty laptop. Go into Project Settings, turn down the shadow resolutions, force vertex lighting and turn off HDR. This can double or even triple your framerate. Make sure to set up baked lightmaps later on, too.

Think it'd work on this?

Also have a screenshot of sprites. Almost done exporting these, I've been able to review them in Unreal and holy god is it very clear this is my first animation attempt. I'm not surprised, but at least I have something I can work with now and I can easily see the animation flaws now.

www.blender3darchitect.com
they don't host anything but link to good shit, seriously, last time i went up to page six and ended with about 10gb of pure textures and this book on how to make my own

The default shadow res is 4K in most cases, so setting it lower will definitely have an impact on low power computers. The HDR is the single biggest performance gain, though. It doesn't seem to be very efficient. If it comes up, Filmic tone mapping seems to tank the FPS a good deal compared to Linear.

Your laptop is better than the one I used. What GPU does it have?

Building on this, combine story elements. Take some stuff from Knights of the Round Table, mix it with Romance of the Three Kingdoms, Warring States, Joan of Arc. Check other countries fairy tales and legends too.

Forget Beam!

Anyone happen to know if Löve uses a different name for it or a different solution for this? Otherwise I'm not sure why it's missing. OpenGL documentation says it's been around since 1.3.

...

...

I'm a dipshit, is this what you're asking about?

best practice for posting specs is installing piriform's speccy and taking a screenshot

...

At least it's not every day that you get quads.

yes, that was quite nice

reminder that the speecy server was hijacked and was serving malware for a while despite being hosted in the avast servers or something like that, i don't remember the full details

(Checked)

Stayed up all night knocking out these sprites, going to bed now so I can wake up later and get started on patching code so I can recompile and get working on animation code again. After that, I'm gonna work on a class puppeting system BEFORE getting on the RPG system. It is very important that I subclass a bunch of stuff so I have a class that is only responsible for receiving player input and being attached to a camera.

Who decided that was acceptable?

Most likely pic related.

Nice to see it in action. Have you given any thought to 3d model/anim? Sprites look like a monumental amount of work.

JUST LIKE MAKE GAME

trips

...

shit attempt, unworthy of those numbers

Time spent getting repeating digits in Lebanese finger dancing forums is time not spent making game

Is that guy retarded or can you actually have a Unity project compile in Monogame?

...

Jacket almost done, need to add details and make the weight painting

day time system works. using a curve to set light intensities/ambient color. basically shadows during the daytime are totally visible, while night time is pitch black unless you have a torch
1 ingame day is half an hour in real life, not sure if too much or not
2 guard patrols along the wall
fucked with splines to make roads
it felt like a lot more progress than it actually is

???

and after 18:00 they go sit in the house for the entire night, until 6 am

A pajeet would never do this, right?

...

He's retarded. If you look now the project has been removed.
I played and downloaded his game and it was actually decent.

Most games have it so an in-game day is 20-30 minutes so you're probably fine.

Turns out just needed a few minor tweaks to get this working. I'll probably end up going with the limitation of not being able to have "windows" in or out of overlapping sectors and that they'll need to not share a wall vertex.

Ah, figures.

Okay so I am going to spend a few days doing the jam, I've decided. Here's the tldr of what happened

>Because of retarded design choices by MS, VS2010 cannot run .NET 4.5 even if the dlls are there

So I'm just using FNA which is okay as long as the submitted game is actually built with Monogame and the source is posted (they're so similar, and I'm keeping my scope minimal so there won't be any compatibility problems, I won't even have a shader).

Anyways, as for my game itself, it will be a simple puzzle-platformer, and you are able to scan and change any object you see, including enemies, coins, rocks, or the exit itself

I've been using MonoGame 3.6 just straight out the box with no extensions for the jam and have had zero fuckery so far.
>Win 8.1, VS 2017 Community
Already have quite a bit of GUI functionality and a basically finished character controller with some basic enemies with a simple AI.
I'm having lots of fun and I like your sprites, user. Especially the snakes

What the fuck.

Yes, but I need a better computer. My USB ports keep turning off which forces a reset from me, using an onboard graphics card that can barely even run Doom 3, plus my hard drive is sketchy as hell right now. All of my progress is backed up, but I really do want to do other things with this project and 3D would save A LOT of time. This could have been done like 2 months ago, I'm just an fgt.

I remembered that a struct will always allocate the memory it needs on creation, unlike a class (which is just a ref, with an integer sized pointer, internally)

So I wanted to see if a struct could contain itself. Somewhat surprisingly, it caught it as an self referential error at design time. But then I saw I could use a generic to sort of do it that way.

Are you using Windows 10?

Another quick test, everything seems to work fine.


I thought you were the user making the boulder dash game for a moment and got confused as to why your sprites changed. You change one type of object into another?

Oh yeah, that's me too. I just figured that since I was off for a week, I can participate in a well-timed jam to prototype a few engine ideas. I'm using an NES palette, and a few psuedo-restrictions to keep it all looking coherent. Hopefully.

There was this old DOS game I played on a shareware demo CD, I can only remember its name as "fgodmom". Basically the same premise: You were chased by crabs, and had to collect silver coins, but you could scan and replace any object, such as walls or creatures. It was kind of cool, I guess. But it was the first thing that popped into my head with the theme. Judges are suckers for pixelshit platformers, aren't they?

On a scale of 1 to lawsuit, how much does this infringe on Kirby's design?

Keep it limb-free and you'll be fine.

it actually reminds me more of the pink bob-ombs from Mario 64 than it does Kirbo.

cool.

here's my super-scary way of doing it (that I have actually used, and it works).

basically, store a list of pointers.

each raw-ass pointer is the memory location of the value you would like to check. Could be anything. float, bool, int are all that I implemented. Check for mission == success by checking each value pointed at by the pointer with a target value. You don't need to know FUCK about the respective data structures they are organized in. Just hope that they don't get released from memory until the mission is finished! Store the pointer and it's target value in a struct. Store a list of targets (the structs). Check them when necessary.

I do the same thing for animating values. Store a struct containing a pointer to the variable to animate and the animation parameters; easing formula type (for a switch statement), start, end, step, speed. I use it for floats, vec3 and vec4 types. The animation value could be anything store anywhere. a color, a coordinate, a volume level, a shader uniform, whatever. Don't care. I just slam in the new value via the ptr every frame.

Downside: need to be very aware of when data is free'ed, deleted, or otherwise made invalid. Because, you know, that big-ass list of pointers will be updating what they point at every frame..

I didn't think this was even possible but somehow there's less useful shit than before.

Alright, art is done, I think.

7, but I've dropped the damn thing like 3 times. One of inserts inside the port itself was tipped out, so it can hold a USB male end but no pins are touching so no functionality. Ever since then, it has only gotten worse.

Animations look pretty decent, but OBS tanks my fps, even if it's idling, and even after I close OBS out, the application stays slow (and using no CPU). Not sure why.

What engine/framework are you using?

Seems like the animations would benefit from following a frame pattern of something like 1-2-3-4-3-2-1-2-3-4… , to return to it's original state before looping, whereas yours loop immediately.. For example, the bat wings take a while to flap down but instantly flap up, it doesn't make much sense. Same with the jellow, it bounces down slowly and up instantly.

I have a craptop, and previously XNA (which worked just fine).

With SFML, my game could use 0% cpu, and when I launch OBS, my game's cpu usage spikes to 25%, even if I close OBS afterwards, while also making the drawing drop to like 12 fps, regardless of what OBS is actually doing.

With FNA, it's the same issue, but it doesn't kill my game's cpu usage

I'm currently working on laying out the medical system of a rougelike. Basically when you get injured you receive a "wound" with debuffs rather than loosing HP, and you're given a guess by your player character as to what they think is wrong. I just realized how funny it would be to model the Dunning–Kruger effect and impostor syndrome with this.

With a medical knowledge of 1 you get a really unsure and frequently wrong suggestion.
With a medical knowledge of 3 you get a very confident and somewhat wrong diagnosis.
With a medical knowledge of 5 you are unsure but more frequently correct.
With a medical knowledge of 7 you are both sure and correct

I'm not going to do this, I'm just going to give the diagnosis with the possibility of it being wrong depending on medical knowledge, but it's a fun idea.

Any system which doesn't provide perfect, numerical information to the player is fantastic to design around.

Imagine a roguelike that had numerical HP but it didn't display them to you, only in chunks of say, 25%. You'd know if you're critical or full HP, but between you'd have to make a guess and hope you don't waste supplies.

The numerical components are in the debuffs, so a sprained ankle will affect speed in a measurable and precise amount communicated to the player, but you aren't going to die from it. It's not that HP isn't being shown to the player, it's that it doesn't exist, only the debuffs exists. The treatment/application of resources is based on how much you want to remove a certain debuff. I could put a number to the amount of blood in the player, but there are still problems. Not all injuries affect the total blood in the player, you start experiencing problems far before you're sucked dry of blood, and some injuries don't result in any blood loss but can still kill or cripple the player.
Resource management is still a problem, I will agree. Would it work to tag injuries with "Mild", "Serious", or "Critical" to illustrate imminent danger? So while you can limp around for a while with a sprained ankle you would want to apply a tourniquet as soon as possible to a giant gash on your leg.

If this isn't a joke, turn Next into an S and value into a T, that type is almost useless the way it is.

I figured out where the bug is happening in my map editor, didn't stop it yet though. Pretty busy…

...

whatever, they are free textures. I'm not sure what the final game will look like, I like these a lot though

I'm not talking about the textures you silly billy.

sorry, I am just tired and dont understand you

I'm taking apart the player class groups I have (Pawn->PlayerPawn->RPG_PlatformPawn->RPG_PlayerCharacters->RPG_LionPlatformer) and moving any player-related functions for input tracking into a single class (Pawn->PlayerPawn->RPG_Controller) who can find puppet pawns (RPG_Puppet, RPG_CombatPuppet) and control them. This way if the pawn dies, the player wont just get a game over instance which would affect player input (triangle on a PlayStation controller would restart the map, many functions trigger). The Controller is invulnerable and untargetable, invisible, and doesn't collide with actors or the world. The camera will move the Controller around, though I have considered just making it so the Controller will spawn in and just not move.

After this is done and working, I need to make a mutator and a menu system. The mutator gives the player an item which will allow them to open the menu system and navigate around different menus, though I am sure I can just bind the menu system to player input somehow. Unreal has a menu that can be toggled by pressing the delete key, so I'll have to ask around about that one.

Is this how the animator in Unity is supposed to look?
This is messy as shit but everywhere I look nobody seems to have a problem with it

Since I have not made any progress since demo day I'm releasing my shit as public domain so maybe one day someone can unfuck my pajeet code
github.com/ostoru/vcsona

you use state machines and blend trees to make it look slightly less retarded

sheathable weapons/shields
player can toggle it by pressing R, or attacking/blocking
ai do it if they do/don't have an enemy
two handed weapons go on the back if the other hand is empty, otherwise they go on the hips
one handed weapons always go on the hips, shields always go on the back

Why is my texture all mangled when i try to display it isometrical?

Did you change the running animation? Looks better
All animations in general seem much better

Because it's already isometrical in the first screenshot?

all my animations right now are from mixamo's site, since they're free and have better retargeting than unity's humanoid system
i did add walking animations for the AI

better than Kingdom Come

How's networking in Godot? Any experiences? Is it easier than in Unreal4 / Unity?

anyone got a torrent for probuilder advanced?
i want to switch over to it for my terrain, easier than going into blender every time i want to make a minor change
but the free version doesn't have smoothing/subdivision/detaching/triangulation
kind of need those for anything other than a blocky mess

Well how do you herd the tiles into a proper formation then?

don't resize/rotate them, displace them by a vector or something

The same way you draw any grid really, with an offset per column and row. What those offsets are depends on how your tile data is formatted.

TileWidth = 32;TileHeight = 16;for (Y = 0; Y < 10; Y++) for (X = 0; X < 10; X++) { drawX = x + 0.5 * X * TileWidth + 0.5 * Y * TileWidth; drawY = y + X * TileHeight * -0.5 + 0.5 * Y * TileHeight; //draw sprite at [drawX, drawY]}
This exact code (with an actual sprite drawing function obviously) created pic 1 using pic 2 as sprite. Yes, I know TileWidth and TileHeight aren't the actual width and height of the tile. It's to not get double outlines.

well this works
not as pretty as the old terrain, but it's a lot easier to make and integrate

I was asking myself where the tactical waifu shooter was.

I don't think I want one anymore

user, how are you supposed to protect your waifus from foreign invaders?

oh no no no

Need some opinions, leave it with long sleeve or make it short and add a bit of armor in the arms?

it keeps happening

...

short sleeve is for fags and numales

Short sleeve looks silly, like a first step toward something a Jap would design.

blogs.unity3d.com/2018/02/13/introducing-2d-game-kit-learn-unity-with-drag-and-drop/
Really makes me think

yeah, im keeping the long sleeves, my bro wanted the armor on the arms for some weird movement options and combat, but i can just make actual armor for that later

Are they just trying to make Valentine's day an even more miserable experience?

Arms bending alright, now to weight paint the rest and then i can do the scarf

Anyone know about multi-threading and opengl? As far as I understand you can't make draw calls in different threads. But what's the work around?

OH GOD MY LIFE IS A LIE

Aside from the graphics, gameplay, and design being absolute shit, I think it might be a good way to learn to make custom editors

Every fucking politician is a feminist

and are feminists not anti-establishment?

No, they pretend to
Feminism is the establishment, unless you live in the middle east or china

...

Well looking a bit into my options for organizing data, I read that I have two options for what I want to do.
Either using combined keys or nested maps, and I think using a nested map might be the best option. The other option seems really dirty. Although maybe that itself isn't the best option either, because I think the other option would let me contain all the information needed for each block easier (should block drop item/should block spawn block/should block spawn entity/quantity to drop or spawn/etc).

...

That's…

Well, using directives have to go at the start of the file before anything else, and then you usually do a namespace then class. That's the usual object hierarchy in C# in general.

I have no idea why it's fucked on the first line, try reopening Unity?

I think this is actually good for Unity.
It gives you an idea of what you can build and lets you play around with it and i assume that you can then dig into it to see how it works.
In their position going with a punk nigress is probably also best, since retards will now give them free advertising and rave how they just like made a game with a stronk female.
It probably doesnt help anyone thats currently using it though and they might fill their community with people that the pandering works on, but that doesnt matter to them since they make money, when people buy the pro version and buy asset packs.

look at it this way though
unity is already the biggest free engine, pretty much no competition aside from the niche that wants foss
they don't need more people, they need to fix the shit they already have so that they don't have the reputation of that engine that easily shits out garbage at incredibly hihg speeds

They already have that reputation and changing that would not make them more money.
With Unreal they can justify actually improving the engine, since they make money from games making money and their asset store is shit.
With Unity everything that they add, that was previously solved by something on the asset store directly cuts into their revenue. Unity needs people to use the engine and buy stuff from the asset store and doesnt need to give a shit about what then happens with the games, as long as new peolpe stream in. Maybe Godot will take some users off them, but lets be realistic it will never get as many users, since the asset store will never be as good and both of them have no buy-in.

I might be biased though since i prefer ue4 to unity and im currently playing around with Godot.

...

trips name my town

faggot's vill

Not Homoville

...

Wholesometown

what have you done, user?

meant for

tumblrville

I hope it's not too late to thank you

Devalda

>XNA MonoGame FNA has a Matrix object for 3d transforms, but it's 16 floats (64 bytes!) in size, so it's unwieldly to pass around or attach to other objects

Oh well, at least the source is there

Wikifags, the Texture Maker 3.0 is broken (Mega):
8agdg.wikidot.com/graphics

I had it downloaded. Updated with my own upload.

Oh SHIT thanks for that
Maybe it's time to finally properly scan these bois

Gonna add this to tuts on the graphics page as well

Why is making games so hard

You put a breakpoint on your rotation code, see if it happens, and then step through the rest of your code to see if it is reset the rotation somewhere after that.

That's because it's going to be calculated into a matrix eventually anyway

What would be the best way to determine what has been hit by a trace in UE4? I need to know if my swing hit world geometry, a physics actor (like a barrel), an enemy, etc. Right now I have a switch case that checks for various things and goes by exclusion (e.g. if it's not worldstatic and it isn't my enemy masterclass but it has physics enabled, it's a physics actor), but is there any better way in UE4? I haven't touched collision channels much, perhaps making everything its own channel (e.g. one for enemies) would be a better idea?

You can assign actors tags in ue4 and just check if an actor has the tag.
The tags are strings though so you will probably want to copy and paste.

aren't there also layers that determine what can get traced in the first place?
unity has that

need to bone the scarf now

hehe bone

is there a link between gamedev and being a massive faggot?

I don't see what's wrong here

...

YOU FUCKING NIGGER NEWFAG USE HOOKTUBE YOU PILE OF SHIT

Gee it's almost as if not spending time with other people and using most of your time at home alone in front of a PC is not healthy

No but there is a link between getting repeating digits and Patrick Bateman's personal happiness.

wew

It's made me a lot happier, at the very least. Time spent with other people to me was always just time spent not doing what I actually enjoy. It's why it took me years to overcome depression - because all the shrinks told me that I needed friends.

If you're prone to this, don't make a videogame. Especially when disconnected from the rest of society to work on it alone. Leave that to professional NEETs.

FULLY BONED

Now need to test on Unity, the structure is different from the previous model, so i need to check

Making a game requires tough skin and a fuckton of willpower and self-discipline, specially if you work at home and you have a huge amount of games waiting for you right there all the time

Not defending that guy, those who quit are weak

I think this should be called "fun prison" or something like that.

Yeah sadly the tags are still the old strings, rather than the new GameplayTag system, on top of forcing you to add said tag to every single actor in the scene, so that's not really an option as it would be a headache.


Yeah that's the collision/trace channels, guess I'll just use those. Don't really need it to filter hits but I can determine what has been hit by the preset I think. Can't really think of a better system right now.

behold, Castle Devalda
no idea how to make this look populated. there's not enough space for more than 1 road, and at the same time it seems like there's too much free space
plus i only have 3 houses that i just copy paste

This probably shouldn't be completely flat.
Make it on the side of a mountain, put the castle on the very top, and make the road go from the main gate to the castle on the other side.

it isn't completely flat
some parts of the wall are slightly elevated
the side near the castle is taller too, you can see it in the towers that split the river

I still don't think the layout makes much sense either way

Those who write their own stuff in C/C++ what compiler do you use?

I think the residential district should be more cramped and crowded. I hate to say it, but the low amount of houses kind of makes it Skyrim-tier.

Happy Valentine's Day, /agdg/!

A-at least I have anime girls.

gcc

Reminder that the one user married his digital waifu

pocc for windows, gcc for linux

thinking about branching out to cl and clang but still not really feeling like it yet

I'm not that far gone yet.

better?
that's about 50 mini houses for the peasants
i think it's around 2-3 times as much living space as whiterun
i can still sacrifice more of the wall for another 20 or so houses, even more if i put these shacks near the castle

looks better but the houses are too uniform and the layout looks a little too organized, maybe a big improvement would be to put some dirty streets in instead of that grass

this is just the layout, i'm probably gonna combine some of the shacks into a 2x1 thing that serves as stores
there's definitely gonna be dirt/road there but i need something easier to paint terrain with (like a splat map that actually works), otherwise i'm gonna wait till i'm done with all the building placement

It just does not look natural at all.
Work from pictures, look at old german villages, build from that.

Also having a mostly flat plane as ground makes it generally not very believable.

it looks artificial with the shacks all lined up neatly.
have a look at gothic 1&2 for some of the best slums you'll ever see in videogames.

better?

Make the ground more wavy and hill-like. Maybe it's a rolling flat plane with a 2 degree uphill incline to the west. Start with a terrain, build it, give it character.

You'll get out exactly as much effort as you put into it. I've found this out with spritework, and modeling or sculpting a little village is no different.

mingw64 on windows
gcc for linux
borland asm & c for msdos

general opinion of gcc is not good. code generation is not great and there are some required flags to use (-fno-strict-aliasing for instance).

I'll probably move to vc++

it's not flat though

can you make a warehouse next to a dock, kinda an active port feel? like have buildings that do something and then surround them with the hovels of people who work there.

go onto a campus or into any silicon valley company and say something even slightly critical of feminism. See what happens.

Will neural network speech synthesizers bring about cheap fully voiced indie games and rpgs? Video sort of related. Few months ago I remember listening to better examples which I can't find now, where I couldn't outright recognize which example is recorded and which is synthesized, though they still were spoken in an emotionless neutral tone. But I guess it's only a matter of time before they can synthesize emotion in the voice too.

CD-quality audio is usually played at 44.1k samples per second at 16 bits per sample. One second of uncompressed, raw audio is 88,200 bytes in size, or the equivalent of 150x150 bitmap image using 32 bits (RGBA).

Storage and playback isn't an issue, it's the generation, but I'm certain we'll get there in time. Remember that there was those facecam videos where it takes an actor's facial movements and transposes them onto a static image to create new motion, and they did it with Trump and others? Audio has to be easier to mess with than visual.

alright, added very slight elevation, pretty much everything had to be rotated
would not want to live here/10

artist here, pls give me a job.

can u post the wireframe?

Holy fuck, I miss SFML badly

nigger no, use bendy bones, same stuff but more manageable.

PSST