Amateur Gamedev General ~ /agdg/ + /vm/

finally got the inventory working as intended with as little spaghetti as possible
also got a little editor extension to make create new items faster

seriously though

I could've sworn that I had seen that inventory somewhere and was going to call you out for using unity store assets. Turns out I was remembering this. Hopefully that's a compliment.

don't be silly, i never managed to publish that to the asset store, unity had fairly jewish standards including such things as "price too low"

checked

...

Stick dick in optical drive

Tried already

Repost from the old topic's end. I actually made progress since but am too lazy to record better footage. Health bar and name displays mostly.

Quads confirm that you can't have a competitive advantage

What is the fucking rush with these threads lately.

Not only did he make a new thread on page 10, he didn't even link to the new thread in the old one.

And Teal is dismissing reports for this as well.
What the fuck is going on?

...

GODDAMNIT I JUST WANT TO MAKE A GAME REEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

Yes, and?

Goddamn forgot to remove this flag

How the fug do I get a evenly baked texture in Blender? It pisses me off my barrel model is brighter at the top and darker at the bottom.

the mesh you're using to bake the texture should be exactly at the same position as well as size/shape as the lp mesh.

But it is already like that.

Portals, I guess

Dear Dairy…
After studying both a kinematic controller from Godot 2.1 and 3.0, i finally managed to create my own that i partially understand. What's great is that i can tweaks it easely to make it either smoother, or snappier.

Fuck dot product for slope detection i manage to do 'em without mathemagics.

Pixelshit placeholders are better than nothing, I guess.

in fact I'd say pixelshit is better than solid color because it holds more info about what you'd like something to look like than that. It carries an idea that can be reimplemented later.

Do you have a light source? If so, maybe it's being included in the bake. Try removing it. If everything gets black, try setting emmisive to 100% in the material settings.

Nope, I used the environment lightning for that, setting it to 1 and to either "White" or "Sky Color" is not giving the result that I want. If I set the material emission to 1 then nothing gets baked on it because it is 100% white now, no details are visible.

Actually I think I got it fixed now. so I have to set the Enivornment Lightning to a bit less than 1 and "White". AO set to 1 and "multiply". And yeah I did have a light source I forget it was there so I removed it. It's better now but some details are not visible or not rendered at all, oh well I will just fiddle around the lightning settings and see if it gets any better.

all mechanics and systems and 0% implementation

Why do so many games use Lua as a modding/scripting language? I was looking at using it in mine since it's so popular and there must be some merit to it, but the more I look into it the worse it looks. Is it just implementation simplicity? There are VMs for all these great scripting languages that you can leverage. This language is complete trash and I don't see why anyone would choose to use it.

is it literally ( say x -> do y ) and as the says cascade the code exponentially degrades in performance?

Someone has said this about every language ever conceived. LuaJIT is fast, readable (especially for small time modders) and easy to expose your APIs to.

tbh if i ever release a game, i'll just give the source code out, it's a lot less work than having to implement modding tools

People might make fun of it, though.

For a completely unreadable abomination, there's something aesthetically pleasing about the way it looks, with the line of semicolons and curly braces and even spacing.

so how does one go about making a scene/environment?
i refuse to use unity's terrain tool, it's garbage and a waste of time
also, would an rpg work without a system to keep track of quests? i need an excuse to do less work

my wife asked a pivotal question: do the in-hand items disappear once you equip another in its place?

sorry about Historia, my dude

well yeah, where else would it go?
i just Instantiate/destroy items as they get equipped/unequipped
if you drag an axe in the slot where you have a shield equipped, the shield will replace the original slot of the axe and the axe will be equipped in the slot where the shield was

Tetrachrome dev used a custom marching cubes implementation for terrain and the Hammer editor with a custom importer for Unity. I think he uploaded the latter to github, but I wouldn't know how to find it aside from digging through our thread archives.

we fucked up. Now I see the icon reappearing in inventory

got too much used to the way it works in Diablo clones

He also has an importer from q3radiant and actually does some shit to import whole quake maps into unity, entities and all. I think this is before the quake maps are compiled, since he should be using unity's occlusion culling scheme.

accidentally deleted the wrong post because i forgot a link
anyways, this is what you can make with the quake editor

you can make the brush size smaller, but modelling detail in quake seems like a hassle

i think tetradev mostly got around it by using some kind of bumpmaps. i'm sure he had it posted on his tumblr or something, but i can't remember the name

I remember seeing some shader tech that randomly replaced the texture of the windows with eg lit rooms, and fully furnished ones

MC is great for this purpose but honestly tetradev's MC unit size was too large (too blocky), and he used tri-planar for texturing.
Best to use some type of projection for UVs, and get a procedural UV map you can paint in-engine for better fidelity.
Tri-planar is okay, but honestly isn't customizable enough (while not becoming a clusterfuck in the shader) for anything beyond procedural terrains.
Also on MC unit size if you do it offline I don't see why one wouldn't just use dual marching cubes (easy to implement with a reference, look at ogre frameworks dual MC source code for reference) so you can get varied unit size (voxel size, so detail where it's needed), and have fewer polys (dual Mc has a more efficient output for poly count); while making chunk transitions easier due to the nature of the dual MC spatial data structure providing easy access to the data u need for seams.

Game for ants :)

I'm doing a Mount&Blade/Chivalry/Mordhau style melee in my game. I'd like the control scheme to extend to everything that hits/chops/smashes/cuts, which includes fists.

Which brings me to this important question.

HOW THE FUCK DO I PUNCH DOWN

working on generating normals for my thing, the only issue is that I need to be able to correctly generate normals that face the correct direction. Right now three of the normals should be reversed.

Oh yeah hey shodanon it's me eschabro, been busy with a new job… With 10hr work days, so haven't had time to do wiki stuff or post much.
I've been lurking but with my time schedule I've got to dedicate my free time to developing.
So unforuntely I won't have time for wiki stuff anymore is what I'm saying; keep up the good work though man.
Just didn't want to leave it unsaid tbh

uppercut is down
overhead punch doesn't exist for sane people
but you could do it like in street fighter

you don't
top of the skull has the thickest bone

TETRADEV DID NOTHING WRONG

Yeah but design consistency, user.

I think it'd look stupid in first person. Yeah, I mean it sounds fun but it's also gotta look good both from first and third person since I'm sharing animation for them to save work during indev stage.

I need to edit "tetrachrome" into the bottom of one of those "video game iceberg" charts

JUUUDO CHOP

On a more serious note, while slamming down with the bottom of your fist may be less effective than a punch in a fight, it may be more useful for smashing shit in general, as it carries fewer risks of fracturing the bones in your knuckles/wrist.

...

Okay, I guess I'm gonna go with the simplest solution here.

I've figured out that its my winding order, so I have to fix that up…

Mecha progress is going good.

Now that I can derive the correct normals, I can use them to generate texture coordinates. The normals shown here are not normalized and are just for debugging.

You could use a terrain editor, but the only decent, recent ones I found is Terragen, which isn't FOSS, and Artifex, which is FOSS, but based on the OGRE engine. The other alternative I can think of is using a level editor like Quark, or going with .

What's there to lose with that, though? Bants only is a trouble if my work is quality, which it isn't.
For example, I have no clue how to convert a float to an int, which resulted in this yandev-tier if if if if if if if if if if storm.
And I completely forgot a 0.12 check, go figure.

I'm not a good programmer, and there's guaranteed to be a better way to do it anyway. Any incoming negative input is likely to tell me a better way to do it, even if it's spiced with "faggot" and "dumbass".

The equation for converting pitch into dashspeed is f(x) = 45 - (250*x). So, dashspeed = (int)(45.0f-(250.0f*pitch));

I also discovered SabreCSG, which is
a FOSS level/map editor for Unity.

ok, well it doesn't exactly match that because you skipped 0.12 in your chain of if statements, but that looks like a mistake.

but can it do curved surfaces :^)

I have begun work on deriving the textures from the polygons, but there are some bugs specifically when I resize the brushes. So, this needs to be sorted out.

well maybe for the design consistency something like a small jump with hands joined and doing an overhead slam

i'd vote for

Also, it will check every single if statement. Even if the velocity is 0, it will still make 12 checks. The way to fix this is to change them into else if statements (a switch/case construct is usually identical to the compiler, use whichever is easier for you)

Also, never check against exact fractional values with floats. They're subject to rounding point error, and even a simple fraction can fuck up comparisons. Floating point is accurate for whole numbers only. You should try to use math to handle it organically, see

As long as u use the right CSG operation with the correct shape then most defintely.
SabreCSG is pretty capable, but lacks runtime modifications.

here is a shot of the map editor now that it can derive basic texture coordinates. The next step will be to start exporting maps with these coordinates and loading them into the engine.

SabreCSG is the best one I found too.
I have it on my todo list to extend its capabilities with my realtime terrain modification framework (dual contouring approach on the GPU).

Oh, that makes sense. Thank you.

I just gave that as an example, but I should probably explain what I'm doing.
In this engine, decimal numbers aren't actually fractions, but are contracted versions of much larger numbers. For example, in trying to get the pitch here, looking straight up (the highest pitch) is -16384, or -0.25. Looking straight down is 16384 / 0.25.
I don't know what correlation the two numbers have together. It's retarded. But I'm not a programmer, so there might actually be a very good reason behind it and I'm actually the retard.
So what I'm trying to do is have a maneuver that throws you straight ahead in a lunge (the distance being what dashspeed controls, about 50 thrust), but the lunge gets less powerful the higher up you look so people don't look straight up and leap 50 feet into the sky. So I'm trying to have pitch act as a negative modifier to the dashspeed, in that the higher you look the less powerful dashspeed is–the problem being that applying -16384 to 50 is a bit of an unwieldy calculation.

Still, I managed to modify this equation quite a bit and got it to work. So thanks very much for that.

So i think i have a few insights of designs in games which i like and hate. Stuff like, where to focus your polish on a save system. What things to avoid and what things to implement from a users prospective. Nothing ground breaking but rather common sense "You'll want to this, this and this, here is why, and here is a game that does this."

Would AGDG be interested in some user posting a webM or MP4 in here and see if i have anything of value to say?

no

go for it

I'm curious, post it.

Yes and no.
Yes, there's a lot to discuss and debate when it comes to game design philosophy and this is the place for it.
No, I don't thing videos would be the way to go about it; at least in the beginning. It would be pretty much the equivalent to avatar- and tripfagging, since your voice would immediately make you recognizable. While I don't have that big of a problem with it, you can expect to be met with hostility from Anons who care more about the anonymity of this community than I do.
Recognition on imageboards is something you must earn in one way or another (think Yandev, Tetrachromedev, …).

I would say that you should make your arguments first, then we discuss/debate them together. And maybe there's room for infographics and -videos for /agdg/ after that.
I'd certainly be down for that. Choose one specific topic and let's hear what you have to say.

It'll get better as you go, I have to derive functions from relations of numbers a lot when programming so i've gotten a little bit of practice. I used desmos online graphing calculator to check my work. I can't find any good worksheets about this but I did learn how to do this in math class.

So it looks like it's using a signed short value, or Int16, which can vary between -32768 to +32767. For whatever reason, it's clamped half that value, or -16384 to +16384.

What you've done is instead normalized it from an integer value to a floating point unit value between 0..1 or -1..+1, which can be a good thing for checking eg percentages

So let's say you're looking up slightly, at a value of -3000.
Or put another way, you're looking up 18% of the way. That's barely moving your head upwards.

So let's say you have 50 thrust. You can apply it linearly, so that moving up one unit proportionally reduces it by a fraction of a percent. We also need to inverse the value - 18% reduction in thrusting power is equivalent to 82% actual thrust. To do this, you would just do:

Most languages work in radians rather than degrees. Another way to do non-linear falloff (eg so the most oomph comes from straight ahead and rapidly declines) is to plug your "thrustFactor" into something like pic related. So what you would do is take your 0..1 normalized value and multiply it by Pi, and use that as a value in eg Sin() or Cos() function

Give it a go, user.
The worst that could happen is people calling you a fag.

How does this look for trap tiles?
Did you think it was some kind of other object?

Basically, it will hold the object firm, until a release button is pressed, at which point it has a few frames to escape

Almost done the tiles, gettin close
Then I'll have to stop being a lazy fuck and actually program things

Okay, I have to level now. I've been putting off the art because I didn't know what direction I wanted to take the mod, but I know what I want now. Think of 'Blade Runner' meets 'The Matrix (1)' meets some of the excess and concepts of 'Warhammer 40k'. Gonna make a kick anim to celebrate, fuck jumping for now.

For some reason, there hasn't been a proper Paper Mario since TTYD. And I have no idea why.
It's not like it's a very difficult game to make. 2d characters in a 3d environment. Not many 3d character models to deal with or animate. Sounds simple enough. But if Nintendo's clearly not gonna deliver, I think I'll take a shot at making my own turn based rpg with "Paper" characters. I'm just not sure where to start.

What would be good, recommended resources or software to use in creating a game like Paper Mario. An actual 3d environment to explore, flat 2d characters, and a turn based battle system. Doesn't seem like much compared to more ambitious concepts and projects imo.

How much of a beginner are you? Do you know enough to know what game engine you want to use?

twitter.com/paperbugdev

The struggle with that sort of game is less a matter of engineering and more a matter of content. (Although if you're a one-man dev team, you may struggle with the engineering as well.)
The amount of content in many RPGs is far more than an individual developer can handle unless the engineering is notched down all the way to minimum-viable-product form.
Furthermore, whereas more action-heavy games can get by with a subpar story in favor of gameplay, a good story or higher-quality dialogue is more frequently demanded by players of RPGs, to help keep things fresh between grinding sessions.
Surely Paper Mario is not an exception to this with its charming characters and playful dialogue.
Do you think you can put together a more charming cast of characters than ?

All of that said, I'm not trying to kill your motivation, only to warn you of potential tall brick walls you might run into.

3D is not exactly easy, even if you use a pre-built engine such as Unity or Unreal as a crutch.
Even your "flat 2D characters" are only actually flat and 2D in the original N64 Paper Mario (and I guess in that WiiU entry as well iirc)- in TTYD, the characters are animated using an hierarchy of bones, much like a 3D character model would be animated.
So to re-create the TTYD experience, you'll need to learn some 3D rigging and animation.

I guess I'll let others recommend you software/resources.
I'm just a crazy enginedev who would tell you to go thoroughly study 3D programming and design so you actually understand what's going on.
Many people will probably recommend you use a pre-build game engine like Unity or Unreal or something rather than try to write your own from scratch.

I think that it would actually be rather easy to make a paper mario styled game engine than other engines, because, there are a lot of challenges that you can just ignore. There is no need for any culling method or even 3d models beyond the levels. If you can make an OpenGL teapot, you can probably make the graphics engine, as long as you don't use some kind of animation system beyond changing between sprites.

I would stick to C/C++.
Actual devs are generally using C/C++ for multi-platform games.
Taking a look at how to program for the Gamecube, Wii, GBA, or NDS using devkitpro might help give you an idea of what programming for a game console is like.
(It's basically like devkitpro except the software libraries you use are more shitty)

For programming in 3D using OpenGL: youtube.com/watch?v=c-0oBXJZwQE&list=PL3wFcRXImVPOQpi-wi7uriXBkykXVUntv

For rigging in 3D using Blender: youtube.com/watch?v=c-0oBXJZwQE&list=PL3wFcRXImVPOQpi-wi7uriXBkykXVUntv

Whoops, for OpenGL: learnopengl.com

Holy, these walls with embedded objects came out looking way better than I thought they would

If you want to release this century, don't make your own engine. Download Unity, Unreal, and Godot and make a simple prototype in each, then from that decide which you want to use.

after some bugs the engine can display the textured maps just fine

What is the purpose of Sigma II if you don't mind me asking? How much of a coding 1337 do you have to be to make something like this?

So Twitter has some pretty good analytics built into it. Pretty fast, and not retarded compared to Tumblr. Easy to get exposure too.

Lots of bots and their video player is ass, though. And it's hard to save images sometimes

The end result is supposed to be a 3d game engine would have been "acceptable" in 2001 or something. I'm also using it as a way to teach myself a lot of new things, like Vulkan. Eventually, I want to be capable of implementing new or just interesting things into my game engines. For some people, engines are a means to an end, but to me making the engine itself is more rewarding than making a game with that engine.

To make something like this you don't need to be that good at programming, you just need to put in the time. For example, making the engine work with Vulkan wasn't "hard", it just took about four months getting the Vulkan renderer working. (as well as asking a friend who is better at graphics programming than me to look at the code a few times) So, the difficulty was putting in the time. If you want too you can look at the source code to my last engine, it's pretty bad in a lot of places and when I went into the project I couldn't even implement a linked list. It's just that I poured enough time into it to finish it with an engine that was able to achieve the goals I wanted it to achieve. I read a ton of tutorials for stuff, and there are a lot of advanced tutorials that help a lot.

...

is there any way i can keep a reference to objects/gameObjects from different scenes in unity?
i'm working on an ai behaviour editor and for example i want an ai to move to a gameObject in a specific scene. i can assign gameobjects from the scene, but when i save the behaviour or change scenes, that gameobject will no longer exist and the reference gets lost
is GameObject.Find my only option?
only other solution i can think of is to create a script that keeps a reference to an object which represents the gameObject i want to reference in the behaviour. but that means i need an additional object for every single gameObject, and it sounds a bit too convoluted

Thoughts? The backhand feels iffy but I'm not sure how to make it better.

Just do it like webm related and it will work fine :^)

Punches start from the waist, backhands especially so

15% traps? kids these days

I wish.

Try to make the inactive hand guard the chin more as well. Also, punches are a full body motion. Any punch that only involves the arms and a bit of the shoulders will look weak. Punches start from the legs and/or the pelvis, they involve footwork, the entire body rotates at times. I'd also change the timing somewhat, to make the actual punch motion execute faster and hold the arm position after the punch a bit longer, but that's mostly preference.

Thank you. I'm going to apply this after I finish implementing the combat.

well the convoluted approach worked
i make an object for every object that has to be referenced in a script regardless of which scene it's in

made a new node to handle events that need to happen in a sequence
example

move to the well
don't do any of the next nodes until it's finished
wait 2 seconds, moving to the well is ignored since it's finished
don't do any of the next nodes until 2 seconds have passed
walk to the anvil, ignore previous nodes since they're finished
when it finishes the anvil movement, all previous nodes in this sequence are reset to be unfinished, so it can start all over again

and technically the script does happen once every few frames, so you could still have something that interupts the chain if need be

every node is handled within itself nice and independantly, with the exception that the script itself stores the list of moves that need to be reset

>Yoko taro's tetrachrome

mister, its been like that for months, when do you plan on finishing your game?

It's a game that probably less than 5 people have ever played, made by a guy who everyone thinks is insane. Basically guarantees it a spot on one of the "level X" below the iceberg slots.

Could you link me your sourcecode or tell me where to find it? I would like to look over it.

I played around with making level / map editors n such but mostly top down 2d Link to the Past kinda stuff.

I want to get more into 3d programming n whatnot and I learn better by seeing an example and working through it on my own.

...

ika.itch.io/redsky
It's labeled as "sigma engine source.zip". Just warning you that this is by no means a good example, pretty much everything I did could have been done much better.

I'm not really if such a thing as "teaching game engines" exist yet, for example do you know how MINIX is a "teaching operating system" in that it exists to teach people how to write an OS. That would be a pretty interesting thing to try and make.

Interestingly enough Sigma II is at its current state slightly over 7800 lines of code, which is larger than the Sigma I code-base.

Say, are you doing any degree of multithreading? It's part of the appeal of Vulkan and I'm wondering. I'm working on a Vulkan application for a project thesis myself.

Right now I'm splitting a "client" and "server" into two threads, but beyond that I'm not doing much. I'm inexperienced in both Vulkan and multithreading and so I'm taking baby steps with it right now. If the rendering gets especially complicated then I will look into that.

Holy fucking shit Unity mechanim is a piece of steaming shit, I have been trying to make it work for the last few hours until I discovered the events at the end of animations are completely unreliable.
I'm about to fucking scrap all this shit and make my own state machine. How the fuck am I supposed to call a method at the end of an animation?

either attach whatever function you want called directly to the frames of the animation, or use state machine behaviours

is there any way i can bake a segment of the navmesh in unity and then stitch them together?
i've exported my terrain to a quarter of it's original polycount and it's still around 130k polys, navmesh baker seems like it's stuck with no sign of progress

user, thank you. I copied the animation from fbx and now the event just works. Who the hell knows why adding events directly on import animations is just so unreliable. Now I can finally sanitize my code.

this is retarded
and if i don't do this, all my shadows will look pitch black

Does Unity not have a quick and dirty preview lightmap generator option?

in either case it ends up giving me errors during the light transport stage about not enough memory

Try using the progressive lightmapper in unity 5, is it gone in 2017? It's a lightmapper that bakes in increments and becomes clearer the more it bakes, but you can stop it anytime. Also, there are shitloads of reasons it could take forever, like wrong texel size or whatever, look at the unity tutorial on light baking with the viking village. You should also not try to bake every little shit like every stone in the scene, keep lightmaps to the bigger objects. Also, playing around with environment light will prevent your shadows from being pitch black and give your scene more color. Don't forget to use some light probes and reflection probes too.

Been fun learning all the setup and steps a GPU goes through.

Anyone here have experience with the Godot Engine?
The time works fine, that's all good.

I just want to know why whenever I press those buttons, it does fuck all when it should be adding 1 to those values.

Isn't _process setting the values of seconds, minutes and hours from elapsed? It looks like you need another variable to store the button click modifications and then add that to elapsed before formatting.

I knew I was over thinking it.
Thanks lad.

i tried the progressive one, first it was giving me errors about not enough memory, and then shit froze for about 20 minutes until i could turn off unity
only option right now is to disable terrain, lightmap it and enable the terrain. problem is that now shadows are all dark blue/purple

what tutorial/books are you using user? There's not many ones for Vulcan so I want to know if there are some better than the ones I encountered so forth.

I've been working through vulkan-tutorial.com/Introduction I'm not the biggest fan of tutorials, but preparing Vulkan seemed so in depth that it made sense to have some explanation along the way.

Sounds about right, my Vulkan-related code is 2935 LOC. So, 37% of my engine code… In comparison my OpenGL 1.1 fallback renderer is 677 lines long.

I started with this and then moved to the tutorial that is using since mine doesn't go all the way into stuff like texture mapping.
vulkan.lunarg.com/doc/sdk/1.0.26.0/linux/tutorial.html

Since I'm writing some Vulkan shit, too, here's what I used.
The two former are good for beginners since they explain a lot, while the Cookbook assumes that you know a little already. However, I wholeheartedly recommend getting it, if you want to work with Vulkan seriously.
Here are some more: stephaniehurlburt.com/blog/2017/7/14/beginner-friendly-vulkan-tutorials
Embed related is what I'm currently at. What you can't see is that the triangles are procedurally generated on the GPU, too. I use a compute shader to write the vertex buffer. What you see in the beginning is my simple display selector routine. I was sick of the fullscreen window opening up on top of the debugger, so I wrote a simple routine. When you start the application and multiple displays are detected, that small window is created. You drag it on the display you want to use and press any key to continue.
I'm using SDL2 to handle the platform specific shit btw. Great library.

I still need to polish the stopping animation, but it will do for now until I decide to revisit the animations again.

>four triangles

update:
i found a script to split terrains
after a lot of fucking around, i managed to get it to split terrains by a variable amount, original was exactly 4. although for some reason my version can't work with uneven numbers. gotta split it by 16 instead of 9

it works though, i disable 3/4 of the terrains and light baking will probably take an hour at best
if anyone's curious, here's the updated script, link to the original is also in there
pastebin.com/GCYGCvkN

T-there's always New Years..! inb4 i barely make it in time for february

Learning Godot. So far works okay, but now I need to implement enemies.

>silky smooth 60 40-60 fps
yay, unity finally works semi-competently in an open world environment
now i just have to actually get combat working before i slack off with worldbuilding

Toaster?

5 year old laptop that wasn't meant for video games

Got some arms done, working on walking backward now.

Question regarding security cameras in gameplay: does Holla Forums like Deus Ex security cameras, or Metal Gear Solid 1 cameras? Instant detection or a warning? I'm gonna rip the turret from AS-Overlord from UT99 and use it as a security camera that can lock doors or activate sentry guns.

hey I think I might be getting there

they're too fast

Well I thought the delay between the input and the actual punch would feel shitty.

REMINDER THAT DEMO DAY IS FEB 2

looks cool as fuck. I hope you're planning to make the different moves meaningful in gameplay

Thanks! Yeah, this is the standard control scheme for melee weapons, and melee when used with firearms. The difference being that you can't parry when armed with a firearm. Attacks will apply damage based on the body part you hit, and can be parried. Parrying is also directional, but unlike attacking you parry a certain angle around the direction. I'm planning on 180 degree angle of denial for now. – Melee attacks cost stamina, but having your attacks parried penalizes you with stamina loss further. Parrying drains stamina over time, but blocking an attack does not. It only punishes you for holding the block. If you're out of stamina, you can no longer attack until your stamina is replenished.

I really, really want to add kicking but I've yet to work it into the model. I won't add it unless I can find a proper gameplay use for it, and with the melee being complex as it is, I'm having trouble finding an excuse.

Oh, that makes sense. I think I understand it now.
I was able to streamline the formula even further now. Now it's just a simple division and subtraction. Thanks for the explanation!

I'm REALLY digging how you've got the different punches based on mouse movements. Or do the punches shift the camera?
Either way, that's gonna lend itself really well to some intricate fights, maybe some enemies are blocking certain areas and you have to punch where they're not blocking?

The punch directions are dictated by the mouse flicks during the animation. This game is multiplayer, and parrying uses a similar mechanic, though, so that's the whole schtick behind directional attacks.

Do you guys look for other creatives to help with the games? I have a background in filmmaking. Maybe I could help with writing and dialogue if there is an interesting story to tell.

Telltale Games would be a good place for you :^)

Really what you are doing is very interesting. I really like watching Youtube videos on GZDoom Builder, Unreal 4 and Renpy. Don't care about telltale games.
It would be fun to get involved somehow.

A sense of cinematography is good for background design and just generally making every scene look good in a screenshot. But most people work alone, or in small teams. Someone like you would be useful in either a really cinematic game, or in a bigger team.

...

My story is kind of a total mess that I just work on when I get tired of writing code, so having someone who knows about writing and story structure would be nice, provided I don't have to sacrifice gameplay for the sake of story.
Hell, if you're a good enough writer I would just say "make some story that fits with these mechanics and game-play structure" and let you do whatever the hell you want within that.

Let's see if I can get shit done by then.

Collision detection working well, Y axis traveling still a little choppy if on a high slope, but it will do for now. Now I gotta add falling animation/state and a sliding condition for slopes that are too high.

user, it was an NES game, you can do it

and now make a recoil game tbh. I would play it
t. idea guy

So I have a world class that is essentially a container game objects and keeps a reference to the current grid (simple two dim array with integer values representing tiles).

Game objects are a base class basically containers for components (like physics, graphics, etc) with an x y position integer that when its update function is hit it goes through the different components and hits their update functions.

My Game Loop hits my world.update method which in turn hits my components update methods.

What is the best way to handle real time input? Right now I have it hard coded for testing but I keep hitting my head against the wall trying to abstract it.

I was thinking put it in the game loop but how do I then send it down-wards to the correct player object. I plan on having multiple players eventually, making the keys rebindable, etc.

Any suggestions?

I'll try and get relevant code put together in the morning if my summaries aren't clear enough.

You receive all inputs in your message pump. Just handle WM_CHAR and WM_MOUSEMOVE. Input is never lost because the queue will never "forget" the messages generated by the user until you handle them.

msdn.microsoft.com/en-us/library/windows/desktop/ms646276(v=vs.85).aspx
msdn.microsoft.com/en-us/library/windows/desktop/ms645616(v=vs.85).aspx
msdn.microsoft.com/en-us/library/windows/desktop/ms644927(v=vs.85).aspx

Also, why are you so concerned with fitting all of your code into an OOP style? It will just hinder your progress and prevent you from doing the easiest thing in cases when OOP doesn't make sense. But to answer that, the "best" thing is to have input handling functions generate messages that then are processed by a function updating the player data structure to alter its movement.

What I do is have xna take the raw input into its eg KeyboardState objects. I made an interface for kb, mouse, gamepad, and chatpad that all have 5 common functions, each one containing two states (prev and current ), and comparing them gives Pressed, Released, Held, WasPressed, WasReleased

That gives me basic button polling against a common interface. From there I have an enum with different gameplay actions, like moveleft or jump. I can make a dict to bind keys from any input stream to a game action

Then, i just compare the game action between its own prev and current state to see whether i keep moving or whatever. You can give this object to various players and have many inputs, or even start to develop an api for an ai from it

Does people actually buy assets from the Unity store and then try to sell the game? Every time I see ads for them, I'm just completely lost. I'm not even an artist, but the idea of putting generic assets in something I make just sounds disgusting.

if it's scripting, you probably won't notice it's from the store
if it's a model you can probably notice, but in some cases it's forgivable
it's one thing to use the free rocks and shit, because how unique can you possibly make rocks look? on the other hand i've seen pretty much everyone using the medieval environment pack, or recently the blacksmith assets. using those is just lazy

Instant detection when sneaking around in a place where no one should be, slow detection in semi-crowded places.

Using asset store assets gets you a publishing deal with Sony, just look at Last Soul Aside.

Did anybody make dorf fort in space yet?

Yeah, there's nothing wrong with using scripts of course, assuming it's free. People that are actually, unironically paying for code, are beyond fucking retarded and should be taken out back and shot in the head.

But yeah I'm mostly talking about pretty important assets like making an entire city in your game out of buildings you bought off the asset store.

Oh, I do plan to make a 3D tank game somewhere down the road, but first I have to get this one to playable state.

Noice. I will be looking into this.

I feel this to such a painful degree that I couldn't even use the UE4 mannequin for prototyping. I couldn't "feel" my game in the generic assets if you know what I mean. And if you don't feel your work, you can't add onto it properly.

i'm having the same issue with animations right now, hence having to do my own walking animations
on the surface most store animations look alright, but when you retarget them, there's always some issue, like the hips not moving, or the hips moving way too much etc.

Anyone have any idea how spyro might have done details like these leaves or the grass that starts and ends at the bottom of the hill (look at the hill in the distance compared to the near ones)? I always noticed how the game has no/almost no stretching textures at all and it looks so fucking good. I also noticed how details like this usually fit on a grid, although I find it extremely hard to believe that they would have cut up the entire level into uniformly sized polygons to do this.

The leaves on that floor could be made in two ways: 1.- have some texture layering, with the yellow leaves' layer having uv's such that they appear sporadically (no way the PS1 could do this, but it's possible now), or 2.- Cut up the floor so that the grass with leaves are little squares with their own texture and the rest is repeating grass texture.

As for the cliffs, I'm not sure I understood what you were saying but I think it was something like pic related?

Forgot the pic.

I was sceptical about this because every programming quote I've seen they've said they're saving literally every polygon. Although I noticed the floor is a bit uneven so I guess it is possible. And what I meant by the hills is the tall hills are 2 textures, as compared to the short hills which are one.

I guess the only way to know for sure if they cut up the floor is to load an emulator with a wireframe hack.

Well damn I guess they did.

I've just noticed that with a lot of my projects I start I'll usually get stuck on the nitty gritty details and never actually get to the fun stuff. I'm not dedicated to the OOP way of doing things and wouldn't mind doing better ways. I just don't know how and I've been trying to go with what I know, if that makes sense?

This project is only supposed to be a 2 to 3 week little prototype.

OOP was what I was taught in school n I use it at work. I've been getting into DOD a little but if I'm honest I've only gotten to like snake and tetris tier arcadey. I haven't gotten deep enough to have confidence in not goin with what I know.

For me there is a distinct point at which I do or don't want to use assets from the store. These assets are from the store, but the assets themselves are just pieces that can be stuck together in different ways. I like these kinds of assets because even with two people buying the same pack, it's clear when effort has been put in.
I can't model or texture, but I can easily design content with the some of the more modular asset packs (I even got it on sale for like 20$, which is pretty damn awesome).
First image is 4 ships that I designed from the parts so far, using a similar pattern when building them. Second image is pre-built examples. I think I've added some uniqueness, and still have more ideas on different ships that I can make with the parts I have. Also, the blue ship is using a custom shader for recoloring I wrote, which I haven't applied to the other ships yet.

Then there is laziness, like just taking the assets out of a pack, and using them without adding your own touch. It's pretty obvious when this is done, and even players can tell (and it's probably why c61277 doesn't even want to use the UE4 mannequin. There's a point at which enough effort goes into making the assets feel unique that makes it less disgusting to use. If the asset can be customized extensively, I feel less bad about using it, especially if I can add my own 'touch' that is distinct from the other instances the same asset has been used.
Asset flips are pretty easy to notice, and same with using default assets. I really like having the asset store around to buy assets to fill in where me or my team are lacking, but I make sure that we always spend time to make whatever we buy our own.

One game I worked on we bought a set of particle effect assets off the store. It was nice stuff, but didn't exactly fit our needs, and was bloated as all hell. Lots of pointless texture data, which I compressed down to about 1/8th the original size. (lots of transparent pixels with different colors, lots of smooth gradients on textures that were very small, light lossy compression is great).
We also customized every instance of the particle effects that we used, and combined them or dropped pieces in some cases, as well as hooked all of the effects up to our own LOD system.
There was a bit of effort we ended up putting in, but I think it has been worth it, because nobody (at least so far) has left comments about it, and it's a pretty common/popular particle effect pack.

So yeah TLDR:
Don't just buy and use assets off the asset store, or people will know you are a hack. Take the time and customize whatever you get, and put some effort in to make it your own, and people won't notice that you're using asset store assets.

I can draw a cube now, too.

Having never played Spyro and source being 240p I can't discern the video, but regarding the cliffs, They're probably two segments.

I think generally speaking you want somewhat regular polygon sizes in games. Some engines can glitch when a polygon extends too far away from the camera. If not, pic is also possible.

Is this Air Rivals/Ace Online?

add an fps counter pls

is onto something when he said that you can get glitches when your triangles get too big. A problem they didn't have back then, but modern engines have with large tris now, is that you can't lightmap them properly.
Maybe they used decals, although I find that difficult to believe. While it would produce less triangles, it also requires you to do alpha blending between the two surfaces, which I assume would be A LOT worse on the original Playstation.

Forgot about decals, but I agree it'd have been crazy on PS1.

Take a look for yourself:
>models-resource.com/playstation/spyrothedragon/
>models-resource.com/playstation/spyro2riptosrage/
>models-resource.com/playstation/spyro3yearofthedragon/
Always look for 3d rips of games if you want to learn how they did shit.
There are also few documentations which go into a more indepth analysis of how to make levels oldschool style like:
>medium.com/@gordonnl/wind-waker-graphics-analysis-a0b575a31127
There is also a really nice Blender addon called Sprytile which is useful in that case:
>github.com/Sprytile/Sprytile

Isn't it a bit wasteful adding a polygon for each texture tile or something?

Yes, but that's how they did it on Playstation because of affine texture mapping, because larger Quad/Tri meant stronger the distortions.
This shader emulates this very well if you want to know more:
github.com/dsoft20/psx_retroshader
If you want more modern techniques that are less wasteful on polys but in return costs more in shader computing, take a look at terrain/blend shaders combined with triplanar mapping. Basically you use colors like red blue and green to paint certain parts of the mesh with vertex colors or an additonal color texture, and then blend textures with these colors, for eample red painted parts is grass, green is rocks and blue is a dirt texture. Terrains are a pain in the ass no matter what.

Tell ne more.

sexy ass splines
say what you will about the asset store, the free shit tends to be pretty neat

I bet a Pajeet wrote that code.

it works though

The Asset never was bad, the cunts at Unity are. They deliberately refuse to insert certain basic features into the engine because they know someone else will do it and sell it on the asset store, which means more profit for them. I mean you can't even make a damn basic dedicated Server with Unet.
Either shit misses completely or it only misses the most useful features and you only realize it hours in. Fuckers at Unity had always relied on pseudo extortion methods like that, first they get you hooked on their tools and then they try to sell you shit once you realize what's missing, and sadly most people buy it because they don't want to waste more time doing it themselves or because they simply can't. I could write a book about how angry these cunts make me. The only reason i don't use UE4 is because fucking chinks are even worse. Even if i would make millions by using their engines i would refuse waste a penny on either of them, no matter how much they sue me.

Will you make me pay for my titty mods?

pixar, fucking hire me, i'm a top tier animator
also you were right, in the editor the spline snaps to the terrain, but when built it follows the original spline nodes 100%

Reminds me of those 90s acrobatic shows.

I thought for a sec that was a Europe Rape Simulator with the naked black guy running around.

IGMC 2017 is over and the judges' feedback for Die Totenmaske is in.
The guy is right about the game being really unpolished. Most items are broken outside combat because they're restricted to the hands and RPG Maker MV really doesn't like that, the dialogue was written by a sleep-deprived user suffering from food poisoning, and there were some pretty big balance issues. Still, working on and finishing a game with other /agdg/ anons was a great experience and I'd like to thank all you wonderful faggots for the help and inspiration you've given us over the years.
We'll be uploading a new DT build with some tweaks and bugfixes soon, then begin work on DT 2.0 in another engine. Hopefully it will be toaster-friendly enough for Ika and anyone else who can't run CURRENT YEAR Unity games well.

Nicely done. It's a big step forward for you guys, looking forward to seeing what the revamped version brings.

For starters, we're moving to a freetard 3D engine and battle backgrounds will be actual 3D models instead of really wide scrolling textures. Since I'll be more involved in level design this time, outdoor maps will probably take more influence from Yume Nikki and indoor areas from Quake.

hey, hope it's financially rewarding at least.

Thanks for all the good work you did on the wiki. I kinda left it alone as well but I'll try to update the thread archives soon.

Hope you'll get to show your next project soon!

Oh, and maybe drop by the discord channel from time to time
discord.gg/9gKG79
I know
it's a chill channel tho and now I use my throwaway account as a promotion tool alike twitter and tumblr, so whatever

What's this Demo Day thing? Some kind of publicity event?

People who want to share playable builds of their game in whatever state it's in.

previous demo day thread:
still available cos it was stickied I think

dudes who haven't tried the demos back then can still try 'em out!

8ch.net/v/res/13774137.html

Alert everyone, upon further inspection mouse flicks to get the attack direction doesn't feel very good. Anyone got any alternatives?

I tried to make a sword game once where you can swing in directions and decided on this:
Try it out, maybe it's fun in your case.

you're wrong
git gud

are you doing it like in Daggerfall or Die by The Sword?

Never played either. Mind describing them?

user, I'm just not really digging it.

Sounds kinda clunky for a PvP game. I gotta parry, too.

someone make a autistic nursing home management simulator and set it in the future where you play as an AI who looks after them with the helps of lesser AIs

...

Yes, I based the designs off of the I-Gear. I'm impressed anyone recognized it.

I'm still in a prototyping phase at the moment, but I have some flight stuff in vid related, though the mech model I have doesn't fit at all.

and here is the other vid it wouldn't let me post at the same time

You're trying to borrow elements from Morrowind, right? I had to run it to double-check if it didn't have the same combat system as Daggerfall. I have Morrowind installed but never actually played more than an hour, which is shameful

Apparently it didn't. So, here's the description:
You hold the attack button (right mouse button is standard) and whichever way you move the mouse, the sword will be swung.
Move your mouse to the left, perform a left swing
Move mouse up, perform a thrust
Move it from top left to bottom right, perform a broad swing across the screen in that direction
etc

You actually had to move the mouse a bit to trigger the swing, it wasn't instant. The attack speed was based on stats. There were only a few possible swings,
mouselook was turned off while performing the attack but you could still move your character

I suggest you play it yourself for a while as it could be a source of inspiration.
en.uesp.net/wiki/Daggerfall:Files

Die By the Sword had a system where the sword swing replicated your mouse movements.

That I-Gear. I used that piece-of-shit Fox Binder until I got enough WPs for another.

Of course, until I realized that evasion builds suck overall.

I'm honestly not trying to borrow from anything. I just want melee to be more involved. My UI is borrowed from 50/50 EYE and Morrowind, though.

The concept is directional attacks that can be blocked with a directional parry. Everything else is up in the air.

I am going to check out Daggerfall, thanks, although I am a bit skeptical because it's a single-player RPG and I doubt they were focusing on fluidity there. Fluidity is really my biggest concern here.

Noticed some stuff after looking over the review again:

Maybe I worded it wrong. "expand upon"? "find inspiration in"? I personally find no shame in "borrowing" but that is probably obvious.

Daggerfall probably doesn't have much of that. You'll just need a few minutes with it to know if it's anything you'd like to try for youself

I'd look into DBTS as well, that is definitely more involved. I imagine a system replicating the mouse input into sword swings would be something cool and novel in first person.

Rune, Oni and Jedi Knight 2 have amazing melee but they are 3rd person.

And there's the modern style of melee combat from Kingdom Come and For Honor. Not a fan of that but my brother plays a lot of the latter and it is very fluid and involved once you get the gist of it

Also, I see in your .mp4 that you don't lock mouselook when flicking the mouse. this might be the reason you dont find it fun. With my mouse sensitivity, I'd probably make a 360 turn before the animation started.

Sounds interesting. I might halve the sensitivity during the hit.

I was thinking about that, but it turns out they did cut up the entire fucking level into squares about the length of spyro. Which I find extremely bizarre, especially in the first game which has large open fields that are relatively flat.

Oh damn, thanks user. How did you manage to dig these up?

Would play / 10

I was curious myself so I loaded the game up in no$psx, which has a cool graphical debugger that lets you see every single draw call that went into a frame. I can confirm that Spyro basically works as you imagined. Terrain is nearly uniformly sized quads. Most of those quads are just a grass texture combined with gouraud shading (not that you can tell in my pics), and for the details like the leaves or whatever they use a different part of the texture page with a different palette (4 bit/16 color palettized textures are the most common on PS1). I was a bit surprised because I went in expecting it to be a decal system, which you can totally do on PS1 (e.g. for bullethole effects), but it makes complete sense why they did it this way.

They most likely used a scheme like this to author the terrain:

Basically most of that work can be automated and rather than fiddling with grass the artists can spend their time placing lights (much of the visual appeal of this game is in the prebaked lighting though again you can't see it in the scene I chose), placing props like trees, and designing custom geometry to flesh out the level designer's rough buildings etc.

As says you had to tessellate everything on PS1 anyway because if you simply slapped a flat polygon over a wide terrain area/wall/etc, the affine texture mapping would look horribly distorted. The devs must have thought, if we have to tessellate everything, we might as well take advantage of it.

...

Try figuring out how to write in a procedural style- it can help your thinking a lot since you'll know how to write code without stuffing it into an "object model".

I also learned OOP in school but I thought it was an unnatural way to program and I just dropped it whenever I wasn't forced too by a school project.

There are also more esoteric programming styles- if you download anything by Ken Silverman, you'll know what I mean.

Oh damn I guess that settles it. It makes me wonder if this is viable in a modern game that isn't minecraft. It just looks so fucking good compared to later games which have to stretch out one large texture.

You know, depending on how texture addressing is done, you can have one giant quad and have the texture coordinates loop back around, right?

Yeah, although I get the feeling it'll lose something. Just off the top of my head flat terrain and edges of cliffs couldn't be as wavey. They also did their shadows that day by coloring the corners of the quads.

This is easily achievable and makes a lot more sense if you look at how fixed-function hardware acceleration worked. The early versions of OpenGL work in this way and make it very easy to do this kind of thing.

>>>/cuckchan/

I know it's achievable. I just don't know how hard it'd lag in comparison to just using giant tri's like 5x the size which is what most games do now.

Unless you are using 90s hardware, its not a concern. Besides your polycounts will probably be too low for this to matter. Most games now do a giant tri and then apply a huge amount of shaders and clutter it with a ton of models- so compared to that, what you're doing is nothing.

(checked)
True. Things like minecraft manage to get away with it.

Minecraft is pretty high poly compared to spyro's graphics. Also its dynamic world system and the slowdown from using Java stop it from reaching the peak rendering speed for the kinds of terrains it generates. If your game is using a (mostly) static world this opens you up to a huge amount of culling schemes that are much faster than what minecraft is able to do. Really you shouldn't use minecraft's performance as a baseline for how fast these graphics are.

Assuming blocks are static (eg no change to them, no new flowing tiles), you could easily flag chunks or groups of chunks as inactive/static and optimize them on the fly.

Also, is there a good approach to extruding a texture? Like you have an arbitrary shape, like a sword, and it generates a flat "3d" object that have "pixels". I'm pretty awful at 3d in general and topology with code

Alright lads, i need your help. I'm making an FPS and i want every single weapon/item in my game to have unique touch/effect. So far i have these:
Weapons
Support Items(Cooldown)

Any games i can get steal ideas from? A weapon/item you would like to see? For some reason i have a really hard time finding things that aren't just reskins of something else or have similar effects to the ones i already have.

If blocks are static you can use a BSP-PVS engine, at such a low polycount. That's optimal for slower computers that actually would benefit from a speed increase, but on a modern game it's a bad idea. But, there are quite a lot of ideas that can be used, a lot of which are just simpler and "do the job" just fine. You don't need an overcomplicated engine for something like a PS1 game.

By "extruding a texture" do you mean having a kind of "billboard" object like the enemies in doom or the trees that are totally flat in a lot of games? Because what that is, is just two polygons with a texture applied to them, the texture has alpha blending which means that it kind of looks like a tree instead of a picture of a tree mapped to a big square.

No, I mean extruding as in it takes a texture, like a sword pixelart, and transforms it into a 3d object with fake pixels with depth. Not a quad/billboard

Most games with large outdoor areas are doing basically the same thing today. The designer "paints" a heightmap and terrain map, and either an editor tool or the game engine itself creates a tessellated mesh from those maps that is used as a starting point for the world. In fact tessellation is built into modern GPUs and can be used to dynamically increase or decrease the tessellation quality of terrain and objects depending on distance from the camera.

For static geometry vertex shaders are usually pretty cheap. A smoothly tessellated object covers close to the same area on the screen as a rougher one, so the pixel shader invocations should be very close (given that the object is close to the screen and you aren't rendering a bunch of 1 pixel triangles). The increase in vertex shader invocation, mesh size, and possibly tessellation shader invocations should not cause a great decrease in performance for these objects.

You should read this article that dissects exactly how MGSV renders a frame to put into perspective just how much work a well programmed game can do on modern hardware and still hit 60FPS. I'll be honest, even I was kinda shocked.
adriancourreges.com/blog/2017/12/15/mgs-v-graphics-study/
Amusingly enough given our conversation, literally the first thing the engine does is generate a terrain mesh from a heightmap and do a depth prepass with it.

BTW, for more experienced devs, please read the section about how they do color correction. Maybe it's common knowledge but I thought it was ingenious. Basically the artist takes a screenshot of the game, does whatever adjustments they want in Photoshop, and a development tool generates a compressed color correction table from the edited screenshot. Saves the programming effort of making color correction tools for the editor and makes the artists happier and more productive using tools they're more familiar with. I'm totally stealing that idea.

Oh, well that's possible I suppose. By "fake pixels" do you mean like how the build engine used voxels generated out of images (I presume) to make "3d-looking" stuff like levers and guns, like pic related?

I personally don't like color correction or putting shaders over everything. I want my game to have the kind of "raw" look that you get from fixed-function graphics.

LuaJIT is very fast. It integrates extremely easily with C (and C++), The language is pretty good actually. I use it for shader and material definitions and initialization. Also started defining all gfx objects, their textures, material settings in a Lua file. Then I added item prototypes into Lua files. Now I'm exposing most of the game logic C++ functions to Lua so all of the baseline AI, encounters, and events can be scripted in Lua. It's easier because no compiles. Just reload the Lua file and re-initialize state.

Support shit?

Bonus points if getting set on fire activates the riveter's burst fire, shooting superheated slags of metal with poor accuracy but extra fire dot damage.

I forgot to mention it's going to be multiplayer, so riveter and quicksilver are probably impossible to balance, but Tether(gravity pull) and Lancer(far and close range hybrid) are good ideas. Thanks.

Just have the quicksilver's speed inverse from the damage and give it a somewhat high cooldown. It would encourage either using it mid fight as a wild card or as a last ditch slow projectile if you're out of ammo. Use in melee range as a kamikaze attack.
Riveter- eh, maybe you'd have to hit unarmored parts of the enemy. Headshots or leg shots might nail them or something.

AKM-74

Health indicator is done.
Following the theme of cellphones and connections and all that shit, it's a phone reception meter. Full health is green, all bars, good to go. Almost dead is one bar, red and blinking.

Now I just need to work on a shader to tint the screen red a bit when you get low on health, just to make it extra obvious.

Reminder to you anons that if your pixelshit doesn't look like pic related, you won't get the soyboy audience.

It's disrespectful that the term has "pixel" in it. Why won't anyone call it shitxel or soylism?

Can someone tell me what the fuck is causing these shadows to look all fucky?

Using Unity

try messing with soft/flat shading

I don't think those are shadows. Looks like directional ambient light not lighting things facing away from the light direction to me.

I am using a directional ambient light, yes. I'd like for this thing to eventually be able to cast shadows with the sun rising and setting but I don't know if I'll ever get to that.

Is there a way to do that in the standard shader or am I gonna have to write my own damn shader finally? Or is that "Shadow Type" (soft shadows/hard shadows) option in my directional light? Because messing with that didn't help.

Playing with stuff further makes it seem like it's just at a funny angle. I don't have time to be fucking making this shit look beautiful, gonna write Djikstra tomorrow.

What's a less retarded way to do this?

I need to check if a character matches one of several reference characters, and keep moving forward until it finds a character that doesn't match any of them. The reference characters may change so I need to loop through them rather than hard coding them.

The for loop is basically the condition to the while loop, but it's not possible to put it there so I feel like I'm doing something retarded here.

That's not to say pixelart can't look good

It's running at 60. And I don't think I'll add an FPS counter, because I don't want to do font rendering. It's out of scope for my project thesis.

Well, a lot of computer graphics were workarounds and doing things deliberately wrong to produce the desired effect. Especially pre-PBR.
You really should use decals nowadays for that sort of stuff. Either decals or multi layered shaders in conjunction with vertex painting.

It's the normals. Like said, you might want to try flat shading. If the mesh is imported, it should be like one button in Blender or whatever the hell you're using. If it's procedurally generated, you have two options:

RESPECT THE IF()

If there are only 3 characters you are looking for, I might use a compound logic statement (in C++, sorry don't know C# that well)

if( char == '1' || char == '2' || char == '3' ) break;

Pretty good attitude. We were all at that point once though, which unfortunately some people forget.

I'd do it like this too. You might also plot out the function on a 2D graph and see if it makes sense. I might even scale pitch into a 1-100 range and use a lookup table with the scaled pitch as the index into the table.

What's painful about this is a little bit of math goes a long way in programming. Modulo is a pretty powerful function. Good for determining even/odd status of a number: as in checking if( x % 2 != 0 )

I would definitely recommend a switch() in a case like that user.
switch (char) { case 1: case 2: case 3: break; //...}

True enough, but the point of the joke is the 6 gorillion identical console writes.

It originally looked something more like this
function isbreak(char){ var is = false; while (char === "1" || char === "2" || char === "3"){ is = true; //move forward index ++; char = source.charAt(index); } return is;}
But the characters I need to compare to aren't necessarily always the same ones or even the same amount of characters. It's such a simple change but the loop turned into cancer as soon as I made that change and I'm not sure how to do it better.

I just told you how to do it better I believe.

Then you need to read it again. I can't hardcode the characters, they need to be in an array or something that can be modified.

Actually check that. In that isolated bit of code my suggestion was correct. Now that I look at the function a C++ string find function would be much more efficient than anything you're likely to handroll.

Isn't
break;

Supposed to go after every case?

how is that relevant. a switch would still be a) superior maintenance wise, and b) perfectly efficient vs. not.

Because a switch is hardcoded.

No, the switch will fall through until reaching a break; Now with C++17 you can be explicit about these behaviors. Otherwise it's an idiom you learn.

Which is exactly what you'd need if you want something hardcoded like 'is this a 1?'

Besides, looking at the simple algorithm a find will be better regardless, as I stated.

the issue is that most pixelshit comes down to using retardedly hueg pixels, significantly reducing the amount of detail that can be put into the sprite and using that limitation as a crutch. The stuff you posted doesn't do that shit.

If what you have is working then use it. Move it to a small function if you want and call it something reasonable.

No. A switch statement is not a good solution for this situation.

std::string is quite literally the worst performing string manipulation library written in C++.

I just told you that I need it not to be hard coded, I need to be able to change the characters without touching the source code, and I cannot use a switch statement if I want that to be possible.

Of course, I could technically make any game imaginable by Just Doing It™. I just don't want to do it in a retarded way because that'll lead to disaster down the line, plus I won't learn anything if I just accept the first solution that comes to mind.

How often are you gonna call that code? Once every frame? If it's not something you are calling VERY often I would not worry about the way you did it.

I often get code epiphanies about how to make things better after I dropped it for some time. It's just how our brain works, go work on something else and when your synapses are formed come back and look again at the code.

Wow this looks great

Add a char[] parameter to the function arguments called "pattern". Then if the array is not null, check each item to see if it matches. Note that you should use a for loop, and not a foreach loop, as the regular for is going to be a bit faster, if you can work with it instead.

See: String.Split, within the C# source code
referencesource.microsoft.com/#mscorlib/system/string.cs

Isn't Console.Write mostly just used for debugging purposes only? Shouldn't you be using System.IO.File.OpenWrite for that shit?

Not a huge fan of the ground and grass tbh

is there any way i can get unity to run multiple lightmaps on one scene? it already takes me at least an hour to bake 1/16th of my map, it would take literal days to bake the entire thing as one. but if i bake them individually, the only way i can actually use the results of those lightmaps is if each map segment is it's own scene, which seems pointless considering that the game already runs at a decent framerate with the entire map loaded

can't you prefab scenes? why not bake from there?

(checked)
The ground and grass look nice, but the tessellation is too obvious. You have a few problems with that in the clouds too. See if you can make different tiles that have identical seams and use them to make it work.

I updated my engine, so now I use the FLIF codec as the primary codec for the engine and have replaced all of the WEBP textures with FLIF textures. I also baked the "missing resource" texture into the executable to make sure that it will never go missing itself (that would be pretty bad). The engine still supports WEBP but I have resolved to stop linking to it statically in the future and move it to a .dll file in the same way that my FLIF support works- I read the functions out of a dll file in the /data/dlls section of my games file tree. Right now I save a small amount of kb since I only switched from using 3 WEBP images which did not even total to an over 100kb filesize, so I saved maybe 5kb of image data and added a 677kb .dll file- but this will of course even out once I surpass 677kb in savings, and I expect to do that when I start seriously adding a lot of textures into the game.

Hey Holla Forums what's the easiest, even if dirtiest ways to save data from multiple profiles in Unity? If I'm getting this right PlayerPrefs are an easy way to store data but how do I locally store, say, usernames, highscores and some options of multiple players?

Interesting. I've never heard of it before, but it looks very good at compression. Is it slower/more computationally intensive to encode and decode though? I couldn't find any data about that on their site.

Why are you assuming your method is retarded?
If you really want to jump down the rabbit hole read the source code for the fmt library. And use it. It will be better and faster than anything you come up with.
fmtlib.net/latest/index.html

Yes, and that is the tradeoff that I wasn't aware of until I managed to compile the WIC module- when decoding an HD image in FLIF, while it was significantly smaller than a WEBP I made for comparison, it actually took twice as long to open in the Windows Picture Viewer, although this is still about one second. It's possible that this slowdown was a "software problem" since the viewer was doing progressive decoding of the image (something that FLIF's creators seem to care a lot about) and this could have been making it slower compared to just decoding the FLIF in full quality with no progressive decoding. Although, I think that it's probably just slower to decode. But, I don't have a way of verifying which reason for the slowdown is true- and this slowdown will only be felt on older systems (I am posting from a machine with a Core 2 Duo processor right now). I think that this also might be due to a lack of optimization in the current implementation of FLIF: if you look at the source code for WEBP, you can see that a lot of it is using explicit SIMD extensions to gain speed, while FLIF does not use any sort of explicit SIMD and has to rely on the compiler (which is pretty unreliable). So, I think that its fair to say that FLIF could become as optimized as WEBP or just more optimized than it is right now in the future.

...

I don't know jack shit about animating, but is there any reason why you can't just set up some kinects and mocap your animations?

/agdg/ are my friends and I'd like to share

I RELEASED SOMETHING FOR REAL

At 5 AM local time my wife gave birth to the future pope/Poland's president/heir to Shodanon's game empire (choose two). I'm really happy but now I wonder to what extent this will influence my already limited dev time.

/blogpost

do i look like i have a kinect?

your wife's son will be a better gamedev than you

I'm not 100% sure what you want to do with this code. Are you looking to find the index of the first non-break character?

The only thing I've seen that discusses encode/decode speed is the comments on this article: cloudinary.com/blog/flif_the_new_lossless_image_format_that_outperforms_png_webp_and_bpg
From there, I got pic related, and the following on decode speed:
Which seems consistent with what you said. I think it's worth looking into in the future; hopefully implementations can be optimized well enough to make it reach WebP's speed.
That slowdown might add up over time; it might not matter if you're only loading one texture, but if you're loading 100 you might start to feel it.

congratulations, but kids are like puberty once you have them your whole world changes, rip your gamedev career don't worry, you will not regret it

Pretty sure your child will hate video games and how sexist they are.

tools.suckless.org/farbfeld/

If I wanted a suckless game engine I would use it, sure, but you need more info. The image format really isn't useable until I have a WIC module and basic tools for windows.

I plan on only decoding game content during level changes, which will have noticeable loading times anyway, so its not a huge concern if my older computers take a little bit longer to decode the compressed formats. That's a pretty cool article, thanks for sharing it. It feels good to be using the most bleeding edge image format… and if any newer, faster version are released, people can just swap the dll files out.

I have to install Linux again once December is over and port all of this windows-specific functionality to the Linux version- right now Linux is being neglected since all new features are developed "windows-first". The next Linux machine I will be using will hopefully support Vulkan, so Linux users can have Vulkan support in Sigma II.

Have a (You).

(You) too.

Pics or it didn't happen.

Dude, the format is literally an 8-byte ID, 4-byte width and height values, and 4 shorts for RGBA per pixel. It's the simplest format physically possible.

mhm I made some minor progress on the second map. The first map is almost finished and needs some touch ups then its pretty much done. Since I have 2 tech styled maps I guess the third one will be more outdoor focused.

I didn't even read it right, when I saw "easy to compress" I thought that it was supposed to show its compression algorithm somewhere.

Congratulations.

Congrats Shodanon. Though to be fair you released something 9 months ago, your wife patched it and did a rerelease.

thanks, I love you all.

Remember when trades were hereditary? I am a traditional kinda guy so I'll share what I know with him. My gamedev skills are what they are so he can always build upon them. I wouldn't want him to become a wageslave dev tho just like my dad never wanted me to become a farmer which I am to an extent

either that or the need to provide will make me actually finish and release a damn game. I wouldn't mind it either way, really. I always say it's a hobby.

I can't plan for anything. It will be his life and his decisions. disinheriting doesn't mean lack of love

took my camera to the hospital, didn't actually take any photos. I don't think pics of newborns is exacyly /agdg/, though

That was a joy to read. Thank you.

I'm late, congrats on getting laid tho

You're already blogposting, don't do things halfway.

fuuuu

Way to show us all up you jerk. Congrats.

Rate this musical style and song? Also, does anyone have any advice for someone playing heavily with sampling?
picosong.com/wRYaL/

Congratulations on having a kid you fag.

forgot to ask the important question, buoy or grill?

Jk, congrats user. Hope you have a happy family.

Being fair, she had to use half of his source code.

hey thanks mang, and a uge congratulations to you and your wife
IP/ID change most likely bcs was on mobile dynamic ip w/prev posts'

I'm hoping for soon too.
I may have at least a tech demo by next demo-day for people to mess around with the OC donut steel mechanics my framework permits

i'll drop by mang

Congratz my dude, i hope your child rolled good stats. As celebration i'll work double as hard today.

Sorry, that wasn't my game, I just saw it on Twitter. It looked like an interesting effect worth sharing

Hey Holla Forums can somebody really explain the easiest way in Unity to implement setting saving with unlimited amount of profiles? Thank in advance

Normal way to implement setting saving:
Unlimited amount of profiles:
As for loading, you might wanna put them in a subfolder with just settings files so you can parse over all of them easier. Or use EditorUtility.OpenFilePanel to have them select the file manually.

So I'm dumb and can't find any answer to this question specifically with google. How do I get sprite animations to transition immediately in Unity 2D? I have it set to change based on what key is pressed for walking animations, but it waits until it completes a full cycle before it changes so there's almost a full second delay.

If you're using an animator to change sprites, you select the arrow that goes between animations and deselect "Has Exit Time", then reduce "Transition Duration" under Settings

That helped, but it still wants to complete a cycle or something first. Is there any way I can make it actually instant so it just switches mid-cycle instead of waiting? Or should I not be using the animator for this?

Forgot to have these settings in the webm.

Uncheck "has exit time"

**forgive me for my faggotry:
just wanted to say that i'm shocked with the amount of work you have to put for a game
i'm not saying "i thought it was easy!", more like "this is overwhelming", it's a lot of disciplines mixed together, like wanting to compose a song but needing to know how to play all the instruments on a intermediate level + being the conductor
it's maddening
and even if i make anything following tutorials, my work is shit compared to people specialized
i try to keep my morale up, but this is draining me
i don't know how the fuck you people can keep up with all of this, and just wanted to say:
i'm proud of all of you for not giving up. godspeed you glorious bastards**
sage for blogposting

and i somehow missed the spoiler too, great

what can you do? we need someone who can code in the near future and people who can model in the far one

You have to be willing to learn user, for me the learning process is very fulfilling. It doesn't matter if in the end I don't manage to finish my dream, at least I tried, and I had a lot of fun in trying.

Does anyone have an up to date tutorial to convert a Blender rig made with Rigify to a format that Unity's Mecanim can work with? all tutorials i found are really outdated and the rigs they show are different from what my rigify generated

generate one form the complex human and delete everything that mecanim doesn't have

What I do is export to .fbx, in unity rotate -90 on x axis and duplicate any animations you wanna use with ctrl+d because using it out of fbx is shit and makes the events unreliable.

Does a good free to play model exist, or should i just sell my multiplayer game for 10-15$ when its done?

Yes. It's called "be independently wealthy enough that you don't need to charge for your game".

Unless this is co-op, or some well crafted streamer bait, your game is going to fail 99% of the time. Indie games are incapable of creating a community because they need to hit a critical mass, and so making your game rely on a community means that your game will die rather quickly.

This.
There was this great looking mech war game that was multiplayer, died within a week despite being good and actually looking nice

there's literally no winning

What game?

...

Hawken, possibly? That's the only one that comes to mind.

should i delete everything that is not a deformation bone?

I hope I didn't get here too late.
github.com/lluisgarcia/UE4-Tools/blob/master/ue_tools_v1-2-4.py
At the top of the screen there should be a bar with "file", "render", "window" and so on. To the right of "help" there is a dark grey button with white squares and rectangles on it. Click that button and in the drop-down select "scripting"
Copy paste the github script into into the script section, which should be the leftmost window. Then go into object mode and select the mesh and the armature, save your scene, then click "run script".
It bakes animations, generates a new armature that is weighted to your mesh but only contains deformation bones then selects just the mesh and the new armature. When you export check "export selected".
If you want to go back and edit the animation/rig/mesh/anything at all, open up the version you saved before you clicked "run script", or Ctrl+z back to before you ran the script. DO NOT just delete the newly generated armature and expect things to work.
It works great for UE4, but I haven't tested it for Unity. I'm thinking it should be better than just using rigify.

Doesn't do anything.

Are you guys gonna be doing any dev over Christmas/Christmas Eve?
I'm honestly tempted just to take a couple days off to relax, but I know that's the start of the slippery slope so I probably won't.

We're on page 13.
New bread:

and a ton of bugs