Amateur Gamedev General ~ /agdg/ + /vm/

Sounds like it renders via a derivative function then. Sounds like fun.

With pixels it is easier to fake a slope since it can be blurred and will appear "smooth" which is not really possible with voxels, assuming it is rendered as cubes only.

feels bad man

Is the minetest multiplayer authoritative? If it's any good, I wanted to spend some time studying it

I've done the first two things. I want Natural Language processing, in particular I want to be able to have the words in a sentence tagged with their respective parts of speech and I want the structure of the sentence graphed in a way that I know what nouns, verbs, adjectives, etc are connected.
The toolkit can solve this problem and it has the MIT license which is great for me. Here's a demo of the toolkit doing exactly what I described.
timan103.cs.illinois.edu/nlp-app/
The problem is I don't know how to use the toolkit. It's not exactly a library, as all the tutorials I've looked up on how to integrate a third party library into a project don't fit the structure of MeTA. The thing has a makefile and standalone executables. It also has an API that is completely unhelpful in actually telling me how to integrate my project with it.

I don't know much about how the internals of Minetest looks like or the multiplayer system for that matter. I just can't stand playing games were everything is made out of cubes only. I know there is 1-2 mod that adds slopes shapes to Minetest but it kinda looks like those mods in question are not updated anymore.

Sounds like you want a parse tree of the text. Not sure if it's called something different in NLP literature, but that's the term in programming for how a program is parsed. MeTA has demo programs and doxygen documentation, so you'd be best off going through those. Since it's a makefile, and makefiles are a fucking pain to write, check if it has output targets for just the library/object files and then compile them into your program. You can just call make on the makefile from whatever build system you're using.

tfw you ruin three threads in one week

whats an amateur game dog look like

We still love you, clumsy.

should have added a link to the demo day thread as well

This isn't my model, I just wanted to see if I could load more arbitrary things into it 900,000 triangles

I only had video concept stuff to give, so there wasn't much for people to do with it. Aiming for a potential playable demo release for 5/5. Possibly sooner, but I might want to ensure it feels good and at least has a little content to toy with. Most the work me and my artist got done was in a little over a month, so it's a possible milestone to hit if things are relatively consistent.
Not sure if I'll use the /agdg/ wiki at some point to post some info here or there. I kind of want to set up a comfy, clean site on NeoCities or something for updates and stuff.

Guys

I think

might be right

What do you mean with the output targets? There's a build file that contains the executables I was talking about, but those executables are all command line applications.

The doxygen documentation is helpful, but not in the stage I'm at with the gaps in my knowledge I have. It tells me what does what, but it doesn't tell me how to integrate my project from scratch with MeTA. For example I want the sr_parser, but when I try to include that with the "#include " including the file path and all that and compile, make can compile the sr_parser.cpp file but has trouble finding everything that sr_parser is looking for. In particular it's looking for config.h, but there's only config.h.in, and config.h is generated by MeTA's makefile.
Speaking of which I should mention that when I initially got the project off of github I compiled and setup the whole thing with cmake, and now I can just call "make" to rebuild the whole thing. I am also doing this with my own project. So I'm not writing a makefile, I am writing a CMakeLists text file which cmake is using to generate a makefile. It's still a pain.

How would I go about calling the makefile from my build system? This may be the solution to this problem although maybe not. I'm using linux and, as I've mentioned, make and cmake. Is this information a search engine can easily find?
Also thank you for taking the time to help me with this.

what have I done

...

So it has a CMakeLists.txt, which is good, fuck makefiles, they're a pain. What I mean by a target is when you run make clean or make install. The targets are clean or install, and you'll want to find one that builds the libraries/object files. It's probably just the all target which is usually the default.

The good news is you can use CMake recursively, so include the MeTA stuff in your project and have an add_subdirectory call from your project's CMakeLists to MeTA's folder. When you run make now, it'll build all of your stuff and all of MeTA's stuff, and you can add dependencies on MeTA's stuff the exact same way you do for your stuff. The aforementioned config.h.in is configured as part of the CMake step and outputs config.h somewhere, so you can just include it as necessary and use include_directories or target_include_directories to where the outputted file lives.

If you don't want to pollute your project's CMake namespace with all of MeTA's stuff, there's always ExternalProject_Add, but that looks like too much work for image board help posts.

You can do pretty crazy shit with actual voxels and todays hardware minecraft is not using voxel rendering but polygons btw. Once i am done with my current project i will look into them and see how far i can mix up modern rendering techniques with old school voxel rendering.

hey, unlimited detail™ is back

I see you've also taken the red pill.
But why remake the proprietary engine when you can BE the proprietary engine?

Yes, that is good news! I'll try what you suggested and if I still have a seemingly unsolvable problem I'll be back in this thread, but I think you've given me enough information to find the rest and get over the smaller problems by searching online. Hopefully I'm posting progress next time I talk about my game in an agdg thread.
Once again, thank you. I really do appreciate your help.

Kek, the reason euclideon will never release anything is not because their unlimited detail™ isn't working, but because such highly detailed voxel visuals require multiple gigabytes to store the information of a single object. Who the fuck would play a game that's a terabytes big? That's the thing they never mention because that's a no-go even for their retarded fanbase.
The downside of Voxes are that they simply take too much storage if too detailed, it's also a big problem to animate them. However, they could prove quite useful if used for distance based games like strategy games where close-up detail is not needed, or for LODs in a poly/voxel hybrid where the undetailed far away objects are voxels and close objects are detailed polys.

unlimited detail™ cannot happen until we have storage of hundreds of terabytes, but Voxels aren't useless right now either IF used under the correct conditions.

Maybe PC plebians can't run unlimited detail, but anyone with access to big iron can handle it. Sorry that you're all stuck on inferior gaming platforms…

It sounds like we would need demoscene level programmers to get it to work on modern hardware.

You can probably mitigate a lot of the problems with the memory consumption just by compressing the voxel data internally, which would fix a lot of the issue, but literally modeling each grain of sand is why its going to take a stupid amount of memory. They might already be compressing the voxel data to save RAM. Any way if you want some interesting stuff about what you can do with voxels, consider the draw distance on this guy's rendering engine for a more practical example of what the technology can offer.

There's also probably some procedural generation going on here, which means that they just have to store an algorithm to get gigabytes of content.

This is not procedurally generated, in one of his video he says that he has created his map out of a heightmap. He explains that at the level of detail he is on, it takes 4.5 GB of RAM to store it.

Interesting, something like that is what i thought of when i meant polyvoxel hybrid. Basically:

Nope. It also shouldn't take too much storage since the voxels are very undetailed close up. See pic related.

Alright. Still pretty neat.

I am retarded and was wondering how games like zomboid do their maps?
Do they just spawn 6 million tiles and then draw them? Isn't that wasteful?
How do tile based RTS games do it? For example sudden strike, wouldn't they also need millions of tiles/sprites just to draw the map?

Seeing this demo day makes me want to finally make something. i'm one of those fags that convinced myself i'd have to become a programming wizard and do everything from scratch instead of using an engine like everyone here is doing How hard is Löve to get into? I haven't really tried Lua yet.

They would only draw the tiles that are supposed to be onscreen, wouldn't they?

What happens is the data in the maps are organized in a way that allows you to easily find out which polygons are inside of the viewable area. This is known as culling. Usually frustum culling, back face culling, and occlusion culling are used. Ok, so in my image about back face culling, the idea is that when rendering a convex shape, this implies that any faces of that shape that are not facing the viewer do not need to be drawn. So, in the image the back faces that are not visible to the viewer are culled. In my image about frustum culling, the way that works is generally that all of the objects in the level have a bounding box, and this is checked against the view frustum so that if an object doesn't intersect the frustum it is not rendered. Finally we have occlusion culling which is just culling that prevents any objects that are behind other objects from being drawn. So in the image all of the red shapes are behind the blue shapes, that means we don't need to draw them.

Now there are also a lot of other methods for culling like the PVS method, (Potentially Visible Set) and the idea is that, if a player is in room A, and you can only see room B from room A, then whenever you are in room A you should only consider rendering things that are in both rooms B and A, since you can't see anywhere else from room A. That means that room A has a potentially visible set of rooms B and A. So, before any other culling is done, since we know the player is in room A in the first place, we can cull the rest of the level and only consider rooms A and B, and then we apply back face, frustum, and occlusion culling to the scene to further cull polygons.

OK so this was just a quick explanation of a few of the techniques used, there are a lot of other tricks and stuff and there are a lot of methods to achieve these types of culling, but this might help you understand a little bit better.

Even if you cull things outside the screen you still need to create millions of tiles/sprites because every tile needs its own sprite right?

No, what happens is the sprite is stored in memory once and then just copied onto the screen as needed, and you can make as many copies as you want. You only create as many instances of the sprite as you need on screen, so if you have like 10 tiles on screen that use the "floor1" sprite, you just copy that 10 times using the same sprite.

If you convinced yourself of that you're an idiot. Just like make game. LÖVE is a pretty nice framework, Lua's also solid but if you want OOP/classes/inheritance you have to build it yourself. Just look it up online, shit's like ten lines of code.

Two things.
First, all games massively re-use assets. Why would every tile needs its own sprite, when you can instead give each tile a type, and assign the type its own sprite? Suddenly you are reusing the same sprite for the quadrillion tiles you have!

Secondly, graphics are made so you batch the shit outta everything. It's pretty standard in 2D engines to draw all the tiles on the screen through a single draw. The cost is only marginally higher than just blitting the screen.

Is there an easy way to batch draw in SFML?

Looking through SFML API, Nope.

However, the API exposes shaders in a way that lets you do it. It even comes with a tutorial on how to do a tilemap!
sfml-dev.org/tutorials/2.4/graphics-vertex-array.php

Nice, i'll look into this. Thanks user

isn't gabe the ultimate neck beard

I bet one of these days, Valve is going to launch a space faring company too

[heavy breathing]
This font looks so good.

Realized that I was so tired yesterday that I forgot to post this. But here's some fancy lighting effects I did. All the data's stored in an 8-number table in the room's data. You specify the X and Y coordinates, X and Y scale and R, G, B and Alpha values for color, and the draw event handles the rest.

What's this fascination with voxels? I think they make things look like shit.

...

Hey, I'm using SFML to do a raycast engine and I'm running into some problems, namely all of the fancy types that don't convert to each other and whatever. That's pissing me off. Should I just make my own vector structure and say fuck it ?

Did you just wander in here from reddit?

twitter.com/tophiadev88/status/959295138789974016
One of the Godot devs retweeted this.

Ricochet?

Demo day has inspired me. I think I've sat on these long enough. It's time to just like make game.

Makes me wonder how much it takes for Notch to run his new voxel-based project. Mostly because the man probably has an ungodly supercomputer that's cooled with the distilled tears of orphans.

He's running it in a web browser which means that it's going to take more processing power than it should in the first place.

This is as bad as the previous one but in a different way.

at least I don't think this one is official.

I bought a few of those courses during the new years sale. I completed the majority of the official Unity tutorials as best I could seeing as how they were all outdated as fuck and didn't know where to turn to next. It's painfully slow going since I have a full time job but I'm getting there and genuinely feel like I'm finally learning. I don't like how slow the blender instructor goes. He repeats himself too much and just generally talks too slow imo but he's competent at explaining how/why shit works at least.

Hoping I can participate in these threads soon with some actual content.

It's a better idea to just google what you need. Like, "movement", for instance. See how a few people do it, modify as you go along, optimize as you get a grasp of what you're doing. That's how I've always done it.

No it's not. They went with an entirely different design that sucks.

This looks pleasing to look at.

Based on this tweet, it doesn't seem like it's official, just that the Godot logo designer happened to make this. But I agree that it looks like shit anyway.

i really hope someone will come along and make a better looking mascot.

i think you are right about that. i was also under the impression, that it now was the official mascot and i think i even read that tweet. very diplomatic, so no one gets upset, but the engine doesnt declare the first shitty mascot to be the official one.

also on the topic of godot: thanks to the guys that explained mouse input in godot to me last thread.

I'm constantly searching up ways to do things, I feel like that's something every dev will always have to do, but it's not a perfect system. Not every question has an answer. Not every question has an up-to-date answer, which then takes a long time to work out what's defunct, needs replacing and how. And the worst offender of all, not every answer is written in normal fucking English.

You can also do some pretty interesting effects with voxels

Looking good you fucker. Did you release your demo on demo day?

...

Are you talking about other languages, or people who don't know how to communicate like humans? I've met guys who can program circles around me, and can speak well enough, but for some reason, can't form a coherent written sentence in English to save their life. I suspect these people are why technical writers exist.

...

let me tell you about blender, son. as someone who's dabbled in it for years (and had to relearn things several times over when I've come back to it) the best way to learn is by doing. There is such a vast amount of functionality, both large and small, massive and fine, commonly useful and extremely situational that someone trying to teach it to you by some voiceover sitdown lecture style video or something is going to be horrendous and painful to learn.

Don't do that. Go use textual blender tutorials that go over how to make a certain something, like a watch. The basic watch tutorial is, off the top of my head as an example, a good way to learn the basics of building a simple mesh, how to use arrays, and at the end of it you'll have a snazzy looking actual object you built yourself you can show to people. Here's the kind of thing I'm talking about blendermesh.dasya.com/creating-a-fancy-watch-in-blender
and pic related it around what it should look like when you're done.

So in summary, learn by doing when it comes to blender. With large overarching functionality, find tutorials (text) and go through them. With small functionality that is highly specific or situational, just google it. Do NOT go along an audio tutorial wherein someone is speaking at you and dragging their cursor around, it is much duller, boring as fuck, and you don't absorb the information nearly as well.

The latter. The monkeys that, for all their intelligence and understanding of scripting languages, can't form a single coherent sentence have caused me many late nights of de-spaghettifying not just their code but their explanations on what the fuck to do with the code. I can respect that English is not the native language for an exceedingly large number of programmers in the world, but holy shit people at least proof-read what the fuck you just wrote before hitting send.

That's good advice, I sucked and I did things. Now I still suck at it but not as much as before.

I'd watch it.

Agreed. There's also some essential add-ons to enable.

There's also a script that estimates animation render times, which may be useful depending on what you're doing. Sometimes it's hard to gauge if you should stick around for it to finish, or go make coffee. Why this isn't on by default, I don't know.
wiki.blender.org/index.php/Extensions:2.6/Py/Scripts/Render/Render_Time_Estimation

This is one of the things I was having issues with the other day, and more or less ended up spending all my lead-up time to demo day trying to wrestle with. Specifically, because I have a memory leak that jumps up by like 80 kb or more every time task manager updates, and I know it's from how I handle my draw calls.

When you call Sprite.Draw in sfml, it copies the internal quad for the sprite to the screen, eg right off to the graphics card (this is where the bottleneck is). So, even though Sprite exists as a convenient wrapper to draw things, it's shitty to use as a batching object because you still have to generate the points and transform them, then assign them to vertices and batch them manually.

And that's where my issue is; I could take a Sprite object and reuse it (pic related) but this feels really clunky, and at the same time, it's wasteful to generate a sprite for every actor/drawable object, even though they have their own positions and transforms (eg some are just rotated sprites). It feels like I should just maintain a massive Vertex[] buffer that I don't entirely use, and then just manage all my coordinates and shit on that, but I have really no idea

Guess it's time to backup my important files to an external HDD and factory reset this fucking thing.

i thought i was having the worst day of my life until i figured out i was rendering the screen 8 times.
now i'm acting like a retard and experimenting with scaling from low resolutions for sweet sweet performance gains(though optimizing how i'm sending objects to the gpu and getting a bvh would probably do way more)

forgot pic

I thought this similar with the Godot logo. I wish they had a mascot by Tyson Tan. They'd probably be able to get something to work.

I was wondering why your thing was so slow , it's good that you're figuring it out

Ultimately you're moving things from CPU to GPU which incurs synchronization. Batching just makes it cheaper.

However, you're probably better off just implementing some simple culling. If you're drawing 2D sprites in SFML yet somehow that becomes your bottleneck, that its very clear you're not doing enough culling.
The best part is that efficient batching requires some culling as well, so it's not like you can avoid writing culling code by just slapping on batching.

You guys are all misleading user. The six million tiles aren't all in memory. They're sitting on a hard drive, and they load a large chunk of them into memory and then, afterwards, only draw the section of that chunk that's onscreen.

i can't shake the feeling that all the artist did was trace a Chaika face onto a standard body outline.

Well basically I'm looking for the best was to manage, batch and prepare the data before I send it off, because spamming new objects every draw call every 60 seconds seems wasteful Or does C#'s GC reclaim shit that was declared in a function automatically?

I have no bottlenecks yet, and I'd be rendering a 16x16 animated tilemap to a 512x512 texture with an equal amount of actors worst case, so maybe only 512 total draw calls plus the UI. But like I said there seems to be a memory leak with making a bunch of objects so I'm proactively trying to stop that from becoming an issue

Pick 0 or 2

I'm trying to break this into chunks and it took me a few hours to even get to this point.

C# will probably do it too as long as you don't make extra references to the object, but that depends on the GC.

Then don't bother with batching until its a bottleneck.

Then you need to fix the actual memory leak, and not minimize it. That said, I have no idea how the fuck you managed that in a GC'd language.

why'd you break it?

Rebuilding the entire mesh the whole thing every time there's an explosion took way too long. The seams aren't supposed to be there.

This has been a great weekend

that's pretty smart actually

Well thats the thing. The GC will run eventually but you can't make any guarantees about when or if it will run; likewise, manually invoking it is bad form. Thus the best design is one that is elegant and works cleanly.

Having a cached object means I won't have to allocate objects every frame, only change them.

My mistake, I meant it depends on the compiler*.

If you don't reference an object, the compiler might arrange code in such a way that its not an object that requires GC to begin with.

nvidia's driver on linux has some odd behavior like a bad gc that causes it to allocate _everything_, but i'm actually leaking memory here. oops.

sounds like you get to take the day off

nicely done user

More like a whole week unfortunately

So how's this work? You have a light at a point and do two or three raycasts against an entity's position (which is just a square box)?

How are you clipping off the texture?

Turns out it was actually a lot easier to extract the relevant data from a sprite than I thought. Good stuff, I can do this pretty easily.

Minor observation is that I have to use quads for my tile map, since each tile is discrete and probably won't share a texture with its neighbours

can I get fucked over by buying assets?
Seems like a better option to buy level creation shit for 65$ instead of paying a 3D modeler over 2k bux

If you're a 1MA, the best thing you can do is keep your scope and assets reasonable. If you're buying premade assets, that's dumb because you're wasting money and using things that everyone else has already used.

The best longterm thing to do is learn it yourself, or get a partner to help

I have 1 artist and a programmer as partners, need to pay them around 200 bucks. I was expecting to be able to do 3D Models, I can do objects just fine but turns out I can't model a fucking cave so I wanted to buy this assetstore.unity.com/packages/3d/environments/dungeons/cave-dungeon-toolkit-79550 I can't afford paying another 200+ bux for a modeler so I thought this is the best solution

You can use a texture atlas for that. Just keep in mind they produce some artifacts along edges when you have trilinear/anistrophic filtering or MSAA.

is it worth it to do vertex transformation and rotation on the gpu and not transfer vertices every frame, there's some potentional optimizations from doing it on the cpu, since you can make triangles into 1 point and 2 edges, and save two subtractions per ray-triangle intersection, but with higher poly counts you could probably manage things easier just by transforming things on the gpu

How would that help? I'm limited to Points, Lines, Strips, Fans, or Quads for the geometry type. They would absolutely have to be different vertices, because the middle points have two sets of texture data (the right side of first quad, and the left side of the second quad)

Am I missing something?

Oh, my mistake. You would have to use quads. Sorry m8. You don't have to worry too much about bandwidth since I assure you; the cost of bandwidth is far lower than the cost of synchronization between CPU and GPU.

Yeah, that's what I thought. Like I said, minor concern, but I just wanted to confirm that I was on the right track.

Although I do want to mention that if you set your texture to be repeating, you could utilize that to merge adjacent quads.
The problem with that though is that you then need to make a 2D texture array object on the GPU, which SFML does not appear to expose.

Is there any reason as to why my scene is suddenly dark after restarting unity? I didn't change anything

Think about it this way.
You don't know how to make it, so you buy that. Suddenly you have an incredibly high quality asset art style that you need to replicate throughout your whole game.
Are you sure you can do that? You might as well learn

Great point, now I don't know what to do

Have you played Dear Esther and Hl2 episode 2? Possibly some of the best looking caves I've seen

You could start from looking at how they did it

My issue is that I don't know HOW to model it, I tried doing it with meta balls but the topology just looks like horseshit and because of that the applied texture has really weird looking spots

(checked)
Thank you. It was a pain in the ass to get to work, and took way too long. Gonna work on some simpler stuff now, thankfully.

That's more or less how it works.
There's not really any ways to remove parts of a sprite in Love2d, so I draw the light on a canvas, then draw a black shape on top of it using the calculations from earlier. Due to the way I had to implement alphamodes, to make the transparency of the lights work fine, and wiping the canvas to black, it erases the color black, so if you draw the shape in black, it erases it as well. (before I did this with canvases, I just had a shader that erased the color white, that did the same purpose)
Since it's drawn on a canvas before being rendered out, it only erases the section from the light, rather than the whole image.

Here's a video of it with the shape being drawn in white.

It says on his personal website, that he does mascots for free and open source stuff for free.
Not sure what anyone not involved in the development of godot would do here.

Are the light sources themselves just drawn from a premade gradient texture?

That should give you a good basic shape. You can continue to modify the base mesh and see what it will look like when subdivided. Then you need to add stalactites and stalagmites, which are essentially pyramids with another subdivision surface modifier.

my dream is one day you can do everything out of voxels but without them looking like cubes(underlying stuff is voxels instead), so everything is destructible and buildable in as many ways as possible

Also, adding the decimate modifier (on the "Colllapse" mode) after the subdivision surface modifier can create an element of randomness and it reduces the polycount.

You better get crakin' if you want that dream to become a reality.

Yeah, seemed like the simplest way to do it. It's actually only three colors, with dithering in between.
I just store the position, scale and RGBA values in a table like so.

Hope everyone had a good demo day. Having people come in and out of the server sporadically through the day was pretty tiring but the feedback was great and the support, despite how bare bones the current state is, was super motivating. Now to work even harder for the next one!

When I made something similar in XNA, I had a 256x256 texture for my lightmap, where one pixel is one tile, then scaled it up with linear interpolation, to 32x larger (view culling would crop out quite a bit of it)

I remember having something like a light curve for the attenuation, and it supported multiple colors for lights (eg it could originate as white, fade to orange, then dark blue, then black) as the distance increased. It was pretty neat. Not ideal, though, but definitely something I want to explore again.

Is 1280x720 the "default" resolution now? Most devices, even phones, seem to use it. It looks very divisible and scaleable (eg 640x360x2, and 16:9 widescreen).

My dev laptop has a height of 768, but between the taskbar and title bar, it ends up being about 64 pixels of dead space, which eats into the bottom of the window, so it's basically unplayable in a window, so I'm just working with 800x600 for the moment

oh jesus fuck

They're rendering the representation of a data set of voxels. That representation being a polygon mesh generated by marching tetrahedra.
This process utilizes voxel data to contour an isosurface, via isovalues (densities, which are volumetric data elements; i.e. voxels) and generating a polygon representation of the voxel data set.

voxels can just as easily mean the representation of data, and not necessarily the exact rendering technique utilized.
In the case of rendering actual voxels (like billboarded SVOs) isn't a mature enough approach for games (not exclusively targeting top of the line hardware) at this point.

direct from (((wikipedia))), but good enough:
So, voxels can also mean volumetric data; i.e. data which represents a volumetric shape/pattern (such as a pattern generated by a noise algorithm).
F.e. the data acquired via an MRI scan can be reduced to a set of volumetric data (set of voxel data, density values, which are generally called a "level set"), and for viewing they're represented with a polygon rendering technique.
However, the actual polygons are built via a contouring algorithm that can auto-magically generate a polygon mesh from the voxel (volumetric) data (such as marching tetrahedra/cubes, and many other contouring algorithms).

they're not rendering "with" a derivative function, but it can be in the process of outputting the data required to render the actual polygons from the data set of voxels/isovalues (and most likely is as it's more exact compared to 2d/3d interpolation).

that brings backs some nightmares
easy enough on the cpu, but hard on the gpu; though it's easier with primal scheme contouring algorithms compared to dual scheme contouring algorithms.

it's a common technique, even minecraft did it with their voxel data

Create the basic shape of the tunnels, and other cave appendages; then apply noise to add rockiness/bumpiness to final mesh (probably some ridged multifractal noise; use in sculpting mode, get a noise brush for blender).
This will add some organic-ness to the final mesh, and will alleviate you of a lot of monotonous work.
A cave is also prime for for harnessing noise when texturing.
Use a pre-generated 2d noise texture to change the actual lookup coordinates of the cave textures as to add mineral striation, per the changing value for lookup in the noise texture (like Y coor of pixel in world space), and the same can be done with a normal map to add crevices and such; which will have an "organic" feel due to using the patterns that are ordered from far away but random'ish up close; which lots of different noise algorithms provide.

What, you've never seen a matrix before?

no it's just that something about having that many assignments in a row hurts me a little on the inside

Can't your [s] series of variables be put into an array?

that's an opencl float16, not my type

I am struggling to make his hair

What frameworks/engines do you guys recommend to stuff like the Raspberry Pi? I was think of making something in opengl through either a python or lua binding.

Funny you should mention. Here's an excerpt from my game's code at the moment. Relax, it's just for testing purposes.

Depends. 1280x720 seems to be a common resolution for VNs. For 3D games on PC or consoles, people tend to run them at 1920x1080. For 2D games on PC or consoles, it seems to vary. I've seen games use 480x270 because it scales into 1920x1080, but it doesn't scale into 2560x1440. 640x360 scales into both, but 1280x720 doesn't scale into 1920x1080. I don't know about phone games.

Asus or Acer?

:^)

What is this and why does it exist?

i just made like heavy breathing noises for 2 minutes after seeing this

Who needs hair anyways

Anybody?

To stuff the raspberry pi?
The fuck do you need to make?

From my experience, 1920x1080 and 1280x720 are the most common. 1080p has basically become a standard, but you shouldn't discount 720p, which is still prevalent on low-mid range phones from up to a few years back. Though even there, 1080p is overtaking.

The resolution you, a30945 , have is 1366x768, right? It used to be a common one, but nowadays isn't as big outside of budget laptops.

With PCs in general, 720p, 768p 1080p and 1440p will cover most machines, but it's important to make sure that other aspect rations, mainly 4:3 and 16:10 are supported.

For phones, you're fucked. For example, I have a phone that's 1280x720 but it has a 96 pixel bar at the bottom, since it has no physical buttons, so it's actuall 1176x720. The iPhone X has a resolution of 2436x1125, for an aspect ratio of 812 : 375, and a bunch of new android phones use some superwide ones of ~18:9, ~18.5:9 and shit like that.

So it's best to just, at boot, calculate the display, internal aspect ration and base game camera and stuff like that on that, like I do.

A game that can run on on at least a RPi.

That's not really down to the engine but rather how you optimize your game.
If you really want to go full performance then no engine and hardcoding would be optimal

Is the Pi powerful enough for snes games? Is there an engine with such low min requirements? RPG maker comes to mind, but I don't know.

At least powerful enough for quake.

That's really cute.

That doesn't sound bad at all.

Correct. It's a refurb Thinkpad T420

Is 16:10 the more common widescreen ratio? I thought it was 16:9 for the longest time.

Either way, I'm not concerned about phones in the slightest. I've more or less settled on a 512x512 region for the main gameplay area. The reason for this is that I can make a drawable region and keep it to a power of two texture size. Secondly, it gives me a 16x16 tile view, which should be plenty (original game used 25x18). I would like to upscale it 2x but it is important that you can see many tiles at a time. Anyways, this is roughly what I'm going to do for an 800x600 resolution, and if the game resizes to a different window size, I might have to pad, rather than scale things

You could use that sony vector library someone linked

If you're using C#, you could try extension methods. It's not ideal, but I've already made a few helper functions for SFML myself.

You could treat every sprite as just a transformation matrix applied to the standard quad in the vertex shader. Then youd only have the vertex data of the unit quad for sprites. I'm doing something similar right now but it's only to render a static image so I make many draw calls.

16:9 is by far the more common. 16:10 was a standard for a brief time between 4:3 and 16:9, and some people still use it.

Tried this out, the tool is powerful but it breaks my lighting for some reason
It's also near abandonware

Trying Realtime CSG now

Do you guys think I should start with 2D games and work my way up with 3D games?

If you're using an engine, 3D games require knowledge of more complex trigonometry, that's about it really.
Depends what you feel comfortable with

Make the UI work on 320x200, then just scale it up if the monitor supports higher resolutions. If you try and play the resolution guessing game you will just get a bunch of complaints from everyone that you didn't support X special resolution, it defaulted to resolution Y and that didn't work, etc etc so just actually query it from the OS so that you can never go wrong. This is how some of those 90's games just work with modern monitors and modern display resolutions directly out of the box with no problems.

So my game just gets a list of supported display modes and I can be assured that in 2070 I will support the 16k alienware whatever.

It can run Q3A

That is probably good if you aren't sure what to expect when making 3D games, since in 2D you have to solve simpler versions of the same problems.

Anyways, I made a QuadBatch object that holds a pre-sized vertex array, and tried about half a dozen different ways, benchmarked, to see what implementation made the most sense, so maybe I can finally work on something relevant for once

I know, there's a way to get a list of supported resolutions and color depths, but my problem is that my art looks absolutely awful if I use non-integer scaling, so I'm just trying to find a nice comfy resolution to design the base around, and then avoid scaling, but using padding for any arbitrary resolution. I might do some tricks that only scale it by integer amounts, and then pad it otherwise.

imagine if some enginefag would put out a good engine instead of making a one off project with it

Is this an issue with your texture filtering or the actual scaling? Try using nearest neighbor filtering and not bilinear filtering if you aren't doing that already.

Also if you are increasing the resolution of 2D game, this is ironically more difficult than in a 3D game because you have to scale the whole view-port yourself, you should never display your 2D objects in an aspect ratio that isn't 1:1 or else you will just stretch the whole thing, so you have to ensure that yourself if that is what you mean by "non-integer scaling"

So, in a tile based game, you would have to display more tiles at 1:1 instead of scaling your tiles to 1:2 or whatever.

I was recently introduced to Henry Miller’s eleven commandments. I've adapted them to game development for everyone here.

Here's the source
stevenpressfield.com/2012/05/henry-millers-eleven-commandments/

Right so I'm finding myself having a lot of trouble building levels even with tools to build those levels directly inside unity

What's the best way to prototype levels? Using 3D modeling software?

Nothing creates a shittier game than working on it without passion

Which you also mention in

So I dunno

Level Buddy for Blender is pretty good, especially if you already know how to use Blender.
matt-lucas.itch.io/level-buddy

He's talking about being in a good mood. Don't keep working even if you feel like you could, stop and rest.

This might be good actually, I'll try it out

Most of making the game is just grinding things out. It's not like you can expect to be excited for thousands of hours.

I have filtering set to NN where possible. I have a separate render target, set to 512x512, where I will draw my playing field. I then render this to the window in some fashion, hopefully preserving the visual quality.

Mostly, it's just me being fussy.

Also if anyone thinks like it would have some value, I can turn these rules into a repost-able image.

Makes sense, honestly for these kinds of games I feel that OpenGL or w/e adds a layer of uncertainty to the whole thing compared to just software rendering. Since OpenGL isn't so precise about where the pixels end up..

Anyway I am sure you can make it scale to almost anything you want, keep it up the work and you'll get it

i'm gonna make a bideo game

...

...

that's a good lookin tree sway

If you don't mind some constructive criticism, the trees are shaking too fast. Unlike grass, they would surely have more weight to their movement.

yet again i'm left wondering why this shit isn't built into unity. you'd think it would kind of be a requirement

Nice

yeah

I meant Portal, but well, what do I know?

user, you can't just one and one like that

...

What?

It's spam.

That's really fucking gross

So you mean that isn't an actual semi-cute 3d girl who's just really into game development and ones to date someone who makes vidya?
Nah, who am I kidding? cute 3d girls don't exist.

Even if you find a cute 3D girl, she'll probably have some extra flaws like "sad all the time" or "wastes money" or "talks"

rpg maker is notoriously low performance for what it technically offers

It's not terrible. I'm pretty sure I used to play some shittier RPG maker games on my PSP.

I did not. She wasn't ready for Demoday.
However at the rate this is going, the update will be out soon. All I'm really waiting on is some sprites for the Cybermancer, a sprite for the Doomslayer in the map screen inventory, and some more polish.
and a trailer.
Have character portraits for character select

What?

anyone understand shaders?
i've got a water shader, mostly wanted it for the refraction/reflection
the waves are scrolling in the world coordinates and i want them to be scrolling along the mesh UVs instead
anyone know how to change it?
i'm almost certain it's in
float4 wpos = mul (unity_ObjectToWorld, v.vertex); temp.xyzw = wpos.xzxz * _WaveScale4 + _WaveOffset;
but i have no idea what to change it to
pastebin.com/rcHy6kdU

Really digging your menu stuff.

Wasn't here for it, did a bunch of taking things apart and putting back together. Making a video today on some progress, then I have a lot of coding and debugging to do. I also need a new map, plus gotta learn how particle systems work.

Wouldn't that just be an offset to the mesh UVs? Not too familiar with how Unity does shaders, but some cursory searching shows that TEXCOORD0 is likely the texture coordinate for the fragment, so adding an offset to that might be the way to go.

Yeah use model coords instead of the wpos.xzxz there in the last line I think, the v.vertex looks like the pre-world-transformed vertex info so use that.
float4 model = v.vertex;
And then use model.xzxz or something like that, I'm not familiar with unity or its types. The basic idea is use coords that havent been multiplied by the ObjectToWorld matrix there.

GO BACK

1) use potions as poisons
2) profit

Knocking out some of the demo day tasks.

Nice, the keybindings work now.

dumb question
when i change the blendshape value, the entire thing translates along Y
same thing was happening in maya, but it got fixed there by unfreezing translations
why is it still happening in unity?

you using blender right?

That's a bug.
I happened to find it a few months back. I don't think there are any solutions.

...

damn, I meant one and done

What did you do exactly?

Run some tests on your code before scrapping it. maybe you're a programming genius?

Top Kek

Got some progress

...

Yes, but im not going for ultra realistic hair, thanks anyways

...

heads up, i'm interested in people integrating a raytraced renderering option into their games if it's possible, preferably stuff with low poly counts.
if you're interested, nothing's going to happen soon, since it's still pretty much in it's infancy, but when it's featured enough to do things like texturing, transformation and rotation, and hopefully soft shadows and ambient occlusion an implementation might be worth it.
i can't actually maek gaem myself so i need something to ducktape this to.

its*
i'm sorry

...

why not play the demo day games?

Soft shadows are one of the reasons people use pathtracing instead of raytracing, there is no feasible method of implementing them with raytracing at realtime speeds unless you only have static lights and scenes whereas in pathtracing they're "free".

Either just start playing some game or fuck off away from the computer, and start doing things that'll reignite your enthusiasm for games and gamedev.

Stop worrying about optimization at this stage. This attitude is a pestilentially common mental illness especially among C types and causes nothing but delays and bugs in the end. First make it correct, then maintainable (you are going to read things more often than you write them), and then profile before you optimize anything. You can get away with badly implemented O(n^7) algorithms if they are only called on small datasets once in the beginning. Most likely the biggest performance boost will lie in some inconspicuous function that just happens to be called all the time, but this will not become clear until the game is complete and your call graph is close to how it will look in the end.

well:
a) i was planning to implement multiple forms of tracing, because idk
b) can't you shoot extra shadow rays to hack this up
c) there's a really dirty hack where you just add a random noise vector to the shadow rays which acts as an incredibly dirty form of soft shadows
oh well fuck me

Extra rays for an area light technically works, but results in banding shadows and massive slowdown with more lights or area. Random noise means you get the same downside of pathtracing (noisy output that converges over time) without the upsides of it.
But if you're going to implement both path and ray tracing anyway, there shouldn't be much of a problem anyway. You'll see what works, what doesn't, learn a lot along the way, and (although unlikely) might come up with some novel new solution.

i am downright obsessed with implementing everything i can
i have problems

well whaddya mean by that specifically? I've posted gameplay in the past if that's what you mean

This pisses me off and I want people to be angry too archive.is/fok4w

he could just like have an array 0-40 and round the result from rand to pick a tile

Would it be stupid to store an array of pointers to my octrees in the "manager" class for quicker and easier manipulation? I hope I'm not somehow defeating the point of using octrees in the first place

There's a ton of garbage programmers out there who write tutorials and post them for new users with horrid programming practices that absolutely no one should use, let alone TEACH to others. Like using a try/catch statement every collision at LEAST once per frame, which throws and catches an error if they don't have a "damage" component. If there's no error? Then you can access it and take damage! This was part of a paid video series I pirated. Just about any game development book or series out there is complete trash that no one should be exposed to, yet people still rapidly churn them out. I'm better than these faggots, and I know that I don't know enough about development to release a book worth reading.

This is why you go to college for this sort of stuff. I thought college was pointless for a long time, until I had to teach myself things. People in the academia may have their own issues and hangups, but at least they constantly push each other to be smarter, better and efficient.

The reason for this is because these programmers want their blogs to be on hacker news and reddit. Both of these sites don't allow self-promotion, which means that everyone just disguises all of their self-promotion as "tutorials" which means that every single thing there is presented in the context of teaching when it is really just trying to promote itself, and not trying to teach you anything.

...

I can't believe how ugly this character is

This inspired me to draw a Love2D-tan that hopefully doesn't look as bad as this.

Yeah, no offense to your technical execution at all, but this design just makes me think of nerd girls back in middle school - bloated, chunky, oily face and greasy hair, perpetually smells of sweat and wears retardedly mismatched clothing just like the character in question.

I just can't look at it without cringing and thinking about the archtype. It's more disgusting to me than even fat male nerds. It's fucking revolting.

cute

I prefer Holla Forums's take.

How about making the software mascot not some dumb fucking animu girl?

I love the style, again - but the fucking design is just gross. What's with the obsession to give her a disgusting hairstyle and nerdy look? Does Godot really make you think of something lame, nerdy and unattractive?

It's either a girl, a trap or a furfag fantasy. You should know this by now.

I know and I'm fucking sick of it. Why can't it be a robot or some dude, or just not fucking have one how 'bout?

Cause making it a dude will bring about accusations of THE GAY COMMUNITY, and people are unwilling to make a non-fuckable-looking robot.

Because despite thinking that they're super special elitist snowflakes, anons really are incredibly simplistic and only like a limited selection of autistic things.

This, too. You can't deviate from the norm too much or you'll be labeled a degenerate, tumblrina, a redditor, a faggot or a furfag.

Sage advice.

It just feels like we've done this all before. We just post the same shit or the same kinds of shit day in day out. It's always the same mascots, the same threads, the same posts.
/kinda blogpost. I'll just leave you guys to it I guess.

but clearly if you do anything different you're not from here so get out

I blame console/os tans and Vivian.

Nah, way before even that, we had board tans.

Godot-tan seems to be hated by most of the developers here. Are we super special elitist snowflakes?

Yeah, Holla Forums and Holla Forums's were male, /x/ was female, and Holla Forums was whatever you could use to describe an ocean of piss at that moment. I don't remember the other board's tans. /vg/ seems to be the only one to have stayed largely unchanged.

The board mascot for 8/tg/ is a rotund necromancer.

When's the next demo day? I'll have started working on my game by then, I promise.

No, in my personal experience /agdg/ has been one of the more levelheaded circles around these parts, probably because people who have the patience to sit down and do something for extended periods of time are less prone to be complete tards. I assume most people don't like her because she's just that fucking ugly.

It's in the wiki
8agdg.wikidot.com/demo-days

You would think so, but then there's that guy over on loomis who's spent the last two months tracing pictures of one girl's feet

Thanks, I must never read the OP because I didn't realize we had a wiki. Time to set a goal that I can realistically accomplish in 3 months.

Well played.

I mean was he trying to learn how to draw feet better? Practice makes perfect.

No, some people are just retarded and have foot fetishes

I don't think having a fetish makes you retarded, user.

having a foot fetish sure does

Do you prefer pic related?

I do not have a foot fetish, but having one does not make you retarded

They're clearly trying to get her head to be the rough shape of the original logo, and are trying to use the hair to do it. I'm no artist, or character designer for that matter, but it seems obvious as hell to me that if you wanted a specific silhouette, but couldn't effectively get it with the natural features of a humanoid character, then you would add something to it, like a hat or other accessory. I don't see how people are going for "blue robot video game girl," but not thinking of a design like pic related instead of what's been made so far. Just add more of those raised bars, and you get something roughly like the gear-shaped head of the Godot logo. Obviously, you can't just lift a Mega Man character, but my point is that the aren't even trying for alternatives to "It's her hair," which is a shit choice in this case. There are lots of other possibilities that could work.

There's not really anything wrong with that. If a mascot robot isn't humanoid, like WALL-E, it not being fuckable is fine, but if it's humanoid, it should be what we consider to be physically attractive, or at least not ugly. The primary purpose of a mascot is to entice interest by being an attractive face, usually by being cool, cute, or pretty. Being ugly is almost always directly contrary to a mascot's purpose.

I think it's because people here are working on the same goal and so we're all connected by our mutual tribulations.

Also, I'll say this: godot-tan has to be a robot who's jaw has the grooves of a gear.

This. I always thought you should just take the robot head from the Godot logo and use it to make a helmet for some Deadmau5 kind of mascot. Trying to get a cute girl out of the Godot robot's silhouette is a lost cause in my own stupid opinion.

But what if it's a cute robot with mask of a godot robot sorta like this?

*cute girl

I'm very tired

...

I just wanted to get mad at pajeet-tier devs making tutorials with people

Other things that could have that shape and aren't blatant copyright infringement include a hairband, a tiara, and a hardhat.

But there's already a qt animu robot grill with a gear-like jaw and his name is Raiden.

That's fine, just stop trying to make her fucking hair work with that silhouette. It's not gonna work. There is no hairstyle that conforms to that shape that is cute.

I posted this last thread. I think it's relevant.

this is acceptable

Oh, that's good.

Last two months have been very productive for me

So are there any good game development resources? Game programming patterns seems to be the only one that wasn't written for retards trying to make mobile shovelware. The vast majority of which I'll never use, but I think being aware of them has improved my code. It's in C++.

pdf related

...

It's all autogenerated :^)

Might be cute, but the godot-ness is just in her clothing instead of her.

This PDF has decent reception, and I've skimmed through it. I think anyone with a decent beginning foundation of C#, an interest in gamedev, and likes to read should at least skim through it. If you understand trig and the basics of openGL, it's probably too basic for you.

just make her slightly younger, give her four highly successful (yet mildly retarded) sisters and you're set

...

See? This works because it's a hair accessory and not some idiot giving her 4 large cornrows or some stupid bullshit.

I've read this book front to back three times and would recommend it to anyone using or interested in C#. It's comprehensive and well-edited, and I'm sure I'll learn something else the next time I read it. It explains the newer language features by building from the old ones, explaining exactly what's happening, and then rebuilding them in the newer features in a much cleaner way.

This isn't explicitly related to gamedev, but that's probably why it's so good

And now for a comprehensive list of books to avoid
and you're set!

I've seen this design before, and it's pretty good. Is there a reason we shouldn't use it? I think it's fine. That's the kind of thing I was getting at regarding adding a hat or accessory for the silhouette.

Isn't that really the case with most -tans?

Unity's official documentation is generally thorough, but doesn't cover the more technical aspects of their engine that more experienced developers would be interested in. It's been improving, and it's generally the first thing you'll want to check.

Their videos are generally dogshit, and take 7-15 minutes to show you something that can be explained in a minute using text. They also don't give you nearly enough information to know how to use them properly, or when to use them. You can see this in their two videos on delegates and events, respectively.

They could have set up a mock UI, had the UI latch onto the game delegates, and explained that the game still works perfectly even if you completely remove the UI. In this way, gameplay and the UI are completely decoupled, and you could completely revamp the UI without even touching the game code. I've also used it on network functions.
Basically, a network manager would keep track of these (mock) functions, and the UI, health script, and gun would subscribe to these events and get passes the information when they're invoked; They get passed the network data struct. In this way, all objects just subscribe to commands, and don't need to know about each other. The network object doesn't need to manage or even know about the UI or player, it just handles the events.

instead its

Video tutorials in general are complete garbage, they get outdated very quickly. In a text-tutorial, you could easily change a few lines to fix a bug or keep it up to date with engine or languages changes. Most tutorials do come out with a few mistakes or bugs, but people rarely (if ever) update videos. They're also harder to follow; Programming is text. Pausing/jumping around a video is more difficult than just following a linear text tutorial. It's just a lot harder to make a video tutorial, in general, and text tutorials are shitty enough >>14275425 Another user pointed out that these are most likely just advertisements disguised as tutorials, as that's why they're so shit. If you're a new developer, steer clear of this garbage and develop very strong fundamentals so you can sniff out this bullshit without wasting your time learning terrible habits.

I'm calm now

I kind of want to see a version of that design with a gear-edged hood attached that can be pulled so far down that the two opposing gear edges make a little mouth. Maybe with a face proxy where her actual face would be.

working on a game engine, it just takes a lot of code to do anything for me because of that

Are you designing it to support larger maps? Once you're more finished with it, I'd really to see how many buildings you can shove into it. Something like pic related.

Vinesauce played my game on stream and it didn't go horribly wrong, which is pretty good.

I had hard drive fail on me, which is less good. Definitely the drive heads.

It's not designed expressly to support large maps but it will support large maps. You could definitely make a map that is a big city, and I don't see a reason for it to not work, however this engine is based on the same architecture ideas as the quake and source engines, so it will run into similar issues that you would run into with those engines. Making a giant low-poly city like that scene is not going to be such a difficult task though, so I wouldn't worry about that. It definitely would be more limited than something designed for large maps like GTA is, so if I wanted it to be able to do that, then I would have to make special engine features designed around that.

Right now the editor is really unsuited for large maps at the time being because it lacks grid scaling. So while I can zoom out I cant scale the grid yet. That will be added in the future and after then it should be possible to experiment better.

What game is this?

but for what purpose?

it's like you want to buy a super computer in order to play games made in Unity or something.

Severe autism, mostly, but it's a hell of a good way to get better at programming. And probably a good portfolio if he's not already a professional code monkey.

You glow in the dark, man. Why should we take your word for it?

Oligocene Snapshot, the game I worked on.

ejew.in/games/oligocene-snapshot/

ignore this guy, he's just an angel trying to fuck with me

Most issues with Unity come from poo-in-the-loo tier devs either coding their shit poorly or using poorly coded third party resources. I know for a fact that I've seen snappy ass Unity games.

UE4 in particular can also be made to run like fucking candy. The thing is, I just feel like making your own engine from scratch is reinventing the bicycle. Don't reinvent the bicycle. Making the game itself is enough work on its' own with the engine already completed.

I suppose it is a good way to understand things better, yeah.

Was this that weird game made with store assets that was posted here a few months ago that I suggested you sent to Vinesauce?

No, this was made within the last month. I did post a thread for it a week ago, though.

The only game that I had a hand in that used store assets to any notable extent was BBA SLAM.

Made a VR pirate game this weekend using an asset pack.

Islands are procedurally generated, got super basic skeletons going, added basic loot, weapons, and both teleport and locomotion with smooth turning.

Might set up a VR Pirate Roguelike.

I enjoy making game engines more than I enjoy making the actual games. It's very fun, it's basically just programming with no tangible end in sight.

No job, although I do have one in the summer I am going to be teaching the "game engines" class for some summer camp. Not trying to dox myself so I won't say which one though

Making a game from scratch is just reinventing someone else's game. Saying that making a new game engine when old game engines exist is reinventing things is just a double standard. Also just because your computer can run unity/ue4 doesn't mean that it's optimized or fast, both of those engines use garbage collection and require high OpenGL versions, yes your gaming PC can spare the ram but I just don't enjoy programming in such a way.

Anyway I am also not very knowledgeable about other game engines, the only engine that I have used to make games with that I didn't make myself was goldsource and that was years ago.

I see you're going with the loose interpretation of Rogelike.

...

Do you have any advice for managing larger projects? I have a very strong grasp on C#, and I'm better than your average pajeet, but once projects get to a certain size the complexity is always too much for me.

Don't work on larger projects by yourself?
Seems pretty simple I dunno

yay blend shapes

Some would call it an insult.
Sounds like he enjoyed it though, so congratulations.

I quickly drew that quite some time ago out of boredom, I'd feel ashamed if someone actually wanted to use it.

Maybe I'll try to make a better design soon.

Do you have some account where people interested in godot will see your design? Because it doesnt help if there are better designs like yours, when only we see it.

oh yeah, I like that android shit and all, also fits better with Godot being a mech, I'd fap to that

hummm yes…?

< the flood gates to game trash
how easy is to spurr out a game in unity really?

I've released shat out a game from start to finish in a single week as contract work for companies using Unity. They're absolute dogshit, but completely to specification and I got paid really well. That was while I was in school, too.

In case you were curious, I did it with a mixture of the engine features (really easy to get shit on a screen) and premade assets the clients picked out. Once, I got to design the gameplay, and hired some user from here to do the graphics. I'm actually proud of that one

looking at my post again it kind of reads as "if you dont have a social media account dont even bother"
What i meant was, that you maybe should create one if you havent already, so more people can see your work.

post more

...

best news in ages
apparently i can just download free animations for mixamo, since my character uses their generic rig
this basically saves 90% of the work i had to do on animations myself, i just have to find the appropriate ones and assign them
yea but their shit is free and looks leagues better than mine

1. thanks for reminding me there's a new episode
2. the joke is meh

Jesus Christ, user, that stock walking animation is the most formal/stilted march I've ever seen

The way I deal with complexity is that I make sure that after I write a system, I can just ignore it while I am working on other systems. So, if I write a bunch of code for the command line, I don't have to think about what the command line is doing or how it works when I am busy writing graphics code, I only need to know that "con_printf" prints to the command line. So, I'm only thinking deeply about one system in my engine at a time. I have huge amounts of code that I wrote and I just assume that it works and forget about it. I have all of my things split into clear components and try and reduce the amount of components that depend on each other. So with my console, I just call a few initialization functions, and then all of the complexity inside the console is being dealt with so I can forget about it, similar to how the system I have that interfaces with windows is working too, I never have to think about my code that starts up a window ever unless I want to change how that works. So, I can just have it sitting behind my call to "sys_init(sys)" instead of dealing with the complexity all at once.

Will the player be able to get his strongest potion at some point?

And which are these?

Battlerite?

Hey, its MoM user again. Based on a few of the common concerns for MoM, I devised a to do list to perhaps make the start of the game a bit more forgiving. I want to thank you all for your critique that you have shared!… Heres what I am thinking of tweaking!

- Have an NPC tell you in the first city that enemies within the Land of Magic are optional, and can be fled from

- Have no Goblin Encounters until the first dungeon is cleared

- Include an Extra Alchemy Potion within the Grave Diggers hut

- I could perhaps have it so that you can Use the Death Cap, Mandragora and Nitro Gel to get minor benefits (10 Mana, 10 Health, 10 AP). This would give them a use if you are not playing as the Imp or Mage!

- IF the Ropers are Still difficult after all this, I will nerf their health down to half… but ONLY for the first area. Afterwards they will still retain their 250 HP. They may be easy after all the above changes

- I may exclude ropers from Spawning in the Land of Magic until After the first Dungeon.

Were they any other concerns I should adjust?… I think people wanted me to raise the platform for the First Dungeon to make the puzzle a bit more easier to solve.

How do I start with Vulkan? What's the best language to use with Vulkan? C or C++?

No goblins might be excessive, although I've only played imp so far and haven't been hit by them even once. You might want to make it more obvious they're enemies, since they were the first thing I encountered and just hopped around being cute. If I hadn't already heard people talk about them I might have actually been hit.
If you make nerfed ropers, give them a recolor so it's obvious they (or rather, the later ones) are not quite the same.
The rest of your ideas are good.
Cooking the hams in the castle seems a tad hard, you have to stand on the edge of the hot plate to throw them on there since it's too high to throw from the floor, and standing on it normally damages you.
Also, someone else doing grammar checks for all your text.

If you have to ask, you shouldn't be using Vulkan. SDL2.0.6 exposes Vulkan functionality, so using that with C or C++ is probably your best bet.

I think having no goblins until the first dungeon will be fine, since as it is, they are technically the first enemy in the game. They can be a bit overwhelming at the start.

Also, it should be noted any fire damage can cook food. So the imps fire tome spell can cook the hams, as well as campfires, etc

Also, I will have an editer go through my text in the final release!

Make sure to be as rude as possible to them. The more they suffer, the better their work will be!

The more the editor hates you, the more minuscule mistakes they look for and force you to rewrite things better. It's one of those rare cases of mutual hatred turning into something good.

...

I think you should have the goblins steal around ~20 gold and have them start to roam in larger and larger packs as you progress.

Looking good, but some of the weight is odd. Fungi are relatively rigid compared to grass, so it doesn't make sense for them to move as much unless you're just doing it so it's cute, which it is.

You got a comfy feel building up in there.

Very good work! Only feedback is that it's a little too jittery, when it should probably be smoother and model long gusts of wind at a time. Or maybe it should alternate from jittery wind to smooth wind.

hair finally done

dat symmetry

i can't help it, i applied the mirror and stared to change the left side a bit, but i hated the result and reverted

Thanks. I'm constantly tweaking the flexibility of objects but in general I'm trying to go for a more lively/cute feel rather than realism. To be honest I'm not really happy with the results yet.

Are the shadows here maps or stencils?

I think a part of the problem the mushrooms have is that while their caps would sway, they wouldn't deform like treebranches or grass would. You'd need to somehow make the cap vertices children of the stem vertices, although at that point it's easier to use bone deformation I think.
Also, everyone knows that to be happy and cute, a mushroom has to bob up and down on the beat of the music.

Game Design Secrets
Design great Facebook, iOS, and Web games and learn from the experts what makes a game a hit!
anyone in the mood for me to dump some of dumpster fire?

Go ahead, user, I'm in the mood for a laff.

im game

...

After sharing your download habits with your friends on facebook, you can then use iPhone built-in face recognition technology to play with your virtual pet!

if you prompt him enoguh, the lemming will mistake the virtual pet for a real pet, and will be more inclined to buy useless toys with real money

...

presented without comment

Tossed out old particle stuff for a better, simpler system and started adding some basic effects. Going to be fun to sculp some gib chunks.

Look up Phil Ryu's history and you'll see he's a fucking slimeball. The number of people he's conned to work with him and then backstabbed or simply ran away with the money is disgusting.

HARDCORE?

I have a mother who plays shit on kikebook, so I kinda know these things. don't worry, I don't use the platform, I just would see her play when talking to her every so often. Most games ask to share your score quite often, and most people dismiss it, but a lot of people do share the scores. It's also common to bother people by asking for extra lives and shit, which are all on a timer otherwise, and if they don't play, that's okay! You can send them invites to play the game, which pop up in the notification tab, right next to friend invites and post responses and shit.

Most of the advertising your kikebook game gets is done for free by people who just really like the game.

I don't get why they claim that this stuff is original. My little cousins play shit like this all the time.

Not unless Apple rips off the Switch and adds physical controls to their tablet.

where's that sound of hand rubbing coming from?

fuck

i'm hurting inside user

there's an entire chapter on designing your game for your future chinese overlords

In twenty years and a migration or two, there will be anons among us that will unironically argue that physical controls are holding us back.

Kikestarter part 1

jesus christ

...

HA

Unity could very briefly export to flash, I guess

how do you forget the entire era of shareware
this never went away

Does this faggot not realize that every big engine can export directly to consoles, including fucking GAME MAKER?

that's good for now

Who wrote all this garbadge?… No, seriously, how old are they; they are so out of touch

Or maybe we're out of touch and the normie takeover of gaming is complete. after all, how many people do you hear calling themselves Gamers just because they play Candy Crush and Farmville?
I pray that that's not the case.

tfw spent yesterday with my 9-year old cousin, who identified herself as a "game freak"
tfw spent the day teaching her the difference between a game and a toy by explaining how we each "won" and "lost" a round of Chutes and Ladders, a "game" that is purely RNG
tfw she drops her mobileshit iPad games the instant she comes over to instead play Megaman, Smash Brothers, Command and Conquer, or arcade flight games like Sky Rogue or Ace Combat
tfw yesterday she got interested in Danmaku

I'm creating the best kind of monster to send into the future ahead of me.

LOL

I said I was to stop this

opencl no

"""""LOL-orific"""""
*screams of pain*

I wish it was that easy for me.
My little sister loves video games, supposedly, but spends most of her time watching people play it. Her favorites are Minecraft, Terraria and FNAF. (and the only reason she plays Terraria is because I tried to get her away from Minecraft)

I've tried to her her interested into all sorts of games with varying degrees of success.
Terraria she liked because it was like Minecraft.
Simcity 4k was way too hard for her, even after I explained every concept to her and spent a couple days playing with her.
Mario Kart and Snowboard Kids she did enjoy, but sadly I don't have the time to play with her daily. And she found them pretty challenging, even when I took it easy on her.
Animal Crossing she used to like, but her Minecraft autism overtook that and she dropped it.
Lastly, I tried to her her into both Mario 64 and Spyro, both of which were far too hard for her, Spyro especially. She did get ~20 stars in Mario 64 and couldn't advance any further, mostly due to lack of will to try again than anything.

I want to get her away from the autism but even the concept of physical controls was intimidating to her, so it's a massive pain.
Doesn't help that Mother keeps convincing her that Candy Crush and play-store trash are actual video games.

Tune in next week for a book by God of War's game designer, Scott Rogers.

I tried to get my sister to like games, but she has trouble moving two analog sticks at once.

Women and girls of today are more tech savvy than ever before and research shows that they currently make up over 52% of Internet users and 70% of casual online gamers. Why then, is the game industry still producing computer games that primarily target males ages 13-25? With this tight focus, game developers are not only sharply limiting their possible total income, but they are losing sight of the bigger picture.

The games industry is currently growing faster than the target market. To keep the industry strong and growing, game developers must start looking at expanding their market, which means designing titles that are accessible to the female audience. Successful entertainment industries have sustained growth for decades because they have considered the diversity of their audiences. Today's blockbuster products, be it movies, recordings or books, are most often the ones with elements that directly appeal to many market sectors, while containing very few barriers to access for others. By understanding the issues and barriers connected to gender, the game industry can benefit from a similar growth strategy.

Gender Inclusive Game Design: Expanding the Market addresses issues that help designers and developers understand the real differences between how the genders approach and resolve conflicts, and what their entertainment criteria and responses are. It also explores the differences in reward systems, game play preferences, and avatar selection criteria, and how these issues all apply to game design, regardless of genre. By understanding these differences, designers can apply this knowledge to the traditional genres that make up the contemporary computer game industry and begin tapping the future market. Perhaps the real question developers need to be asking themselves is, "but what if the player is female?"

why are we doing this

you mean you aren't interested in acquiring the PYRAMID OF POWER?

Same here.
Slower games like Animal Crossing were easy, and I think she did tolerably in Mario 64 just because it technically uses one stick, and she'd rarely turn the camera.
She learned Minecraft on a fucking touchscreen though, which seemed like a pain in the ass every time I tried it, so I think all you really need is encouragement from your friends to push through the barriers. I remember she cried over the Skeletons killing her every time, but she never quit Minecraft, whereas everything else she drops due to difficulty.

Just to suffer.

My younger siblings can handle Wipeout HD on the lower speed levels and we used to play stuff like SuperTux and Hedgewars together. They're slowly getting back into regular vidya after realizing mobileshit is boring and maybe we'll get a Nintendo Switch when the PS3 finally croaks.

all i feel is suffering

sorry user, I'll move onto a numale game designer

just kidding, he can't contain his autism on twitter

Use that as motivation to make a good game. If you don't, who will?

Have they considered just implementing the controls and trying out whether they feel natural?

fucking hell

my interests align with designing technical stuff(renderers, cpus, etc.), so that's what i end up doing. i suck at like doing the actual game part so i just do that. it would be cool to have somebody eventually use that work.

What about a journey to the Pyramid of Power?
zelda.gamepedia.com/Pyramid_of_Power

...

>women apparently make up about 52% of all gamers according to factoids if you include mobile shit and so on

Considering this faggot's take on pause menus and QTEs, god of war's design makes a lot more sense

...

...

Computer games are targeted, not all games. The reason is simple.
In general men will give the time commitment to sit down and play a game on a console or computer, while in general women want to play their games whenever they want in short bursts and won't give a large block of time to a game, meaning they dominate the mobile market. Men and women who break from this mold are the exception, not the norm. Marketers are just targeting the majority on their platform.

MAKE THE TEXT MATCH THE AUDIO AND LEARN HOW TO CAPITALIZE PROPERLY YOU FUCKING NIGGER HOLY SHIT GOD DAMN

There was a fucking bug that cost me a weekend, I couldn't properly fix it, so I've made a shitty fix that involves checking if the algorithm actually worked and if not, run it again with the shitty fix that makes it shitty but works.

Thankfully it's very hard to actually step into the situation where the bug happens, so the 'shitty mode' shouldn't run very often. Although it's very likely that the shitty fix don't actually always fix the problem

The webm starts with the algorithm running in the 'shitty mode'

You do realize it's just debug text to test functionality, right?

I found the "game" I made in Blender when I was 12 and finally webmed it. It didn't even pick up on OBS so I had to use Fraps. if you look closely i had my real full name on the title screen and censored it in the video

That's pretty neato, user.
When I was 12, I was just making bad romhacks of Mario World.

I might still have some of the Super Mario Flash levels I made around the same time.

The first "game" I made was a Where's Waldo clone made in PowerPoint

Holy god, I need new USB ports bad. I keep having to disconnect/reconnect my plugs a lot. I hope this tax return is at least half way decent for a new dev PC. Anyways, here's the video.

Warning: commentary.

What do you mean you made it in blender?

Blender actually has a game engine. blender.org/features/game-creation/

Rewrote some of my skill casting to lessen the Pajeet-level "just make it work" code and removed a feature I had that didn't add much. Also added a small message system for skills finally. Doesn't blend well in the background with colors yet but mostly wanted it to work.

Pretty good level design advice if you don't take it literally. If there's 2 or more exits from a place, it should be easy to tell the difference between them somehow.

What comes next? "American Soldier is shooting" for iPad which will be now considered as a state of the art gaming experience on it?

Working on jumps now, finished one direction. Working on forward right now.

So I'm actually having an issue with how I'm going to make it so player characters can use items outside of RPG combat. I think I'm just gonna make a single 'standing still and throwing item upward' animation.

I gotta be honest user, that retarded dance he begun doing halfway was funny as hell
btw, are you making skyrim? asking for a friend

CDRX9-43CVM-LXR7M - Steam
free shitty "indi game" that i won in this year game jam
WARNING: it's shit

Oh god, it's like I'm back on newgrounds

oh god, that is pretty much like that isn't it?

in any case i picked out better animations, although they definitely weren't meant to work together

That's an insult to the genuine creativity that existed on NewGrounds. It looks more like the shit games that advertised on NewGrounds, like AdventureQuest.

lol, moon strafing always look shit as fuck
make sure that diagonal movement isn't faster than unidirectional
also did you voice on your own?

any anons around ever done custom plugins for gimp?
I know its python, but it looks a bit complicated, how easy is it?
I need to process batch operation on several layers at once

The voices are from this video.

The pants are following the body perfectly, no longer clothes clipping through skin! Im loving this model remake.

How do you do the actual 3d shooting inside the raycast ? It should be touching the target no matter what as long as you're shooting the same column.

My entity collision shapes are cylinders with height rather than being infinitely high like Doom. Unless I'm misunderstanding your question.

So how do you do the look up and down ?

You mean for the raycasting? I use the pitch of the camera along with the camera height and surface height to calculate the position of where it would hit. Starting from the player's current sector I check for any circle-line collisions with entities in that sector. If there is a collision I know how far along that collision happens compared to where the cast would hit the surface, so I can just lerp the height and check if its between the top and bottom of the collision object.

Added some flowers… Not sure if I will keep this art style. Need to keep experimenting.

Shadow map.

It's looking really cute to me. Are you not satisfied with the movement or the overall result?

I don't know, it doesn't feel right to me. Maybe I need to try making a different scene, something with architecture. Perhaps it's just hard for me to imagine actual gameplay in a forest environment like this.

pretty sure both speeds were the same, but it's most likely a difference in offsets between the animations, left-forward worked, while right-forward didn't, even though right was just a mirrored left
anyways i found more diagonal animations, so as an 8-way locomotion it works. they look too silly though, probably just gonna use them as half-decent placeholders

neat stuff

The problem is that in the diagonal animations, his arms are outstretched, but in the cardinal directions they're closer to his body. Plus the forward run has too much exaggeration.

that is much better user, it usually will always look weird because objects move at a constant pace, while realistic you would expect foots to be fixed, (eg all games are weird like that)

but what I meant by diagonals, is that if you move x at += 1, and y at += 1, if you move them both on a diagonal your actual speed is 1*sq(2)

uh, well i'm assuming that actual diagonal animations kind of account for that.
4-way didn't

Might help to try arranging all the objects into an actual gameplay space and not just this cluttered space that's obviously just meant to show off the tech.

That's beautiful, user. Any webm with the grass integrated?

Thanks. If you mean the lush grass thing I've shown before - no, that was just a stress test, I'm not actually going to have that much grass in the game.

This is why I've enrolled in a CS school, because I'm retarded and can't be trusted to learn this shit on my own.

Say is the grass actual polygons? Not that I know that much about gamedev, but could it be remade as a pixel shader? Maybe then it wouldn't hog many resources.

Yes, the grass is polygons. I optimized it so it doesn't hog resources even in that stress test, it's just not the art direction I'm looking for.

My graphics cards have fucked drivers right now, so while I was installing them in safe mode I tried to mess around with my engine and it was totally fucked, it turns out that some drivers don't give you any display modes. Also my framerate is unacceptably low, granted that is because this is literally the safe mode "GDI Generic" OpenGL driver, but maybe once my culling works I will go back into safe mode and debug every issue that i'm having on this driver.

sounds like someone should have bought a mac

MacOS doesn't even have Vulkan support. Granted it would be fun to mess with Apple's Metal graphics API, and I do want to port my engine to Mac, but I just don't have one right now. Anyway messing around with drivers helped me solve some bugs.

I was always kinda partial to Albino Black Sheep as far as flash content went, personally. In the beginning I used that shit site ebaumsworld though.

transformation on the GPU!
all that needs to be reuploaded is an array of 4x4 mats now per frame.
also kind of fail
oops

ikeatrace

I had a problem with unity in Windows 7 where outdated USB drivers caused a reproducable BSOD on launch.

Welp, setting this aside until I can afford a devphone or something.

Also it's page thirteen, anons.

i am too retarded to create op for new thread

How do you even breathe, user, let alone just like make game?

i mash my hands on the keyboard and rays come out or something

Last time I did it I fucked it up and forgot to put the title in, and haven't dared to try again since. it's been three months.

NEW THREAD

how the fuck do you manage to survive in the wild?

and why do you assume he would?