Beyond the waves of the sea

Good morning again. Yesterday was one of those days where I did make a little progress building the world, but most of the effort went in conceptual improvements. Not do much actual code. 

Firstly, what did I make? Well, this.

What you're looking at is the beginnings of some ragged mountain range. For testing purposes, I have these coloured textures that indicate height on the level. Brown posts are the higher regions. In the brown part, there's a bunch of spikes popping up. These are random points that I rose to a level where they're easy to see. They are going to be the basis for more ragged mountains. Those are the peaks.  I'll write more about that once it gets under way.

Secondly, the ideas. I found out a problem in my wrapping system. When the left tile its not nearly the same height as three rightmost one, there is a lot of space to cover to make the right one connect. This can give weird results:

The leftmost tile is a sea tile and this tile on the picture is highlands. To connect, the needs to be a steep drop which does not look right. For that reason, I'm changing the wrapping system. If you look away the world map (of earth) there's two points pg interest here: the sea East of Iceland and the Bering Straight between Siberia and Alaska. From there, there's a pretty simple line of ocean going from the north pole to the south pole.

What I'm going to do is stimulate something like that. The world I'm generating will be surrounded by low regions on the east and west side. That'll be oceans. I think this will fix the problem described above without making it look unnatural. I might even put water around there poles to, but probably with extra generated land masses or ice sheets. Have not decided yet.

I do know how to make the surrounding ocean. That'll be by mixing the height map with a noise that has low values around the edges and high ones in the center. I'm not sure about the algorithm yes, but it should be feasible and I've seen this before. I'll probably make that first because it's not very difficult.

I'm also going off for a long weekend from Friday to Monday, so there probably won't be much work then.  Maybe a blog when I have one of these Eureka moments, but don't count on it. You'll never know, but the laptop stays home,  except if you're a thief. I'm taking my laptop (which I always refer to as"slow" and "crappy") totally with me.

The world is round

Good morning again. Weekend's over so I'm in the bus to work again. This weekend was quite busy, but I did have some time to work on Holocene. As may buy now be known, we're procedurally generating the world and for three nth time I restarted the project to set it up better and now I finally get the idea I start to understand how it's supposed to be structured. So even if it seems like I'm not making progress, I do.

First, I've been thinking about the structure of the world building script and the order of operations.  We'll start with the broad outline of the world that we'll have.  I use small plane objects to set it up. This broad outline right now had one simple Perlin noise, but it'll probably have four. My aim will be to have big low regions that'll be the oceans and big higher regions that'll be the continents. These continents will have features of their own.

Once this is done, I'll see if I need more pronounced mountains. If I do, let there be mountains. I think I'll use mid point displacement to generate them.  When that's done, I'll focus on the rivers again. The aim will be to erode the mountainside so that there will be room to drop rivers there.

Maybe unclear when I write it like this, but there point is that I want to have a playing field where I can drop a river on a mountain and it'll naturally find its way to the sea. Last time I tried that, rivers would get stuck in dips along the mountainside. Now I'll try to carve a path so that the dips have an exit.

When I finish the rivers, I'll add a heat map and a moisture map and that'll bring biomes. Any biome had certain vegetation types and animals. When I drop them in, this should be it.

Right. That's the end aim. Right now I have this. This it's a 4 by 4 grid of tiles with a Perlin noise on the 3 left columns and a special row to the right. This row is a special kind of Perlin noise, because it mixes two. It connects to the one top the left, but it also connects to the left most to its right, in other words: the world is round. I decided not to do this to the top tiles, so it'll be a cylindrical world, rather than a round one, but I'm pretty pleased with this. You have no idea how long this took.

Anyway,  i think I'm starting right. Next task will be to make the work larger and more natural by adding more Perlin waves. Hopefully that connecting time column will look better by then. We'll see Howe long this all will take. I also have a job and a wife and I know this is all going to be a big task, but it's begun!

Still grinding

Good morning. Yesterday I found out that I still don't have what it takes to pull off a good procedural world generation. That's ok, because I can still learn and I already started following yet another massively large course on it. The idea is that in time all puzzle pieces will come together and I get a system that actually works and gets us a good and believable world. In the end, the world is one of the main elements for Holocene. I need it to be good. No concessions allowed.

I told you rivers don't flow the way I like. I think that is mainly because of the noise I'm using for the landscape. There are hills around where rivers can form. At first, I tried having them just spawn somewhere and then flow to the lowest neighbouring area until it hits the sea. Well, that meant that rivers went to a low point and start going around a bit until it surrounded itself and had nowhere to go.

Well, that's ok if that happens sometimes, but certainly not if it happens to 90% of all rivers. Now how can I fix that? I think that's a difficult question, because it's hard to see such a thing coming without making huge analyses.  A river should in principle look for the low point and go there and it's good if it goes meandering down a bit, but at some point the river makes a sort of U-turn and closes in on itself. That often spells misery.

What I think I should be doing scares me a bit, because I think I need to procedurally create river beddings and leech the water flow in afterwards. It would give me power to force a way to the sea.  I still need to have it find a way to avoid the weird u-turns, but maybe it works. What I really need is a way to make sure that from any point on the map, water regions are mostly reachable by only going down. Again, it's ok if it is not 100%, but over half the rivers should flow to the sea or a lake. That's what I'm going for.

The river just stops somewhere

Good morning everyone. I'm in the bus to work again.  This weekend was quite productive while in the end, I only got a little progress. The problem was that my game totally broke and I hadn't made the effort to back it up yet, which was stupid. On the other hand, there were things not working correctly which I were trying to fix and which I could not, so it's ok. I'm not beyond that point anyway.

I'm still working on the procedural level generation for Holocene and I'm using a Zenva Academy course to guide it. I'm not just following it, but have it guide my progress and do the refactoring that it's screaming about while I'm writing the code. Instead of linking noise generation scripts without attributes to every single object, I'm using static ones for that kind of purpose, because a larger level might turn a slow and laggy level if I'm not careful.

Right now, I have a grid of tiles which in turn are grids of vertices and blocks. Any vertex had a height and that stands for the height of a given place in the world. It also had a heat and moisture value which make for a biome value together. Right now, any biome can have one kind of tree. I'm going to change that later. There's no reason why tropical rainforest can only have one kind of tree. 

Last thing I added was rivers. While I was following the course, I kind of knew this was not going to work correctly:

The river gets a random origin point somewhere above a certain height threshold. On the right on there picture. Then, for every step of the way, it finds its lowest neighbouring vertex and goes there until it reaches the sea. Here is the problem. The way to the sea is hardly ever a simple downhill path. There are bumps on the way.  It the real world, rivers overcome those bumps over time, but here they don't. This makes that the river will just stop in a valley, zig zag around a bit and look weird like in the picture.

Seeing this thing is easy, but I have not yet come up with a solution. It seems like there needs to be more planning, although that feels counterintuitive. The origin is fine, I think and the first downhill part is too, but then we need to make sure to keep some longevity. We need to make sure the river does not stop in a valley (or it should form a lake) and we need to make sure the river flies not surround itself.

I'm thinking of a system that had the rivet have a general direction to go, to a shore somewhere. I'm not sure how to make sure it does not have to be a straight line, but I'll manage. The other thing I'll probably do is have the river have a force to be able to lower vertices around it. When the river has gone down, it should be able to dash through small uphill sections by lowering that section.

I'm thinking about this some more. Hopefully, I'll have a good solution soon.

A big procedural overhaul

Good morning. Let's get to it.  Yesterday, while I was working at home for my day job a mechanic showed up (it was an appointment) to change some electric stuff and that means that for a while we did not have any electricity. I decided to spend the time on thinking about Holocene. This is what I came up with:

This is what it kind of looked like before yesterday and I backed up my project because...

Well, I agree. I owe you an explanation.  Well here it is. I'll try not to make this too technical. I made some design mistakes. The most important one is that I never considered a world with any size above one Unity terrain which is about 500 meters if I'm correct. I'm sure most of you know that setting up an ancient civilisation will not work on so small an area.

What I did now is plan for a system where the world can be really big. The above image is a tile. It looks like nothing worth spending time to discuss, but it'll change. In the game, we'll have a couple of these tiles rendering while we're walking around. This fits excellently with the first person camera view, because you can only see a couple of tiles at any given time, so why render all of it?

Second thing is that with Unity's terrain system,  we'd have a landscape that does not fit well with the low poly models that I have in mind, like the mammoth. I think it'll immediately look like a weird mash up of different art styles. With this regular geometric shape landscape, I think I can prevent that and keep the artistic coherence.

Lastly, and this is tied to the big world tile thing, I need rivers and seas to have this game the way I want it to be. Water its pretty easy to implement. Just have a landscape and drop a plane with a shader on a certain height et voilà. Problem is that that'll make for ponds and lakes, and not for seas really.

What I plan to make is a system where some of the tiles  (don't be surprised if that's over half of them, like Earth) sea tiles. I'll probably use some smooth Perlin noise to get some magnifier for the heights on that tile. Need to find out how that'll work in practice.

Rivers will be something else. I'm not sure about that either, but in my mind, I'll have something decide a random place above some height threshold and drop a river there. That river will flow down the slopes of the landscape until it hits water. Maybe it'll have to push through when it hits a depression with hills in all sides. Maybe that'll be a little lake.

Whatever I do there, I'll have to be mindful of the time it takes to generate the terrain. It's ok to have to take 20 seconds if need be and I'm sure the generation of the world will be one of the heaviest processes I'll ever have made until now, but let's bit have it take minutes unless it's absolutely necessary and let's certainly not have it take time during runtime.

Anyway. I thought I needed this blog post to explain myself and get my thoughts to paper. When I told my wife yesterday, she was not pleased by what she saw. I'm sure she is not the only one.  Besides, the Perlin noise thing for the seas, I came up with that piece of genius during the typing of this blogpost.

Generating the Big Apple

So. Back in the bus again. It's been a couple of days. I made some progress in refactoring of the object generation for my terrain and I also now have a little apple model to have the characters interact with. This is important because the plan is to have default behaviour for the characters to try and satisfy their needs. Apples are here to satisfy the need for food. Can you believe it?

Before my refactoring, apples were actually functionally the same as rocks. They appeared on the terrain in a pseudo random way. This made for weird situations where there were apples around and not a single tree anywhere. To fix that, I moved an object generation method to the object script for the tree object. Trees spawn apples like they're supposed to. No trees, no apples. I don't know yet how I'm going to trigger the generation, but for now, yes get one chance top drop an apple and that's it.

I also decided not to repeat myself instantiating objects, so I moved all instatiation methods to a central script and run or with parameters. That took some debugging, but I think it works well now.

There was something off with this image, but the generation worked alright. All I had to do is make the model smaller in the prefab.  In the old system, spawning apples worked well because the scale was set from the script. Now there is not going to be much variation in apple size, so I'd better not do to much with it. So I fiddled a bit and this is the result:

Those characters better like these apples if they know how much work it was to get them there. 

They have character

So.  Back in the bus to the office. They say protesting farmers are blocking the roads, so we might be in here for some time. While I don't share their aims and reject the violence they use, they have the right to protest. For now, no right wingers blocking the road. We'll see...

Yesterday evening was quite a productive evening for Holocene. I finished the first version of the starter script. It generates a terrain, populates it with trees and rocks and places the player object at a random position on the terrain.

The player can walk around and discover. When they press the space bar, two or three white capsules appear. While they don't look like it, they'll be your first tribe members.

There will probable be some work to be done left. The characters (I'm calling them characters) are rigged to be able to move around on the terrain with certain constraints like the steepness, but they're generated within a random distance from the player in any direction. This means they could instantiate over the edge of the terrain. That could pose a problem because for now I only have one terrain and I don't want the characters to spawn in the endless void of Unity.

While I'm planning to fix that buy generating a very large or even endless terrain, another issue needs to be fixed anyway. The characters can spawn anywhere and that includes places where the steepness is so high that they can't move and even at the same place as trees, rocks, water or anything else. That'll have to be fixed and I think that should not be a problem.

With that out of the way, I also started the scripting for the characters and their well being. The idea is to have anyone have general health, things like energy, sickness and injuries and a set of needs like hunger, thirst, sleep and the like. These will all be properties of the characters that I'm thinking of how to implement.

Next there will be skills. Any character will have a skill set with skills and their amount of mastery. That'll decide what the guy can do.  I think this will be the next thing I need to lay the groundwork for.

Right now it's time for work in the office. I don't think I'm seeing many tractors blocking the highway this morning. Which is a good thing, but it means this blog is done now.

Rolling Rocks

Good morning. It's Monday again so I'm off top the office. Time to share what I've been up to over the weekend. I said I'd be starting game mechanics and I really did, but that did not result in game mechanics but in some changes to the level generation. That is because I need rocks and other small objects to appear and these need to have some physics about them and a way to interact with them. Just the meshes of the Unity terrain are not enough for that.

What I did was create a couple of rock models in Blender and hook them up to an empty game object in Unity. The model gets a collider component and a rigid body for the physics a and the parent gets a worldobject script. Then the objects are dropped onto the terrain by the generation script.

What we see here is the generated terrain with static trees and rocks that are dropped just above the ground. Some of the rocks roll down the slope and that it's intended. Buy three end, rocks will probably not be instantiated on the higher steepnesses to avoid them from rolling down, but for testing purposes I did not change that yet. Same applies to the ground textures and tree placement. That'll be fine tuned later. 

I need to make some adjustments to the rock generation. I find the rocks to be mostly too big, so I'll find a way to have the variation in size and have some big ones, but not that many. I have some ideas to get that.  

Next, I think I need to lift the terrain up a bit. Right now, most of the time, I have terrains with huge flat parts that turn into lakes. That is because the altitude script checks for a lowest possible level and sets all lower values to that. I think I can make the terrain more interesting by raising it a bit. 

After that, I hope to finally be able to start with the real game play and add characters. Those are going to be the beef of it all. 

Talk about trees

Good morning. I've been quite busy working on Holocene over the last couple of days and I think I'm making progress finally. The procedural terrain generation is not finished yet, but for now I'll leave that there for a while to focus on what I'm actually really trying to do all along, the game.

I made a simple first person controller with a camera and some moving mechanics that I picked from my RPG project of half a year ago. Next I tried to hook it up to the trees and things on the terrain and that it's where it goes wrong first.

The reason for that it's actually pretty simple. Trees on a Unity terrain are not really objects in a way that one can add scripts to them and interact with them individually. At least, I have not been able to find a way to do that. For that reason, I cannot get a vision script to work. With a first person camera, I might be able to have the player look at the terrain as a whole, but that won't help to much, because we're not cutting the terrain to get wood. We're cutting trees.

To fix that, I'm rewriting past of the terrain generation script to allow for tree prefabs to be instantiate as separate objects. We'll have to find out if this approach will make things slow. If it does, I'll look for ways to not render objects that are not in view.

A tree object will have a parent object within a graphical element and one or more colliders. The scripts will be on the parent object or the collider, if there is just one thing. Once the player looks at the object, the GUI will give a message and certain functions will be unlocked.

I tested this a bit with a hand placed tree and that works alight. I'm now working on the generation script to allow for placement of the. The same will apply for rocks and the like, because those will be interacted with too. I think I'll be able to finish the script part today or in the weekend.

Right now, I'm using placeholder meshes to speed up development, but the plan is to replace them with my own creations. That'll take time and effort,but it'll make the game better, I think.

Procedurally generated frustration

Good morning everyone. I'm in the bus to work again. This weekend was pretty rich in game development. I managed to spend quite some time on the procedural terrain generation course that has had my attention for quite a while now. I learned a lot about Perlin noise and midpoint displacement algorithms and about how top make editor scripts in Unity.

Because I don't want the game to be too complex, I'm not implementing erosion for now. For that reason, I decided to just put all the scripts on my game project and get them to work before removing unneeded stuff and tinkering with the other stuff. It seems to work for a part.

This island is procedurally generated by using a midpoint displacement to generate the heights, a water plain at a certain height level and a couple of textures like grass and rock, which seem to blend too much right now. The white line is an animated shore line and the green stuff at the bottom... those are the trees.

Somehow the trees behave differently than they did in the course, so I must have done something wrong. At 11pm yesterday I decided to delete all the procedural code from my game project (only there) and start  from scratch. I am a lot more at ease with scripts with many methods than with scripts with a few large methods, so I'll have an Awake() with just about ten lines and keep it the way I like and the course teacher does not seem to.

My point is... sometimes you spend a lot of time on something without anything to show for it. Whatever I did, at least I learned from it.  I do now get that annoying feeling that I want to start working on the actual game mechanics instead of vital but secondary stuff like the terrain. I hope I'll ferry this thing working this week, because this way it does not feel like a game.

Mountains and valleys

So. It is taking longer for me to write new articles on this devlog. That is mainly because I am studying ways to improve on the bumpy world that I made in the previous version of Holocene. Right now, I'm working on procedural level generation again and I think it's starting to dawn on me.

In the beginning of this year, I bought a couple of Unity courses on Zenva.com to get me started on this new years resolution that I got myself into.  Now that we're 9 months along the way, I saw quite some of the hours of videos and completed some of the courses.

One of the courses is on procedural level generation.  It uses a Perlin noise to make a height map and move vertices on a grid. Then it has a heat map to make sure stuff at the equator is hotter than around the poles. That is mixed with more noise to have some randomness. Another map brings the moisture and that comes together to create biomes. All very interesting.

I am going to finish the course,  but I am not going to use it straight away,  because I want to improve it.  Firstly, water is just blue land. Making a game about guys walking around the world,  water cannot just be blue land. I know what that is about and I think I can fix that, by having the water a separate object that the navmesh can now cross. I'm not sure how to do this exactly, but I'll manage.

Second thing may be more difficult to do and less important. I don't really like the way the ground is textured, by a colour map. Some parts of the world are "boreal forest" (which is a difficult term for us Dutch right now, by the way) and that's a specific tint of green for the ground. There will be trees and plants later, but I'm not sure green floors per se will be right, and especially not the red deserts that Zenva came up with.

Next, I am not sure if the course will cover this, but I will need more  detailed stuff on the scene to interact with. That'll in itself probably be like trees which will be in the later part of the course, but I don't know yet.

Probably even more important than all the other things combined is preventing the game from taking too much resources generating the world and keeping track of it. I'm planning to have a huge world, kind of like Minecraft. Problem is that the procedural process takes quite some time on a 10 by 10 tiles grid. I know this is mainly because of building a navmesh for the tiles. That should go better later. I hope I can manage that without clogging the PC totally or else I might need to change my scope to smaller levels which would really be a shame.

Whatever will be, first I am going to finish the course and restudy the code because some of it is very poorly explained.  Then I'll tey to find the solutions.

Trying a different way

So the magic phrase is "procedural level generation". I opened a whole can of worms when I told myself that Holocene needed a ground that would be different from a flat plane. It got me into a whole range of tutorials for making something that looked exactly like I needed, but really wasn't.

I learned to work with Perlin noise maps  and how to use those to get a 2D configuration that looks random enough for my purpose. That's nice and it's also nice to use plane objects of home made meshes to get the level rolling,  but probably not in Holocene,  because I could not get the navmesh working. And by now you should know that this is a no-go.

The problem is that, no matter how hard I try, navmeshes end up below the surface of the level and not on top. That means that the characters would not be able to walk over the surface of the world in the way that I planned. I couldn't get it to work.

Yesterday I found another tutorial on putting what I learned to practise on an actual terrain mesh and that works if I use the special navmesh scripts that I found on the Unity github. The scripts allowed for baking the stuff at runtime which seems to work. All I need to do now is bake the mesh after generation.

I have a working navmesh now and I can now add the features to it. I am still not completely satisfied with the way the terrain looks. I think I need to tweak the noise script to allow for more flat regions, but for now it works. Next is to turn this into a game.

Building the world

Alright, I've been quiet on the blog recently. That's because I mean to have a big change on the game. The change is about level building. Holocene is a Real Time Strategy game and those games should be replayable, but that only works is the playing field is more interesting than just a flat plane and if it's different every time you play the game.

For that reason, I studied procedural level generation for a couple of days. It works by putting a plane in the scene and having vertices inside of it go up and down according to some noise thingy that I do not yet totally understand. The idea is to get the plane to resemble a sort of low poly terrain.

Next, there'll need to be some changes in the different scripts to allow for the third dimension that we have now. In fact, Holocene may look like a 3D game already, but it's the famous thing with the ant that quantum physicists always say when they explain the multidimensional world they think of. An ant is walking on the pavement like its 2D. There is no up or down for it. Well, the objects in Holocene should turn from ant to human.

I'm starting to get the gist of the level creation and I think I might be able to use the same script with some modifications once I build a campaign mode, but right now, I am not yet satisfied with the way the terrain looks. It's too bumpy and not mountainous enough. It's too all-the-same to my taste and I'll need to change that, because we need valleys and mountains, not just one 6 meter high hill following another.

I need to get the noise changed. While there's the Perlin Noise, I need to tweak that. I need lower the frequency so that we don't get hills everywhere we could get. We're trying to run a civilisation there, right? Next, we need higher amplitudes, because we need mountains and valleys. Lastly, I need to add more noise within the noise so that it does not look the same everywhere we look. We need hilly parts, flat parts, sea and mountains. And everything in between.

Looking at the code I have now, that must be feasible.