Browsed by
Month: December 2018

Atmocity progress report #14

Atmocity progress report #14

It’s been a productive week. Mostly under the hood stuff and mostly optimizations and bug fixes.

I’ve redone the control scheme slightly, moving the demolish action from the right mouse button to a separate button in the UI. Right mouse click now anchors the camera to a structure so that players can more easily get the camera around corners or rotate around buildings. Rotating the camera has been moved from left click to scroll wheel click. I’ve also tweaked the “floatiness” factor a bit more and this will be implemented into the options menu in the upcoming week, allowing players to adjust how rigid or floaty they want the controls to be.


I’ve added better lighting to the “shady side of the world”. This has been an issue for a while, that certain areas will always be worse lit. I’ve tried to compensate this by adding a dimmed light that is always opposite the sun. It will need further tweaking, but should end up solving the issue. This also works as an additional “night light”, making building easier during night time, if players insist on keeping the day/night cycle on at all times.


I got the save/load menu fully implemented too. The UI still needs work, but in terms of function it’s all there, thumbnail of the city and all included. I also managed to clean up some lousy placeholder code because of this too as well as removing what used to be a “loaded city”-scene and a “new city”-scene. This is now just one gameplay-scene, if it’s loaded with no city loaded, all the new city stats and resources will be loaded as well as a solitary building block placed in the center.


As an additional visual feature, I will implement several different skyboxes and day/night gradients. Exactly what types there will be, I haven’t yet decided, but there should be quite a few to add more variation.


Having done quite a bit of code optimization (some scripts were lagging behind in terms of optimization), next week I will get into refactoring some of the newer code, like drag and drop for building blocks etc. This code is rather messy atm and requires a good cleaning. I’ve also done quite a bit of UI-optimization, splitting up large canvases and also getting rid of some graphic raycasters on canvases where there are no clickable objects.


Local economy should also be finished enough to get it into proper testing next weekend, after which I will be able to set up conglomerate and citizen demand systems. Overall, I am behind schedule, but should have time to catch up after christmas. The game is not quite as finished as I would like for late alpha/early beta, but it’s good enough I reckon.

/ Dispersing Minds

Atmocity weekly devlog entry 13

Atmocity weekly devlog entry 13


Dreamhack came and Dreamhack went. I’d like to start out thanking those that dropped by and checked out the game. It’s always nice hearing that the game looks good or looks cool, but a completely different thing seeing people actually try it out and responding to things. Many people seemed very intrigued with the main concept of the game, being able to build really inventive cities while also managing complex economic systems.


Of course, I also noticed some things that need fixing and also received some very good feedback. I will detail some of the fixes I’ve made, feedback received and some other fixes I’ve done since the event and also what will go into the game as I move forward to the early beta build.

Unlike most city builders, Atmocity requires a fairly complex camera system so that players can navigate around the city. Most builders fare well enough with just a top down camera, that can move east/west/north/south. In Atmocity, players need to navigate within a volume in full 3D. With this comes a bunch of difficulties that need to be overcome. Most of the people who tried the game out did ok with the camera, but I did get some good pointers on how to improve it further. So in the upcoming few days, the options menu will be fleshed out with a Controls-panel. Here, there will be a sensitivity slider, since some players wanted the controls to be more rigid, while others felt they could be softer and floatier.

Aside from that, there will also be a type of sideways rotation or “lean” feature implemented, so that players will be able to easily setup the camera on level with the overall area they’re viewing. Currently this requires a bit of a specific click and drag movement to do. Zooming will be implemented on the scroll wheel, and clicking the scroll wheel will initiate the camera drag movement. Currently the drag movement is assigned to left click, but this will be used solely for building selection.

At the moment, demolishing buildings is weirly assigned to the right mouse button click. This will, as has been planned for a while, be moved to a hot key and button in the building panel. Instead, the right mouse button click will be used to “attach” the camera to buildings, so that it can rotate around them easier.

These changes should be implemented in the upcoming few days. Navigating mazes like this should become easier:


I discovered a handful of minor bugs, among them a null ref due to cars not despawning correctly when buildings were demolished. This has been fixed, as has the efficiency display shader not displaying correctly on skyscrapers.

I’ve also continued optimizing the game during the week. The reflection probe now uses time slicing so that it isn’t quite as performance heavy as before. I’ve managed to cut some shadow casters, removed hidden faces and made some other minor (per object) optimizations that make a rather sizeable difference in bigger cities. I’ve also managed to cut down on the tri count from shadows by cutting shadow casting distance a bit. This scene maxed out at 1.5 million tris and 1000 shadow casters before, I’ve now gotten it down to 900k tris and 300-400 shadow casters (depending on the number of cars in the scene):


Further improvements will be made making far LODs even more low poly, since they are hidden in fog and blur anyway.

An idea that came up at Dreamhack, to make building weird cities more of a functional choice as well as visual, was to make pollution not spread through blocks. This will be implemented in an early version too. This means that unlike now, when building factories within a specific radius from Residences will cause a penalty, you will be able to plan cities so that factories can be built on one side of a block and residences on the opposite side. There will probably be some sort of balancing issue with this, and I will need to implement it well for it to work well and not make city planning too easy.


My initial plan for the upcoming beta was to have everything more or less implemented. Sadly, some systems are still a bit too underdeveloped to be used properly, so some will be left out during the early stages of the beta. But I am hoping to get them done early to mid january so that they can be properly tested in good time before EA release too. Local economy system is still proving quite difficult, but should get there fairly soon.

The remainder of the week will be spent optimizing traffic further and adding the connection testing that buildings will do to ensure connectivity with key buildings like power plants and water absorbents. On top of that, the save and load manager will be started so that more than one city can be saved and loaded.


There will be Another update on Sunday. For those interested in trying out the beta, beta signup information is available on the Steam page for the game (

If you want to more closely follow the development of the game, you can also join Discord,

(Also vote for the game on IndieDB:
/ Dispersing Minds