Well it turns out that there was a major issue with the new shadows where they didn’t maintain the correct positions if the parent was moved. It was visible in the previous video as the parent was situated at 0,0.
A little annoying but glad it reared it’s ugly head sooner rather than later.
Onwards with pathfinding, currently working on a navigation mesh for path finding.
Its been a few months since any sort of update, mainly due to other commitments but I was also struggling with getting my head around how to tackle randomised building generation. At this point my plan of action is…. I don’t. I can get away with having a few different variations of each building which will do for now. In the future I could always create an algorithm to make it completely random but that’ll be saved for another day.
A couple of days ago I began writing a separate prototype for the buildings which includes a new method for handling shadows. The new shadows can now work with any polygon which should work nicely with the effect I’ve got in mind for the game’s atmosphere.
Below shows the prototype in action, it achieves the effect I was after and only uses 162 triangles but this will likely multiply five fold once the textures are sorted out and extra elements like furniture are added.
I know the textures are ropey but it’s just a quick prototype to demo the progress. It’s also a good testbed for building related feature development.
Over the last month I’ve been working on a new project. Usually I’d try to resist starting a new project as long as possible but feel that this could be a rather quick game to create.
It originally started off with trying to think of other ways I could reuse my city generator and I came up with a city where enemies come in from all exits and your squads have to fend them off. There will also be survivors in the city which can be rescued for cash. If I were to give it a genre it’d have to be an action strategy with hints of an unconventional tower defence.
Progress is coming along quite well, although there’s not a whole lot to show at this point. Units are controllable with pathfinding and enemies will wander around until they see a unit and attack.
There’s still a lot to get done but its a good start, I should have something to show shortly.
I’ve been taking a look at the game and thinking about pieces which dramatically increase development time and not really essential to the game. The key piece is the 3D side of things which although is a rather nice feature, is somewhat expendable.
Looking at games like Hotline Miami and Teleglitch, we can see that they’re still amazing games even though they’re of a top down 2D perspective.
With this decision in mind I started looking into the Starling 2D framework which integrates the 3D stage to utilise the GPU in 2D rendering.
After getting Starling setup its really quick to start working with as it uses the same variable and function naming convention as core AS3. Unfortunately it became apparent too quickly that my intentions for the framework were more than the intended use. Once the city generator was implemented, the performance went down to half which was seen using the core 3D previously. I’m not saying Starling isn’t good, for most projects its ideal, but unfortunately its not cut out for this project.
This has resulted in me creating my own version of it. It’ll conform to the AS3 functions to behave in a similar manner to using the inbuilt 2D engine but for it to easily extend the standard model to easily create complex sprites.
So at the end of the last year I entered the Oryx Trial with a somewhat poor entry and unsurprisingly didn’t win. I didn’t expect to but the point of entering was to push myself into writing some classes which would be used in my other game. I must say it’s also inspired me to get back into the game development side of things.
Since the new year I’ve made a big push and managed to get the random city pretty much sorted and integrated with the 3D stage. Now although this is pretty much regurgitating previous posts, a lot of work went into rewriting big ol’ chunks of code for it to work more accurately and quicker.
Now the main thing that’s up with this example is that there’s a delay when you hit the edge of the city block, this shouldn’t be a huge issue as I’ve purposely scaled the map down to down to show this. The map is really ten times the size so by the time the player reaches the edge of the map, the neighbouring block would have already loaded (in theory =P) The way the map works is on a 9×9 grid, when the player reaches half way through a block it starts loading in neighbours. The delay is present as threading in AS3 is still in beta so I’m currently daisy-chaining the processing off the main frame rending. When the player moves to a new block, the loaded blocks shift over so only new blocks have to be loaded. This now does mean that you can walk infinitely in one direction and the city will keep generating buildings around you.
There is a minor issue where some buildings don’t get loaded which I shall have to look into but I’m happy with the progress at this point. There are still a few pieces which can be optimised at this stage but I thinking I’ll work on a different piece of the game for now. I’m really hoping to have something fresh for the next post as this is yet another post which contains the same as previous entries but with minor updates.
Well I submitted by game last night but have mixed feelings about it.
I’m pleased with the amount I got sorted with the limited time available but I’m not happy with the quality, especially towards the end as I was having to cut a few corners in order to get core functionality in place.
For example, the inventory & UI in general was very slapdash, levelling and experience of the player and items was somewhat unfinished & untested and there were still quite a few noticeable bugs in place.
The main bug is still the shadows, they still allow defogging of tiles behind walls and there’s a weird ghosting issue where you can see ghostly enemies on the fogged tiles but they’re not there when you get to the tile.
Since the last update I’ve added in enemy AI, they’ll intercept you when they see you and go to attack. If they lose sight of you for too long they’ll lose interest and stop chasing. I quickly added in food to heal the player.
Anyway, for anyone wanting to see what was submitted, you can download it here
Progress of the roguelike has been slow but steady, not really had a huge amount of time to spend on it but happy with what I have managed to do.
I thought I’d post as a slight milestone has been hit and that’s shadows & fog of war. I remember spending many hours before trying to get a good 2D shadow algorithm created and failed due to heavy resource usage but I’ve come up with a new method which works quite well.
It still needs a little tidying up but it’s a start.
While I was looking at sprites for my rogue like game at Oryx Design Lab, it turns out that they’re doing a competition called Trials of Oryx which started on the day I started looking into doing the game, what are the chances?
I’ve taken this chance to accept the obvious fate and join in with the challenge. Could be tricky with a minuscule amount of free time I have for such things but perhaps it’s the inspiration I need to keep at it.
I’ll try to keep updates coming but for now, I’ve got a game to make =)
I love playing the majority of video game genres but think out of all of them, RPG games are probably my favourite game genre and delving down, rogues would have to be my favourite sub genre.
For those not in the know, rogue like games are classic dungeon crawlers which are randomly generated and notoriously difficult.
It’d been a while since I’d played one and once I’d delved in, I became obsessed. After a week of being reminded of the unforgiving difficulty it dawned on me that a started building a dungeon based game many moons ago. I checked my project folder to make sure it was still in existence, luckily it was.
Although it’s in the very early stages, the key functionality of the seeded random dungeon generator was pretty much in place. Other pieces in place is the beginnings of the path finding and sprite loading.
As you can see it’s all rather barebones, the same one sprite is used to represent the player and the enemies, stairs are yellow tiles & doors are green. Where I left it, it was mainly the dungeon generation which was being developed with everything else to be thought about later on.
I’m quite keen to pick this project up again so hopefully I’ll have more updates in the near future while the enthusiasm is in full flow =)
Not too much has changed over the last couple of months, nothing at face value at least. I’d been doing some testing of the game on a tablet and wasn’t completely happy with the city transition. On the mobile device there was a slight lag when moving between the city blocks as it was generating each city block instantly which had an effect on the performance. In usual cases this would be solved by threading but unfortunately in AS3, threading is currently only available in beta for android applications and currently in development for iOS.
After some time thinking of solutions I came up with changing the city generation from using standard loops which would run through everything there & then to using the game’s frame rate as the loop instead. This would result in spreading out the processing over multiple frames allowing to maintain a good FPS. The only drawback of this method is that it will take longer for the city blocks to be generated which shouldn’t cause too much of an issue, the city will be large enough that the generation would have finished long before the player reaches the edge of the city.
There’s been two other relatively small changes, the first is random buildings which is quite significant for a random city generator. The second change was so buildings at the edge of city blocks get grouped together into one building instead of fragments.
This is a rough example just implementing the classes used for the city. The different coloured squares represent the different building types and the roads poking out the sides are the roads which link up to neighbouring blocks.