Due to a dose of Real Life™, I did not have it in me to do updates or socialise much in the past few weeks. Thankfully, things have a progressed and much of the stress is gone.
On the bright side, one way I deal with stress is to distract myself by coding… a lot. I’ve done everything from updating more loadout and UI related code (booo!) to gametypes and AI (woohoo!).
One thing that can happen with game development that spans a lengthy time period is art done early in the process can start to feel outdated and fall behind the current standards. This isn’t always a bad thing (It’s the gameplay man!) but I felt like the weapon models/textures needed an upgrade since they are so visible. The work previously done by Snowfella was top notch and served its purpose very well. So again, my hats off to him…..and we are still using some of his work in the finish game as well.
So let’s start off with the high res version of the AK-74. This model is used to bake Normal and Ambient Occlusion maps for the in game model. I’ll do some more updates down the road to show some of the process.
New UE3 build this week, which delayed work in other areas a bit, but no matter.
After finishing an update of this kind, I like to check things by making a quick map in the editor, chucking in random stuff and running it with multiple players. I was already in the process (before the update) of more tests involving loadouts, so I thought I’d combine the two.
Basic bot code has been sitting there for ages. I just created a nav mesh, told them to use a loadout and let em’ rip. Considering the level of asset optimisation (none) and the age of the machine (2008), I’m happy with the results.
Also been testing and updating firearm related code.
With any luck, we’ll be shooting at each other soon, which will be nice.
You know, read out of context, that could sound weird…
Pretty rough (lighting etc), but the information travels across the network correctly, which is the most important thing.
Loadouts are saved in the JSON format to make sharing and editing them a lot easier. To make this easier to send across a network, the client will take this loadout and reduce it down to references from a item list that has been synchronised between the server and client, then send that to the server.
After verifying its contents, based on admin, gametype or scenario restrictions, the server then distributes this to all the clients, as required.
This allows for plenty of player customisation, but ensures that the server remains king.
Another decent step towards a playable build 🙂
Working on replicating loadouts – both player and item – from server to client.
Since items can attach to other items, there is a chance that attachments will run several layers deep. To test this, I added an attachment point to the ACOG and created a new weapon loadout to abuse it.
The following is the in-game results…
I had to laugh.
Needless to say, no, you won’t actually be able to do this in the final game 😛
Either way, it feels good to be working on gameplay related stuff.
… and turn it into something playable.
We have no plans to release the build shown here to the public at this time, but it is making the rounds with donators from Ground Branch’s early days. Once it has been abused and patched up for a while, we’ll consider our options.
It is really starting to shape up.
Engine updates, bug fixes, fiddling with motion capture files, UI work – all quite boring.
I spent quite a bit of time this weekend tracking down an issue that developed – which was my fault – and migrating to a new UE3 build.
Seem to have nailed the problem, hopefully for good.
I hope so, otherwise John will slap me upside the head again.
Speaking of John, he’s recently been replacing the
crappy awesome programmer art I created for the UI with proper assets, based on this mockup.
Jeremy continues creating new animations, based on earlier motion capture work.
Nothing particularly fun or interesting, but it is progress.
Currently, we have eight (8!) weapon positions planned, excluding sprinting or similar.
Why so many?
A few reasons.
What follows is essentially a cut and paste from our dev wiki, minus requirements and references.
When discussing features in Ground Branch, I often bring up points based on the idea of n00b vs pro players. Either the need to leave out or expand a specific aspect of the feature, or simply making it more open then it originally was.