No more updates?

[rdstudio]

Hello, Im wondering if there will be more updates or project just died? Last time was at 27 august and I was happy that I can play with some few new features. But from this time - nothing. Dont get me wrong Im not complaining becasue I know how complicated is game making but guys! If you wanna our support give us something at least once a month !! Some new feature to play with ...

[Arvarik]

Are you blind?^^
http://automationgame.com/2012/11/09/update-schedule/

[Daffyflyer]

Yup, this current update has taken a while as it contains a ton of new/revised stuff. We've also rewritten a lot of the calculations in response to player feedback too

[irimiadan]

Can i be a beta-tester ? Is there a section for beta testers around there ?

[Killrob]

There will be a new round of beta testers put into the beta for the start of the next subproject (turbos, and final updates to the engine designer). I'll put up an announcement when we get close to that point.

[rdstudio]

Yay new update ! Thank you guys.

New TEST MODE with manual throttle controls is AWESOME!

And I need make medium quality setting because on my high end comp maxed work about 1-2 FPS Are you guys render all to 4kx4k and make rescale to user screen becasue I dunno why this details cost so much gpu/cpu power.

[Daffyflyer]

The highest viewport slider setting is quite slow yes, but I still get 40FPS+ on my ATI 6950.

The main reason its quite slow is that we just don't have the huge teams and millions of dollars dedicated to engine performance optimization that bigger studios do.

[rdstudio]

you don't need have millions of dollars. Just good engine. Unity3d for example. Can handle all you need and keep nice framerate and have decent price.
Trust me I'm game developer with more than 10 games released.
And I know - now is too late to change engine.
Good luck and waiting for next updates

[Daffyflyer]

Oh indeed Unity would be more optimised, but it didn't do a bunch of other stuff we needed

[rdstudio]

What for example? If you not in the mood to speak about technicalities - forget about replying

[Daffyflyer]

Speaking as the artist im not the best person to answer that, but certinally having used an in house engine has meant that pretty well any feature we want can be implimented by our engine programmer. The main performance sink is the SSAO really, but that is a major reason that it looks quite good.

Regarding what unity can't do that we needed I'm sure zeussy can give a better reply, but the choice of engine wasn't made lightly.

[Daffyflyer]

Zeussy's comments on engine choice: Unity was less capable and far less common when we started 3 years ago compared to now, it also (at least at the time) was not great for complex Uis

In terms of rate of iteration Kee engine is somewhat faster to program for than unity.

Unity wouldn't offer great deal better performance than Kee engine for what we are doing anyhow, the SSAO and IBL are just kinda slow anyhow, such is the price of making pretty games on a budget

[zeussy]

Actually now that I have had sometime to think about it.

Automation was first built on the Irrlicht game engine (about 4 years ago), and Automation was always going to be a very heavy UI based game. With the floating windows etc. Most UI api's out there sucked, CEGUI, Crazy Eddies gui. All had no pipelines or workflows, and somehow were made to be far too overly complex to implement. So I knew I was going to have to write my own to get what I wanted. Now our UI pipeline is a one click export process out from Expression Blend. So we have a very powerful/professional UI tool. Ultimately we would of implemented something flash based. But Scaleform has just ludicrous licensing costs, and we have no-one that knows flash.

Unity back then (3 years ago), and still is today, is known to be pretty bad for 2D. It is mostly fixed now with various API's, but not back then. Jayelinda/Nortala had also been using it professionally for a while, so we had a fair bit of knowledge about it in the house, and she had many gripes with it back then for various things that just put me off.

Automation itself was always going to have static scenes, rooms etc. With dynamically loaded engines, so the big plus with unity, the scene editor is wasted on us. Unity's sound implementation is horrid, just basic FMOD with no designer support. So would have to make a full fmod implementation like we have anyway.

With the Kee engine we have just as high, if not higher iteration rate than Unity anyway. All game code is implemented in lua (yes that could be done with Unity pro), so code fixes/data changes can be loaded on the fly. Almost all exceptions caused in lua are non fatal, so can also be fixed while the game is running. Our asset database system allows for hot loading of art assets while the game is running as well.

We made our choice, and we are happy with it. There are some fairly straight forwards things we can do to optimize the Kee engine graphically, in the new year I hope that Nortala will be able to implement them, now she has more free time available.

[rdstudio]

Yes all those answers guys are all right for me and I understand it perfectly. I go myself through few engines too and I know that if you will start work on some engine - build tools and get know it perfectly - its hard and uneconomically switch to another one. And yes - 2d in Unity is one big mistake I didn't check in 4.0 but from what I read - they still didn't change it. And your game uses 2d quite hard.

Still I don't understand this big slowdown in maxed quality. I'm making garage simulator game now and we have very heavy models in it + all modern post process effects like SSAO, AA, DOF, BLOOM, COLOR CORRECTION, REALTIME REFLECTION and some custom effect write by me (glare effect in style of Battlefield 3 for example). All this works on about 60 FPS stable on my computer. With realtime shadows (we don't really need them) - about 20 but its very very heavy stuff. We using lightmaps for garage for this and custom shadow for car.

Sample scene (1.2 milion vertex) :


Car itself has 200000 tris. Front light detail example :



I hope you will guys optimize your SSAO shader and I will play on maxed details We all hope I think
I'm just curious - what is average vertex count in your garage scene with fully shown engine?

[fordford]

wow wow wow.great work rdstudio.

how about a camshaft and rdstudio Collaboration