A bit It does look a lot like the real world. But it still looks good. I have to make my own showcase too.... soon.

Thanks guys ... the game's completion is a way off yet but who knows, hopefully a playable level might be on the cards in the not too distant future!

(04-02-2014 07:04 AM)Rubeus Wrote:  The picture on the wall is freaking me out.

lol ... it freaks me out too, that guy's got such an incessant stare!!!

(04-02-2014 07:16 AM)AndrewBGS Wrote:  I have to make my own showcase too.... soon.

Look forward to seeing that Andrew.

I wish others would post more too, helps people feel that it's a vibrant community and acts as inspiration.

Well I would, but my world really doesn't look like much.
I focus on functionality, I usually play on a small square empty world with just my character and a couple of objects I test on it. I kept thinking one of my friends would help me with level design... fat chance.

Anyway I started working on that also, I'll try to make a showcase as soon as it starts looking playable.

Definitely, to be honest it can be the prettiest level in the world but ...

no functionality == no game!

I'm far more impressed by functionality than I am graphics, combine the both and things really start to look good, but my advice to anyone is to put time into developing the functionality first, which it sounds as if you are doing, the rest can follow.

My village and surrounding terrain acts as my test ground, the level will need to be expanded massively once I have all the game mechanics in place and working.

So I look forward to seeing some of your functionality demoed, and anyone else's for that matter!

A bit more atmosphere

   

These screens look great but I can't say I'm surprised. Your work is inspiring Pixel.

That looks on the outset to be a relatively simple scene. Any idea why your FPS is dipping into the 30s?

Thanks Rofar, that's appreciated

(04-02-2014 11:26 PM)Rubeus Wrote:  That looks on the outset to be a relatively simple scene. Any idea why your FPS is dipping into the 30s?

Yeah, pretty much. There has been no attempt at all to optimize the scenes at this stage. It's a test level to refine all of my game functionality really. I left physics based rain on in that shot which I know takes a good 15 fps straight off the top, there is also a lot of AI running with guards patrolling the village walls and externally in the level. I've noticed the shadows cast by the sun take a fair bit of frame rate off as I look into the village (largest poly count in the frustum); as when I switched the day night cycle on again I noticed the frame rate shoots up at night for the same camera view.

I really need to experiment at some point with the level and get to grips with what factors are key in dropping the frame rate and try and optimise it all. Whilst it's important to do that it's a way down on my list of things to do right now!

Ok, that makes sense. My strategy has been to develop every feature in my game as optimized as possible straight from the beginning. I found doing it this way makes it easy to see how much each system added affects the FPS.

Sounds like a good strategy

My code tends to be pretty well optimized. I profile things during development to ensure my code is reasonably efficient; which most of the time it is.

The physics based rain is an exception to that as I've not really looked at it since I introduced it, I guess I figured it was always going to be fairly computational due to its nature and probably most of the time consumed is by the physics engine. However, I really should do as there could be room for improvement. Altering the number of rain drops makes quite a difference of course, it's set fairly high at the moment.

I suspect the main area for improvement regarding frame rate lies in the level geometry and the game engine graphics settings, I think I could improve the frame rate quite considerably here. I was planning on experimenting with settings once I have the game mechanics in place so I have a reasonable idea of constraints I face when I come to expand the level design.

I'm happy with the premise that if my code is efficient then the rest is down to fine tuning the engine settings and optimising the geometry in order to achieve the maximum frame rate for this code/engine combination.

Hope I'm not being too naive in my approach as I have no previous experience of bringing a game to completion and no one to call on for advice other than you good folk

Indeed, this thread is very inspiring, and insightful, thanks for sharing Pixel.

You're making good progress all in all. Nice scene in the last couple of shots, the candle light makes all the difference to the atmosphere, looking forward to seeing it in motion.


Regarding fps.. just speculating but I suspect EKI One has a not negligible base performance hit, even if you're running an empty simulation (which you don't), and then it slowly rises as you add complexity (assuming you're not cpu bound yet).

That physics based rain can eat up a lot of fps as you said, and it seems it's not only affected by it's settings, but also by level geometry and other physics stuff happening around (from my limited observations). While you can't directly optimize physics calculations (other than decreasing the amount of them), the rain offers an indirect optimization opportunity in the use of fog, limited visibility, skipping of effects like bloom, atmospheric scattering, cloud shadows etc., and don't know if it's actually doable or if it looks right, but maybe turning off shadows for directional light, and going with lightmaps and/or AO only. (when it rains it's usually overcast, very soft, diffused shadows similar to what you can get with ambient occlusion).
And of course, culling the rain altogether when indoors, ..maybe placing billboards just outside the windows facing inside could do the trick in some situations? (assuming your roofs don't have any holes in them)

If you're using a lot of arteria3D models then yes, you'll be able to gain some performance by optimizing the geometry and textures/materials. Unfortunately, it's quite a bit of work.

Thanks again Para, you always provide great feedback and advice.

I set the maximum time slice that EKI One is allowed per frame which sets a limit on the amount of time that the AI engine can utilise per frame before being forced to return. Currently it's not even approaching this although as you rightly say as I ramp up the concurrent AI routines this will increase and I will need to finally reach a compromise if I'm not to impact acceptable frame rates or risk the AI not having sufficient time to execute completely.
Fortunately this is not an issue as yet. EKI One seems to be a pretty lean and mean machine, which it needs to be given its role, but clearly it will always have some impact on frame rate.

Regarding your comments on the rain and the opportunities for optimisation that arise from that ... great advice. I've taken that all on board.

Shadows definitely have a significant hit and again I think it's a question of fine tuning and reaching a compromise. I love shadow mapping and would hate to have to do without it as it adds so much to the scenes.

Geometry optimisation is the bit I'm dreading as I'm sure it's a significant factor in this and for a programmer non artist this is a daunting prospect. However, I've noticed in the past that it appears larger combined meshes seem to give better performance in the engine (I think I read recently a comment from Greg saying that the engine will break large geometry up as required and perhaps does a better job). So there main be something to be gained from simply combining groups of meshes into larger ones. Some experimentation should reveal the validity of this or not!

It's all smoke and mirrors at the end of the day and anything goes really, so long as we achieve the desired end result. It certainly pays to keep open minded about solutions and techniques for all aspects of game design.

Animated quads are just so versatile. I've mostly used them for flame and spell type effects, here I'm using one to simulate rain splashes on the ground to compliment the particle based rain:





This was set up by simply defining a static animated quad and laying it on the ground. It's not necessarily the most practical use of them but demonstrates the differing application of these visual effects.

I have support for static, and two variants of bill boarded animated quads now which pretty much gives me the scope I need for most effects. For any one who is interested below is a screen shot of how I set this up in the editor, really simple to create and apply:

   

Nice features as always Pixel Perfect, have you considered selling some of your features, such as your weather system on the store?