by date
Plan for Melv May
Plan for Melv May
| Name: | Melv May | ![]() |
|---|---|---|
| Date Posted: | Nov 30, 2003 | |
| Rating: | Not Rated | |
| Public: | YES | |
| Comments: | YES | |
| RSS Feed: | or Subscribe with . | |
| Profile Page: | View profile page for Melv May |
Blog post
Strategem development is coming along really well, having had the last week off work specifically to tackle some the most technically challenging aspects of the game. I didn't get through as much as I would've liked but I can't grumble at how much I did get through. We've got an excellent AI working now (thanks to Craig) and it's looking like it will kick your butt.
Currently, alot of the screens prior to actual gameplay are functional but without GUI skins or misc artwork. This is intentional as it really is a low priority to gameplay at the moment but we should be starting to tackle them immediately after Christmas. I think I'm finally happy with the gameboard/entity animation system/networking now as I've been looking for ways to make it much more efficient in terms of network bandwidth.
Over Christmas, I'll be working on the powers which bring much fun/chaos to the game. Our artist (who has lots of web experience) is taking a timeout to do the web-site which unfortunatly means a delay on some of the games artwork but we all thought it was about time to get the web-site up and running and should be looking at a grand opening fairly soon. We'll endeavour to keep everyone updated as to the progress on a weekly basis.
Alongside this, I've been working on refining the new Particle-Engine model. For those who don't know, Strategem will require a pretty sophisticated array of particle effects for battling/powers. To this end I've undertaken the task of creating a neat data-driven particle-engine which will be showcased in the game but will lead to a standalone product for the community to use sometime after that.
This week I've streamlined the design a little as I found that the reusuable library model I had was way OTT. It's certainly much better and simpler now which should naturally lead to better performance. At the moment, I'm dry-running the model to see if it's capable of creating the wealth of different effects that I've been thinking up.
Help would be much appreciated here in terms of submitting different effect ideas so that I can go away and see if my engine could create them. Perhaps some may be easy with a minor design change, some may be too far outside of the design; either way, an email or post here would be much appreciated.
... back to it. :)
- Melv.
Currently, alot of the screens prior to actual gameplay are functional but without GUI skins or misc artwork. This is intentional as it really is a low priority to gameplay at the moment but we should be starting to tackle them immediately after Christmas. I think I'm finally happy with the gameboard/entity animation system/networking now as I've been looking for ways to make it much more efficient in terms of network bandwidth.
Over Christmas, I'll be working on the powers which bring much fun/chaos to the game. Our artist (who has lots of web experience) is taking a timeout to do the web-site which unfortunatly means a delay on some of the games artwork but we all thought it was about time to get the web-site up and running and should be looking at a grand opening fairly soon. We'll endeavour to keep everyone updated as to the progress on a weekly basis.
Alongside this, I've been working on refining the new Particle-Engine model. For those who don't know, Strategem will require a pretty sophisticated array of particle effects for battling/powers. To this end I've undertaken the task of creating a neat data-driven particle-engine which will be showcased in the game but will lead to a standalone product for the community to use sometime after that.
This week I've streamlined the design a little as I found that the reusuable library model I had was way OTT. It's certainly much better and simpler now which should naturally lead to better performance. At the moment, I'm dry-running the model to see if it's capable of creating the wealth of different effects that I've been thinking up.
Help would be much appreciated here in terms of submitting different effect ideas so that I can go away and see if my engine could create them. Perhaps some may be easy with a minor design change, some may be too far outside of the design; either way, an email or post here would be much appreciated.
... back to it. :)
- Melv.
Recent Blog Posts
| List: | 06/10/08 - The long overdue blog... 10/24/07 - New Baby Girl... 03/19/06 - The GG Journey... 03/05/06 - Just arrived in Eugene... 11/18/05 - Plan for Melv May 10/10/05 - Plan for Melv May 09/01/05 - Plan for Melv May 07/27/05 - Plan for Melv May |
|---|
Submit your own resources!| Phil Carlisle (Nov 30, 2003 at 19:51 GMT) |
Engine trails.. kinda like a bunsen burner flame.
Really nice looking muzzle flash (look at the max payne slowmo demo in the bit in the lobby of 3dmark 2001? or 2? where they slowmo the gun firing and the muzzle flash slowly creeps out of the weapon. I want that nice sorta volume type effect.
Some way to apply randomness to different elements, for instance, having random muzzle flashes of slightly different looks.
I'm having the ships break up.. some way of describing "fault lines" which can then emit glow/fire particles would be nice.
Volumetric lighting on particles.. I've mentioned this already, but I think its doable using a texture lookup for the lighting, just have a gradient ramp which is oriented correctly as you emit particles it looks up the particle emmission colour in the texture.
Actually.. melv, if you want to see effects, I really cant think you'd get anything anywhere near as effect ridden as Zone of Enders on the PS2. Its got LOOOOOTS of stuff.
Emitters which emit trails.
Must be able to attach emitters to anything
It'd be nice if we could script or code the emitter movements so that for instance, we could do a missile "swarm", with emitters which contain trails, glows and maybe smoke etc..
A nice volumetric sort of explosion system (nice one in ZOE2) would rule..
Man, there are so many different things I can think of.
Phil.
| Johnny Hill (Dec 01, 2003 at 03:07 GMT) |
www.gamemyth.com
Its a Chinese online action MMORPG the effects are unbelieveable. :)
Edited on Dec 01, 2003 03:09 GMT
| Melv May (Dec 01, 2003 at 12:47 GMT) |
I've looked at both these sites and most of the images they contain and I've yet to see an effect that couldn't be produced with my current configuration. I don't want to get the particle engine confused with the attaching of objects. It's not the particle engines job to attach to objects, It's the other way around.
There needs to be a more efficient and less complex method in the engine and I intend to give examples on how to do this without modifications and with, which will be an example of how I did it in Strategem. I could easily put in a method to attach to particles but it would be much more beneficial if I showed a system where any object could be attached to any node rather than it being specifically for particles.
There will always be script/c++ methods to control the particle engine with basics such as moving/reorientating/time-scaling them but more complex ones such as actually generating the effects from code if required. Basically, the API that is used by the pre-baked particle editor itself.
You'll be able to set-up complex swarming-like behaviour using the graph system using offsets and rotations. The inputs to these graphs can be randomised over time to produce very complex bahaviour. Shield effects, sword motion effects, beams, explosions, fire, rain, snow, primitive shapes, volumes, effects spawning particles and other effects are all easily done.
The more specific effects you describe would be scripted such that you design an effect such as a glow/fire effects (from a ship breaking up) and generate/orientate them as required. This is going to be made very easy and I've got a simple system devised to make generating effects on the fly much easier, including a simple to use trigger system for controlling the effects through various pre-defined stages e.g. starting shield effects, producing shield collisions effects, shield failing effect all from a single effect but controllable through user-define keywords via script such as:-
shieldFX.activate();
shieldFX.trigger("shieldon");
...
shieldFX.trigger("shieldhit");
...
shieldFX.trigger("shielddamaged");
...
shieldFX.deactivate();
The approach I've taken would allow most effects to be data-driven only but ones that require specific controls to be scriptable/C++. This way, the data-driven model is not overwhelmingly complex for users but still offers the power for much more complex effects via scripts if needed. I suspect that 99% of the time, the data-driven model would be more than sufficient and triggering would be limited to starting/pausing/stopping.
I estimate that before Christmas, I should have the data-model for it fixed and validated.
Any more examples are most welcome. :)
- Melv.
| Jarrod Roberson (Dec 03, 2003 at 20:14 GMT) |
Just about every non-melee attack is nothing but a massive conflaguration of particle effects! LOTS OF EYE CANDY FOR SURE!
| Mike Stoddart (Dec 04, 2003 at 17:12 GMT) |
My brother is incredibly keen for me to write it, but he couldn't help with the code! :P
You must be a member and be logged in to either append comments or rate this resource.



Not Rated


