Which engine for melee based game
by Charles Kimbril · in General Discussion · 05/03/2010 (2:15 pm) · 8 replies
My Question is, which engine would be best?
I would like to build a small mulit-player melee based game, where players can fight both other players and NPC's.
The map would be small.
The focus on melee.
I own Torque 3D Professional, which comes with TGEA and TGE.
So, which engine has the best melee support?
I plan to incorporate other stuff later, but I want the focus to be melee based combat.
I am asking here as well as Torque's website, just to get more opinions.
Thank you,
Charles
I would like to build a small mulit-player melee based game, where players can fight both other players and NPC's.
The map would be small.
The focus on melee.
I own Torque 3D Professional, which comes with TGEA and TGE.
So, which engine has the best melee support?
I plan to incorporate other stuff later, but I want the focus to be melee based combat.
I am asking here as well as Torque's website, just to get more opinions.
Thank you,
Charles
About the author
#2
05/05/2010 (2:42 am)
just wondering, are the melee true melee (based on hit box collision) or simulated melee (shoot a short range projectile). I think sean's implementation belongs to the later... 
#3
The default gun mounting doesn't pull in a fully collide-able object.
But yeah, if you were willing to recode a lot of those base features, I'm sure you could get a melee game with good detection going.
Which engine? I'd say T3D, if only because the development would be quicker and easier. Plus, it's their currently supported engine, likely to get attention if you find any really tricky bugs or whatever.
In reality, NONE of them support melee (with true hitbox detection) by default, so I'm suggesting you use the shinier engine you spent good money on.
05/06/2010 (1:25 pm)
The problem is that your need to re-write the code for mounting weapons. Currently, if memory serves, collision information doesn't translate to mounted objects (in the weapon slot) because it uses an "image" of the object to cut down on CPU usage. The default gun mounting doesn't pull in a fully collide-able object.
But yeah, if you were willing to recode a lot of those base features, I'm sure you could get a melee game with good detection going.
Which engine? I'd say T3D, if only because the development would be quicker and easier. Plus, it's their currently supported engine, likely to get attention if you find any really tricky bugs or whatever.
In reality, NONE of them support melee (with true hitbox detection) by default, so I'm suggesting you use the shinier engine you spent good money on.
#4
05/06/2010 (3:56 pm)
one method is the invisable projectile the other codes a ray cast between two nodes on the mesh and applies damage when that ray crosses the player...it is also fairly easy to use with the hitbox resource.... 
#5
I've got plans to implement melee at some point in the distant future, using either polygon collisions (create polygons using the starting and ending positions of a blade edge, see if other polygons intersect them), or raycasts (but with multiple swept raycasts to catch collisions against small/thin objects)
05/06/2010 (5:42 pm)
Quote:In reality, NONE of them support melee (with true hitbox detection) by default, so I'm suggesting you use the shinier engine you spent good money on.Agree!
I've got plans to implement melee at some point in the distant future, using either polygon collisions (create polygons using the starting and ending positions of a blade edge, see if other polygons intersect them), or raycasts (but with multiple swept raycasts to catch collisions against small/thin objects)
#6
05/07/2010 (12:57 pm)
So far I have seen 2 attempts on hit boxes; Davide Archettis and Josh AlbrechtsQuote:, using either polygon collisions...Sounds like way more precise impact determination on the model hit, but you would still need some sort of model subparts to allocate damage for the various parts wouldent you?
#7
The advantage of polygon collision is just that it catches small or thin objects, especially accounting for very fast-moving blades. If you raycast along the blade once per tick, the faster the blade moves, the more likely that some thin collision mesh will just be skipped.
The second implementation I mentioned solves that by casting multiple rays each tick, at a rate determined by how fast the start and end points are moving, so you keep a constant resolution at the fastest-moving end of the casts.
05/07/2010 (11:45 pm)
Oh yes - I'm implementing hitboxes, basing them off Davide Archetti's resource. But that's a different (if related) kettle of fish.The advantage of polygon collision is just that it catches small or thin objects, especially accounting for very fast-moving blades. If you raycast along the blade once per tick, the faster the blade moves, the more likely that some thin collision mesh will just be skipped.
The second implementation I mentioned solves that by casting multiple rays each tick, at a rate determined by how fast the start and end points are moving, so you keep a constant resolution at the fastest-moving end of the casts.
Torque Owner Ken Johnston
Indiestructable Systems