any advice welcome from those who tried this to accomplish
I think there’s a need for at least 2 attack animations (mirrored) standing and half-height and blocking animations (for these 2 for left and right) and some moves between to assume attack position.
so animation charts to center position but always goes back to starting attack position (and maybe after time to idle)
am I thinking right? is upperBody mask the good route or doesnt matter, not a factor in transitional moves
What kind of sword fighting – European? If so, what kind? Fencing foils? Zweihanders? East Asian? Flashy wushu with butterfly swords? Japanese iaido? South Asian? Other?
What perspective? 2D like Niddhogg? First person 3D? Third person like Soul Calibur?
What kind of input? Mouse? Gamepad? Mocap? Mobile accelerometer?
good questions.
third person 3D and maybe japanese, and one sword.
control by gamepad.
not just a slash-and-die type of thing.
but mostly I’m interested in the way the moves / attacks / blocks make up a swordfighting system.
I did an IK-controlled sword wielding before on mobile, but it was kinda hard to control, so I think initiate-move-and-do-the-footwork mocap
some links I found while typing
http://www.akban.org/wiki/Japanese_sword_combat_stances
http://wychwood.wikidot.com/fighting-sword
How does it compare in gameplay to games like Bushido Blade or Hatashiai? While you don’t want to clone another game, it’s important to know what’s already been done, how it succeeded, and how it failed.
I’m curious as well.
In my case, the combat is similar o Kingdom Hearts. I’ve played the game “What if Adventure Time Was a 3D Anime” by Mike Inel, which is quite close to what I want.
http://gamejolt.com/games/what-if-adventure-time-was-a-3d-anime-game/132926
However, melee combat still puzzles me. I know that in the Unreal Engine 4, you can set up a warning in a keyframe and reference this warning in your code. would it be similar in Unity?
@ProxyDoug - In Unity, you can add animation events to keyframes.
@dibdab - How would camera control work? Over the player’s shoulder? Static side view? Adjust/spin to get the best angle on both fighters?
As for animation layers, stances, attacks, and defenses incorporate the whole body in the style you’re describing. You’ll probably want to have a large variety of whole-body animation clips instead of compositing them from separate lower-body and upper-body clips. You could still have a separate layer for the head, but it’s probably best to simply use IK for that.
@TonyLi
Thank you very much.
Hatashian prototype looks pretty good.
lets start from the basics. I think there’s no such prototype on the store that would have a more complex footwork.
if I’d want a character to turn 90 and continue to do so
but if only using one turn animation (here it is mixamo’s standing turn right 90)
it keeps going back to default position (either root motion or not)
so probably have to go with capsule rotating, but then what about other directional changes? am I missing something?
I think perhaps your root motion must not be configured correctly. Mixamo animations with root motion will rotate the character correctly.
no-no. root motion doesn’t work like that.
for rotations it doesn’t keep orientation.
unity 5.5
It may be helpful to review Unity’s page on root motion. Best of luck!
Any sort of realism in a sword fighting game is pretty hard to manage. There is a lot going on in any sword fight. The average sword fight involves a stance, foot work, and attacking and blocking with your sword. Combine that with an off hand weapon or shield which can do all of the same things. Then add in spatial awareness and navigating the environment. At the same time you have to react to an opponent doing all of the above. There simply aren’t enough buttons on a keyboard or controller to map the required input.
Most games deal with this by grouping all of this complex input into a series of moves. One move might represent a high strike, another a low strike. Other moves represent blocks. By combining the various moves in different ways the flow of combat can be simulated.
At the heart of almost every fighting game is a series of moves and counter moves.
That’s putting it politely. The fact that most armored sword fights (not that you want to even fight a guy in full armor with a sword to begin with) ended in a wrestling match should make it evident that this shit is way too complicated. Sport fencing can be simplified enough to work, but actual rapier, where you might run someone through to the hilt and finish him off the the main gauche, is still too complicated.
And those moves come down to three variables: pacing, spacing, and damage properties. Good variety makes for a much more engaging fight.
it’s pretty evident it is hard. Hatashian been workin on it for 1 year, and plan to release it 2017 Q3.
@There simply aren’t enough buttons on a keyboard or controller to map the required input.
yes, the control scheme is a challenge too.
@By combining the various moves in different ways the flow of combat can be simulated.
okay. that’s where it starts to get interesting.
I was talking technical. can root motion be used for footwork. (?)
when you have a slash anim, you crossfade to next anim, or somehow else. like full slash upwards, then do a slash back to left or choosing other anim, like stab (its about controls too).*
*probably mecanim charts with triggers (like sequence tree), triggers by control
I think the footwork could be managed if grounded by modifying animation by ik. If ik follows a circular curve > for rotating footwork (the weights curves will factor in) and straight walks leave uneffected.
then capsule rotation go by hip and set rotation limits
might work if weight change curves are right
radius * Vector2D so radius * x * y , * weight
PathMagic would be best choice for this
Both of these games look interesting.
I know about the famous bushido blade game.
And I have heard about the Hatashiai game, for a while now.
Speaking of Hatashiai, it does look kinda interesting though.