Archer grabbing bow from off their back

[greebo]

It's a comfy feeling that you're still working on animations, Noisycricket. Good to have you on board, I must say.

[Noisycricket]

Uploaded.

 

Same goes to you Greebo. You and the other programmers have made it crystal clear why programmers make a higher wage than other game development positions. Forums seem pretty quiet lately though, hope it doesn't stay that way.

[greebo]

Same goes to you Greebo. You and the other programmers have made it crystal clear why programmers make a higher wage than other game development positions.

Wait, we're supposed to get a wage?

 

Forums seem pretty quiet lately though, hope it doesn't stay that way.

I hope so too, I hope the team will gravitate towards each other again after release to push the mod further. The (hopefully) positive feedback ought to give us incentive.

[Noisycricket]

Wait, we're supposed to get a wage?

 

Of course.... thats why I signed up! No doubt in my mind...

 

 

 

We get a wage right?

 

 

 

...

[Noisycricket]

Just so you all know, i didn't give the aim animation any movement, its just a pose. I could have given it some movement, but in testing the ranged A.I., I realized the player would be 1: running, and 2: facing away from the guard. Nobody is going to stand there and get shot in a gameplay situation, thus taxing the graphics card for that didn't seem worth it. Also, i assumed the code will have the A.I. track the player, thus granting some movement.

[Crispy]

The graphics card doesn't care whether animated AIs are actually moving or not, and I'm fairly sure the animation code (which runs on the CPU) wouldn't either. It'll still go through the same sequence of transformations every frame even if the sequence of transformations is the same all the time. In any case, this is extremely premature optimisation and not at all likely to have any noticeable effect.

 

Please just base your animation decisions on what looks good, not on vague theories about CPU/GPU loads.

[Noisycricket]

I know, but it all adds up Crispy. Theres no point in me not making considerations like this. If you want a moving idle animation, I'd be happy to put it on my list, it would be pretty easy. Also, I hate people trying to regulate my behavior, unless your comment is the result of some internal discussion, kindly put it where the sun doesn't shine.

[Crispy]

Also, I hate people trying to regulate my behavior, unless your comment is the result of some internal discussion, kindly put it where the sun doesn't shine.

 

I'm not necessarily agreeing or disagreeing with what you did, I'm just pointing out that you don't need to worry about performance loads when creating animations. Sorry if my comment rubbed you the wrong way.

 

I know, but it all adds up Crispy. Theres no point in me not making considerations like this.

You appear to have missed my point. If doing "A" causes performance load "X" and doing "B" also causes performance load "X", then performance shouldn't be one of the factors that causes you to choose A over B, or vice versa.

 

In this case, A = held animations, B = moving animations, and X = small anyway.

 

This isn't intended as a criticism, just advice for next time.

[Noisycricket]

Well, Crispy, your post came off to me as "Hey lacky, don't think, just animate, K dumbass?

 

Are you sure about the performance? I mean, each vertex has to be moved around the local axis while tracking the player, but then for the bow_aim_idle, isn't each vertex further calculated for adjustment around the torso joint? Either way im fine. Hell, i'd prefer an idle myself. It just seemed from some of the discussions that processing the A.I. is hurting some of the framerate, as mentioned in the conversation wherein someone mentioned using the monster-kill (or whatever its called) command solved a greatly lagging framerate. Who knows what % of that is from animation or path-finding or rendering, etc. I don't, but i would like to know out of curiosity.

 

I'll create a combat_bow_aim_idle in addition to renaming the current one, combat_bow_aim_pose so people can choose for themselves, but as far as what should be the default animation, think about what your going to do the split-second you clearly see a guard aiming a long bow at you.

[Crispy]

That certainly wasn't my intention, and I'm sorry that it came off that way. I have great respect for the skill and thought that goes into animating. Especially since I've dabbled myself on occasion, so I have some idea of the amount and type of work involved. I'm just trying to contribute help from my area of speciality, which is programming (and, by extension, performance analysis).

 

I don't know for sure what % of the time spent on AI goes towards animation/pathfinding/thinking/rendering, since I haven't measured it, but I can make some educated guesses.

 

My comment about animation performance load being identical regardless of movement is based on a few reasonable-sounding assumptions. Specifically, that transforming the vertices from the AI's neutral pose into its current pose is by far the biggest contributor (since there are lots of vertices per AI). This transformation is done by chaining together a bunch of matrix multiplications (a matrix multiply can be viewed as a transformation - due to linear algebra, which I won't go into). My guess is that these transformations are not cached between frames, since 99.9% of the time they change every frame anyway. Therefore the mere fact that a mesh is being deformed by a skeleton causes a performance hit, regardless of how the deformation changes between frames.

 

My second guess is that even if my first guess is wrong, then the difference in performance load will not be noticeable on a real map. (i.e. It won't have an effect on the frame rate which is greater than the background noise.)

 

Granted, I could be wrong. In truth this should be settled by benchmarks, not words. If someone wants to set up such a test then by all means. I'd do it myself but I don't have time... in fact, I'm late for a lecture right now. Oops.

[Noisycricket]

Is there any need to have a "cancel attack" animation for bow-firing (i.e. once the bow is in the drawn position, relax it and um... I guess drop the arrow), or does the AI not have such a behaviour anyway?

 

Do we want a cancel attack animation, now that the code is changing?

[greebo]

It depends on whether we want to have AI cancel their bow draw when they get hurt by a shortsword. I'd vote for yes.

[Noisycricket]

Its so easy to do, ill make it and upload it.

 

EDIT: along with the moving bow idle.

[Noisycricket]

That certainly wasn't my intention, and I'm sorry that it came off that way.

 

Its so easy to misjudge the intention of posts over the internet without seeing the person speaking, I just read one of mine that sounded strange upon second glance. Perhaps i jumped to conclusions too quickly... anyway, I'm going to click "add reply" and start forgetting about it.

 

EDIT: OH! and about that comment going up you know where, lets remove it: *snaps latex glove and motions Crispy to turn around and bend over while nodding his head just ever so slightly too enthusiastically for Crispy's comfort*

[Noisycricket]

I searched youtube videos for reference of the bow movement and tried to stay away from professional athletics, the only video that showed any significant bodily movement was of a small 8ish year old girl using a bow. But, as promised, I made at least some exaggerated movements that will be visible from across a room, which I uploaded as aim_idle. I also renamed the non-moving version to aim_pose.