Character Animation

[]-UnderTOW-[]

What's the plan for the main character animation?

Since we have an awesome character model and there's going to be no third person (thank you!), i suggest we go the way of Thief 3, in one of the few things ISA had the right idea (but still managed cocked it up. Riddick's was done well), and have no view weapon models/animations and have the camera in the player model attached to specific camera bone. This way we can have full immersion by being able to see yourself and we can also have a player shadow.

To get this to work nicely there'd need to be some extensive animation blending. What are Doom 3's capabilities of animation blending? Doom 3 uses IK so that's good (no feet cliping through ground).

[Ishtvan]

I would wait for someone higher-up to post, but I seem to recall we decided against "body awareness" (ie, camera stuck to head bone of model and all player model animations used) in the past, due to the fact that we'd need to "suck" the player into place before starting the animations, like T3 did, in order to start off with the right initial conditions so that the arm/leg animations look right for stuff like climbing ladders, picking locks, etc.

 

No one was a big fan of the "suck into place" animations of T3, and we'd rather be a floating armless/legless character on a ladder than get sucked in to the ladder at the start of climbing

 

I don't know the details (not an animator), but I think D3 does have animation blending. I've messed with some of our animations and cut them in half, and D3 would move fairly smoothly from the ending coordinates to the start coordinatesof the next animation.

Edited March 15, 2005 by Ishtvan

[Renzatic]

If I remember correctly (need tha clarification) we decided to nix body awareness because we thought it'd take too much time to coordinate everything to coincide with the animations...it was a "time best spent elsewhere" issue.

 

But if you can do it and can avoid the problem with being sucked into place that Isht mentioned above then I wouldn't have any complaints. I like the idea of body awareness, it can add alot to the mood and it'd be cool to see in the mod if it can be pulled off well.

[]-UnderTOW-[]

Well, the problem of using the normal first person style is that we have a great player model and it'll go to waste (since you're never gonna see it), and we wouldn't be able to have a player model shadow. It doesn't bother me that much, cause there are plenty of other characters to do .

I don't mind a "sucking" effect if the transition is nice and smooth. I can't recall what it was like in Thief 3 but i don't remember it bothering me too much, what i didn't like was some of the other things about it which is mostly accountable by just not doing it propperly (such as turning, if you look 90 degrees to your side and move in that direction he'll run in an arc to that direction. Often causing you to run off a ledge).

It is a difficult thing to pull off correctly and i understand not wanting to go that way

[New Horizon]

Ahhh, but we can have a player shadow in doom 3 without using body awareness. However, it will involve creating animations for the Thief model.

 

If you play around in Doom 3, you will find that you can play in 3rd person if enabled through the console...you can also enable the player shadow in both 1st and 3rd person through the console.

 

When playing with the shadow on in 1st person, the game just creates the shadow from the 3rd person model.

 

Perhaps you could run some tests in D3 by working in body awareness. The assets are already in place. In the long run though, I don't think body awareness is worth the losses..unless it could be proven that we can somehow avoid all the problems Ion came up against. It might also contribute to a loss in performance as well.

[Renzatic]

In my opinion it should be something you attempt on down the road when alot of the other animations are done and some of the central gameplay mechanics have been finalized.

 

As cool as it could be, there are too many potential problems that could arise or guff up other areas of the game..for something that seems so simple at first glance, it makes itself central to the gameplay just by it's presence. You'd have to redo alot of stuff just to make sure the animations are played in key, like setting ladder rungs exactly the right length apart so that the hands will match up with them while climbing, lockpicking would have to be done with your nose smacked against the door, and..well..alot of other things.

 

If it can be pulled off dymanically without changing the pace and feel of the game then I'll be all for it, but like we decided before I think it's time best spent elsewhere...at least for now.

[]-UnderTOW-[]

Having a player shadow without seeing the player model is going to look silly (which is why iD didn't have it enabled in Doom 3, i'm guessing), and doing player animations just for a shadow would be a bit of a waste of time. As i said, i don't mind not going with it. I'm happy to go either way.

 

Also, there is the option of th way Halo 2 did it. They had normal first person camera with view model. It worked, felt and played just like a normal FPS camera, but when you looked down you could see the player model (i imagine it'd only be the lower half, but the shadow was the full model). So it's just a visual effect and didn't screw with the movement, although when you looked down or up, it seemed as if the camera was attached to some part of the player model.

[Renzatic]

You're welcome to experiment with it if you want...but it'll be alot of work for something that might end up on the cutting room floor when everything is packed together and ready to roll.

[oDDity]

It's all rather pointless unless we have third person (which we definitely are not) How much of the player model or anmations could you see by 'looking down'?

Anyway, his huge shlong would constantly get in the way so you wouldn't be able to see much.

We'll see the player model in mirrors and in cutscenes, so you'll have plenty of animation to do there anyway.

In fact, by the time you see just how much animation needs to be done all over place, I doubt you'll be volunteering any extra work like this)

[]-UnderTOW-[]

I think the Halo 2 style thing would be very simple to implement.

I reckon it'd just require player model animations, 2 versions of the model (one which is just the lower body and part of the torso which is seen by the player, and a full bodied one which is just used for the shadow and mirrors). Then have a normal first person camera and have the half player model visible to the player. Could just do that to test it out first off with the doom model.

[Macsen]

I'd like to be able to see the legs when I look down. Especially taut, manly legs.

[New Horizon]

I'd like to be able to see the legs when I look down. Especially taut, manly legs.

To Quote oDDity.

 

Anyway, his huge shlong would constantly get in the way so you wouldn't be able to see much.

 

Sounds like our guys steals the booty rather than looty.

[]-UnderTOW-[]

*slowly backs out of the room*

[Darkness_Falls]

Along the same lines as everything else spoken, some of us like that locks can be picked from the side and farther back in Thief 2. If we all the sudden have body awareness like T3 and try to pick a lock, though, we'll have hands flailing in the air at nothing; assuming we tied lockpicking animation to arms/hands. Being sucked into the lock (like T3) would of course fix this, but that will be the day I don't play Dark Mod

Edited March 15, 2005 by Darkness_Falls

[Springheel]

The design decision was that we would not be supporting body awareness at all. The Thief model will be seen in in-game cutscenes and mirrors, but that's it. Absolutely NO 'sucking in'.

 

I think working on the animation for the other AI would be more useful than starting with the Thief.

[Macsen]

But all we need are legs, not entire 'body awareness' (such a silly term, sounds like he's going through puberty or something doesn't it? 'Oh, I'm becoming aware of my body! Those spots weren't there before!') It would also be to the master thief's advantage, because you could look down and see exactly where your feet are landing, and when. It will also look strange if the Thief has a shadow but doesn't have a body to go with it; the shadow will just emerge out of the floorboards.

[Springheel]

But in order to achieve that, Macsen, you have to do all kinds of coding to make sure that those legs are in sync with your surroundings--with ladders when you climb them, as an example. This can only be achieved by various forms of 'sucking in' which we don't want.

 

I suggest people do a search for 'body awareness'. We hashed out all these issues months ago.

[]-UnderTOW-[]

But in order to achieve that, Macsen, you have to do all kinds of coding to make sure that those legs are in sync with your surroundings--with ladders when you climb them, as an example.  This can only be achieved by various forms of 'sucking in' which we don't want.

Not really. You're not going to be able to see your legs hit each rung on the ladder (your perspective wont allow it), all you would see is your legs doing a ladder climb animation. The only coding required would be to allow the player model to see the specific player model. Shouldn't be difficult to test out now with the doom player model.

 

BTW, why are there ingame cutscenes? That's one of the things that totally ruined the feel of Thief in Thief 3, the fact you can see Garrett at anytime and they display him in cut scenes. In Thief 1 and 2, you only ever see parts of him in those cut scenes, only ever see him in the shadows, there was still some mystery about him there.

Edited March 15, 2005 by ]-UnderTOW-[

[Springheel]

would see is your legs doing a ladder climb animation.

 

Well yeah, but you have to make sure your legs are actually hitting the rungs properly, and not clipping the sides, etc. And ladders are easy. Once you get to rope arrows and climbing vines it gets even more complicated.

 

BTW, why are there ingame cutscenes? That's one of the things that totally ruined the feel of Thief in Thief 3, the fact you can see Garrett at anytime and they display him in cut scenes. In Thief 1 and 2, you only ever see parts of him in those cut scenes, only ever see him in the shadows, there was still some mystery about him there.

 

There's nothing about in-game cutscenes that says you have to show the thief in bright light with his face exposed. We wouldn't do the silly comic-book style stuff of T3.

[]-UnderTOW-[]

Ah i see what you mean about the rope. I think it'd have to be normal first person camera, no player shadow (incedently, how difficult would it be to get the AI to recognise the player shadow and would it ruin gameplay?) and no mirrors for this to work propperly, first off. Then later down the track, player model animations for mirror and player shadows. Then further still, maybe an attempt at "body awareness".

 

And i'm relieved about the cutscenes although i prefer Thief style cutscenes over any ingame one. I really loved that style and it didn't break the immersion.

Edited March 15, 2005 by ]-UnderTOW-[

[Springheel]

And i'm relieved about the cutscenes although i prefer Thief style cutscenes over any ingame one. I really loved that style and it didn't break the immersion.

 

I think we'll be doing that style for our own campaign, but other FM authors may wish to use in-game cutscenes, since the ability is already there.

 

Ah i see what you mean about the rope. I think it'd have to be normal first person camera, no player shadow (incedently, how difficult would it be to get the AI to recognise the player shadow and would it ruin gameplay?) and no mirrors for this to work propperly, first off. Then later down the track, player model animations for mirror and player shadows. Then further still, maybe an attempt at "body awareness".

 

That sounds like a reasonable workflow. Although the player shadow already exists...the only issue is that when you look down it seems to come from nowhere. If that could be addressed we wouldn't have a problem.

 

Anyway, as I said, getting the animations for AI is by far our top priority.

[sparhawk]

Well, the problem of using the normal first person style is that we have a great player model and it'll go to waste (since you're never gonna see it), and we wouldn't be able to have a player model shadow. It doesn't bother me that much, cause there are plenty of other characters to do .

D3 can do the shadow of the player model without the need of placing the actual model. And there are also mirrors in D3 which we certainly will have in TDM maps as well, so the player can see himself even if it is not fully animated in all situations.

 

And of course there can be in-game cutscenes as well where you can see the model.

[sparhawk]

-UnderTOW-[,Mar 15 2005, 01:58 AM] Having a player shadow without seeing the player model is going to look silly (which is why iD didn't have it enabled in Doom 3, i'm guessing), and doing player animations just for a shadow would be a bit of a waste of time. As i said, i don't mind not going with it. I'm happy to go either way.

Having a player shadow still serves a purpose. The shadow can extrude into another room where the AI can notice it. Or on walls if you have a lightsource behind it. I would like to see this in our mod so we can still make use of the shadow even though we don't need a player model in-game.

[sparhawk]

Although the player shadow already exists...the only issue is that when you look down it seems to come from nowhere.  If that could be addressed we wouldn't have a problem.

That's not really a problem anyway. It would be nice to have this, but it's not a urgent requirement. And if we never get it solved, then I still would have no problem with it.

[oDDity]

Never mind all this nonsense for now.

The first thing to do is get the base_guard class of animations done.

That's runs, walks, idles, pains, surprise, searches, etc, that can be shared among all of the various guard classes, and then the more specific weapon attacks for the various weapons the different factions carry.

There's still tons to do after that, but that should keep you going for quite a while.

You can just use any generic human model for now and then attatch the real meshes to the animated bones later.

From Maya, the models must face the positvie X axis, you can download the model from the 'model for scoobs' thread in this animation forum to see the general scale required. We'll want to use the same rig and animations for all of the guard class.

Edited March 15, 2005 by oDDity