Getting The Thief Model In Game

[Springheel]

I'm slowly adding Odd's prop models to CVS, but there is a lot of stuff in there that's out of my league, like the architectural models, animated files, etc.

 

Oddity apparently had the Thief character in game already, at least on his system. How hard would it be to get him on CVS and swap him with the marine player model for the May 31st deadline? BT?

 

We need the thief ingame before we can address the AI size issue, and before we can really test a size standard for future maps.

[New Horizon]

I'm slowly adding prop models to CVS, but there is a lot of stuff in there that's out of my league, like the architectural models, animated files, etc.

 

Oddity apparently had the Thief character in game already, at least on his system. How hard would it be to get him on CVS and swap him with the marine player model for the May 31st deadline? BT?

 

We need the thief ingame before we can address the AI size issue, and before we can really test a size standard for future maps.

 

I'm pretty sure the basic animation files are ready. I seem to remember odd making walk, creep, and lean animations. I've been really anxious to see how the new lean animation affects our lean since it's tied into the model.

[BlackThief]

hard to tell - I just browsed through all the farrell files and beside some specific anims like he showed to us in his video there seem to be various walk-cycles, crouch and lean-anims. One thing I noticed was that in some files the thief has arrows and a blackjack but most models only had the sword and bow - that's not only a modelling/animating issue, but also a an issue we probably have to discuss with the programmers - the d3 marine never had any weapons connected except the current weapon he's carrying. I'm not sure, how we're going to do that with our thief - we have the blackjack, sword bow and quiver - when these belong to the thief(body) model and are animated with it we'd have to do all animations for every possible combination of blackjack, sword and bow.

 

if they are connected to the player model in another way we'd need only animations for the body + the attacks. I've no idea if this will work properly - it obviously does in most of the RPGs I played, but I can't remember seeing an object connected to an AI like that in d3. Most likely this won't be a big task for our programmers though.

 

so all in all as a realist I'd say, no - it'll take longer to get the player ingame.

[Springheel]

Hmm, I actually wasn't thinking about the animations. We've talked about wanting the actual player model so that we can see how it matches up with the lightgem diamond, as well as to see how the height of the player model matches up with the camera. We don't need animations for those things, right?

 

Since the player animations are only going to be used with mirrors and cut scenes (and posssibly shadows), they really are pretty low-priority, IMO. But testing the size of the model against the camera is something we want to get out of the way sooner rather than later.

[New Horizon]

Hmm, I actually wasn't thinking about the animations. We've talked about wanting the actual player model so that we can see how it matches up with the lightgem diamond, as well as to see how the height of the player model matches up with the camera. We don't need animations for those things, right?

 

Since the player animations are only going to be used with mirrors and cut scenes (and posssibly shadows), they really are pretty low-priority, IMO. But testing the size of the model against the camera is something we want to get out of the way sooner rather than later.

 

Hmmm, I'm not sure Spring. I think the animations are tied into the player behaviour more than people realize.

For example, the model actually has to lean and the walk animations have to play for the footsteps to work..etc...

 

It's basically a hybrid of first person and body awareness...but the body is invisible. I think that's how they did it in FEAR. T3 just strapped the camera right onto the model, but both FEAR and doom 3 use a first person system with a 3rd person model just tagging along. Makes for a much more traditional feeling control scheme.

[Springheel]

T3 just strapped the camera right onto the model, but both FEAR and doom 3 use a first person system with a 3rd person model just tagging along

 

It doesn't sound very different to me, if the camera can't move without the model also moving.

 

I guess my main concern is just getting this size issue taken care of. It's kind of hanging over everything at the moment--if the AI have to be resized, it affects Dom's animating, if it turns out the AI I've been using as a size standard for the models is actually wrong, then the models may also have to be changed, etc.

 

It would be far too easy to have the issue fade from people's memory and just continue on as normal, but the longer we do the more things that may potentially have to be changed.

 

Step #1, (as I see it) is getting the thief model into the game (or at least the editor) to compare the height to the camera. Because if it turns out the camera is actually over the model's head, or in his chest, that's a problem, correct? If the model matches the camera perfectly, then we can use it as a size standard for all the other models.

[New Horizon]

It doesn't sound very different to me, if the camera can't move without the model also moving.

 

The camera can move without the model moving, you just won't hear footsteps. It's the opposite approach to T3. In T3 the camera is welded to the third person model, there are absolutely no first person remnants, and the model determines how the camera reacts in every circumsntance...it's third person zoomed in basically. The model in Doom 3 is controlled by the First person system...instead of the other way around. First person rules apply to how the player moves, camera bounce...all that jazz...but the model is pulled around by the camera in a sense, instead of the other way around. That's why the doom 3 player doesn't fall off of ledges so easily...or wobble when turning around. It's really how they should have done it in T3. Anyway, really not an issue for us...I just find it interesting that id did it the right way and it's there for anyone who wants to take advantage of it.

 

I guess my main concern is just getting this size issue taken care of. It's kind of hanging over everything at the moment--if the AI have to be resized, it affects Dom's animating, if it turns out the AI I've been using as a size standard for the models is actually wrong, then the models may also have to be changed, etc.

 

It would be far too easy to have the issue fade from people's memory and just continue on as normal, but the longer we do the more things that may potentially have to be changed.

 

Step #1, (as I see it) is getting the thief model into the game (or at least the editor) to compare the height to the camera. Because if it turns out the camera is actually over the model's head, or in his chest, that's a problem, correct? If the model matches the camera perfectly, then we can use it as a size standard for all the other models.

 

We could just pop the lwo of the Thief model into the editor...go into game mode and do a pm_thirdpersonplayer 1, and tdm_lg_player 1...and compare the doom marine to the Thief in size. That would easily tell us where the camera is in relation to the size of the Thief because the tip of the lightgem model rests in the doom marines head.

[sparhawk]

IMO the model of our Thief should go ingame as soon as possible. I was discussing this with Oddity, since we need it for a lot of tweaks on the playerstuff. Maybe some of the animations can wait, at least the details, but we should have basic walkcycle, leaning in all directions, jumping, mantling and crouching.

[Springheel]

Those basic animations already exist--well, maybe not jumping or crouching, but for sure there is walking and leaning. It might be worth posting about that weapon issue in the modeling forum, however.

 

We could just pop the lwo of the Thief model into the editor...go into game mode and do a pm_thirdpersonplayer 1, and tdm_lg_player 1...and compare the doom marine to the Thief in size.

 

Ok, I'll try that tonight, unless someone beats me to it.

[New Horizon]

Those basic animations already exist--well, maybe not jumping or crouching, but for sure there is walking and leaning. It might be worth posting about that weapon issue in the modeling forum, however.

Ok, I'll try that tonight, unless someone beats me to it.

 

Yeah, I've been trying to do it...but I'm having some issues. I'll keep working at it though. Have to learn how to do this stuff.

[Springheel]

Yeah, this is clearly beyond me...I can't make heads or tails of the farrel.lwo files. Someone else will have to take care of that.

[Springheel]

update: Ok, farrell1.lwo, farrell2.lwo, and farrell3.lwo all look like they are only part of the whole model, as far as I can tell. But I imported farrellposed.lwo into the editor and it was the entire model. Strangely, the arms, weapons and legs were already textured--I guess they're referencing the texture files for the onscreen hands and weapons?

 

Anyway, I was going to compare their heights in the editor, but I can't find where the D3 player model is kept--any ideas?

[New Horizon]

update: Ok, farrell1.lwo, farrell2.lwo, and farrell3.lwo all look like they are only part of the whole model, as far as I can tell. But I imported farrellposed.lwo into the editor and it was the entire model. Strangely, the arms, weapons and legs were already textured--I guess they're referencing the texture files for the onscreen hands and weapons?

 

Anyway, I was going to compare their heights in the editor, but I can't find where the D3 player model is kept--any ideas?

 

Looking through the editor, I don't see any references to the standard player model...but I do see a few variations with _player on them. Go through the delta listings in the menu, I think there is a player model there you could pop into the editor to compare.

[oDDity]

I'm pretty sure Farrell is already rigged in a maya files. There are MB anims for hims as well, they just have to be extracted to md5anims.

[New Horizon]

I think I have pretty much figured out how to do it. Could someone convert them to md5's for me? I'll get what we have so far ingame.

[Ishtvan]

Re: Weapons

Those will all be done with AF attachments, so the final animations probably should have no weapons attached. And yes, we can attach more than one AF AFAIK, so we can put the bow on his back when he's not using it and put it in the hands when he's firing. It might be good to do some test animation runs with the weapons in tho so you can make sure there won't be clipping issues.

 

Re: Relative AI size

I agree that getting the player model ingame is important, but I think for the AI size issue, the standard to go by is the camera height.

 

The player model height is a bit arbitrary and might have to be resized to fit the camera height, but I think the camera is the standard. If we change the camera height, then all units effectively change and the physics will no longer be correct because all forces will still use the old units. It doesn't make sense to fit one arbitrary value to another one, when we already have a camera height standard, and D3 character heights to go by.

 

As for the whole size issue, I was playing thru D3 again and noticed that when you look straight ahead, most of the characters eyes are above your eye level by a good foot or so. It didn't strike me as unnatural though the first time playing thru D3, they still psychologically seemed to be the same height. Maybe it's something about how we usually frame people in FPS when talking to them? It seems I tend to frame people so they take up about 3/4ths of the screen, with their eye level slightly above mine. Maybe when we talk to people in FPS our human eyes naturally look a little above the center of the monitor?

 

I tried noclipping up to where my eye level when looking parallel to the ground was the same as their eye level, and they looked really short because they only filled the lower half of the screen. Maybe there is a reason why Id put the camera in the player neck/chest and made the AI models the same height as the player?

 

I'm not saying everything's perfect as-is, some of the AI still feel way too tall, but the builder guard seems about right, maybe just a tad too tall.

[New Horizon]

Maybe there is a reason why Id put the camera in the player neck/chest and made the AI models the same height as the player?

 

Saywha' now? The player camera only drops into the chest cavity when the model is crouching, otherwise it's right inside the head where it should be. You can see this by going into thirdperson mode and using tdm_lg_player 1 since the tip of the lightgem model sits at the camera level.

[Springheel]

Ok, so we should assume that the camera height is the player's eye level? What was that again, 74 units or something? I forget where it was posted.

[Ishtvan]

That's what I would argue, yeah.

 

@NH: I thought the camera was in the neck/chest the whole time? If you put in a mirror surface like in soundtest1 for example, and run up to it and look straight ahead, you'll be looking at the player's chest/neck. I'm sure a mapper will come by and correct us at some point.

[Springheel]

Ok, NH posted earlier that: The default camera height is 68, player height is 74 (doom units).

 

So what height would that make the player in real life?

[New Horizon]

Ok, NH posted earlier that: The default camera height is 68, player height is 74 (doom units).

 

So what height would that make the player in real life?

 

Since we're assuming a 1:1 ratio instead of dooms default 1:1.1, then 74 doom units equal 74 inches. Roughly 6.1 feet tall. I think that our Thief model is 82 units high...although I can't confirm that since I only looked at it in Lightwave, but I think it's the same grid setup as Doom 3. We'll need to resize accordingly once we get him ingame. I wouldn't have the Thief be anymore than 5'6" to 5'8"...especially if our Guards are supposed to be fairly tall.

[New Horizon]

That's what I would argue, yeah.

 

@NH: I thought the camera was in the neck/chest the whole time? If you put in a mirror surface like in soundtest1 for example, and run up to it and look straight ahead, you'll be looking at the player's chest/neck. I'm sure a mapper will come by and correct us at some point.

 

Just pop into a map and take a look at the 3rd person model to get a glimpse of what I mean. Make the gem visible and you'll see that the very top of the light gem model lines up with the eyeheight.

[Springheel]

Since we're assuming a 1:1 ratio instead of dooms default 1:1.1

 

I thought we decided not to do that, because it would mess up the physics calculations?

[SneaksieDave]

Wasn't it decided that we'd better not go with 1:1? Physics (and Ishtvan) spoiled the party.

 

Edit: erm yeah what he said

[Springheel]

So if we're using the default 1:1.1, does that make the thief shorter or taller? I keep getting the ratio confused.