Jump to content
Mizzu

Dev's stance on enemy model, animations, hitboxes, ragdolls

Recommended Posts

So basically, does anybody know what the devs think about these topics?

1. Is the current enemy model final? If no, will we eventually get something that has a more nuanced outline than "a giant rectangle"?

2. Similarly to that, will there be animations that derive further from Keel, who was pretty much just conveniently standing around in his cilindrical hitbox, the upper body staying completely rigid? (I'm thinking some max-payne-esque jumping animations, or something like this, from a stance viewpoint)

3. How does the current hitbox look and is it final? Does it coincide strongly with the model's geometry (cs:go style) or is it just a big cylinder (q3-style)?

4. Ragdolls for non-gibbed enemies pls?

I think these are hard questions to answer since something like an animation and model stance/movement is quite important from both an aesthetic and gameplay perspective. I doubt people will ever say that the game really looks like it's from 2015 when the enemy model is a slightly sculpted and decorated cylinder that looks the same when it's walking, jumping, being midair, moving on stairs, double-jumping, firing, etc (look up some CS/CoD/Battlefield gameplay footage and imagine all the complex animations gone if you don't think this is true). But if it was given complex animations, the hitboxes would have to be adjusted as well (staying vertical all the time would not be an option for obvious reasons), which would mean that suddenly the art has a profound impact on gameplay beyond how easy it is to see something. Certain moves would trigger certain animations which in turn would have effects on the hitboxes, which in turn would have an influence on how hard or easy it is to be hit.

Edited by Mizzu

Share this post


Link to post
Share on other sites

I think it would be cool to have a thread about devs opinion on various popular topics such as these. Stuff that they have planned to come so people dont freak out. For example ive seen some ppl going crazy on current current models and some of them think they are final, others think that are placeholders for the new art direction but nothing official. This is an example

Share this post


Link to post
Share on other sites

So basically, does anybody know what the devs think about these topics?

1. Is the current enemy model final? If no, will we eventually get something that has a more nuanced outline than "a giant rectangle"?

2. Similarly to that, will there be animations that derive further from Keel, who was pretty much just conveniently standing around in his cilindrical hitbox, the upper body staying completely rigid? (I'm thinking some max-payne-esque jumping animations, or something like this, from a stance viewpoint)

3. How does the current hitbox look and is it final? Does it coincide strongly with the model's geometry (cs:go style) or is it just a big cylinder (q3-style)?

4. Ragdolls for non-gibbed enemies pls?

I think these are hard questions to answer since something like an animation and model stance/movement is quite important from both an aesthetic and gameplay perspective. I doubt people will ever say that the game really looks like it's from 2015 when the enemy model is a slightly sculpted and decorated cylinder that looks the same when it's walking, jumping, being midair, moving on stairs, double-jumping, firing, etc (look up some CS/CoD/Battlefield gameplay footage and imagine all the complex animations gone if you don't think this is true). But if it was given complex animations, the hitboxes would have to be adjusted as well (staying vertical all the time would not be an option for obvious reasons), which would mean that suddenly the art has a profound impact on gameplay beyond how easy it is to see something. Certain moves would trigger certain animations which in turn would have effects on the hitboxes, which in turn would have an influence on how hard or easy it is to be hit.

​1. The current enemy model isn't final, but I don't expect the silhouette of the model to change significantly. This week though, we did implement special "hero lighting" for the players that helps bring out recognizable internal shapes in the character, regardless of the scene they're standing in*. After a certain point, these internal shapes are much more useful for reading a character than just the silhouette.

2. The animations are definitely going to be redone, but we don't have an ETA on them yet. There are technical reasons why the Keel model in Quake 3 is "conveniently standing around in his cilindrical hitbox" and gameplay reasons why 99% percent of players forced their playermodels to Keel, instead of something that backflips everywhere. 

Hitboxes in Quake 3 are axis-aligned bounding boxes, meaning they don't rotate on any axis, regardless of what the player does. AABBs are used it most games when possible because it's extremely fast to calculate collision and hits against them. QuakeLive moved to a cylinder hitbox for hit detection but I'm confident it always remains flat on the ground for similar reasons. 

So, most players used Keel because he was one of the few models that actually sat nicely within that AABB and didn't leave it. Quake 3 had some characters with far more interesting animations, but they were generally avoided because of how inaccurate they were. For example, the Doom model occasionally does a backflip. When he's upside down at the apex of his jump, his entire legs are outside the bounds of the hitbox (so if you shoot them, nothing will happen). But then at the bottom, there's half a player worth of "empty" hitbox (so if you shoot the empty space below them, you'll get a hit).

While we want to make the animations in Reflex look good, they need to still consider these issues. We have a very small threshold for what we consider acceptable when it comes to player animations straying outside the hitbox, so we need to work with that.

3. The current hitbox for you guys is a standard AABB, same as Quake3. As you can see, there's issues with it. It's easy to have the player rotated in a way where he doesn't fill out the hitbox nicely and (since the hitbox is axis-aligned), there's certain angles that give a greater surface area to aim at than others. This was annoyingly inconsistent in Quake 3, so Quake Live eventually replaced them with cylinders to minimize those issues.

The current hitbox we've created for the next update is a capsule shape.* This significantly reduces all of the issues above, to an even greater extent than cylinder hitboxes do. The hostile was actually designed around this hitbox shape and if you look at one of the early concepts created literally years ago, you can see the capsule design in the background.

So yes, as of the next update, the hitbox will "coincide strongly with the model's geometry" -- although not CS:GO style. Despite seeming good on paper, capsule-per-bones systems like that have a ton of technical issues to go with them too. The most obvious (because there's 36,000 threads about it scattered around the internet) is that once you're sending everything over the internet, it ends up feeling less accurate as there's far more to get out of sync / incorrectly extrapolated between client and server. One potential solution to that would be client side hit detection, but that generally doesn't feel nice to be on the receiving end of (and is a cheaters wet dream). Doom 3 actually had per-pixel hit detection, but it was offline only for exactly these reasons.

And really, a system like that doesn't even make sense in a game like Reflex. Given the movement speeds, shots to the lower half of a player would basically be a coin toss, even if we could magically have the hitbox immune to network issues. I'm pretty confident that if we implemented it, neither us developers or our players would be happy with the way it felt.

4. Ragdolls for non-gibbed enemies has been planned for a very, very long time. We were going to put it in before Kickstarter but we ran out of time and ever since then, it's been pushed back in favour of more important tech and art work. At the very latest, it will come when we update the characters and animations. It will also magically fix a bug with floating corpses (which is the reason why we don't use the death animations we've got and instead always gib players ;)).

--

So yes, there are ways in which Reflex will never look as polished as something like the next CoD or BF title. But it's import to remember that there are solid reasons for this. The slow transition from AFPS to modern military shooters can be easily understood through these art concerns being a higher priority than gameplay and that's not the way we work -- gameplay comes first, even if it looks a bit stupid. Bunny hopping will always look a bit stupid, so you're not likely to see a modern AAA title that has bunny hopping. You can't have accurate crosshairs unless you fire from the middle of the view, so you get aiming-down-sights. It looks wrong if you instantly transition to aiming down sights, so you get the delay as the view lerps. It looks wrong if you instantly trigger animations, so you get less responsive input. It looks wrong if you switch weapons in 100ms, so you get extremely long switch delays.

There are hundreds of examples of this, and the core difference between Reflex and other games is that we will always put gameplay first and do the best we can to make it not look weird not creating gameplay around what looks best.

* All of this is a work in progress. Please don't send me giant angry rants about things that are a work in progress. Most of the time, the issues you've sent me a small thesis on are already well understood and simply not implemented. All the premature rage does is make me less inclined to show anything we're working on.

Edited by newborn

Share this post


Link to post
Share on other sites

THANK YOU​

 

Bunny hopping will always look a bit stupid

Bunny hopping / strafe jumping is a very elegant way to represent Running, it's just the case that most games didn't anticipate it and failed to animate for it.

It doesn't have to look stupid at all as long as you've planned your game around it.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×