(WIP) - Ninja Arcade

I finally recorded a bit of new video for my game. It’s been a while but I’ve been working on the weapons pack and taking a break from Ninja Arcade for just a bit, but I have done plenty of things on there over the last few months. A lot of what I do is non visual and either tweaking or experimenting with things and not things I intend to show as is. But I’ve also worked a bit on the next step in adding game play elements.

So to start with I wanted to make the fighting work pretty good, which I think is basically there. It needs more final polish on certain things but I think the majority of the work is out of the way. Of course I do a lot of graphical work as I move along. I tend to jump around the project and work on what I feel like, jumping between graphics and blueprints, so my time doesn’t go directly into one area.

So anyway, check out the video, I’ve added a few game play elements recently and since I haven’t shown anything in a while I thought I’d record a new video to show that plus one or two other small things.

Another video update. I’ve got the sword power up and health spawning systems now pretty much fully set up. The effects are looking much better as well. That place holder stuff was just bugging me… :slight_smile:

So this game is made only with blueprints? Looks awesome, can’t wait to see more! :smiley:
Kinda reminds me of Spartan:Total Warrior, from the guys that are making the TotalWar series, it was a great game :smiley:

Looking good as always Obihb, keep it up!

Thank you…

Yes, I’m making the game only with blueprints… blueprints are awesome… :slight_smile:

Hey (: Looks really nice. I watched the “Tutorial” part with the combos and will probably use this setup.

May i ask you how your Animations are setup? Is Attack2 directly starting from the Endposition of Attack1? Because the Couch Knights
Project uses a short Animation after each attack and while this is active, you are able to launch the next combo attack. You do this with the
Gate and the Delay or? (:

Thanks,

Yeah, my attacks don’t work like that, I haven’t seen the Couch Knights stuff but from the sounds of it, that’s not what I do. I rely a bit more on the engine blending animations than making perfectly “tillable” animations, meaning my animations do not have a clean start and end based on what came before or after. So the animation blending takes care of that, to make them seamless.

So, I do call my animations using montage, but I don’t actually use montage to make loops of any kind. You can call any section at any time and that’s pretty much what I do. They don’t specifically string together, they just blend based on the blend timing I set in montage.

The way I animate the moves is with consideration for a combo window and a blend period. So at the end of each animation after the follow through, is a “hold” to allow for this. The start of each animation is actually a short hold already in the extreme position of the back swing to allow a blend in. So I don’t animate the back swing, that is taken care of by the blend in. This way I have way less animations to deal with, no transitions, just major moves, and don’t have to care too much about starts and ends being perfectly aligned. It take a load off animation work.

The way I control combo timing is through a delay, so as soon as I press attack, the delay starts and will reset all booleans associated with attack combo if it gets to finish the delay, unless I press attack again, in which case the delay restarts and never calls the reset. It’s a Retriggerable Delay, so as long as you keep pressing the attack, it can not reset and the combo is valid. The combo flow is controlled directly on the moves, each move in turn enables the next.

What the Gate actually does in the context you refer to, is control how quickly the next attack can be activated. Since you don’t want immediate input to activate the next attack, if you tap the attack key rapidly, you don’t want overlapping moves, you want that attack to play out and then you are allowed to do the next attack. Though I’ll mention that I’ve also changed that setup to go through a Branch in stead of a Gate, it still does exactly the same thing. So that’s also controlled through a delay, but this is not a retriggerable, it’s a straight delay. The input is actually disabled for a short delay to allow the move to complete, meanwhile the combo delay is still active and leaving a window open for a combo to be put into motion.

I hope that makes sense, I know it can sound convoluted trying to describe it… :slight_smile:

Yes, that’s a pretty good description. So you only have 2 different kinds of length for the delay in front of the gate (or branch now). This means
your Animations are pretty much of the same length? And the short TimeWindow for the player to really make a combo is the

0.5 Retriggerable Delay in the Sheathe Sword part?

If yes, i fully got it and i really love this setup more than the Couch Knight one, because
it needs so many unnecessary animations :smiley: Thanks so far.

Hi

I did download the demo and I really liked it. It’s looking and feeling good, and that’s what an arcade game should do :slight_smile:

Just one small suggestion would be to tweak the FOV a little bit. The main character looks good etc, but with this level of zoom you see very, very little of the level around you. It’s always some kind of trade-off, but I think that the main character could be little bit smaller in favour of a wider field of view. Maybe simply going with a user controlled FOV is the best option for the future.

@eXi - Yeah, that’s pretty much it, my animations do match in length for the major part of the combo sequence, where it matters, and then the gate delay / attack delay is 0.2 I think (if I remember correctly, not looked at that in a while) and in the sheathe sequence, the combo reset is 0.5 which makes a valid combo a 0.3 second window.

@rimau - Thanks for trying it and taking the time for some feedback, I really appreciate it.

Though the camera setup is probably not 100% set in stone, I feel that this starts getting into personal preference and I can’t really cater towards that. I basically have to do what feels right for me. If it were a matter of badly influencing game play, it is another story. But I think it works pretty well. Small changes one way or another isn’t going to do much and large changes, well, that starts affecting things in a bad way.

Giving some amount of player control, that is possible to some extent, but I’m not sure I would do that. I have considered it in the past but ultimately, with the implications it has, I don’t want to do that. I can certainly look at giving a little more open view, but it probably won’t be too much different from the current view. The height of the camera also affects it quite a lot and I’ve changed all this a few times over and I feel like I have a decent sweet spot on there. But as I said, it’s not set in stone, there’s room there to play, just a little bit at least.

Sure, this is very, very personal, and… it depends of the gameplay, and as I do not know about the stuff that’s going to come, and how the overall experience is going to shape, all I could share is some of my (again) personal feelings. If the gameplay is going to be totally focused on really close combat, than I guess it’s fine (personal). And if there will be an option of a more tactical approach to the combat (like using the surroundings to your advantage, etc) and if moving around the level itself is going to be a bit more important, than I would personally opt for a bit wider FOV. Or a dynamic one (an idea that has literally popped into my mind at this very moment) a bit wider on the move, and a bit closer during the combat, when all the eye-candy animations are used.

Anyway, this will be one of the last tweaks you will probably do, so no real reason to think to much about it now :slight_smile:

Sure, totally understand your concerns. And to be honest I do think about all this stuff quite a bit. Nothing I do is just off hand. I take lots of things into consideration and make choices I feel work best. Even small things, which never turn out to be small things… :slight_smile:

I have plans for a dynamic system and I’ve got parts of it already set up so where you go in certain areas, the camera will move into position to accommodate those areas, stuff like that. The Boss video I posted is also a good example where the camera moves up and back to open up the view and also readjusts to float on a 50/50 weight between the player and the Boss, keeping the Boss in view, even if partial at times, depending on where you go. That is a very special case but that’s how I intend to deal with the camera, basically just automatically placing it in the right places and you never have to worry about the camera. The other part of this setup is not yet in where general combat plays a role in camera position / zoom. I tend to like moving the camera in stead of using FOV. But this part of it requires more focus / care to make sure it works well or to make sure it doesn’t, and use it or not. I haven’t gotten to it yet, it’s on the long to do list… :slight_smile:

The default camera, like you see in the demo, is essentially the wide view. and yeah, it is sort of close but I felt that it’s wide enough for a default view, though that can certainly be a bit wider, and I’ll know more for sure when I have my proper game play set up. It’s a pretty good bet though that current view is about where I’ll keep it, even if I make it a small percentage wider in the end, it’s pretty close now I think. The level itself will not play such a huge role in the final game as far as game play is concerned. It’s mostly about fighting game play and the level is mostly scenery.

Looks awesome. How do you get such a clean blend from ragdoll to animation? This is the only video I could find that demonstrates it. I’ve tried nearly every combination I can think of (moving the mesh/moving the capsule/fixing the root bone/letting it simulate). Whenever I start blending back, there is always an offset between the ragdoll and the mesh origin, so it always slides back as it blends. I’d appreciate if you could give me a quick overview of your ragdoll/blending setup, so I can see what I’m missing.

I think the important thing is the version of the engine you’re using. The ragdoll setup I created, that actually worked, was done with 4.4 and since 4.5 came out, my setup was quite broken.

So, as you mention there, the offset and stuff, yeah I get some screwd up results as well after 4.5. I looked at it quite a bit during the 4.5, before 4.6, and I could not make it work again. With 4.6 and now 4.7 I have not looked at it again. I’ve currently got it just disabled. I really should look at it again. I think I might just wait until the final 4.7 is out and then take a look see.

With that said though, the way I had it working before (if I can recall now). When the model goes to ragdoll, it basically detaches from the parent, the capsule. So at some point you have to attach it again. I found that doing this will kinda twitch the ragdoll. So what I did was to immediately attach it after going to ragdoll. So even if it caused a twitch, you’ll never see it because the thing is busy falling.

After that, the next step is to move the capsule to the correct position. Again I found that snapping it to position, twitches the ragdoll. Because now it is attached to the capsule again, it in fact will affect it, even though “technically” it shouldn’t. So I interpolate the capsule to the ragdoll. I get a position and rotation based on the pelvis bone on the ragdoll skeleton, then do the interpolation using a timeline. I think it was maybe 0.5 sec or something like that. So because of the smooth interpolation, the ragdoll does not twitch at all.

So, now it’s good to go. The root is essentially lined up because the animation has the feet come to the pelvis and then stand up. Any offset is smoothed out by the blend over.

I think that is the basics of the old system. I tried various combinations of this with 4.5, but could not make it work. The best I got, if I remember correctly, is the stand up animation would first move above the ground as it blends in and then come back down. So it “almost” worked, but not quite… :slight_smile: In most cases it would jsut shoot off to some direction or do some completely weird thing. I think it was the root bone that caused it, it should stay with the capsule but did something else. I can’t remember now. I think it moved to the pelves or something, it was really weird. It cause the animation blend to go all wacky.

Don’t know if this helps you any, but when I take another look back at that stuff after the 4.7 final, I’ll post back if I find a working method. Who knows, maybe it’s easy to do it, I just haven’t actually tried again since ver 4.5.

Thanks for that, the description was quite helpful, so at least I know I was on the right track. I messed around a little more, seeing if maybe I could detach and reattached the mesh (in case it was the old behaviour), but that wasn’t it. It just seems to be an issue with the way fixed/unfixed bodies are handled by physics: you can’t move the fixed root without it moving all the other bodies too. I posted on answerhub, and made a simple demonstration video, so hopefully one of the devs will see it and can explain the system, and how to get around it. I’ll let you know if there’s any answer Issue blending from physics into animation - World Creation - Unreal Engine Forums

I’ve done a bit of an update on my combo system. The problem I had with the old system, which became more and more apparent, is just that, even though it was nice and fluid within itself, within the actual combo, it felt limited in the fact that it was a five move combo. So after five hits, it breaks out and you start over. So that didn’t feel so fluid when you were fighting loads of enemies. And it especially felt “wrong” when trying to break the gate.

I can’t really set up a full fighting combo system with many types of combos. I just don’t have the time to do that. It’ll take forever. So I changed my current system to be more “manual”. So it’s not a straight combo, the same length all the time, you can control it.

It ultimately feels much better, cutting through crowds of enemies… etc.

So, here’s a new video just to show that in action.

Nice game!

I’m also trying to make a game inspired by Ninja Gaiden and Original Devil May Cry series, using C++.

Do you know if there are many other developers here that could be interested in discussing how to make hack 'n slash games like NG and DMC?

The ones I have noticed are you, President & RhythmScript…

If you got any tips, then feel free to drop by:
https://forums.unrealengine.com/showthread.php?62452-Advice-on-how-to-make-a-freely-rotating-amp-jumping-character

Thanks!

So I finally made a new video showing some stuff I’ve been working on. I haven’t made a video in a while and that’s more because of the type of stuff I worked on is not really possible to show, it’s more underneath, reworking and tweaking mechanics. I’ve gone over my fight mechanics about a million times just trying to get rid of any possible bug or anything that felt weird. The basics have always worked fairly well but diving into it and really trying to break it, I found many little things that wasn’t working so well. On top of that just going over things to get them more optimized and just cleaner and of course working on new content… etc. It’s a crazy project for me… :slight_smile:

Since this is my first lone wolf game project, doing any type of “scripting” whatsoever (blueprints of course), I learn a lot of lessons on the way that kinda force me to go back and re-evaluate things. So, this is both fun and frustrating at the same time… but mostly, it’s just fun being able to do this. This isn’t an ad for UE4… but UE4 rules… right… :slight_smile:

So anyway, I hope the video is bare-able to watch. It shouldn’t be any worse than previous ones with regards to me rambling on about something, I don’t think. I kinda record these on a whim and don’t edit so… yeah… cool… :slight_smile:

Thanks… :slight_smile:

As far as engine updates goes, I’ll say in general it’s not bad for me. I’ve had my bad moments, especially with the 4.7 preview builds, which of course is my own fault for jumping over too soon. But now, having jumped over to 4.8 Preview 4, it’s about as smooth as a transition can be. Some small things changed because of how things work in certain cases that I had to go back and update. Not really all that different from older versions, having gone through every version. They’ve all had their things I had to deal with but nothing major I can remember. But that is the price to pay for using an engine under development. With that said, it’s not that bad, not for me anyway. The good far outweigh the bad. I’ll say in the worst cases, it was actually my own fault having the trouble I had.

With the dynamic lighting, well that actually has a very small difference in the way that I use it. It is dynamic but it is not animated. The biggest impact dynamic lighting has in my game is on the interior scenes. With exterior you basically have a sun and the sky. With a static lighting setup, the only difference is the sky, because the sun is basically still a dynamic light, but it does cast static shadows on static meshes, so that is cheaper. So I also have smaller light on there but they have very little overlap and no shadow casting, and really shadows is where you can probably lose more fps. So the overall exterior impact on my game, with dynamic lighting, is minimal.

With interiors, it’s a different story. There are more lights used and there are more shadow casters. So I use some old “Doom3” lighting theory, keep it simple and keep the overlaps to a minimum. In some key places I have blueprint lights that will basically scale out based on distance so in the back, they cost nothing. For the “deeper” scenes, that gives me back some fps, especially more noticeable in higher resolutions. So speed wise, I just have to keep a close eye on it, compared to static lighting. So with static lighting you just need to deal with using stationary or static. Though I always just used stationary lights, which is also why my transition to dynamic lighting went as easily as it did. Stationary lights are basically dynamic lights with static shadows on static mesh and can contribute to bounce light, but you can only have limited overlaps, I think 3 or 4, can’t remember now exactly. So my interior lighting was close to working with dynamic setup even as a static setup. All I really had to do is deal with the fact I have no bounces, which I do with sky light, and some optimizing of overlap sizes in my major lights.

Comparing to UDK, just in general, without dealing with animated lights… etc… it’s way better. In UDK I was never able to successfully go with fully dynamic. It never worked for me, no matter what tricks I tried. I’m a bit **** with fps, and I couldn’t get satisfactory results. It just simply cost too much. I can’t really speak towards matinee animated lights because I have not used it. I have lights that technically move, like the health pickups, the sword power up, that kind of thing. They don’t seem to do much in terms of impacting fps but they are fairly small and not many of them. The lights I have that moves the most is the lava rocks from the Boss attacks. The big rocks emit light and when they break, the smaller ones also emit light and you can kick them around. To be honest, they did not impact fps much from what I could determine, but they are pretty small lights and as long as they are small you can get away with a lot there. The bigger the light, the more easily you can run into fps problems with regards to having more than one light and of course shadow casting. But matinee animated driven lights, I’m not sure. I think anyway if I had to animate a light, like time of day… or something like that, it would just be a blueprint with a timeline to drive it, not with matinee. I can’t say if that will have the same implications to fps impact or not.

Anyway… I’m typing a lot so I guess I’ll leave it there… :slight_smile:

Your project’s looking great boet! :wink: