You NEED this if you want to make a multiplayer game in unreal
2024 ж. 14 Мам.
77 705 Рет қаралды
Wishlist our game Spanky! store.steampowered.com/app/17...
GMC: www.unrealengine.com/marketpl...
Smooth Sync: www.unrealengine.com/marketpl...
In this video we address the topic of multiplayer games made with unreal engine. Blueprints are an amazing tool for developing games especially for those who prefer the visual layout over written code. There are many amazing courses and tutorials to begin learning blueprints for free, however many of those that are made for multiplayer functions aren't developed with real world latency conditions in mind.
00:00 Introduction
01:02 the trap of unreal engine
01:59 custom movement
05:53 Solution 1 - C++
06:20 Solution 2 - Smooth Sync
07:11 Solution 3 - GMC
09:06 Conclusion
i did not see that you only have 500 subs and 200 views until the end of the video. So good quality!! couldnt see the difference to a big youtuber. very good
Thank you for doign this video! I hope everyone learning Unreal stubble across this. For I was in the same situation that you were in and had to go through all of the headache that you did.
I think Tim and Epic need to buy GMC and provide it for free to the users. would be amazing :D
No, they would rather hire 20000 more workers to do fortnite skins.
lol i did not expect GMC to be the first comment. Shout-out to GRIM, doing gods work with custom movement frameworks
What does that exactly do? I read their store page, and it lists a bunch of stuff that is built into unreal engine as "Features", I get that its supposed to be a more general movement component, but why does it come with a "steam integration with UI backend"? like, even the confusion is confused here on what this thing is supposed to be.
@@MrSofazocker I've never even seen the steam UI feature of GMC? It's god tier for extending the character movement class though. Try and make a character with directional gravity and you'll see why this plugin is lauded so much
@@sundaysquire4013 It's at the bottom of the featurelist, they call it "Steam UI backend", how a UI can be backend idk, if he means they made bindings you can call from UI, then that's pretty weird, as the Steam integration already gives you those!? But you can extent anything from the character movement component using your own components, or even make child class of it and expose the things you want? Like to donwards vector? I still don't get it, sry. Like I do not doubt it will speed up the setup of characters, animations etc, it seems to support a wide-range of them. But depending on how you want to trigger those and how they affect gameplay, you will end-up making your own system either way. Nothing can fit your game perfectly, or elese you'll end u making your game conform to the tools you used. But the notion that this is somehow not at all possible using plain Unreal Engine is myth.
loved your video, noticed you don't have a lot of subs but keep going, you already have that professional editing style, now you just need some time to grow it out! best of luck to you
Not developing in unreal yet, but THANK YOU, this video was accurate from 0:00 to the end. Tired of people saying multiplayer is not accessible.
im aspiring to build my dream game and im so grateful for this video! subscribed ✅
You are awesome dude , continue working .I'm gonna be your constant viewer)
From personal experience, I can tell you that Unreal's movement component is very, very rigidly coded in certain ways. This is just as big of an issue in C++ as blueprints. As an example, I was doing a project with spherical worlds. Unreal's movement component however is hard-coded with gravity going in a specific direction, up being a specific axis, etc. So custom rotation? Yeah... no. You basically have to rewrite entire sections of the component to get it to do things like that. And that movement component is freaking massive. Latency may be a bigger issue in BP, but the movement component is thousands of lines of rigidly structured code.
I'd say its not that hard to fix the gravity yourself in a child class, but they seem to have changed it so the get gravity function is no longer public. Hooray for maintaining a custom engine branch.
That pretty much a non-problem. You also dont have to recode everything, the movement component isnt a component by mistake. Components enable you to compose (duh) functionality. Instead of being stuck with the functionality, you can simply extend the character component with your own stuff, by utilizing all the replications already setup from the character movement component, as it is server authored, you can let the server handle everything, The players only get their updated position from the server, they dont move themselves. This is also to prevent cheating. For things like speed hacks, etc. Imagine you could tell the server what ever you want, and the server updates your position for all other players, simply bcs you said so... That's not how you do multiplayer games. (Tell that Bethesda... yeah Fallout76 movement speed was tied to your fps, and was not server authored, so you could sprint at mach2, not fun)
@@Gorgonzeyeyou could just copy-paste the source code of UCharacterMovementComponent into a new class and modify the parts you need.
@@jacobsoares7068 That method would cascade into having to replace ACharacter, and then recreate everything that normally references AQCharacter.
kinda sounds like a game I like called gods of gravity. I always wondered about their gravity system.
Really useful information. Good stuff! I was worried there for a bit, but then realized that it doesn't concern me since I have in mind making a shooter with very grounded movement speed with the only multiplayer being co-op.
This is awesome man, great video. I ran into this exact issue in my first game that I wanted to be co-op. I'll be implementing GMC for my game. Yes I know its not required since its just a simple co-op game but I may expand on the game and open it up to a larger amount of total players.
Thank you so much! I literally was working on my online game i don't even know this could be a problem
Thank you, great video and this will really help me, the timing of this on my game is perfect.
I came here to learn your solutions instead found myself wanting to play spanky 😂 brilliant love it subbed 😁✌️ I'll give ur game ago when I get chance 😁✌️
The fact that Smii7y played your game should make you feel amazing
what? only ~200 views on such a quality video? hope you get more popular
I just want to make good videos that one day I can look back and see the progress I've made 😄 thankful for anyone who sticks around for the journey!
If not already said in the comments: You might want to start using multiple event graphs in a single blueprint. Having all that logic in a single graph slows down the blueprint editor.
that gmc is realy expensive but i can see why its good and the tutorials for it are rather simple too after a short glance, this to me is a 5 star it is rare to get decent documentation and even more rare for them to do a vocal tutorial and even more rare for it to be as clear as they have done if you have the money go for it well worth it money spent on that free up your time to focus on something else
Legend man, thankyou for this video. About to dive into some multiplayer stuff so i am glad I watched this! All the best with your game too, it looks like a lot of fun haha 🍑🤚
Badass dude congrats on your publisher.
I'm currently facing this same issue in my game. I've gotten as far as I have by coding it all in the character, but even in "low" ping cases the player will rubber band. Especially if they spam the sprint key, like you showed. I've had my eyes on the GMC for ages, and it's good to hear from another dev that it's worth it. Once I can scrap together enough money to get it, I will definitely be buying it, as I think it'll truly make my game WAYYYY smoother for clients. Though I am still worried if it can replicate some stuff that's specific to my game. Either way, I appreciate the insight! Spanky looks really interesting, you've got my sub, and as a fellow multiplayer dev, I can't wait to see further progress!!
I really appreciate it! Your game looks like an incredible amount of work, I'll definitely be keeping up! The GMC has definitely saved our project as more then half our movement is custom (and was broken with latency) but it does have it's learning curve. Although I am still learning about the GMC each day I'm always happy to chat if you wish through discord- Rory#5727
@@BoryDev hey I know this wasn’t intended for me but I also just added you on discord, much love thank you so much for the amazing video, I am extremely early in the process of working on my first (kinda I have some past experience but not in unreal and not for a long time) project and it is intended to be multiplayer and I am so grateful to have this brought to my attention this early. Much love 💚 keep making incredible content
@@BoryDev Talkin about GMC and Smoothsync. Non Competitive Multiplayer game comparable to FiveM Lobbies : Smoothsync or GMC ? Which is easier to setup in the long run? I may have to redo everything, but i just want to start with the basic movement there is in UE5 and then redo it later on when i got enough experience.
@@DarkSession6208 If you aren't worried about cheaters *at all* then smoothsync will work. But I will say, using smoothsync is basically heaven for cheaters, it is pretty much entirely syncing data from your PC, which means a cheater can manipulate data as they please. You can also use smoothsync in conjunction with the regular movement component as it is not really adding movement, but syncing location & rotation from your local PC to the server (and everyone else) so its useful for physics objects too if you need. I wouldn't just throw it on everything though, performance is still necessary. GMC is great if you are ok with rebuilding all movement and finding solutions to movement problems yourself (most people will help in discord but its still all you) and it is also much more authoritative in the sense that the server will correct you if there are location or data differences. So it will protect you more than smoothsync. If you are still early stages of figuring it all out, I would recommend staying with the default movement and smoothsync if you run into issues. When you have more experience or know *exactly* what you need to make, you will understand if you need the GMC or not - until then smoothsync would be ok.
@@BoryDev i know your game was very specific, but i thought some months about this whole problem and still did not come to a conclusion. Do you think it would be a good approach to use a C++ Approach for Sprinting and slap Smoothsync on everything else ? (Vaulting for example) ? My goal would be to rule out velocity changes and teleport, so speedhack and tp. Im just not sure what Cheaters can use to hook on, i mean from what i know the walking state, jumping state is well replicated and not the problem, its just the custom movement different from the CMC that is causing hackers to hook onto this right? So if i would use C++ for those 2-3 custom states i have and keep everything else default or use smoothsync for non critical stuff? Im asking because 1) i looked into GMC and hell people say it took them as long as if they would learn C++ and with V2 they raised the price to 700€ .
thanks for this video, i was about to switch to unity watching this video until you got to smooth sync and general movement component. I'll stay in Unreal
0:40 o o ... o k :( (Nice video and good quallity keep going)
Fantastic. Thank you
I hate that the only fixes for any of these problems is always "Oh just go buy a $300+ plugin" Good video i just hate that this is the only fix I keep seeing for replicated movement (For me im making a marble game and the clients are very stuttery when moving)
I hate it too! I totally understand where you are coming from. That's why I tried to suggest Smooth Sync plugin depending on your type of game (seriously if big game with online cheaters may not be an issue for you, then smooth sync is a decent option). In the original recording of this video, I had a section towards the end saying that I really don't like that the only solution I can provide to people is to spend more money, but at the end of the day that's what we need to accept when it comes to these topics I think. Either we gain the years of experience ourselves (which I think is astronomically helpful in the long run) or we can pay for the work of others, which is what we have decided is best for this project given its timeline.
thanks Roy love your work this video is treasure
Thank you!!!
sick video, what multiplayer series do you recommend for unreal? trying to find one that does things through steam using blueprints and it's a bit of a struggle lol
Thank you 👍🏻
You are a true gamedev ❤😊
thanks for this video 😍😍😍
That is a phenomenal video.
Great video to stumble upon. Ive been working on a multiplayer game for the last several months and I figured Id hit this wall eventually.
But now what? How are you gonna find GMC tutorials for this new movement system?
I think tutorials gloss over the nitty gritty stuff because theyre meant for new people who just want something easily thrown together without being discouraged. We are in a niche area where we actually wanna do the complicated stuff without using the complicated language. BP tutorials are too basic, CPP tutorials are too advanced. But this very much helped, thank you!
Thank You
Currently learning Unreal, sticking with singleplayer games for now, so I've focused only on blueprints. But if I ever need/want to work on multiplayer, good to know the 4 years of C++ I have will come in handy, thanks! Good luck with your project.
Holy fuck what a perfect time to find this video, ty so much
Great video.
I needed to watch this video a couple times round because of all the DISTRACTIONS
your game looks crazy af!
thanks
yay!!!!!! i hope this game becomes very cool someday!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
You deserve a subscribe❤
It's unreal to me (pun intended) that this is such a MAJOR flaw in the engine, and nobody talks about this. If you want your game to have any kind of uniqueness/sauce at all, it's gotta have its own movement system. I feel like Epic is putting all of their resources into impressive graphics, and forgetting about all the actual GAME pieces. I see so many show off videos of rendering things in UE5, but almost nobody has a handle on how the CMC works, and how to extend it. In general, the design of the CMC is abysmal, and I can't believe Epic hasn't provided a better solution or at the very least more insight on how to make something that goes beyond the standard movement modes and simple "go to position" type movement. The GMC from what I can tell is superior in every way, but it's so expensive. I don't understand how Epic hasn't bought it out and adopted it. I personally can't spend that kind of cash on an asset right now, but it seems incredible.
This is true. Nobody tells you this. I assumed the same thing about custom movement and now am balls deep into a game which works well EXCEPT for the damn movement. Good video thanks for this.
Update: got my character movement working like a dream using the smooth sync plugin suggested in this video. Thanks
very very useful,. I v got 2 questions :) plsss. How much did you take to learn all the programing part to get to this game? (with ur animation and 3dmodeling base)? and a curious question,. if its 200 ping, why did at 9:50 ur left player takes 500 miliseconds (30 frames at 60fps) to actually hit the player on the right screen
Took me a decent 6-8 months to really feel comfortable with the programming. I was up and running about 2-3 months in, but to really understand the flow of everything without needing tutorials and such took some time. I think it's important to highlight that, even though I was (and still am) deep in developing our game, I am always learning and getting better. Second, the ping is a little bit misleading on games that are peer-to-peer. And there is no way around it, the simple fact is when you are peer-to-peer the actual network ping is pretty much double. Because there is the time to the server player, the processing time, then the time back to the client player. It's not exact but as a general concept the ping takes roughly double the time to process on peer-to-peer. Glossing over that fact, we display the ping of just your machine in our game. We could "fake" it and double it to give a better representation, but at the end of the day as long as it gives a general idea of your connection then that's what's most important. It would also be a little misleading for people to see themselves on 140 ping when connecting to close friends, they might think something is wrong when really its just the peer-to-peer ping.
I can’t wait to play this! Do you have music for the game?
working on it!
@@BoryDev i make music! If you need help I would love to lend a hand
Several years ago i tried to make my game on unreal with custom movement and i didn't know what to do. May be now is a chanse.
thankfully, the game I'm currently trying to make is only a Dungeon Master vs Player type of game. A 2 player card game with no movement. Just need to figure out how to give them each separate controls
Aus? The latency problem is real!! Haha Rollback Netcode!? Damn! Even modern AAA games, ie. fighting games, are finally only starting to realise it's importance. Seems you got thrown in the deep-end, and it's a success story! Really great video, happy to watch. Thanks for sharing! Inspiring and hopeful 🥺
I also find it amusing that the whole unreal community sees this as the hugest problem that they use Client Authoritive Data Streaming when basically GTA V, DayZ in its early stage, EFT etc... all had absolutely ZERO anticheat and everything was client side, and now they added at least anticheat and people are still cheating because you just can't prevent Aimbots and Wallhacks. I don't see any reason to focus so much on anticheat when everyone will cheat anyways. CSGO would not need to have VAC at all because everyone who wanted to cheat could cheat without problems in the past. So where to draw the line...
I took a break from watching UE5 videos because the text on their blueprints became so F'ing small I counldn't see on my TV screen I use on my computer. And I need a good channel to teach me about multiplayer and how to prevent hacking and so forth. Ill watch more of your videos
Dude, the game looks genius. Whatever it is today or whatever it's supposed to be one day, it is already a bomb! 😂 Shake it
So correct me if i am wrong. If you dont need to dont add any special movements to the game, develop game to its full potential. Then add GMC redo characters blueprint and you are good to go? of course there is going to be some troubleshooting and stuff but did I got it right ?
Thanks for this video it was super informative, but any chance you can confirm if i couldnjust add the rollback code to the c++ side of the project and leave it at that?
would you be able to expand that question a bit? in what aspect do you mean just "add the rollback to c++ side"? Do you mean generally for everything, or for the example implementing the GMC solution?
The character movement can be client predicted and server authoritative with GMC but what about other actions such as weapon simulation, is there any solution for that?
Have everything server authored, you dont need any plugins for that to do so. Weapons are similar, let the server simulate everything, only send the results to the player. Idk why the creator of the video makes such a great fuzz about it. If you programm the thing like you should (using the tools epic gives you) you wont have any problem. Not with custom movement, custom items, physics replication, everything works flawlessly, if yuo do it like epic wants you to do things, aka the easy way = Server authoritative. You can of course execute everything the same way client-side optimistically, and update from the servers result, so things happen immediately, but end up using the servers result.
Do you have tutorials on how you created your servers? Begginer dev looking to start with simple multiplayer.
200ms is ok for fun game, but if u want competitive multiplayer, for exemple like valorant, CS:Go, Call Of Duty, it's too much delay :/ we can't grab C++ code and add to our project like plugins ? and then have same performance than people who use C++ ?
GMC looks great, and $600 is nothing if you are actually working on a commercial product. That's a chunk for those of us doing this stuff as a hobby though. That said, $600 would buy a good bit of C++ training as another option. Not necessarily "better" option for everyone, but should be considered. As you mentioned, some people would still have trouble achieving these results in CPP, so it's not an obvious decision. That said, it would be more enabling overall to learn CPP. I recently had to use to just to get modifier keys working (think P vs ALT-P), so it's pretty common to need to use CPP in UE. You can do A LOT with blueprints, and, as you have demonstrated, you can complete a game with only blueprints, but that is not the intention of the engine developers. The engine and toolset are developed around the idea of using a combination of CPP and Blueprints, and if you are able to do that it's a whole new world vs fighting the natural flow.
I agree! 100% no doubt. In my case, the only limiting factor was time. When money is no longer a barrier, time is what makes the world tick and with limited time solutions like the GMC saved me. After this game's launch calms down I will definitely be investing my free time into CPP courses before I commit my time to another project.
@@BoryDev Isn't it awesome to have the options we have today? If AI lives up to the promise, how many more people will be able to create things that seemed out of reach because of time or whatever before? Amazing time to be alive.
Regarding GMC, What about other actions? (except movement) such as: punching, kicking, shielding. is the prediction & rollback still applies?
I'll try to answer my best however I don't want to confirm anything as I'm not the developer. From what I can understand the GMC is meant to handle everything that 'moves' the player as it is essentially made to replace the movement component only. However It's up to you if you want to handle a punch or kick anim through the GMC, we do our slap animation through it (doesn't need to be) but you need to make sure any logic that happens as a result needs to be handled accordingly. Our slap is on overlap outside the GMC. Other people have asked about knockback etc which needs to be handled outside the GMC too but then 'tied' into it properly since knockback is a form of movement. So it's a very situational answer, it will handle prediction of any variables you bind inside it but you need to be smart about that. I think if you want fast rapid kicks and punches, it could be good to handle the replication and prediction of those inputs and drive the animations, then any resulting gameplay logic done outside. Don't put things like health, stamina, or shield etc meters inside it.
I give a LIKE because of your characters butt still thank you for tips
so in other word, if i dont use custom movement i can only blue print and create multplayer game no problems?
How did you make your blueprint connections look so clean and linear?
Will someone direct the dead by daylight devs to this video😂😂😭
Something I am curious about - IF you had the time to learn C++ What resources would you use to learn this in particular?
I already have a few courses on gamedev.tv to watch, I just need the time. Any course that specialises in C++ and multiplayer. I wouldn't be too picky about the "game" they teach to make or who the teacher is, at the end of the day the more knowledge in my brain the better. I think a cohesive full course is more beneficial than scattered KZhead tutorials. Other than courses I could go back to a degree or something but I already got student debt so I ain't spending more money on that
But for learning the networking features "in particular" I don't know I haven't researched that far hehe
@@BoryDev as an experienced c++ coder, I can tell you right now that proper character movement netcode is HARD.. can't just learn that from a c++ course unless your desired custom movement is simple.
I had the fortune of witnessing the state of modern AAA releases which instilled the fear of networking in me early on, allowing me to avoid getting blindsided by it later down the road.
Could you maybe make a tutorial on blueprints with this add on
Does this framework work on console? or will this limit a game to PC only?
About six months late here, I fear, but... While the framework only ships with Windows binaries (and an AllowedPlatformList of only Win64), I can attest that it works fine on macOS and Linux as well; the Windows limitation in the .uplugin and on the store page isn't because the code is Windows-specific, it's because Grim can only build and test it for Windows. Given that it *does* work fine on macOS and Linux once you strip the AllowedPlatformList out (and build the plugin yourself), I would assume it would also work fine on consoles.
Bro really be making Thugg Shaker: The video game
How did you learn to right the character with GMC
Look, I'm a noob at cooding, im curently using godot engine, but I just want to say that I really understand what you say, It's litteraly disclose verry little known publy information that is actualy verry important at some point , I mean for multiplayer game servers , the network sync is verry important, and even for godot 4 and probably for all engines , the "movement sync" is requiring a hella of C++ data script from both server and client in order to be perfecly synced. So from my point , I appriciate your explenations and you got a like and sub for these "small" acctualy important remarks. Good job brother, keep going this way.
make a story game with Spanky, doubt that's the players name, but I was thinking it'd be cool if he had his own story game like Sackboy's Big Adventure on the ps5
You make some tutorial series about your game?
How did you get those straight connections?
Would Blueprint Nativization help with this?
Very interesting video, unreal is very complex and making games is hard, don't worry about "not finding solutions earlier"... Btw what the fck is that game
For a space game I'm using SixDof Movement which includes network replication. Does that mean my game is pretty much safe from huge headaches?
I don't think any game is safe from huge headaches in regards to replication unless you have enough experience to say that for certainty yourself. Best I can say is to prototype as much as you can and at each stage always remember: launch 2 separate instances/windows in standalone with one as a host (or host in background) and play with "NetEmulation.PktLag 500" where 500 can be any number you want to test in milliseconds of network delay. I did not know to do that myself, and that's when I ran into the realization when testing with friends that my movement is more complicated than I thought. So always test with emulating pktlag and just look for anything you wouldn't want to see in a real game
Dude, how did you get that awesome-looking pin shape? How are those diagonal???
Plugin called "electronic nodes" :)
@@BoryDev Cool! Thank you!
As a free alternative, wouldn't ALSv4 Replicated achieve this as well? I'm interested in trying that as GMC is prohibitively expensive for me
It could but I haven't touched any versions of ALS. They seem great but over bloated for smaller games. I also enjoy having the creative control of including exactly what I want and learning how to implement that, rather than removing code I don't want. I agree the price is on the steep end, there's nothing wrong with trying premade things (als) before committing to the pricey solutions. For me it was easy because I had a lot riding on this game, so I did what I thought was the best in the long run (spending company money) but I know it's a bit different for personal cases. In the long run, I now have the knowledge to make almost any movement type I want for all future games, rather than be locked in to what a premade solution can give me
have you tried the community made, "advanced steam subsystem"?
Can you say something about the performance of GMC? How many players can be handled when a tickrate about 30~ is targeted on a dedicated server?
Sadly I can't yet. We haven't really dove into performance for our project yet, and our lobbies hold 8 max so stressing the system out hasn't been on our to-do yet
Eos integration system helpful for this custom movement? Can anyone explain
EOS or any other online subsystem don't really have much to do with the moving or replication. They handle the sessions, lobbies, friends etc. and other online systems. But the replication of data is the same with all - so using different ones shouldn't affect movement or gameplay at all.
Would you recommend this for vehicle movement?
We don't use any vehicles but I know its possible and there are others in the discord developing vehicles and ships with it. Depending on how realistic you want the physics interaction will dictate the amount of work. It's not a plug n play solution especially if you want to build a physically simulated vehicle (looking at the discord this seems possible but is some uncharted territory), but for arcade feeling cars I think it would be similar setup to characters. Overall it seems like physics is the biggest hoop for vehicles, I would imagine the replication and prediction part would be just as stable as anything else in the GMC
I feel like since im going to use the basic movement component, I dont have to wory about thsi
Similar to early 2000s games having tons of hackers or lagswitches to win.
how do your blue prints have those lines, where can i get them
Plugin called "electronic nodes" :)
Im doing a turn based game.. would that be easier to implement?
Probably not needed for turn based as you can mostly hide any kind of latency with all kinds of visuals, player turns, animations etc. The components / issue that this video discusses is more for real-time action and response games
@@BoryDev That was my thinking and that's so re assuring to know thanks :)
how did u add network delay manually?
There are a few commands, but I mostly use: NetEmulation.PktLag 200 To toggle the command window in-game press the ` key (button below escape key). You can replace the "200" with any number you like, that's how much delay in milliseconds will be simulated.
@@BoryDev I am from unity, being able to use a command in game costs $40 as a plugin for unity. Thx for the info!
When you showed yourself turning the speed of the video to 2x I had a laugh. Me too m8!
As an engine used in one of the biggest multiplayer games (fortnite), I feel like they should have this NAILED DOWN somewhere other than custom addons.
Most unreal tutorials are like this, they lead you down the wrong path just to get something quick to put in a video. Then you go to make a real game with those techniques and it falls apart. Epic themselves are the worst for this, they do those short tutorial streams where there's no way they could cover the topic in enough detail to be real. It's a big trap indeed.
I struggled with this for months in the past. I tried smooth sync, but it wasn't compatible with root motion animations (my melee combat system mostly depends on it). But then I found a tutorial with a c++ example from Reids Channel in youtube, and that tutorial saved me. I avoid c++ if I can, but you can copy that code, and adapt it to your requirements (I just removed the wall running logic). The video is titled "UE4 - Advanced Networked Movement Tutorial (Sprinting & Wall Running)" by the way. After that, to fix root motion rubberbanding, I enable client authoritative movement only during the montage and that works fine.
That's really cool! I should have a look at this when I get time. Other than the setup, have you found the need to use C++ in any other movement aspects or are you able to handle everything through blueprints?
How hard was it to transition your game to GMC? Did it take long?
Does anyone know if GMC supports rootmotion?
Hey, for anyone else looking for an awnser on this question, i found a video made by the developers for GMC that goes over implementing rootmotion, very easy to follow as well
does this plugin support android build?
It's only officially supported for windows at the moment
Sorry for my ignorance. I'm still just looking into game engines. Can someone explain the following. In the GMC plugin page, it say it only supports windows platform. What does that mean? does it mean that any code I make cannot be then exported to game consoles in the future? sorry, noob here trying to develop my own game.
Many plugins will say the same thing, it just means that the developer is only making it to work for windows but for a lot of plugins you can still export for other platforms, sometimes you may have to recompile the plugin yourself. Essentially the plugin developers don't guarantee anything will work on anything other than what they support, but that doesn't mean it's impossible for you to do yourself :)
@@BoryDev Thanks, and really cool game you are making, I really enjoy the videos! Just to clarify what you are saying. This 'limitation' would limit me on the dev side, or you are saying it could limit the final product side (eg it won't compile on PS5). Because a quick look though plugins, I don't see a single plugging that shows PlayStation support per see. It seems to me its more for the dev purposes. On a different topic. Can you make a video of the plugins you have used/ recommend? Thanks again!
@@user-zy6wg5qt5f I'm not sure if this is what you mean, plugins shouldn't affect you as the dev working on your project unless perhaps the plugin doesn't natively support Mac and you are working on a Mac. In that case it would affect you because you would need to rebuild it. But otherwise, if you worked on a PS5 project, it wouldn't affect you until you need to actually build and test the project on a PS5 The way I understand it is that Plugin Developers will only list what platforms are supported based on what they are able to test with. If a plugin developer doesn't have a Mac, they would not put Mac on the list of supported platforms. Same goes for PS5 or Xbox, if the plugin developer can't test their plugin specifically on those platforms then they would not say it can - BUT in a lot of cases the plugins will work perfectly fine (might still need a quick recompile to allow that platform) because the code that the plugins provide are just C++ that the engine exports to those platforms normally anyway. However if the plugin uses an external library / dependency (idk what the correct language is) that is not native to the consoles or unreal and is some custom library of data that's when it gets tricky.. I think .. don't quote me on that because I am not 10000% sure but that's my understanding
To add to this thread, I can personally attest that GMCv2 builds fine on Linux and macOS. The plugin is set up for only Windows (because that's what Grim can build and test on), and only provides Windows binaries, but the dev does make an effort to ensure that everything still works pretty much out-of-box on macOS and Linux if you strip the "AllowedPlatformList" out of the plugin file and are prepared to build it from source yourself. In testing on Linux and macOS for Grim, I've also tested on Android and iOS and can confirm that (not shockingly) it also works there. As such, I figure the chances it would compile fine for consoles as well are fairly good.
Thank you so much, this video is really great, but GMC is not support for Android/iOS. Do you have any suggestion for mobile game?
Sadly I don't have any experience with mobile. It would depend on the type of game you are making and the security you need. The smooth sync example I used in the video works for android and IOS, but I don't really know anything about cheats and exploits on mobile. Otherwise you could look at the GAS system unreal has, I've heard good things about it but I haven't used it myself.
@@BoryDev thank you Rory. I hope you can making more excellent content like this
i dont get why epic games dont just buy gmc and include it in the engine, or make something better (if thats possible). its difficult to understand how they can focus on metahumans and all that stuff but networking which is arguably one of the most important parts of games is so weak without a plugin like GMC. is it that they dont want multiplayer competition for fortnite from indies?
5.4 has the solution, also ability to change collision shape at runtime.
Is it better than this?@@EnigmasGravitas
@@EnigmasGravitas How so, please elaborate
Kinda bummed that the answer was a 350$ plugin... very informative though.
it was 350$ when you checked? 10 months later is 682$!
bro when u finally release your game, sure as hell you getting a pg restriction on it
So.. if one was making a PVE add on for an existing game (that is multiplayer and works okish) and wanted to ensure I don't have to recreate everything if I decided to do multiplayer - would it be advisable to use this from the get go. The cost is not prohibitive at all (for me) and on the off chance I decided to go that route (or if I reuse what I create in another project in future) would you recommend just starting with this plugin instead?
Is the existing game made by you or is the PVE add on kind of like a user mod? I guess this is a bit of a larger design choice, if I thought that there's a chance that movement in your pve add-on could become a limiting factor then I would personally just start with the GMC. In my current project I spent longer than needed making my movement only to redo it all with the GMC. And unless I learned more about the unreal movement component and C++ I don't really see myself needing to go back at all for this project.
@@BoryDev We have been playing/polishing (to our ability, I have videos uploaded) multi for a few years and tinkering with the idea of doing something that involves single player. I guess my ultimate concern would be doing this, being mildly successful and people asking for coop. So if this is some middle ground for doing that.. it might be useful. To clarify this is more for the enemies and not the players.
I think I understand a bit more now (your movement looks very nice btw). The GMC is good for singleplayer too and code-wise is pretty much exact same workflow as multiplayer GMC. So if you made singleplayer with GMC then want to convert that to coop - you shouldn't have much issue at all swapping from single to coop as the only major difference is just making sure the variables are replicated properly, that's it. However if this is just for your enemies (supposed they're AI), I can't really vouch for AI yet because that is our next months task we have yet to undertake and I don't see a lot of people posting AI showcases / W.I.Ps in the discord. I know some people are doing it, I just don't see it posted much.
@@BoryDev Thanks for taking the time. Much appreciated.
damn boi he thicc