No announcement yet.

Has anyone played around with Unreal Engine 4?

  • Filter
  • Time
  • Show
Clear All
new posts

    Has anyone played around with Unreal Engine 4?

    I was just wondering anyone's thought are on UE4?

    * I think Epic killed UDK before UE4 was ready... So much so, that some days I have trouble believing that people have to pay for the UE4 license in its current form...

    * You can see the pot holes by going over to the new forums and reading posts by smallB (stability) or Tegleg (car mechanics) for example... Read the Roadmap thread and follow other user threads giving their impressions of UE4 too. But pay particular attention when Epic's Developers reply. Its telling how even they are making new discoveries about UE4. This confirms that the engine is still very much evolving...

    * I started out with 4.3 and it just crashed randomly and incessantly. Recent versions have been better but not stable enough, not for high-end PC development IMHO... Add to that, the Trello Todo list is massive and further changes like Matinee to Sequencer will herald endless change, surely at the cost of stability...

    * Prior to UE4's release from private beta I assumed the product was going to be very complete because of all the hype. But after using 4.3-4.6 I'm not as convinced. It reminds me of the release of MS.Net. Microsoft decided to change a lot of things for sake of change because they were jealous at the rise of Java. It was at the expense of Visual Studio, their killer product at the time. I think Epic too are guilty of changing stuff for the sake of it. I believe they're being influenced by Unity also. The new UE4 editor UI is strikingly similar to Unity's UI, and the Marketplace is clearly an attempt to create a homegrown Asset-Store. But change is most visible in the fact that UE4 & UDK are completely incompatible. All that change is going to take time to get right. Its been suggested meantme that we're all Epic's guinea pigs, its food for thought!

    * For that reason, I fear UE4 is 2-5 years away from being a complete, thoroughly debugged and finished game engine as per UDK (if Epic would only fix UDK's known bugs). That's a controversial view possibly, but I'm sure UE4 is at least another 6-months to a year away from being stable and useful. That said, some like 'Fighter' tend to disagree and he's completed a game in UE4. So why not research past posts as there's some interesting threads UDK vs. UE4 vs. Unity vs. Cryengine etc. But if my view is on the money, we're in for a long wait for a complete UE4, which is the future of course, or one possible future along with Unity 5 and several others etc..

    * The Marketplace has undoubtedly got a long way to go too, years in fact, to catch up to Unity's store... Right now, the cupboard is pretty bare and its expensive to get access to even basic starter kits etc... The approval process is also a minefield, with new assets only being released on a drip feed... That's in stark contrast to the rich supply of free assets available to the UDK community today, or the pay-for Assets on the Unity Asset-Store etc. So if you go straight to UE4, don't expect a wide selection of free vehicles, assault / sniper-rifles, characters and terrains etc ready to play around with. What's there is limited, so you'll need to be able to build your own or pay for everything up-front...

    * What can you do while UE4 evolves meantime... There's lots of overlap between the UDK and UE4 editors, so if you're just starting out it shouldn't matter which you learn first. But if you stick with UDK eventually you'll be faced with the unrealscript dilemma. As its been deprecated its wasteful to learn it now with all sights set on UE4... Then there's Kismet, but its just too basic compared with Blueprints and requires custom Kismet nodes to do anything useful, which in turn requires unrealscript, so you're caught in a trap!

    * Overall, its a tricky decision, but I think it boils down to your chosen game type... For my money UE4 high-end dev for PC has some way to go, but I can't speak for mobile...


      well put frankit,I agree 100%.


        agree with frankit

        just to add to it, there are often times a blueprint becomes corrupt due to something simple like renaming a variable or just a random crash. this essentially corrupts the entire project, lost count of the number of times ive had to start again.
        i really don't know how people have managed to complete a project with ue4 in its current state.
        though it is an amazing engine and hopefully it will be more stable and complete sometime in the future.

        as for mobile support/performance, dont know for ios but android is worse than terrible.


          I already tried UE4 myself, while blueprints seem for me something like very stupid and how dumb people seem to treat them , " like some sort of super powerful tool ", it's just archetypes, kismet together really... it's 100 times better to code in c++ than have 1000 boxes around so blurry and confusing. "I have tried them and i got to get results really fast like i was just unrealscripting not talking blindly ".

          In the c++ side, with the vector math, i think that compared to unrealscript, things definitvely changed " but not so much ".

          Im talking about vectors since like we all know they are a real important " math branch" for the gameplay programmers.

          So well if in unrealscript you typecasted to for instance translate a vector to rotator, now you just need to take the vector as a instanced vector, and it's already translated " i guess that in runtime ", to a rotator. MyVector.Rotation(); and vice versa with the rotators.
          Then they implemented, add local rotation or add local location, somehow, " 2 different actor functions " i don't even know how to add local rotation in unrealscript or location, while when i tried UE4, i finally learn how to in c++...
          Maybe everything it's even more well structurated.
          And you no longar have set base or attach ... and 100 attach functions, you just have AttachTo, that works for sockets and pretty much everything. " maybe better" why would you have 3 different functions to do the same, that just confuses the programmer using your product. .

          i don't see how EPIC killed UDK, that's not true.
          They could implement deferred lighting and all that stuff, that really i just know that it's cool for performance and that would be so cool for PC devs, but hey it's just fine with DX9 i guess im not an expert with the graphics and all this stuff.
          UDK is great engine, very versatile to make your graphics look like you want, and your gameplay to be like you want. You don't have that in other engines. It's very monolithic in other engines. Though the limitations are there for depending what project. First and the most obvious one it's
          how expensive dynamic lighting is.


            Originally posted by Neongho View Post
            i don't see how EPIC killed UDK, that's not true.
            Hi Neongho, I agree UDK is a great engine. By killed, I meant Epic killed off 'support' for UDK... i.e. There's no more fixes / additions in the pipeline for Desktop / PC games. This means Epic will never address well documented bugs such as Crowd Agents etc... This hurts my game as I invested heavily in CA's and now they won't replicate in multiplayer at all. I spoke with Epic about this directly and they bluntly said, migrate to UE4 because this will never be fixed! Epic's website also states that there will be no updates to UE3, and they've removed prior UDK builds from the download links (although the binaries can still be found by amending the URL for now)...

            BTW: There are workarounds for less expensive dynamic lighting in UDK. A lot of the overhead stems from using Matinee or unrealscript in such a way that changes to Brightness & Color incur a heavy penalty. However Obihb and others have suggested workarounds...


              Thanks for UE4 nfo. I guess UE4 is the way forward definitely for big companies who have the people and intimate cooperation with Epic. For someone like myself UDK even with its faults works. Also @frankit appreciate your work with Crowd. I've been thinking about an alternative for multi online crowd and think I'll be using a simple AI with low poly - 1/2 doz or so random wanderers. Shootable and dieable and screaming muha.


                Smart thinking Snipe34. I believe that's what others have done to get around the problem, especially if hundreds or thousands of CA's aren't needed on-screen at the same time.

                CA's don't have proper collision either so while they can be shot at, you can drive or fly right though them without contact which isn't credible. But they're easy to get up and running and can perform simple behaviours and flocking, so they're attractive from that point of view i.e. acceptable for prototyping / demos.

                For now, I've got around the problem by using one listen-server and one client. CA's show up on a listen server, so they can be used to hide or mask a Client player (terrorist) from a Listen-Server player (security guard). But obviously this isn't ideal except for a prototype / demo set-up....


                  Well after reading what everyone has said about UE4, I think I will stick with UDK instead.


                    For what it's worth, I used to be a UE4 subscriber--I joined in March 2014, the month the new engine became available--and have come back to UDK. Questions like this don't have a right answer, but I wanted to briefly make some points that may help others who are pondering where to start. Without a doubt, UE4 is the new shiny, and they are investing heavily in its success, which is why UDK is on life support and no more. If it makes those of you who feel left out in the cold feel any better, CryTek did the same with their free CDK when they moved to their Engine as a Service. That said, before you jump, there are a few things that I think you should consider.

                    First of all, we need to go back to March 2014, when the engine was made available to the masses. At that time, Epic warned new users that the engine is "powerful, but not very polished, and it requires a beefy desktop computer." They mentioned there were some rough areas, including Mac OS X and Android support and sparse documentation, saying, "If you're looking for a more polished product, please check back in 6 months." Well, almost a year later and they are still making major changes to the engine, and to say that documentation is still "rough" is a bit of an understatement, especially if you're looking for C++ documentation/examples. Blueprints are getting a lot attention, but C++ is relegated to the status of poor cousin. Epic has put more effort into updating the documentation, and creating new, but it is slooooow. I will hand it to them, however; their marketing effort was superb and they easily outclassed their competition. Whoever is in charge of this is a freakin' genius.

                    It has gotten to the point where, when they release a new version (i.e. 4.5), you are better off waiting for the soon-to-follow 4.5.1 that fixes some of the new things broken. With 4.7 they are making a more concerted effort to provide a better polished release, but even here we see the underlying business model--people are paying Epic to be beta testers. Prior to 4.7, only those who downloaded the source from GitHub and compiled it themselves had access to the latest changes--and these individuals typically know what they are getting into. With 4.7, someone at Epic had the bright idea (no sarcasm, it's brilliant) to make these "preview" releases available in the launcher, thereby greatly increasing the number of beta testers. When Chance Ivy announced that Epic would be taking more time with the 4.7 release, the fan boys fell all over themselves thanking Epic. What? Shouldn't this be the case for all of their releases? In addition, the release of the source code means that individuals can make changes to the engine, or add new functionality, and request that Epic make these changes part of the the code base. In the release notes for each major release, Epic acknowledges such individuals and it is informative to look at the number of names listed. One of the recent releases had somewhere around 80 names! Wow, that's a lot of free programming help.

                    A recent post (27 Jan 15) on the UE4 forums ( gives you an idea of just how much things can break although, in this case, Epic said that they were really sorry and promised not to do it again.

                    Why some groups are switching over to UE4 now (you frequently see the announcement on the UE4 home page) is a mystery to me. Anyone with a game in progress would be foolish to switch to an engine that is still in development. I also found that, like UDK, Epic still seem to focus on introducing new functionality, rather than fixing known problems. For an example, go the UE4 forum page and look at the new build preview pages. There are a lot of cool features in UE4, and it will be a great engine someday, but not this day. I could go on (and on) about how C++ feels neglected in favour of Blueprints, and how inconvenient questions go unanswered, but I won't. The bottom line is that Epic is in business to make money. Period. If they can make you happy, then you're less likely to go to CryEngine or Unity. They only have so many engineers and programmers, however, and UE4 is the future for them. The fact that the next Batman game, Arkham Knight, is made with UE3 is somewhat telling in regards to UE4's readiness for prime time. A year later and they are still breaking projects when they release a new build. If you want to be part of the new shiny, then you should definitely go for UE4. Their subscription model is excellent and even if you opt out, you can still use what you have and can opt back in at a later date. But make no mistake, Epic is doing this for Epic and you will have nowhere near the number of examples, tutorials and sample code with UE4 that you have with UDK. If you just want to get your feet wet with game programming, you will find more resources for UDK. When the engine quality catches up to the marketing quality, then switch.

                    Just my $0.02


                      Originally posted by Djohn View Post
                      people are paying Epic to be beta testers
                      - Ouch!

                      Originally posted by Djohn View Post
                      One of the recent releases had somewhere around 80 names! Wow, that's a lot of free programming help
                      - Sucker Punch!

                      Originally posted by Djohn View Post
                      Why some groups are switching over to UE4 now (you frequently see the announcement on the UE4 home page) is a mystery to me. Anyone with a game in progress would be foolish to switch to an engine that is still in development. I also found that, like UDK, Epic still seem to focus on introducing new functionality, rather than fixing known problems.
                      - Bat to the back of the head!


                      * Excellent stuff Djohn! Pity you don't post here more often! Question: Why aren't these Epic 'Rock Stars' getting something for their time? As you point out, this is a commercial project, not 'open source' love. Are any of the helpers getting free licenses for life?

                      * Where have all the former UDK game devs gone? Several well known names have moved over and have become moderators or champions of UE4. But others have simply disappeared from the face of both forums. What happened to 'the others' and what tools are they using meantime while they 'wait' for UE4 to be 'ready'???

                      * Many game studios claim to be using UE4 to create their next game, and are solely hiring UE4 devs accordingly. Question: Do they all have ample money-in-the-bank and 'floating' release dates 2-5 years from now? Its baffling! How can this be?

                      Originally posted by rodstone
                      Is it weird to be in love with a Game Engine Development Company?
                      Is there a UE4 Zombie virus...?


                        I dont think is a virus but it coud become someting like an Apple syndrome...You buy it just for the majic of the name.As you have told me Frankit things about the ue4 marketplace its no wonder they will try to copy Unitys idea.From what i see unity wins a lot of money without moving a finguer just because they take % of what people sell at their marketplace and they have a hell of a lot users(even if the engine cant match unreal).In the end ue4 is going to become a great engine and its name is going to be famous as saying Coca Cola but is buisness and people have to eat.
                        I must admit that i was behaving selfish that udk got left behind but the truth for me and probably other users was that its hard to let go of all the knoledge gained to try and master the thamn thing and now start all over.But as humans we are capable and we shoud find time and continue to learn.This udk thing opened my eyes that so much existed.I knew other engines but really didnt prested much atention to them.In the end of the day you have to move with the heard and try not to be left behind as in this buisness is very easy to loose your way i supose.

                        I was going to say something very motivating here but im going to stop as it will sound cheesy


                          Just to float a couple more glass half full thoughts into the mix.

                          I think with any engine, as an indie dev (particularly with limited access or just capability to modify the engine) you really need to design a game while already having a good idea what does and doesn't work in the engine. I haven't used UE4 to make a full game yet, I've just toyed with it and kept an eye on others' progress, but a lot of the complaints I see are related to a certain feature that isn't available or not knowing how to use a certain feature. If you have one idea that you absolutely need certain features for, you're probably going to have to code them yourself or make heavy compromises no matter what engine you go with.

                          Once I finish my current UDK project I'll be going to UE4, but for the above reason, I have about 20 or more ideas on the backburner and I'll be doing a significant amount of testing/research to find out what works before picking the most suitable project.

                          From what I've seen, Epic have been pretty keen to fix any genuine "release blockers", so if you target basic working features it seems like you should be ok.

                          I've seen a lot of objection to Epic's model of "user pays to test/fix their stuff". Regardless of whether we're guinea pigs or not, I wouldn't not use an engine just because that doesn't feel right - what it comes down to is whether you believe you can find better for your money. If I paid a million dollars to license the engine I'd have a right to complain about having to code my own features and fix everything, but I think the UE4 revenue model is designed with "there might be problems" in mind, because it's pretty reasonable for what you get even at this stage.

                          I think it's fair to say you'll definitely want to invest a fair bit of time in learning c++ basics (or make sure you have a capable programmer) because you'd be lucky to complete an entire game in blueprint unless it's super simple.


                            Originally posted by majic12 View Post
                            about the ue4 marketplace its no wonder they will try to copy Unitys idea.From what i see unity wins a lot of money without moving a finguer just because they take % of what people sell at their marketplace and they have a hell of a lot users(even if the engine cant match unreal).
                            actually this was a direct suggestion from the UE4 beta testers, not Epic's own intention of copying Unity's idea 'just because it was cool or profitable'
                            also on Epic's approach of the marketplace they do have to move more than a finger, as they manually go through all submissions to ensure they match the quality standards. there's a lot of back and forth between Epic and the marketplace developers so it's quite some work for them.


                              I don't like blueprints either. I think they are fine for small things, but if the thing is complex, becomes a labyrinth of boxes and lines unmanageable.

                              I started with EU4 when it came out, and I saw that Epic did great efforts to facilitate the programming to people without programming experience, using blueprints.

                              But I saw that they had completely forgotten those who had experience or simply wanted to continue programming in the traditional way and not piling thousands of boxes connected by lines.

                              For example, things such making a modular pawn were extremely easy to make using blueprints, but found nothing to manage it from code (necessary for player customization, for example).

                              With that kind of stuff I ended up very very frustrated.

                              Maybe I wanted to do with EU4 the same thing as with UDK too fast, but some documentation (without blueprints) would have helped.

                              That happened long ago, and now probably there are more documentation for C ++ programmers. I don't know, because I left EU4 and I decided to finish my project in UDK, hoping to return after to EU4 and had more documentation available.

                              I really like what can be achieved with EU4, I give much importance to the graphical look and EU4 is incredible, but the new programming paradigm, the blueprints, kill me.