Professional game developer since 20+ years here, worked on quite a few AAA and AA projects. I understand that it may be frustrating to update the tools to newer versions of the engine, but a major version is still a major version, and Unity explicitly states that previews and experimental systems can be reworked and even pulled at anytime. It has to be like this, and in case this is not working for some, wait a few releases until it has stabilized. Working on early packages or systems is usually a risk a responsible developer should really consider. I would never use these packages in production. I have been working in several versions of Unity, as early as Unity 2, 15+ years ago. In general, in production, we only work with LTS versions, even if they are far from the latest tech, pre-production are usually over a long time ago. I do understand that indie developers with lower risk assessment would like to be on the latest tech streams, but in real productions these should be handled with caution. If you are lucky, you can be on the same LTS throughout production, but in many cases, longer production require you to bump the major engine version. However, you never bump an supported LTS to a non-LTS, and the transition is planned and carefully executed. I do understand that you as a tools developer is frustrated, but working on supporting and catching up a the latest versions unfortunately is a lot of work. No to mention when the engine is releasing competing features. We are exploring Unity 6 now and it is a stellar piece of software. I would never release a game in Unity 6 today, In this case I would have to still go for 2022.3 LTS. Sure, the engine is far from perfect, in cases overselled, and countless mistakes has been made. Especially when launching ECS as production ready, some 5 years ago, even though it was quite pre-mature. But in general, I would choose Unity over any other engine I have been working with, including Unreal. I could spend an entire evening to endlessly speak about the issues and clusterf*cks Unreal have had or overselled. Unfortunately that is a part of development, so it boils down to risk assessment and risk mitigation, making responsible and well researched decisions. Working responsibly on the LTS versions and with a good support agreement in place, you are as a developer very safe with Unity. Also, it would be interesting to know where you got the information regarding the demo not being real-time. I would question the truth in that.
I agree with you, but still see a hole here and that is that Unity 6 is now called an LTS version. I understand that want to let us know they are going to support it for a long time, but LTS stands for more than just long support, it suggests that this version IS ready to be used for serious projects. Like you said, you wouldn’t, so they shouldn’t have said this. Because they do this, I get a lot of customers that come to me with bugs, with the claim that it comes from my asset, instead of a buggy Unity version “because it is LTS!”. On your last question, they indeed claim it to be realtime, they seem to show that off in the a keynote. But I have an i9 cpu and 4090 rtx and as I said in the vid (in text), I can just about run that character with its hair and shader goodies at 60 fps. Meaning that the whole scene and all sfx will never be something that you could publish as a game. Unlike those unreal demos
If you are close to releasing a game today, you should not be on the newly released "Unity 6 LTS", you should definitely be on 2022.3 LTS for many more months. The meaning of LTS is that the feature set is locked down and stable for production, and it is planned to be stabilized over 2 years. A LOT of things will be fixed and stabilized over these years, so obviously there is still a lot of work on it. I would be so bold to say that Unity 6 is ready for pre-production, and we can plan our features and system architecture based on the information available. If the game is scheduled to be released later next year, we will get a lot of stabilizing fixes until then. If the game is smaller, with lets say 6 months of production, I would as mentioned earlier, go with 2022.3 LTS. Even in the LTS, they are bumping the SDK's of console target well in time before their sunset. When starting to work on an LTS, the major feature set and API should be pretty much locked down, so that you can make a qualified decision if these features are in line with the requirements. The package ecosystem in Unity allows for new major package feature versions, and you can always expect to update your systems to suit these, but this is what the release notes are for. In some cases, we use reflection to reach package internals, and here things can change without notice, but that is the risk of accessing internals. Usually easy to fix though. For packages that receives major updates or introducing features that are not compatible with the stabilization phase of the engine, are instead flagged as unsupported for that major engine version and instead available in the next major engine version. Such as the later versions of URP are unavailable to us when we are on 2022.3 LTS Yes, Time Ghost was shown at the keynote. It is a technical demonstration, it always is about pushing and selling the engine. In a few cases I know that tech demos (not mentioning any engine vendors) has been running on a modified engines to reflect what the developers should expect. It is important to understand that these tech demos is always a sales pitch that shows the potential. When starting a preproduction of a game, you do not necessarily targeting the current hardware, you are targeting the projected hardware at the game release. If I'm allowed to use my imagination here, a game like Time Ghost would take somewhere around 3-5 years to develop. In this time, you will have a lot of more compute than what you have today, and not only for early adaptors. For console targets, the specifications are set from the start, but again they are usually scaled to, and optimized for the specific hardware. This takes a lot of effort. Looking at Unreal and Nanite for instance, it is not magic, you still have the storage and memory bandwidth to adhere to. They are not full games either, but this is nothing that they show of during their keynotes. And that is okay :)
Everything you're describing here assumes a team following normal production & project mgmt approaches. The issue is that the ecosystem includes a lot of teams who jump on the latest shiny thing, with no understanding of the risks or headaches involved. Sadly, this isn't only indies, but also established teams who just haven't had it all blow up in their face... yet. There's also the issue of ppl not reading Unity's own docs & website to grok basics like Tech Stream vs LTS, or URP vs HDRP, or even what "experimental" or "preview" mean. The return to pure semantic versioning is nice, but so many don't even know how that works. It's not a language barrier, or a tech problem, it's an attitude problem. Loads of folks figure they can "skip the boring stuff" and then come at the asset creators (or the Unity forums) with anger & entitlement when their lack of preparation becomes an emergency. It's the number-one reason I've avoided making assets.
@@mandisaw Yes, this is the troublesome truth. It becomes obvious when looking at the Unity community forums and a vast majority of all replies are just poor development practices and guess work. I never use these as a reference, but fortunately I'm also on a support agreement with access to the engineering teams. Ultimately Unity is an advanced tool, and game production is a hard earned skill. It may look easy on the surface, and Indie development is awesome (and something that I would like to do at some point), but as a large portion of these developers have never worked in a reasonable large production, learning the ropes by an already established and experienced team, this is the outcome.
@@keytronic5631 You are more forgiving than I. Most modern "get started in game-dev" books and even Unity's own talks & workshops all refer to and emphasize these basic hygiene practices. You get a lot of variance with indie dev-tubers, but I would say most of the ones that write code do mention these fundamentals - solo or team, yoe measured in decades or 6-24mos. Most experienced forum commenters and asset-creators will mention it as well. People are just glossing over that stuff. Unity is uniquely generous with their doc materials, before you ever get to a point of needing an enterprise-support plan. I pulled an entire eBook's worth of "Shaders: Noob to Expert" material out of their _old_ manual, for free! But RTFM is still gonna be the #1 issue. See it in bedroom coders and in Unite case-study talks.
I love Unity, its modular, accessible and far easier to get crazy ideas out there, if you are willing to sacrifice visual fidelity. A good concept of a game is much more valuable than copy-pasting meta-humans for a demo. However, getting the hang of Unreal using blueprints is a nice way to kickstart your ideas and the very Tripple A, enterprise games are just much more suited for Unreal. Unity has diversified, with mobile apps, URP, HDRP, rendering for movies and so one and kind of lost its track when it comes to their core identity of traditional PC games. Now I believe many, many features and even visual fidelity can be achieved and sometimes outperformed by Unity but their marketing is just bad, instead they seem to prefer filling the headlines with greedy corporate decisions. I just hope, Unity will set their focus straight and hey.. maybe even idk.. developer their own game, to fix some of the fundamental work-flow blunders to make our lives easier :)
It would be very interesting to see what you could create in Unity if you have a serious budget, 4-8 years time and a seasoned team… Man, that sounds like something i’d really would love to be apart of…
I migrated my ongoing project to Unity 6 and after a couple of days messing with it I decided to staty on 2022. First, the editor is slower (yes, even more). Second, of the new features, the only one which actually caught my attention was the possibility of interpolating between several sets of baked lighting and, to be honest, is something that I was already doing in Unity 2022 by other means. Same happened with the automatic light probe placement. The GPU resident renderer didn't give me any performance improvement. I also tried HDRP which still performs poorly. I got better framerates using Unreal with Nanite and Lumen enabled. Ah, yes. The package manager was better. Congratulations.
Yes, I had the same experience! I was so surprised that the srp batcher didn’t actually seem to do anything! The fantasy kingdom demo ran 1.5 times as fast on 2021 than it did in 6. Now in all fairness, some 6 features like AVP didn’t work ofc. But still, it (again) feels as if we’re spammed with useless new features instead of the important ones like: fix the damn editor regression!
I'm still on 2022 the best stable version so far. Have test Unity 6.0 , not really buying the up-hyped around it. Its nothing new that is useful except than the baked light probes (and hopefully but haven't seen that yet, animation tools). The rest then ? I wouldn't even use Muse , after 15 days I have to pay for it, and not only that, it gives me crap textures , I could generate these for myself, for free in another AI. So please someone, explain, what I have been missing except for the up-hyped "cool" demo that you could make in HDRP already ? I might have, but can't see that. For me it looks the same.
game engines arent cars. i get that, when people see a new feature they run to it. but that's a horrible idea to start with. the latest the version, the more bugs it has. for a general rule of thumb, use the lowest possible version of unity with the features you actually need. unity 6 released not even 2 weeks ago, anyone complaining about it being buggy, for their production product. has a few screws loose or something
“Game engines aren’t cars” that is the quote of the day!! 🤣 In all ernesty, the issue is that Unity is calling 6 an LTS. Claiming that it’s stable (the most stable) and production ready. Which I think is not very smart and hence this vid!
@@The-Naked-Dev fair enough, but. seeing a new engine release. doesn't matter what they say really, its going to have bugs, issues. changes. that should be expected., no one. should be doing any serious finacial project on unity 6. regardless. not for a little while yet. not to mention, they don't say, it wont have bugs. the wording used. is "...significantly more stable and performant than our last major release..." which is true. it IS more stable in comparison to unity 2023 when it first came out or 2022. but more stable, isn't "stable" I think the direction unity is heading, is great though. last year, it really seamed liked they were ready to just focus on mobile garbage. so, for now. ill play with unity6, more as a toy, while I use 2022/2023 for projects. gamejams and such probably use unity6. but I wont look at unity6, for anything serious until next year
@@littleowlgaming-unity-tutorialoh I absolutely agree, Unity 6 is to play around with and not to use for production. But 90% of the user base will see the LTS term and go for it, which is a problem. And yes, the current direction is, as I also say in my vid, way, way better it was going a year ago.
I have little choice but to update to Unity 6000.0.23f1 from 2022.3.10f1 as there are a few issues some new Unity components fix that only work in Unity 6. I have been going through the process of updating my scripts and sadly I have to say, I am disappointed. Like majorly. I was hyped for Unity 6, but it feels, performs and behaves the same, and a little worse. For instance, light baking is messed up. I have groups of scenes that bake together as to not run out of VRAM. In Unity 6 I have to restart the Editor everytime I bake a group because unloading a group and then loading a new group causes their bake to hang the same way as when you bake too much and run out of VRAM. I was hoping the light baking would be better, but it looks identical and performs worse and FFS Unity, at least spit out an error that you ran out of VRAM versus the indefinite light bake. Another new issue is that for some reason when I do a build and run it crashes on my Meta Quest, but relaunching it works just fine. Never seen this issue before. There are other issues too some noted in this video. Overall, I just feel like Unity is a horrible business partner who does not respect their customers. I guess heres to hope Unity 7 is the one. Hopium is my drug of choice. For those still reading here is a tip that may help you out when updating your scripts to Unity 6. You see ChatGPT 4o, o1, Claude and others do not know anything about Unity 6 as their knowledge cutoff is October 2023 or worse. I tried Muse AI and even it does not know Unity 6. WFT??? In fact, it did not even know the namespace for Cinemachine change between 2.x and 3.x which is not even that new of a change. This means Unity did not even update Muse which is a paid service they are hoping we will all use. Thankfully, there is an OpenAI GPT trained on Unity 6 docs which has been extremely helpful (need the plus paid account to access). For those considering this path, I found feeding the script to the Unity 6 GPT while asking it what needs to change and then feeding that answer and the same script to be edited to o1 preview or mini worked the best as the GPTs seem to have a token limit in their replies so if you feed it a 1,000 line script to edit, it can read it, but not spit it back out at you, whereas o1 can, but o1 does not know Unity 6. The combo of the two has saved me many hours of work as who wants to manually go through many lines of code to just change names and API calls etc. o1 is very good at just spitting out the edited script if you tell it what needs to change.
That is an amazing tip!!!! Especially when using URP Rendergraph the current documentation is absolutely horroble, being able to use ai for it would be great. Many thanks for sharing! I’m very sorry to hear about your issues… All I can suggest is opening many, many bugreports so that Unity actually learns how bad 6 is actually doing… GL with your project though!
Unity on Linux still has some glaring issues, and it's Vulkan support is still not 1:1 with DX12 on Windows. I wonder when they're gonna officially implement XeSS, DLSS 3 (with frame generation), and FSR 3.1 (with frame generation), the FG stuff specifically requires changes to how swapchains are managed. How does URP not have any sort of temporal anti-aliasing out of the box? That feature of Unity 7 where the entire project doesn't freeze while it imports scripts and assets can't come sooner can it? Legit alt-tabbed out of the editor to do something else while I wait for it, just to then alt-tab back into it when I think it finished, just for it to do even more stuff.
I douby vulkan will ever get full 1:1 feature support, but that is more due to nvidia refusing to. I highly doubt Unity will ever get official XeSS, FSR 3 and DLSS support on all RP’s but that may be mostly because I am already selling those assets!! Not having any freezes in editor while code is compiling sounds like absolute dark magic and can’t come too soon!!
@@The-Naked-Dev Interesting about the whole NVIDIA thing. Is there a good summary on specifically what might be the issue? Does it have to do with Vulkan extension related stuff, or something else?
@@KingKrouchyeah, it has something to do with that nvidia never wanted to fully support opengl, due to that it was an AMD thing and they didn’t want to help the conpetition. Vulkan is basically a rewrite of opengl
I feel like there's just sadly no real alternative to Unity for ""medium"" development still. Unreal is a behemoth of a monster I don't need. And Godot is just sucky with missing features and performance of a dying horse after you put in a few hundred nodes/scenes. Honestly, my eyes are on O3dE or Bevy more than the other players, they've been doing some nice stuff lately but still a long way to go. More than Bugs though I am weeping over the Editor's performance itself. Slower and slower with every release.
I guess it is easy for us to bash on Unity, but they really are in a bit of a pickle… But yes, I think they shouldn’ed focussed on Editor perf and stop its regressing
I am also facing project-breaking bugs in Unity 6, They are adding and removing stuff with no proper documentation, It's fine for a normal user but for those who create their custom pipelines and custom features it's a nightmare, of course, I haven't switched my main just yet, but I am testing it by making a duplicate copy of it for unity 6 and so far a ton of custom features are broken, I have fixed 50% of those but it's a tone of work and I won't switch until I know that 100% of my custom things will run on unity 6
I won't be touching it until they fix thsoe long load times and making everything separate such as packages is cancer. even godot and flax comes preinstalled with a shader graph. unity makes them into packages and making the project heavier.
To be fair, the package concept does make the editor faster, due to the assembly definitions of the packages not needing to recompile. But yes, the editor is still way slower than it should be…
@@The-Naked-Dev I know but after adding all the packages, those rebuild times and launching the editor times sky rocket specially that assemblies rebuiild times. flax is a new engine sadly not much about there unless you study the source code yourself and know how to work with game engines as the documentation is extremely limited but even that is more stable than Unity. I got unity 6 for testing today as I am a shader developer but that too was screwed and didn't launched so I am redownloading 6 now and test it out. hope they fix the packages stuff too and only add special stuff as packages such as networking and other stuff. what I dislike is having basic features such as a shader graph, URP, HDRP, vfx graph being packages which is bs.
@@The-Naked-Dev got unity 6, package manager was indeed a pain it at least they fixed the shader graph. they also added a default bloom in volume and some good features, the input system is simplified too. Hope they get rid of the specular based environment reflection as well it is outdated and unreal got rid of it in version 4. hope unity does the same to make graphics look better.
Unity 6 crashes the first time I used it (last week). Didnt even get to do anything. Started it up, went away from my PC for a few minutes. Screen saver had kicked in - logged in, and boom. Crash. Reported it to Unity. Response back; Won't fix. Your GFX card goes into powersave mode during screensaver mode, and that crashes Unity. That is to be expected. WHAT?....? !!!
Join the club… I reported two gamebreaking bugs (crashes) to Unity. Both they say they will not fix. I mean, one of them is during building for pc, default build setup…
It seems like unity 6 has a lot of heavy lifting to do. It has to compete with the godot and unreal, but also pull the company out of the dumpster fire they put themselves into. I'm not sure who this demo was for , really. I didn't see anything that made me want to come back to it. maybe if they focused on stability and fleshing out tools unity already has, like the landscape and open world system, I would have been tempted to come back. To me as a dev, flashy new features and graphics just mean more headaches.
Yeah, they say to focus on performance and stability but only came up with new and unstable features instead of focussing on improving the stuff that is already there
Glad you liked the song, I have so much fun making these! We knew the shift was coming, but many things that they are claiming about Unity 6 and its “stability and performance” are plain untrue.
Hello, congratulations on a very good job. Maybe it is not related to the subject but after I updated my project to unity 6 there is no menu “windows > rendering > rendering pipeline convert ”. I will convert materials to URP but there is no menu “windows > rendering > render pipeline convert”. There is only “Lighting, Light Explorer and OC,”. What could be the reason? Thank you.
Why are you upset that they were breaking your assets by updating the beta. Why were you using a beta for serious development, and then continually updating? That's 100% on you, dude. Use LTS for serious projects. Convert to a beta for fun, and don't keep updating beta versions every time a new one is released... you are dev-ing wrong there...
I think you misunderstood. I’m not using beta software for serious development. Unity is marketing their Unity 6 beta versions as the most “stable and best-performing”. Convincing my customers to upgrade to these versions. Additionally, now that it is out of beta, they also claim it to be LTS. While it is not.
Switch to another engine, do not trust Unity, they are bad people, theirs license plans are not certain , even worse publisher in many talks to me and my colleagues at my company told me boldly they will not invest in any new Unity based game project!
What's the song that starts playing at 03:24? I've been racking my brain trying to figure out where I've heard it before. Is it something by Kevin MacLeod? It sounds like it could be.
It is a song I had made years ago, you can here it on spotify here: open.spotify.com/track/51oXX70vtxQvHLsYCuTAlt?si=gzhjsKdwTr-Zn0OtGiIeCw&context=spotify%3Aalbum%3A3nieXDxNOel4gpePtk0jUJ
Indeed it does, and LTS means it won’t! Unity 6 has now been officially released but is still receiving big changes, which go against the concept of LTS.
0:25 why so salty? sure it's marketing bla. But it's just: yeah, we fixed bugs so it is more stable by definition. "the most" is not really saying anything really. 2:00 this is more an issue of using/supporting preview versions. Expecting preview versions being as stable as an LTS would defeat the purpose 6:00 Time Ghost did run in real time. Fair: on 4090 but it also has an amount of detail that is higher than 90% of games. Also Epic put's marketing lies even into their docs. They also don't talk about any of the drawbacks of nanite, lumen, and mega lights. (especially in presentations, and why would they?) 6:40 Yeah, because PostProcessing effects are not costly at all. and I guess you profiled it on Mobile and Web against the new version, so be in an actual hardware constrainted environment? 7:18 ... first video I was watching from you ... is you video a 10-minute-hoax for the youtube algorithm?
The current Unity 6 version is less stable than 2022, 2021, 2020 etc, claiming it’s more stable is not true 🥲 About the 2nd thing I agree but its out of preview AND called Unity 6 LTS, so they claim it is as stable(or more so) than previous LTS versions. Which again, is untrue.
As an outsider I see a lot of similarities with the music industry. Unity is like Yamaha, their production equipment is top notch. Unreal is like Guitar Center, everything has a price tag and they love store credit. Godot is like a school band teacher, who'll hand you a cowbell or ukulele just for fun.
@The-Naked-Dev in my 20s I wanted to make music, but didn't know where to start. Most advice was to get a guitar and all these courses and equipment. I took a piano class, bought a $50 ukulele, and played pretty much nonstop for a few years. I'm closing in on 40 now, and when I look for game dev advice, a lot of it is similar to making music.
While unity doesnt lead with innovation, it still leads with simplicity and stability. Sure some of these updates will cause issues especially to asset store devs but... I been using unity for 8+years now and I never really had major issues with the engine. If you try something like unreal you will actually feel blocked
I’m an Unreal Engine dev. I am often blocked by finding solutions for seemingly simple issues or implement something simple. UE 5+ is hyper focused on third and first person game genres. It’s difficult to find reference for something simple like Grid Movement. 2D projects are a massive fight.
@@fragileglass9622 I hear about this quite often. For more “indie” like 2D games Godot or Unity do seem an easier match. That doesn’t mean that Unreal can’t do it ofc!
If you really have been using Unity in a commercial setting for 8 years and have never had any issues, you are part of an extremely small minority group. On the other hand, feeling 'blocked' when using an engine that gives you the *entire* source code for everything inside, is simply a skill issue.
@@darak2 I guess it depends on the tech requirements of the commercial projects. I can think of many smaller project I’ve done is the past that had 0 issues. Only the more complex projects, and other platforms than win only that came with issues. But I have to disagree with you on Unreal/source code access. Ofc it can be a skill issue, but in a serious commercial situation, it is all about time, not skill.
@@The-Naked-Dev a lot of unreal's features in 5.0 were... laughably bad. World partition has been revamped almost entirely (the only thing that i can think of that *hasn't* changed much since 5.0 is how you assign an actor to a grid) Nanite performance in 5.0 vs 5.5 is night and day, lumen is already being mixed in with nvidia's magical tech (megalights are, afaik, the result of nvidia's NVRTX branch efforts in coordination with epic) the list goes on. and on.
@@The-Naked-Dev you should watch ThreatInteractive's epic rants about nanite and TAA hahaha("Epic's Unreal Optimization Disaster | Why Nanite Tanks Performance!")
WE NEED MORE OPTION AS STOP PROCESS COMPILE COMPILED SCRIPTS OR NO NEED TO COMPILE , SEPERATE INPUT SETTING FOR EACH SCENE SEPERATE ,AND LIGHTING SETTING INSTEAD REPLACE SETTING WITH OTHER SETTING ,SAVE EACH INPUT SETTING SEPERATE INTO FILE , AND MORE AND MONO DEVELOP FOR SCRIPTTING AS UNITY 5 ,AND DISABLE PLUGINS IF NO NEED AND OR NEED TO USE OTHER PLUGINE(SWITCH BETWEEN PLUGINS THAT PROBLEM WITH OTHER PLUGIN)
After spending the last 1.5 years on an Unreal Project. Be careful, coming from unity. You'll encounter a bad awakening. Having a physics based renderer, but lumen not supporting realistic light intensities (like 1500 lux and more) but instead getting bloomed out of the universe if you try. But going too low gives you spots that make the screen become pitch black. Also it means that you can't have 1 light setup for fast iteration with lumen and when finishing things relying on light baking - you'll have to setup everything new for that, but than it's also not dynamic RTGI for higher end hardware. Also Lumen completely relies on temporal AA techniques (or at least TSR), but those introduce a lot of ghosting. Having a documentation that tells you "every mesh is improved by nanite" and in reality most meshes have WAY less overdraw using good LODs with nearly no visual impacts. fancy new features like nanite tesselation looking like a glitchy mess in like 80% of the scenarios. Don't ever use "ChildActorComponents" they will mess up whatever you're trying to do, when you least expect it (latest example i encountered: breaking in a shipping build where suddenly their transforms where interpreted in WorldSpace instead of ParentSpace) For the programmers: unreal enforcing stupid naming schemes even for local variables and parameters and their compiler will error when you have a member and a parameter named the same. The Actor->ActorComponent system is also a mess coming from Unitys super clean GO->Component separation. It's also "just an engine" and it has some pros over unity but also some cons. don't get blended by marketing bla. some lessons learned with unreal from the last 1.5 years: notion.so /Unreal-Engine-1b131f85e86d4f75b41945072fa6a0de
@@Soraphis91 im fully aware of its quirks etc., rendereres pros and cons, each engine has its own bits, at end of the day it's just a tool, i ain't willing to build my own so i choose which one felt best and has actual c++ support. you could go on for days about things that arent perfect, missing, badly supported in any of the engines. the rest is on the developer to know what their building and what they're going to utilize to achieve that goal. (nanite vs no nanite, to use lumen or not to etc, etc) use what works for ya and stop blabbering.
Professional game developer since 20+ years here, worked on quite a few AAA and AA projects. I understand that it may be frustrating to update the tools to newer versions of the engine, but a major version is still a major version, and Unity explicitly states that previews and experimental systems can be reworked and even pulled at anytime. It has to be like this, and in case this is not working for some, wait a few releases until it has stabilized. Working on early packages or systems is usually a risk a responsible developer should really consider. I would never use these packages in production.
I have been working in several versions of Unity, as early as Unity 2, 15+ years ago. In general, in production, we only work with LTS versions, even if they are far from the latest tech, pre-production are usually over a long time ago. I do understand that indie developers with lower risk assessment would like to be on the latest tech streams, but in real productions these should be handled with caution. If you are lucky, you can be on the same LTS throughout production, but in many cases, longer production require you to bump the major engine version. However, you never bump an supported LTS to a non-LTS, and the transition is planned and carefully executed.
I do understand that you as a tools developer is frustrated, but working on supporting and catching up a the latest versions unfortunately is a lot of work. No to mention when the engine is releasing competing features.
We are exploring Unity 6 now and it is a stellar piece of software. I would never release a game in Unity 6 today, In this case I would have to still go for 2022.3 LTS.
Sure, the engine is far from perfect, in cases overselled, and countless mistakes has been made. Especially when launching ECS as production ready, some 5 years ago, even though it was quite pre-mature. But in general, I would choose Unity over any other engine I have been working with, including Unreal. I could spend an entire evening to endlessly speak about the issues and clusterf*cks Unreal have had or overselled. Unfortunately that is a part of development, so it boils down to risk assessment and risk mitigation, making responsible and well researched decisions.
Working responsibly on the LTS versions and with a good support agreement in place, you are as a developer very safe with Unity.
Also, it would be interesting to know where you got the information regarding the demo not being real-time. I would question the truth in that.
I agree with you, but still see a hole here and that is that Unity 6 is now called an LTS version.
I understand that want to let us know they are going to support it for a long time, but LTS stands for more than just long support, it suggests that this version IS ready to be used for serious projects. Like you said, you wouldn’t, so they shouldn’t have said this.
Because they do this, I get a lot of customers that come to me with bugs, with the claim that it comes from my asset, instead of a buggy Unity version “because it is LTS!”.
On your last question, they indeed claim it to be realtime, they seem to show that off in the a keynote. But I have an i9 cpu and 4090 rtx and as I said in the vid (in text), I can just about run that character with its hair and shader goodies at 60 fps. Meaning that the whole scene and all sfx will never be something that you could publish as a game. Unlike those unreal demos
If you are close to releasing a game today, you should not be on the newly released "Unity 6 LTS", you should definitely be on 2022.3 LTS for many more months. The meaning of LTS is that the feature set is locked down and stable for production, and it is planned to be stabilized over 2 years. A LOT of things will be fixed and stabilized over these years, so obviously there is still a lot of work on it. I would be so bold to say that Unity 6 is ready for pre-production, and we can plan our features and system architecture based on the information available.
If the game is scheduled to be released later next year, we will get a lot of stabilizing fixes until then. If the game is smaller, with lets say 6 months of production, I would as mentioned earlier, go with 2022.3 LTS. Even in the LTS, they are bumping the SDK's of console target well in time before their sunset.
When starting to work on an LTS, the major feature set and API should be pretty much locked down, so that you can make a qualified decision if these features are in line with the requirements. The package ecosystem in Unity allows for new major package feature versions, and you can always expect to update your systems to suit these, but this is what the release notes are for. In some cases, we use reflection to reach package internals, and here things can change without notice, but that is the risk of accessing internals. Usually easy to fix though.
For packages that receives major updates or introducing features that are not compatible with the stabilization phase of the engine, are instead flagged as unsupported for that major engine version and instead available in the next major engine version. Such as the later versions of URP are unavailable to us when we are on 2022.3 LTS
Yes, Time Ghost was shown at the keynote. It is a technical demonstration, it always is about pushing and selling the engine. In a few cases I know that tech demos (not mentioning any engine vendors) has been running on a modified engines to reflect what the developers should expect. It is important to understand that these tech demos is always a sales pitch that shows the potential.
When starting a preproduction of a game, you do not necessarily targeting the current hardware, you are targeting the projected hardware at the game release. If I'm allowed to use my imagination here, a game like Time Ghost would take somewhere around 3-5 years to develop. In this time, you will have a lot of more compute than what you have today, and not only for early adaptors. For console targets, the specifications are set from the start, but again they are usually scaled to, and optimized for the specific hardware. This takes a lot of effort. Looking at Unreal and Nanite for instance, it is not magic, you still have the storage and memory bandwidth to adhere to. They are not full games either, but this is nothing that they show of during their keynotes. And that is okay :)
Everything you're describing here assumes a team following normal production & project mgmt approaches. The issue is that the ecosystem includes a lot of teams who jump on the latest shiny thing, with no understanding of the risks or headaches involved. Sadly, this isn't only indies, but also established teams who just haven't had it all blow up in their face... yet.
There's also the issue of ppl not reading Unity's own docs & website to grok basics like Tech Stream vs LTS, or URP vs HDRP, or even what "experimental" or "preview" mean. The return to pure semantic versioning is nice, but so many don't even know how that works.
It's not a language barrier, or a tech problem, it's an attitude problem. Loads of folks figure they can "skip the boring stuff" and then come at the asset creators (or the Unity forums) with anger & entitlement when their lack of preparation becomes an emergency. It's the number-one reason I've avoided making assets.
@@mandisaw Yes, this is the troublesome truth. It becomes obvious when looking at the Unity community forums and a vast majority of all replies are just poor development practices and guess work. I never use these as a reference, but fortunately I'm also on a support agreement with access to the engineering teams. Ultimately Unity is an advanced tool, and game production is a hard earned skill. It may look easy on the surface, and Indie development is awesome (and something that I would like to do at some point), but as a large portion of these developers have never worked in a reasonable large production, learning the ropes by an already established and experienced team, this is the outcome.
@@keytronic5631 You are more forgiving than I. Most modern "get started in game-dev" books and even Unity's own talks & workshops all refer to and emphasize these basic hygiene practices. You get a lot of variance with indie dev-tubers, but I would say most of the ones that write code do mention these fundamentals - solo or team, yoe measured in decades or 6-24mos. Most experienced forum commenters and asset-creators will mention it as well.
People are just glossing over that stuff. Unity is uniquely generous with their doc materials, before you ever get to a point of needing an enterprise-support plan. I pulled an entire eBook's worth of "Shaders: Noob to Expert" material out of their _old_ manual, for free! But RTFM is still gonna be the #1 issue. See it in bedroom coders and in Unite case-study talks.
I love Unity, its modular, accessible and far easier to get crazy ideas out there, if you are willing to sacrifice visual fidelity. A good concept of a game is much more valuable than copy-pasting meta-humans for a demo. However, getting the hang of Unreal using blueprints is a nice way to kickstart your ideas and the very Tripple A, enterprise games are just much more suited for Unreal. Unity has diversified, with mobile apps, URP, HDRP, rendering for movies and so one and kind of lost its track when it comes to their core identity of traditional PC games. Now I believe many, many features and even visual fidelity can be achieved and sometimes outperformed by Unity but their marketing is just bad, instead they seem to prefer filling the headlines with greedy corporate decisions. I just hope, Unity will set their focus straight and hey.. maybe even idk.. developer their own game, to fix some of the fundamental work-flow blunders to make our lives easier :)
It would be very interesting to see what you could create in Unity if you have a serious budget, 4-8 years time and a seasoned team…
Man, that sounds like something i’d really would love to be apart of…
I migrated my ongoing project to Unity 6 and after a couple of days messing with it I decided to staty on 2022. First, the editor is slower (yes, even more). Second, of the new features, the only one which actually caught my attention was the possibility of interpolating between several sets of baked lighting and, to be honest, is something that I was already doing in Unity 2022 by other means. Same happened with the automatic light probe placement. The GPU resident renderer didn't give me any performance improvement. I also tried HDRP which still performs poorly. I got better framerates using Unreal with Nanite and Lumen enabled. Ah, yes. The package manager was better. Congratulations.
Yes, I had the same experience! I was so surprised that the srp batcher didn’t actually seem to do anything!
The fantasy kingdom demo ran 1.5 times as fast on 2021 than it did in 6. Now in all fairness, some 6 features like AVP didn’t work ofc. But still, it (again) feels as if we’re spammed with useless new features instead of the important ones like: fix the damn editor regression!
I'm still on 2022 the best stable version so far. Have test Unity 6.0 , not really buying the up-hyped around it. Its nothing new that is useful except than the baked light probes (and hopefully but haven't seen that yet, animation tools). The rest then ? I wouldn't even use Muse , after 15 days I have to pay for it, and not only that, it gives me crap textures , I could generate these for myself, for free in another AI. So please someone, explain, what I have been missing except for the up-hyped "cool" demo that you could make in HDRP already ? I might have, but can't see that. For me it looks the same.
Nope, nothing to explain here, you are absolutely right. 😭
I downloaded and installed Unity 6, but 2022 LTS remains my main version for now.
Probably a smart move! Ofc we should check out 6, but I don’t think it’s a good idea to switch to yet!
@@The-Naked-Dev I've already tested my asset for compatibility, and it works in Unity 6. 👍
game engines arent cars. i get that, when people see a new feature they run to it. but that's a horrible idea to start with. the latest the version, the more bugs it has. for a general rule of thumb, use the lowest possible version of unity with the features you actually need. unity 6 released not even 2 weeks ago, anyone complaining about it being buggy, for their production product. has a few screws loose or something
“Game engines aren’t cars” that is the quote of the day!! 🤣
In all ernesty, the issue is that Unity is calling 6 an LTS. Claiming that it’s stable (the most stable) and production ready. Which I think is not very smart and hence this vid!
@@The-Naked-Dev fair enough, but. seeing a new engine release. doesn't matter what they say really, its going to have bugs, issues. changes. that should be expected., no one. should be doing any serious finacial project on unity 6. regardless. not for a little while yet. not to mention, they don't say, it wont have bugs. the wording used. is "...significantly more stable and performant than our last major release..." which is true. it IS more stable in comparison to unity 2023 when it first came out or 2022. but more stable, isn't "stable" I think the direction unity is heading, is great though. last year, it really seamed liked they were ready to just focus on mobile garbage. so, for now. ill play with unity6, more as a toy, while I use 2022/2023 for projects. gamejams and such probably use unity6. but I wont look at unity6, for anything serious until next year
@@littleowlgaming-unity-tutorialoh I absolutely agree, Unity 6 is to play around with and not to use for production.
But 90% of the user base will see the LTS term and go for it, which is a problem.
And yes, the current direction is, as I also say in my vid, way, way better it was going a year ago.
I have little choice but to update to Unity 6000.0.23f1 from 2022.3.10f1 as there are a few issues some new Unity components fix that only work in Unity 6. I have been going through the process of updating my scripts and sadly I have to say, I am disappointed. Like majorly. I was hyped for Unity 6, but it feels, performs and behaves the same, and a little worse. For instance, light baking is messed up. I have groups of scenes that bake together as to not run out of VRAM. In Unity 6 I have to restart the Editor everytime I bake a group because unloading a group and then loading a new group causes their bake to hang the same way as when you bake too much and run out of VRAM. I was hoping the light baking would be better, but it looks identical and performs worse and FFS Unity, at least spit out an error that you ran out of VRAM versus the indefinite light bake. Another new issue is that for some reason when I do a build and run it crashes on my Meta Quest, but relaunching it works just fine. Never seen this issue before. There are other issues too some noted in this video. Overall, I just feel like Unity is a horrible business partner who does not respect their customers. I guess heres to hope Unity 7 is the one. Hopium is my drug of choice.
For those still reading here is a tip that may help you out when updating your scripts to Unity 6. You see ChatGPT 4o, o1, Claude and others do not know anything about Unity 6 as their knowledge cutoff is October 2023 or worse. I tried Muse AI and even it does not know Unity 6. WFT??? In fact, it did not even know the namespace for Cinemachine change between 2.x and 3.x which is not even that new of a change. This means Unity did not even update Muse which is a paid service they are hoping we will all use. Thankfully, there is an OpenAI GPT trained on Unity 6 docs which has been extremely helpful (need the plus paid account to access). For those considering this path, I found feeding the script to the Unity 6 GPT while asking it what needs to change and then feeding that answer and the same script to be edited to o1 preview or mini worked the best as the GPTs seem to have a token limit in their replies so if you feed it a 1,000 line script to edit, it can read it, but not spit it back out at you, whereas o1 can, but o1 does not know Unity 6. The combo of the two has saved me many hours of work as who wants to manually go through many lines of code to just change names and API calls etc. o1 is very good at just spitting out the edited script if you tell it what needs to change.
That is an amazing tip!!!! Especially when using URP Rendergraph the current documentation is absolutely horroble, being able to use ai for it would be great. Many thanks for sharing!
I’m very sorry to hear about your issues… All I can suggest is opening many, many bugreports so that Unity actually learns how bad 6 is actually doing…
GL with your project though!
Unity on Linux still has some glaring issues, and it's Vulkan support is still not 1:1 with DX12 on Windows. I wonder when they're gonna officially implement XeSS, DLSS 3 (with frame generation), and FSR 3.1 (with frame generation), the FG stuff specifically requires changes to how swapchains are managed. How does URP not have any sort of temporal anti-aliasing out of the box?
That feature of Unity 7 where the entire project doesn't freeze while it imports scripts and assets can't come sooner can it? Legit alt-tabbed out of the editor to do something else while I wait for it, just to then alt-tab back into it when I think it finished, just for it to do even more stuff.
I douby vulkan will ever get full 1:1 feature support, but that is more due to nvidia refusing to.
I highly doubt Unity will ever get official XeSS, FSR 3 and DLSS support on all RP’s but that may be mostly because I am already selling those assets!!
Not having any freezes in editor while code is compiling sounds like absolute dark magic and can’t come too soon!!
@@The-Naked-Dev Interesting about the whole NVIDIA thing. Is there a good summary on specifically what might be the issue? Does it have to do with Vulkan extension related stuff, or something else?
@@KingKrouchyeah, it has something to do with that nvidia never wanted to fully support opengl, due to that it was an AMD thing and they didn’t want to help the conpetition. Vulkan is basically a rewrite of opengl
I feel like there's just sadly no real alternative to Unity for ""medium"" development still. Unreal is a behemoth of a monster I don't need. And Godot is just sucky with missing features and performance of a dying horse after you put in a few hundred nodes/scenes. Honestly, my eyes are on O3dE or Bevy more than the other players, they've been doing some nice stuff lately but still a long way to go. More than Bugs though I am weeping over the Editor's performance itself. Slower and slower with every release.
I guess it is easy for us to bash on Unity, but they really are in a bit of a pickle… But yes, I think they shouldn’ed focussed on Editor perf and stop its regressing
O3DE looks promising, though clearly in early stages of adoption.
I wonder if those upcoming engines will ever be able to get enough funds and traction to really become serious contenders though…
wowie, I can't believe that Unity is finally doing the industry standard. hwo exciting
What standard are we talking about? Lying about every feature the engine can(t) do? 🤣
I am also facing project-breaking bugs in Unity 6, They are adding and removing stuff with no proper documentation, It's fine for a normal user but for those who create their custom pipelines and custom features it's a nightmare, of course, I haven't switched my main just yet, but I am testing it by making a duplicate copy of it for unity 6 and so far a ton of custom features are broken, I have fixed 50% of those but it's a tone of work and I won't switch until I know that 100% of my custom things will run on unity 6
My suggestion is to wait at least a few more months untill actually investing time, big chance that the stuff you fix now will be broken again later…
@@The-Naked-Dev I will take U up on this advice, U are right!
I won't be touching it until they fix thsoe long load times and making everything separate such as packages is cancer. even godot and flax comes preinstalled with a shader graph. unity makes them into packages and making the project heavier.
To be fair, the package concept does make the editor faster, due to the assembly definitions of the packages not needing to recompile.
But yes, the editor is still way slower than it should be…
@@The-Naked-Dev I know but after adding all the packages, those rebuild times and launching the editor times sky rocket specially that assemblies rebuiild times. flax is a new engine sadly not much about there unless you study the source code yourself and know how to work with game engines as the documentation is extremely limited but even that is more stable than Unity. I got unity 6 for testing today as I am a shader developer but that too was screwed and didn't launched so I am redownloading 6 now and test it out. hope they fix the packages stuff too and only add special stuff as packages such as networking and other stuff. what I dislike is having basic features such as a shader graph, URP, HDRP, vfx graph being packages which is bs.
Yeah they may have gone a bit overboard on the packages manager.
@@The-Naked-Dev got unity 6, package manager was indeed a pain it at least they fixed the shader graph. they also added a default bloom in volume and some good features, the input system is simplified too. Hope they get rid of the specular based environment reflection as well it is outdated and unreal got rid of it in version 4. hope unity does the same to make graphics look better.
I do think they plan on making URP much, much better.
Unity 6 crashes the first time I used it (last week). Didnt even get to do anything. Started it up, went away from my PC for a few minutes. Screen saver had kicked in - logged in, and boom. Crash. Reported it to Unity. Response back; Won't fix. Your GFX card goes into powersave mode during screensaver mode, and that crashes Unity. That is to be expected. WHAT?....? !!!
Join the club… I reported two gamebreaking bugs (crashes) to Unity. Both they say they will not fix. I mean, one of them is during building for pc, default build setup…
Unity 6 is awesome, I have no issues with it
Awesome, I’m happy for you!
It seems like unity 6 has a lot of heavy lifting to do. It has to compete with the godot and unreal, but also pull the company out of the dumpster fire they put themselves into. I'm not sure who this demo was for , really. I didn't see anything that made me want to come back to it. maybe if they focused on stability and fleshing out tools unity already has, like the landscape and open world system, I would have been tempted to come back. To me as a dev, flashy new features and graphics just mean more headaches.
Yeah, they say to focus on performance and stability but only came up with new and unstable features instead of focussing on improving the stuff that is already there
What's with the music at the end bruhh...!!🤣🤣🤣🤣🤣
I just love making tunes!
Heyo i don't have a Important take on the topic but still wanted to make a comment for the algorithmey
Much love!
but really love the ai song
Have you ever tried to make AI songs? It’s seriously fun 😂
Calling a shift in direction, that we all knew was coming for a year, a hoax, is a bit of a stretch. A.I. song was hilarious though. 😋
Glad you liked the song, I have so much fun making these! We knew the shift was coming, but many things that they are claiming about Unity 6 and its “stability and performance” are plain untrue.
Hello, congratulations on a very good job.
Maybe it is not related to the subject but after I updated my project to unity 6 there is no menu “windows > rendering > rendering pipeline convert ”.
I will convert materials to URP but there is no menu “windows > rendering > render pipeline convert”. There is only “Lighting, Light Explorer and OC,”.
What could be the reason? Thank you.
Thanks! I actually wouldn’t know! Maybe they moved the menu location of the feature?
Why are you upset that they were breaking your assets by updating the beta. Why were you using a beta for serious development, and then continually updating? That's 100% on you, dude. Use LTS for serious projects. Convert to a beta for fun, and don't keep updating beta versions every time a new one is released... you are dev-ing wrong there...
I think you misunderstood. I’m not using beta software for serious development. Unity is marketing their Unity 6 beta versions as the most “stable and best-performing”. Convincing my customers to upgrade to these versions.
Additionally, now that it is out of beta, they also claim it to be LTS. While it is not.
Switch to another engine, do not trust Unity, they are bad people, theirs license plans are not certain , even worse publisher in many talks to me and my colleagues at my company told me boldly they will not invest in any new Unity based game project!
I’ve heard this aswell… publishers not liking the risk Unity games bring with them…
Lol 😒
What's the song that starts playing at 03:24? I've been racking my brain trying to figure out where I've heard it before. Is it something by Kevin MacLeod? It sounds like it could be.
It is a song I had made years ago, you can here it on spotify here: open.spotify.com/track/51oXX70vtxQvHLsYCuTAlt?si=gzhjsKdwTr-Zn0OtGiIeCw&context=spotify%3Aalbum%3A3nieXDxNOel4gpePtk0jUJ
"Preview" means it's going to change a lot...
Indeed it does, and LTS means it won’t! Unity 6 has now been officially released but is still receiving big changes, which go against the concept of LTS.
@@The-Naked-DevIf it’s not preview anymore and it’s receiving a lot of breaking changes (I don’t know) then it’s a different story
0:25 why so salty? sure it's marketing bla. But it's just: yeah, we fixed bugs so it is more stable by definition. "the most" is not really saying anything really.
2:00 this is more an issue of using/supporting preview versions. Expecting preview versions being as stable as an LTS would defeat the purpose
6:00 Time Ghost did run in real time. Fair: on 4090 but it also has an amount of detail that is higher than 90% of games.
Also Epic put's marketing lies even into their docs. They also don't talk about any of the drawbacks of nanite, lumen, and mega lights. (especially in presentations, and why would they?)
6:40 Yeah, because PostProcessing effects are not costly at all. and I guess you profiled it on Mobile and Web against the new version, so be in an actual hardware constrainted environment?
7:18 ... first video I was watching from you ... is you video a 10-minute-hoax for the youtube algorithm?
The current Unity 6 version is less stable than 2022, 2021, 2020 etc, claiming it’s more stable is not true 🥲
About the 2nd thing I agree but its out of preview AND called Unity 6 LTS, so they claim it is as stable(or more so) than previous LTS versions. Which again, is untrue.
As an outsider I see a lot of similarities with the music industry. Unity is like Yamaha, their production equipment is top notch. Unreal is like Guitar Center, everything has a price tag and they love store credit. Godot is like a school band teacher, who'll hand you a cowbell or ukulele just for fun.
I wouldnt know, the music industry is liie magic to me! But I’ll take your word for it 😀
@The-Naked-Dev in my 20s I wanted to make music, but didn't know where to start. Most advice was to get a guitar and all these courses and equipment. I took a piano class, bought a $50 ukulele, and played pretty much nonstop for a few years.
I'm closing in on 40 now, and when I look for game dev advice, a lot of it is similar to making music.
@@dfunited1Learn by doing! I am a big fan of that approach, wasn’t a big fan of having to study the subject at a university🎉
While unity doesnt lead with innovation, it still leads with simplicity and stability. Sure some of these updates will cause issues especially to asset store devs but... I been using unity for 8+years now and I never really had major issues with the engine. If you try something like unreal you will actually feel blocked
I’m glad to hear you haven’t had any major issues in the past 10 years! I definitely had and still have 😓
I’m an Unreal Engine dev. I am often blocked by finding solutions for seemingly simple issues or implement something simple.
UE 5+ is hyper focused on third and first person game genres. It’s difficult to find reference for something simple like Grid Movement. 2D projects are a massive fight.
@@fragileglass9622 I hear about this quite often. For more “indie” like 2D games Godot or Unity do seem an easier match. That doesn’t mean that Unreal can’t do it ofc!
If you really have been using Unity in a commercial setting for 8 years and have never had any issues, you are part of an extremely small minority group. On the other hand, feeling 'blocked' when using an engine that gives you the *entire* source code for everything inside, is simply a skill issue.
@@darak2 I guess it depends on the tech requirements of the commercial projects. I can think of many smaller project I’ve done is the past that had 0 issues. Only the more complex projects, and other platforms than win only that came with issues.
But I have to disagree with you on Unreal/source code access. Ofc it can be a skill issue, but in a serious commercial situation, it is all about time, not skill.
UE5.0: first time?
What do you mean?
@@The-Naked-Dev a lot of unreal's features in 5.0 were... laughably bad. World partition has been revamped almost entirely (the only thing that i can think of that *hasn't* changed much since 5.0 is how you assign an actor to a grid)
Nanite performance in 5.0 vs 5.5 is night and day, lumen is already being mixed in with nvidia's magical tech (megalights are, afaik, the result of nvidia's NVRTX branch efforts in coordination with epic)
the list goes on. and on.
Hmm, that feels like Unreal also doesn’t really go for an developer safe approach..
@@The-Naked-Dev you should watch ThreatInteractive's epic rants about nanite and TAA hahaha("Epic's Unreal Optimization Disaster | Why Nanite Tanks Performance!")
finally somebody said it!
Let’s hope that someone at Unity is listening!
So glad I finally made the decision to switch to Godot 4 a year ago.
Happy devving in Godot!
WE NEED MORE OPTION AS STOP PROCESS COMPILE COMPILED SCRIPTS OR NO NEED TO COMPILE , SEPERATE INPUT SETTING FOR EACH SCENE SEPERATE ,AND LIGHTING SETTING INSTEAD REPLACE SETTING WITH OTHER SETTING ,SAVE EACH INPUT SETTING SEPERATE INTO FILE , AND MORE AND MONO DEVELOP FOR SCRIPTTING AS UNITY 5 ,AND DISABLE PLUGINS IF NO NEED AND OR NEED TO USE OTHER PLUGINE(SWITCH BETWEEN PLUGINS THAT PROBLEM WITH OTHER PLUGIN)
Inproving editor workflow and speed would indeed be a very welcome upgrade…
@@The-Naked-Dev yeap
Unreal.., cough.... 😂 ill stay with the dark side they got cookies
Be gone demon! 😈
After spending the last 1.5 years on an Unreal Project. Be careful, coming from unity. You'll encounter a bad awakening.
Having a physics based renderer, but lumen not supporting realistic light intensities (like 1500 lux and more) but instead getting bloomed out of the universe if you try. But going too low gives you spots that make the screen become pitch black. Also it means that you can't have 1 light setup for fast iteration with lumen and when finishing things relying on light baking - you'll have to setup everything new for that, but than it's also not dynamic RTGI for higher end hardware.
Also Lumen completely relies on temporal AA techniques (or at least TSR), but those introduce a lot of ghosting.
Having a documentation that tells you "every mesh is improved by nanite" and in reality most meshes have WAY less overdraw using good LODs with nearly no visual impacts. fancy new features like nanite tesselation looking like a glitchy mess in like 80% of the scenarios.
Don't ever use "ChildActorComponents" they will mess up whatever you're trying to do, when you least expect it (latest example i encountered: breaking in a shipping build where suddenly their transforms where interpreted in WorldSpace instead of ParentSpace)
For the programmers: unreal enforcing stupid naming schemes even for local variables and parameters and their compiler will error when you have a member and a parameter named the same. The Actor->ActorComponent system is also a mess coming from Unitys super clean GO->Component separation.
It's also "just an engine" and it has some pros over unity but also some cons. don't get blended by marketing bla.
some lessons learned with unreal from the last 1.5 years: notion.so /Unreal-Engine-1b131f85e86d4f75b41945072fa6a0de
@@Soraphis91 im fully aware of its quirks etc., rendereres pros and cons, each engine has its own bits, at end of the day it's just a tool, i ain't willing to build my own so i choose which one felt best and has actual c++ support.
you could go on for days about things that arent perfect, missing, badly supported in any of the engines.
the rest is on the developer to know what their building and what they're going to utilize to achieve that goal. (nanite vs no nanite, to use lumen or not to etc, etc)
use what works for ya and stop blabbering.