Search Unity

  1. Megacity Metro Demo now available. Download now.
    Dismiss Notice
  2. Unity support for visionOS is now available. Learn more in our blog post.
    Dismiss Notice

Official Lightweight Render Pipeline is Evolving!

Discussion in 'Universal Render Pipeline' started by Tim-C, Jul 9, 2019.

  1. hugokostic

    hugokostic

    Joined:
    Sep 23, 2017
    Posts:
    81
    Yess done, with a little of pain, but if you want a good exemple of render feature, buy the L(ux)LWRP assets, its better than a simple example and it will be update, actually there is a URP preview package in it and it work fine.
    Some times you just have to kill the render features asset, which can bug after an upgrade and recreate one, at least render feature are base on a script so its easy to just replug it in a rendrer asset, same for some shader function in the Water system in the Verasl's Boat's attack demo who break after update but correctable.
     
    Opeth001 likes this.
  2. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    I just watch that, and while the guy is certainly brilliant in term of photorealism and optimization of shader, he is scared and clueless about Non photo realistic rendering, that's especially apparent when he is asked question about stuff that should be there in the first place.

    He doesn't seem understand that, like photo realistic light, we can build any style from a set of few principles and primitive (texture, color, shape, contrast, line), WHICH are all there in the set of instructions of a shader. SO his concern that it will impact performance isn't justified because the same concern could be leverage against ANY shader the shadergraph already does ... We just need to access to the light parameters!

    For example he say some people would prefer a -1 to 1 ndotl over a saturate one, which is weird because we can just applied teh saturate if we want later, and any post lighting process is equivalent to texture process anyway, which they provide, we just need the light data exposed!

    WORSE IS THAT UNITY JAPAN HAVE MANY VIDEO WITH EXTENSIVE ANALYSIS OF TRANSLATING NPR STYLE! yeah it's not about the manga style, shape language is what differ from other style, and that's geometry not lighting!

    Where is the cry in unity emoji? They need to talk to unity japan asap. Unity japan also gain from optimizing unity chan toon shader (which allow to go beyond toon in the good hand).



    Also I'm gonna diss how wrong his toon specular implementation is.
     
    Last edited: Oct 2, 2019
  3. larsbertram1

    larsbertram1

    Joined:
    Oct 7, 2008
    Posts:
    6,893
    think of hdrp and urp deferred: there you do not have any access to light... so shader graph just lets you create a surface description (next to some vertex animations) and will do the rest automatically.
    custom lighting always has been tricky and needed you to write it around the data the engine provided (_LightColor0, unity_4LightPosX0, ...). not trivial. so hiding this behind an hlsl door seems rather fair.
     
  4. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    Okay fair, up until I know about that exact issue I'll shut up :oops: Now it seems that ther render pass abstraction make more sense (in that they might try hide forward and deferred by unifying them behind common abstraction).

    edit: would passing a single direction as uniform to fake light per material be a reasonable anticipatory hack for deferred?

    Anyway that seems custom srp all the way down beyond their lighting solution. Best is a better pillaging API to transpose their library and features to a custom RP. It's seems structured enough so that we can "compile" to a custom render pipeline (declaration are all other the place right now, from a profane view at least).
     
    Last edited: Oct 2, 2019
  5. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    We don't even have vertex color in shadergraph oups ...

    Not so universal, I mean it's a very fine lightweight pipeline, but we need a middle ground too. Now it is can't do anything pipeline.

    They also clearly gave an answer about culling (we can only sort object by big monolithique tag they control, it's hardcoded) Culling is fine, I just want more fine control of the elements after (per group and not just layer) so I can optimize the sort.
    It's mostly damning for procedural game.
     
  6. larsbertram1

    larsbertram1

    Joined:
    Oct 7, 2008
    Posts:
    6,893
    Unity 2019.1., LWRP 5.16.1.:

    Bildschirmfoto 2019-10-02 um 11.09.37.png
     
  7. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    THat's what they said in the talk LMAO at the end when teh question was asked! I mean changing the vertex color, not just reading it.
     
  8. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    This is a neverending nightmare.
    They should have perfected the current render pipeline instead of making TWO new ones from scratch.
    It will take YEARS at Unity's pace for the current SRP to have the same flexibility as the built-in forward renderer, and by that time, it will be deprecated for some new shiny thing.
    I'm sticking with built-in forward and surface shaders...
     
  9. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    The only thing that built-in forward needed to be up to speed was to be able to render 8 lights in a single pass, and Valve already did it for their VR lab experiences.
     
  10. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    MY bet is that in a few years, ie at least 2 high end console machine generations, we will be doing non polygonal rendering so ...

    But URP I don't know, maybe as a library for you own SRP, which seems to be the way to go. I expect SRP to be easier and easier to build a high level integrated description ... Can't wait for AMPLIFY custom SRP editor.

    Better control on filtering/sorting and I'm happy. I don't mind Hlsl, as long I'm not blocked by shadergraph.

    Right now I'll be testing implementing stuff I describe above (rendertexture) on built in then port to srp to see what's in.
     
  11. AljoshaD

    AljoshaD

    Unity Technologies

    Joined:
    May 27, 2019
    Posts:
    216
    Hi all,
    Thanks for all the feedback on URP and its flexibility!
    We have posted a new article on the evolution of URP here. From the article:

    In 2020, we’re working towards feature parity with the Built-in renderer, including deferred renderer and camera stacking. You’ll be able to use the same set of features in your project, but not necessarily the same way.

    While developing the Scriptable Render Pipeline, we took the time to redesign our rendering system. Some previous features will remain deprecated as we have developed better alternatives. Some examples of features that changed in Universal Render Pipeline are the rendering callbacks, grabpass, and camera stacking.

    As some of you point out, one of the great things about the built-in renderer is the ability to easily extend it without needing to know the inner details. With URP, you have the source code and can modify but not everyone has the time or expertise to do this. The URP team has introduced the concept of renderers and renderer features to allow you to customize URP without actually forking the URP source code.

    The presentation has been posted here already but for completeness, here's Felipe's talk at Unite 2019:
     
    Last edited: Oct 3, 2019
    noio and elbows like this.
  12. LaireonGames

    LaireonGames

    Joined:
    Nov 16, 2013
    Posts:
    698
    *Sigh* I guess I will jump on the bandwagon first.

    So lets start with the good. The renderer system does sound great and looking forward to eventually using it.

    Now back to the usual story, I really hate Unitys marketing department.

    "if you want to develop once and deploy everywhere, this is the rendering technology to use"
    "Production-ready now"

    Production ready now and yet announcing a huge new feature.........
    The feature is of course full of experimental sub features........
    2020 is the year they will aim for feature parity.........

    If a product is still missing big features (like point light shadows) then its not production ready. Who knows what will change or break once they are added?

    If all of next year will be about adding more big features to get feature parity then it sounds like a year of instability to me cause again, who knows what will break/change once camera stacking, deferred rendering or missing feature X is added?

    The biggest thing that really annoys me about all this is the marketing, it wouldn't have taken much for me to be happy and on board. If built in was still being properly supported and the wording around URP etc was more honest then instead of complaining my attitude would probably be more along the lines of:

    "This is looking awesome, can't wait to use it, shame its taking so long to get there but they are adding SO much new stuff so I totally get it"

    All it would take to get there is to stop talking about SRP as though its done and production ready, this is a blatant lie and it feels like Unity is taking advantage of more beginner developers (probably most of their user base) who don't know any better.

    For the record I feel any software product can start to be considered production ready once you stop adding huge/core features to it. With URP that wont be the case until we have feature parity.
     
  13. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    It's funny because that's the article that prompt many of us to come here :p I we have been looking at the source code and current documentation. Lightweight was a neat tech package, focused and delivering on his promise very sharply. It wasn't built for flexibility, it was designed to pack a punch where it needed, and it delivered. It was design to have performant lighting, with narrow constraint, to deliver the most FOR compatible styles. It was HDRP lite, it does one thing very well, using it you knew what you would get, if you didn't want, you had alternative. Clear price, clear reward.

    Evolving to URP removed that sharpness in goal and the impact of the solution, let's see how that go. It's obvious that as such URP isn't mature and need to reconsider its legacy as a sharp solution, I hope that legacy don't actually create baggage for you while working to that new goal, which is the reason built in is seeing his decline as support.

    The new flexibility of URP on the pipeline is neat
    , but also do so on a design philosophy that is at odds with how certain effects are achieved, and we lost important tools in the process, which is sad and contrary to his new universal moniker. It's NOT MORE flexible, as much as flexible in a very different way. Right now, I'm expecting that instead of unifying stuff, we will see an explosion of SRP and a fragmentation of solution.

    URP will be a default for people who don't have particular visual ambition, and made everything more difficult for visual people, I'm not sure people presenting the project actually understand the stake or the problem (based on presentation and answer to audience, so far). Fun fact HDRP and LW/URP are surface like, they just don't want you to write to it ... that's tone deafness. They didn't break down how people THINK about their job to build a tool for it, they presented another of unity's silver bullet. Yeah it killed wereworld pretty quick, but not all monster are werewolf, it's useless against ghost and vampire ... :p
    Feature parity mean nothing if you don't understand the feature to begin with.

    I hope you the best though, I'm still processing to the impact of that shift. Still had many sleepless night to compile all the data, maybe I'm cranky and just need some rest. :rolleyes: I still have some reading and chart to do to get SRP under control.
     
    Last edited: Oct 3, 2019
    atomicjoe and JoNax97 like this.
  14. AljoshaD

    AljoshaD

    Unity Technologies

    Joined:
    May 27, 2019
    Posts:
    216
    Thanks for your feedback! The development team is aware of these concerns and a lot of effort is put into addressing them. I'm not from marketing but I think with production ready we wanted to signal that you can actually build and ship games using URP if its current feature set matches the needs of your project. I completely understand your concern with the impact of adding big features to URP. Once a package is out of preview, the goal is certainly to not break existing features when new features are added.
     
    LaireonGames likes this.
  15. LaireonGames

    LaireonGames

    Joined:
    Nov 16, 2013
    Posts:
    698
    Now this is getting closer to the line the marketing team should follow. If the line to the masses was like: Our core is stable, if you want the performance gains etc now you can ship with version X but be aware we are still adding new features.

    Rather than:
    Everything is awesome!! Look at how fast this is, upgrade NOW.

    I get that Unity wants users moving to the new versions so there is less of a trail of old tech needing support. I get that Unity wants to always be shown as always moving forward. I get that RP will be the future so everyone is keen and excited about the new tech. But it just feels deceitful to always show the glowing picture without being honest, it betrays my trust in Unity when you understand what is going on.

    Of course but this is exactly the point, you don't and can't possibly know that :) Pretty much no one can and the mad thing is that Unity already understands and follows this principle with the LTS versions.

    LTS versions are (or should be) production ready by design since they will only make small bug fixes. Once URP makes it to a stable LTS stream, then marketing can start shouting about production ready to their hearts content.
     
    sonofbryce and transat like this.
  16. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    They better not deprecate built-in renderers.
     
  17. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    MY concern is raised more when they people say thing like that "we want feature parity". Feature parity isn't really the goal really because:
    We don't want feature parity with built in, we want the ability to reach desired result by our own mean, and being able to evaluated cost benefit on the project basis by ourselves.

    I mean, the motto "performance by default" is very good! That is when I open unity and don't have precise goal, I should be rolling with a good starting point, ie the default. But as soon as we start needing specific feature, we shouldn't meet a wall, and understand that, as we stray from the default, we may lose performance.

    One example is fur shading:
    - Currently the easiest way to do it is through multiple pass, on the transparent queue, ie multiple drawcalls and provoke massive overdrawn, it's not performant, especially on tile base mobile.
    - People learnt that pass is discouraged and set out to be depreciated in the future ... they rant about pass being canned ...
    - but is pass really the problem or the fact that they want a simple solution to fur? I mean sometimes you just need it and are willing to sacrifice over aspects (nintendogs). It just solve many problem, like authoring fur, also despite being costly it work everywhere equally, unlike advance fur rendering method.
    - therefore Pass isn't really the problem, but ny depreciating it, you remove the single tools people reliably had to achieve a goal. What other unstate goal pass solve? Also what future innovation do you remove by depreciating it.

    It's similar about lighting, I rant more about the philosophy than the implementation for reason, it's the promise that thing will be evaluated and conclusions will be reached in a certain way, it's the root of misunderstanding. For example all doc point to a certain conception of shader and rendering pass, that conception will prevent the actual concern to be answered because, fundamentally, the concept needed to process them don't exist in the logic.

    That's why feature parity is concerning to me, it tells me "we have no idea, let's redo the old stuff they like". I'm basing my concern simply by evaluating what is presented to me with Motivations, Methodologies, Evaluations and Purposes, ie how is the form emerge from and implement underlying assumption.

    I say standard had a good philosophy, because when they added surfaces, they didn't remove the ability to do custom shader, they provided a "default alternative" to quickstart problem that come from practical reason, and it was great because it understood that light implementation and light resolution was kept different thing altogether, so they solved the implementation and left us with the resolution.

    LWRP was genius too, because it solve both, BECAUSE there was a trend of people trying to replicate advanced lighting on small hardware, it solve that SPECIFIC problem, and also why I refer to it as mini HDRP. URP is where everything collapse, it remove more than it extend, but that can be fixed by let it mature, sure, but more important it's not future proof, as in fundamentally it has an identity crisis. It was supposed to solve the implementation but over reached.

    Built in main problem was that it was the "default", with bad looking starting parameters, and it's inefficiency propagated to non technical project. Surface was great, because it was a tool to support people, not solving stuff for them. And that difference between SOLVING and SUPPORTING is the major philosophical break we are having with unity at large and URP in specific.

    I mean it's like the asset store, I don't use it anymore, I can't scan items fast enough, especially item I bought, and they take so much screen estate I can only see a very few at a time! I have 500+ items I won't start labelling them one by one. What were they trying to solve, what is the purpose of the current form, how did they evaluated problems, what methodology did they used to understand and solve that problem, what were their motivation for that change? It screams brand continuity, because evaluated against usability, I have question.

    Unity on the other hand, has muffle identity, it was democratization, but that's not much heard these days, it seems that with diversifying market reach they abandon that front facing slogan. A lot of game in their made with unity reel have artsy, cute, warm, personal and diverse style, while unity's demo are mostly drab, edgy, cold hard surface, which mean they solve visual problem for that target, which is at odds with what its audience seems to be targeting, what they select in their reel, and even at odd with their Disney and Marza collaboration. And there is a palpable sense of polite frustration when an internal artist get on stage and inevitably allude to gentle friction with the technical team, be it the Sherman demo or the heretic demo.

    The thing is that unity's is (seems) designing demo to prove itself to the industry at large, that they are part (perceive) of the elite graphic community, it's like an inferiority complex. And you can feel that with what get care and not, realistic surface get details 30 pages docs that have a sense of history of photorealism, and a sense of direction of where to go. You know and feel what they want to solve (shout out to the amazing Natalya Tatarchuk and Seb Lagarde).

    NPR shading meanwhile is like a 5s eyedrop (the botw like shader) that miss the point of the style, and provided very approximate solution, of something the unity's twitter hashtag have already solved through countless in depth discussion, so they aren't even paying attention (they might have seen it as a cursory glance), ie there is a genuine sense of distance and a lack of history and context.

    They would probably had a standing ovation if they showed you can do a temporally stable 3D that is perfect imitation of Mike Mignola, Mark Shultz, David Mack (watercolor) or Frank Frazetta (to stay in a "toon" like NPR), as it would show ambition and understanding, while keeping their internal cultural drabness. That is, it would be the Adam of NPR. They need a spiderverse moment, as in not replicating that style, but oozing that sense of genuine love for the art and technique, what define spider verse amazing look was just that, something that say that, yeah, you can do EVERYTHING with unity. But they don't have the Natalya and Sebastian of NPR.

    In some way, unity get compared a lot to unreal, because unreal has many good default (plop some asset rip, and hey presto thousand view on youtube with zeldo next gen vs unity asset flip cheapness), and a far end advance goal, with plenty of example to validate that. It has fault, but it's also well positioned as a brand, it's for advanced people, and advanced artist. They don't have silver bullets, but they have a gold plater, ie multiple implementation of the same thing, with imperfection laid bare for you to choose (you are advanced, deal with it like all advance people). They also had a clear clean slate when going from 3 to 4, with complete solution right of the bat. They were very resistant at the start about non realism art, but finally fully commit to it having huge success in that department (fortnite).
     
    sonofbryce and Ruberta like this.
  18. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    The main problem here is they are throwing away everything that came before them and redoing it from scratch without any attention to compatibility.
    All my concerns would instantly go away if URP was compatible with surface shaders.

    Just speak with @Aras on how to implement surface shaders on URP and make it compatible!

    I have a feeling that Unity is just making excuses and throwing PR canned answers waiting for people like me to just deal with it.
    Those are NOT growing pains, this is just throwing away what Unity was to make something different in it's place.
     
  19. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    They have to though, all that investor money wasn't free. With the influx of new market to cater, that don't pipe game dev lingo and culture, don't like to tinker, and want thing exactly as they use to, so they can do their 3D presentation hassle free on time at 14h, feel that they are serious not hippie like game dev, and pay good money for stuff game dev want free because they take it as granted.

    Of course it's going to be different, they don't even use that word democratization anymore. The brand at to be forward facing professional and safe, so that they are culturally accepted in these new market that don't know any better.

    We are niche now.

    I'm personally just going through the 5 stage of grief to let go asap and move on.
     
    sonofbryce likes this.
  20. elbows

    elbows

    Joined:
    Nov 28, 2009
    Posts:
    2,502
    This demonstrates one of my big problems with these sorts of complaints.

    Many people have a tendency to describe the evolution of Unity in grandiose, earth-shattering terms, but the whole thing turns out to be unhappiness over one or two features they need in particular, that they dont feel are getting sufficient attention. Well, if all your concerns can be eliminated by fixing one area of Unity, then that just demonstrates that the broader Unity doom narrative is overblown and inappropriate.

    Dont get me wrong, its important to speak up about specific weaknesses and failings in Unitys plans. But for years I have been really tired of some 'zero sum game' thinking, where some users blame a particular focus of Unity at that time for them not getting the features/fixes they really need.

    It happened when Unitys public-facing presentations and marketing were heavily focussed on VR, for example. Some people here used to bring that up and blame it for Unity failing to do something they desperately needed instead. Never mind that dev work done on VR can, in theory, lead to other performance optimisations that will benefit the wider game development community. Never mind that even if they hadnt focussed on VR, there was no guarantee they would focus on what you wanted instead.

    Well, I expect the same with all sorts of other stuff now that Unitys focus very much includes other industries use of realtime 3D, and the marketing focus on the term RT3D to describe what Unity offer.

    When, in fact, many of Unitys current initiatives will yield results for both the gaming and other sectors. DOTS performance-related stuff has non-game applications but certainly not to the exclusion of games.

    And looking at it from a different direction, the failings, omissions and unresolved bugs in various Unity systems have always been there, and if Unity had never focussed on anything other than games it may still have made no difference at all to any of this. There was certainly no end of issues back in the 'game only' days with neglected stuff that people could not believe Unity didnt improve sooner.

    Anyway I just used your comment as a starting point, pease dont think all of this is supposed to be directed at you, it isnt, its a response to other messages in this and other threads too.

    Personally my main complaint to Unity about the way the next gen of Unity is being made is mostly about perceptions. Its clear we are in the middle of a long, painful period of change where we have to go for some prolonged periods with some big gaps in what is available. I dont think that could have been avoided, but it could have been done in a way that allowed more developers to avoid the bleeding edge for longer, have a clearer sense of the destination, and a better sense of when the new generation of stuff would be have feature parity+ with the old ways.

    For example, in my head I think it helps to think of the modular, render pipeline & DOTS based Unity as a different product/toolset, that is still some way from being ready. One that we have access to for years of its development, rather than just at the last mile. If it had been marketed more this way, and they had retained more engineering efforts to improve the 'old Unity' for longer during this period, user perceptions might be better/clearer/less anxious.

    Because certainly in my mind, no matter what pain, what missteps, what dead ends happen along the way, the ultimate destination still looks excellent for both gaming and non-gaming applications. High quality, performant systems, better tools, far fewer black boxes.
     
    Last edited: Oct 5, 2019
  21. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    You kind of misconstrue the nuance of the criticism here, and it's definitely not about a single feature to be fixed, you single out a comment to illustrate your own feeling, falling in the same trap. You are basically saying "it doesn't matter to me, therefore it shouldn't matter to anyone, why are you disappointed?".

    Sure they are adding good stuff and lead some in a good direction, how does that help thing that get worse or plain don't work? And unity being (now) a big corp, who derive part of his market on the quality of the tools, has a stake into that, these criticism are signal to them.

    I'm for my part have many instance of thing being pull under my feet, making entire project fall apart, so the doom is not only real, it has already happen with financial repercussion. None of my issue so far as ever been resolved. SO I allow myself some grandiose statement (which I point in my own criticism, in the middle of more reasonable analysis) because that's a way to cope, and I would rather laugh at that than fall into despair. I can't even afford a new screen and unity start to feel cramped at 768p.

    My current project is entirely based on visual exploitation of shader, the change I'm seeing can threaten like 10 years of work. Of course I'm going to be scared.

    And i'm sorry the asset store especially is not all up to standard, especially where I need to filter paid asset in unity itself. I'm glad they didn't failed at the renewed unity's interface, it's flat but functional and sexy.
     
    sebastiansgames likes this.
  22. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    Lack of surface shader support is not just a missing feature, it's a huge regression over all other Unity versions that came before and a big middle finger to compatibility and flexibility.
    The problem here is the lack of continuity and compatibility with current development workflows. One that actually breaks 90% of the custom shaders made up until now and, what is worse, doesn't offer a comparable solution in terms of performance, flexibility and convenience (because that renderer feature of URP just adds passes over the current renderer, which is a big performance concern and hasn't half near the flexibility offered by current surface shaders)
    If breaking compatibility with everything without a way to migrate or even replicate current results and the shadow of imminent deprecation of current built-in renderers isn't a concern for you, I don't know what more to say.
     
    sonofbryce likes this.
  23. elbows

    elbows

    Joined:
    Nov 28, 2009
    Posts:
    2,502
    Well at no stage did I say these things were of no concern, or that those things in particular dont matter to me.

    I could construct a long list of asset store criticisms, but this isnt the thread for that. I dont have good things to say about it, and several times in the past on the forum I complained at the way Unity was handling the render pipelines from an asset store developers perspective.

    'breaking compatibility with everything' is too broad and sweeping a statement for me to touch on sensibly. Since I first read about the scriptable render pipelines years ago, I knew that shaders and things would not be easily convertible, and that there would be some losses in functionality here and there, not just gains. I already had a long time to cope with my concerns, and embrace the new ways (or stick with the old ways for projects where that makes more sense). And I certainly save my concerns about deprecation of the built in renderers for much nearer to the time when it actually happens - because (a) they acknowledged certain missing features from Universal and know that they need to get those working before they kill off the old ways, and (b) part of the pain of deprecation is lack of new features/improvements/bugfixes/general engineering efforts directed at the old system, but really we have been living with some version of this reality for some time already. Some bug fixes etc still, but overall the built in renderers are not the priority at all and have not been for a long time now, and we knew it.

    A lot of my stance and attitude on this is simply based on what I expect to happen, not what could or should have happened in the past. I expect some shortcomings of the new systems to be addressed, which will help, but it wont always be done in a way that satisfies everyone, nothing ever is or can be.
     
  24. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    So because this was coming for a long time we have to just deal with it?
    But at the same time it's too early to complain about things that haven't happened yet?
    You act as if everything had been said and done, when in fact NOW is the time to complain!
    I NEVER criticized LWRP or HDRP while they were just another option to the user.
    I honestly thought they would just implement surface shaders down the road, just as they did on Unity 5 with all the new physically based rendering shaders.
    Now that they make clear that they REMOVED options and force a worse implementation than before is when I have a problem with it.
    No amount of faster rendering will justify losing surface shaders. Specially when the "alternative" is to just draw more passes via renderers.
    Issues can and will be fixed if we say them out loud and make it clear.
    I'm not giving up on this.
     
  25. elbows

    elbows

    Joined:
    Nov 28, 2009
    Posts:
    2,502
    No, I did not say it is too early to complain, I said that I am not worrying about the deprecation of built in renderers until nearer the time that it actually happens. Mostly because by then it will be much clearer what is still missing from Universal.

    Clearly surface shaders not being part of the new way is whats bothering you, and I am not asking you to stop going on about that. I cannot judge how many other people are bothered by that omission in particular, if there are lots and it is technical feasible to solve it, then perhaps Unity will. If its just a few people repeating themselves on forums then that is no reason to change their plans at all. Time will tell eh, in the meantime I will leave you to it, since it is not productive for me to continue this discussion, we will just go round in circles.
     
  26. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    It's better to let them know now than later, anyway there is some sort of course correction already happening, so it's not just some small sample of people complaining on a forum. Also the complaints are plenty detailed, so it's not just doom spelling. Also the concern are diverse, but with a common core as the same root affect us both, so they have touch to a pretty fundamental aspect if different interest are equally affected. I kept cutting features to fit unity, and this one was a very big one, because i'm now relying on graphics to set the appeal of the project.
     
  27. Peter77

    Peter77

    QA Jesus

    Joined:
    Jun 12, 2013
    Posts:
    6,586
    I was looking into integrating UniversalRP into my project, basically replace built-in with URP.

    I've several custom shaders in my project and I wonder what's the better, perhaps more future proof way of converting those shaders to URP. It seems I've at least two options:
    • Use Shader Graph
    • Write shader code based on the shaders found in the URP package
    I'm a programmer, so writing code feels more natural and is easier for me. However, if implementing those shaders through Shader Graph has any advantages, I would probably go this route.

    Which of these options to do you think makes more sense? What do you think is less likely to break between Unity upgrades?
     
  28. larsbertram1

    larsbertram1

    Joined:
    Oct 7, 2008
    Posts:
    6,893
    shader graph.
    however adjusting HLSL shaders from LWRP to URP is not that difficult either (search and replace).
    this will change of course when deferred is added – as well as other features such as motion vectors.
    HLSL on the other hand will give you more freedom and lets you access all vertex to fragment interpolators.
     
    Peter77 and neoshaman like this.
  29. larsbertram1

    larsbertram1

    Joined:
    Oct 7, 2008
    Posts:
    6,893
    terrain normals are flipped (unity 2019.3.0b4, urp 7.1.2)
    (Case 1189337)

    terrainNormals.png
     
    Last edited: Oct 7, 2019
    florianalexandru05 likes this.
  30. Peter77

    Peter77

    QA Jesus

    Joined:
    Jun 12, 2013
    Posts:
    6,586
    Thanks for your answer! Then I go the ShaderGraph route.
     
  31. Voronoi

    Voronoi

    Joined:
    Jul 2, 2012
    Posts:
    580
    This is a good point that I think sometimes gets lost in the complaints over things not working with URP. It's certainly irritating to have stacked cameras taken away (with a promise they will come back at some point). OTOH, developing for VR exposed some glaring problems with the built-in renderer that was holding back performance. Most notable for VR work, but it would be true for everything.

    As much as I complain about URP, I recently wanted to explore what they are doing with DOTS and so I downloaded the project from this thread. On a 2015 Macbook, I am seeing 800K animated 2D sprites at 60fps, DSP 0.1% and 1 setpass call. That's in the Editor with Game and Scene view open. A standalone build would be much better I'm sure.

    That performance is just ridiculous and opens up all kinds of possibilities for games and non-games. A DOTS game could be wildly more complex and/or just not drain the battery so much. For my future project research, I found this PointCloud project making heavy use of DOTS and Burst. I've been working with medical research and the issue of accuracy vs. game-ready models has been a stumbling block for many years. With DOTS, I can now see a viable path towards visualizing highly accurate medical scans for XR purposes on any device. That's HUGE, and I think a direct result of this painful (but ultimately good) process!
     
  32. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    Yeah DOTS is cool, It's a takedown, no question asked here.
     
  33. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    ...as long as they don't remove monobehaviours.
    If they do, it will be a huge regression.
     
  34. neoshaman

    neoshaman

    Joined:
    Feb 11, 2011
    Posts:
    6,492
    That's true that LWRP was a takedown too until URP
     
  35. sebastiansgames

    sebastiansgames

    Joined:
    Mar 25, 2014
    Posts:
    114
    This all makes me sad because I fell in love with Unity (way back in V3) when it just worked. It was relatively easy to make games! Now it's hard to wrap your head around just opening a new project... what pipeline should I use? The features are different. The assets aren't compatible. I've gotten accustomed to relying heavily on camera stacking as a design tool over the years and now it's gone in the new pipelines (I know they're working on it -- they've been saying that for some time now). But the legacy pipeline doesn't support the new features (shadergraph etc). What a mess. I trace this back to when Helgason (a game developer) stepped down as CEO and was replaced by a guy who started at Clorox and Pepsi. My hope is that somehow Unity can find a way to get back to basics -- make game dev fun and easy (it's never easy but you get my meaning) and make it work. Then sure, some bells and whistles would be awesome. Right now this is all very confusing and is getting in the way of my projects, not enabling them.
     
    MaxPirat, Ghosthowl, chingwa and 6 others like this.
  36. Zephus

    Zephus

    Joined:
    May 25, 2015
    Posts:
    356
    I really wish I wouldn't say this, but I feel the exact same way. Unity has become such a confusing mess that I'm honestly thinking about switching to a different engine.

    I started using Unity because it was simple. Most of the time you had a very straight-forward way of doing things. What the hell happened? Now we have all these fragmented features, most of which don't even work properly. Or they only work in certain rendering pipelines. I just love how every new feature that gets introduced requires the SRP, which is definitely NOT ready for production yet, making Shader Graph, 2D Lights and so on completely useless for at least a few years. Great.

    And nobody at Unity can give a definitive answer on what to use for production. It's always just 'it's up to you'. No, I don't want it to be up to me. I'm using your engine because I wanted to cut all the crap and just make games, without thinking about rendering pipelines, entitity component systems and burst compilers. And now we have dozens of ways of doing things, and not one answer to what the recommended procedure for new games is. Everything's always just an alternative.

    Unity was so great because you didn't have to weigh all of these options against each other. It felt like a complete engine that just let you work on your games. Now I'm spending more time trying to get behind what the hell Unity wants me to use instead of actually using it. And don't even get me started on teaching people that are new to Unity all of this stuff. This has become way too convoluted and confusing. Unity has become a gigantic mess, and I have no idea why it seems to be so hard to understand for the developers how what they're doing is perceived by the users. I honestly don't.

    Rant over.
     
  37. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    INDEED.
    Let's not forget that's the same guy that pushed the renting model for unity "just as an alternative, nothing to be affraid of guys" and then removed the pay to own option.
    The same guy that pushed for DLCs and microtransactions in their games and made electronics arts the most hated company in the US.

    He's making a great work with Unity. o_O
     
  38. gamejen

    gamejen

    Joined:
    Sep 18, 2019
    Posts:
    3
    Unity will be offering IPO next year and seems they just want to show 'flashy' things and give the illusion to the unsuspecting public that they are 'revolutionary' and future proof and working on new tech. This is major snake oil tactics, Unity as it stands is an absolute mess and an embarrassment.

    Where are you unity? 26000 views on this post and counting. Any serious PC / Console developer will certainly choose Unreal Engine, not this shenanigans and convoluted development environment. The Features of HDRP are also years behind other competing products, Unity is always playing catchup but never seem to get there.

    Unity used to be good for Mobile development but even that is a joke now, nothing ' Really works ' everything is half-arsed.

    Look at the post processing issues thousands are suffering from, LWRP is now deprecated and PP2 is not working (PP2 in general is crap anyways, only a handful of effects that are actually wrong in their implementation )

    Here is an idea, get rid of the current CEO, hire someone like John Carmack :) Get rid of URP/ HDRP / Standard and do ONE, guess what, call it, Unity, to Unify.

    If the SRP is so great and customisable, why not ' Customise ' it to include the important features from each version into a single product, AND, do updates every three months without breaking everything! Just like …. Unreal Engine
     
  39. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    That's actually a very good question.
     
    MrPaparoz likes this.
  40. Ruberta

    Ruberta

    Joined:
    Mar 5, 2019
    Posts:
    114
    URP is good for support all platforms that Unity support. For the features, people tend to move to HDRP instead since it's feature-rich.

    Sadly, Unity currently doesn't have options to switch between pipeline. You gonna lose low end pc target and some platforms(Nintendo Switch and low end mobile) when using HDRP.
     
  41. suyashtnt

    suyashtnt

    Joined:
    Nov 26, 2018
    Posts:
    6
    they actually are planning this to be the default render pipeline, but since it does not have all the feature of the standard one, they cant do it yet
     
    Opeth001 likes this.
  42. sonofbryce

    sonofbryce

    Joined:
    May 19, 2009
    Posts:
    111
    Thanks for the great discussion, you guys! I'm glad I'm not alone in my confusion.

    I've recently just had to accept that for AT LEAST the next year of Unity is gonna be shaky and I don't intend to adapt to the SRP until the dust settles. I got frustrated trying to start a new project with it and it not being clear which features are supported. I had a "WTF??" moment when I realized camera stacking wasn't available (never saw that mentioned in any talks or blog posts). I don't want to get deep in a project to find out something else basic doesn't work. And I REALLY just want to use Shader Graph but it's impossible without transitioning.

    I was at te Unity Lhos Angeles Meetup earlier this week and others voiced similar concerns. This big headache really motivates me to start exploring other engines than try to sort through the mess of Unity right now. I hope they get their shiet together.
     
    neoshaman and mleodev like this.
  43. mleodev

    mleodev

    Joined:
    Mar 4, 2015
    Posts:
    7
    How frustrating this situation is, Unity is in a very messy state right now. The idea of learning other engine seems more and more necessary and urgent, sigh.
     
  44. liiir1985

    liiir1985

    Joined:
    Jul 30, 2014
    Posts:
    147
    For console or high end pc, maybe.
    But for mobile game, I'm afraid you have to stick with Unity, it's definitive the most suit engine for that for at least the next years
    We had a new project which uses UE4, and caused us far more issue and headache that what LWRP did. You'll even have problem to get it running on mobile devices at all, it's full of pain everywhere(Everything looks good and fancy in your editor, but once you build it to mobile device, everything changes. Many things just won't work on mobile, some are buged, once you got everything fixed, it looked sometimes even worse than what you'll get with Unity5), You'll need a far more experienced and talented team than what you'll need with unity.

    Oh,btw. you won't have problem with camera stacking in ue4. because it doesn't support it and screen space camera ui anyway. UE4 only has Overlay. so no 3D elements on ui, only render texture is possible
     
    Last edited: Oct 12, 2019
  45. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    You know something is terribly wrong with Unity when veterans and new comers alike share the frustration with SRPs
     
    nasos_333 and mleodev like this.
  46. liiir1985

    liiir1985

    Joined:
    Jul 30, 2014
    Posts:
    147
    Yeah, you are right.
    But I'm quite confident that unity will solve this problem within the 2020 cycle:), despite the issues mentioned above, srp and urp actually did a really good job, we are already enjoying the new features and possibility it provides(with some extra work we've done to the URP at source level)
     
  47. atomicjoe

    atomicjoe

    Joined:
    Apr 10, 2013
    Posts:
    1,869
    Could be, but I'm not that sure anymore.
    The whole surface shader deprecation along with incompatibility with all my current shaders is a big NO-NO for me.
    As the time passes by, I have the feeling that Unity has been kidnapped by corporate mentality.

    EDIT: and I say that with the biggest sorrow. :(
     
  48. Ruberta

    Ruberta

    Joined:
    Mar 5, 2019
    Posts:
    114
    I waited for Unity to be completed this whole year. What a waste. Now, I cannot wait any longer. So, I use whatever it has now. (Monobehaviour + HDRP) Who know? Would it be more waste next year?
     
  49. nasos_333

    nasos_333

    Joined:
    Feb 13, 2013
    Posts:
    13,269
    Anyone else have problem with quads with Lit material set to render both sides ?

    I cant get directional light normally, back side is treated as always looking away from light and is black

    Tried Unity 2019.1 and 2019.2 with latest LWRP pipelines

    This is huge issue as would essentially double all my vegetation and completely kill performance.

    Made a thread about it
    https://forum.unity.com/threads/dual-side-quad-is-not-receiving-directional-light.759875/
     
    Last edited: Oct 13, 2019
  50. nasos_333

    nasos_333

    Joined:
    Feb 13, 2013
    Posts:
    13,269
    But take our development years back, but this is not really a good option for everyone