Official What is next for us at Unity with Scriptable Render Pipelines

It would be nice if the API for SRP's was actually a usable option, I mean besides lights you can't get any information of about culled objects.. not even id's..

 

It would be nice if we had actual full documentation for SRP.
I sometimes feel like they fired all the people that made Unity in the first place and replaced them with random ppl.

 

“It’s not you, it’s me”

 

They didn't fire them, a lot just left.

 

Which is a pretty common occurrence when companies shift mindsets to a business one and care more about the bureaucracy than the actual work.

 

According to LinkedIn, Unity has about the same number of employees as Epic Games, even though Epic runs a B2C storefront, develops Fortnite, and publishes 3rd-party games.

 

That says a lot, with out having to say anything, one just needs to think about that for a minute and get it.

 

It's certainly been taking too long, this shader API should have been the focus before shader graph was. Now progress is being made, albeit slowly... Unity has a massively overgrown codebase, and for better or worse a wider focus than any other major engine, so with so many moving pieces, disparate teams and countless communication channels, throwing developers at features doesn't necessarily make development faster, in fact it will often make it slower. See The Mythical Man Month. It's the reality of any development, but not an excuse for hasty planning: Unity executives and/or directors (I've no idea who or in what position(s) exactly) some persons made some dubious choices early on to cut corners, put focus on the high level tools, pushing their devs to crunch out "all the things," in particular a barely usable Shader Graph on top a pile of sand (tech debt) instead of first building a powerful mid level shader API with the SRPs as a strong foundation to then build tools for rapid prototyping and artist workflows on top of.

This (as we know) backfired spectacularly with the Unity developer community at large rallying with negative sentiment around this and related SRP workflow pain points. I'm sure many Unity Tech developers actually doing the internal work and other mid-level Unity employees and managers themselves echoed the sentiment upwards to get traction and approve resourcing to make this right. It just seems to take forever to gain momentum for these initiatives at these huge companies, and this one had an especially late start. Anyway, hindsight is 2020 --we can't go back and "do it right the first time," so here we are, and the people who mote it be have finally gotten the message and greenlit this to be worked on beyond it being something "Under consideration" that was still up in the air.

The "Surface Shader" like API and Shader API improvements were added to the Roadmaps ~3 weeks ago, by our champion "on the inside" @ali_mohebali. At the same time, Ali responded in the forum post linked above by LaireonGames.
As for where this lives on the roadmap, see the "Planned" sections under the URP and HDRP tabs for "Surface Shaders", and "In Progress" section in URP tab for "Shader Library and API Documentation Improvements"
Main Rendering roadmap page: https://unity.com/roadmap/unity-platform/rendering-visual-effects
Direct links to roadmap items: https://portal.productboard.com/8uf...er-library-and-api-documentation-improvements
https://portal.productboard.com/8ufdwj59ehtmsvxenjumxo82/c/621-surface-shaders
https://portal.productboard.com/8ufdwj59ehtmsvxenjumxo82/c/628-surface-shaders

I encourage everyone to "upvote" these items on the roadmap by selecting one these buttons here:


Also linking to this excellent document on "How to evaluate a new Surface Shader system design" written by our champion "on the outside" @jbooth in response to the roadmap update. This document covers many key points that I couldn't have said better myself: https://twitter.com/slipster216/status/1408859016047046656

Stay vocal and hopeful folks.

 

AMEN!

Seriously, make this.
This way, even if Unity Tech screw us again in the future, we'll be able to fix it ourselves instead of having to beg for years in the forums.

 

I don't think there is much to be hopeful about lately.

Unity has tried to do everything and as a result all it really works for is casual/mobile games. Unreal is literally defining the next generation of graphics, meanwhile Unity is still playing catchup to the previous generation.

 

I don't disagree, but this sentiment is kinda sad. We want Unity to give us features that will save us from Unity...

 

Some of Jason's recent thoughts:
@JasonBooth

https://docs.google.com/document/d/1S0n5vlCUG1zv4pvr7V2gxyvWbpsd-7CbOgOEkq96J5w/edit

 

We have to keep it real, yes.

 

better shader by jbooth

 

I know. I bought it, of course. But that's a bandaid, we need official support inside the engine. (and it's on the way)

 

I've been using Better Shaders for a few months now and have shipped a few assets with it. The way it works is incredibly nice, and the stackable feature is a game changer. However, I can't expect users to buy an $80 asset just so they can make use of this stackable functionality, or even to be able to modify the shaders.

I wish Unity would buy out Better Shaders and integrate it themselves, because honestly I have 0 confidence they would make anything as nice as it. Even if they copied the design exactly, I envision that they would omit support for BIRP or something equally frustrating. In a perfect world they would then rewrite Shader Graph to emit Better Shaders code, so everyone can live in harmony and make great games, but I just don't see this happening unfortunately.

 

While we don't know what the plan is exactly for this new Shader API beyond the broad strokes on the roadmap, it is a critical time for feedback on shaping development for this API, so please, everyone flood the roadmap "how important is this to you?" feature with this type of feedback, critical deal breakers especially. The product managers do actually read what people type into the roadmap feedback feature as 1. It's newer and not as many people know about it 2. It gets neatly organized by product and importance, unlike random forum posts that are just another drop in the ocean of noise on the forums. Not that Unity staff don't read or respond on the forums it's just hit or miss, as you likely know.

Having said that, something we do know is that Shader Graph support for Built-in Render Pipeline / BRP / BiRP was quietly added in Unity 2021.2, currently in beta. By quietly I mean it isn't mentioned specifically in the Unity Blog post, which only mentions the somewhat related BRP -> URP conversion workflows:

• A new converter framework from Built-in Render Pipeline to URP makes the upgrade tooling more robust and supports more than material conversion.

but it is mentioned in the Release Notes.

• Shadergraph: Added: New target for the built-in render pipeline, including Lit and Unlit sub-targets.

I don't think anyone was expecting this, but it's certainly a step in the right direction. And we need to make it known that these are the types of cross render pipeline features and support we expect going forward.

 

We don't need a step in the right direction, we need all of them, there has been so many steps in the right direction undone by countless steps in other directions.

 

I don't use URP myself, but I have checked some benchmarks versus Built-in in youtube (recent ones) and it's actually A LOT slower than built-in for the same or worse visuals on Android. The moment Unity removes the Built-in pipeline, I'm done.

 

Well this is the big fear, right? Unity has a long history of getting things to the 80% mark and then essentially abandoning them before they are really usable - no new features, very slow fixes if any. It just kinda dies before it reaches its actual potential. The original surface shader system, as I understand it, was entirely Aras's doing, and he followed through on it adding the things people needed because he cared about the system and had control over his own schedule to a large degree.

When I was in charge of the engine design at Harmonix, we did a cycle of 3 two week sprints working on features of the engine. Each feature had an owner responsible for it, always - and if that person left to work on a project, those features would be reassigned to have a new owner (with appropriate ramp up/hand off/etc). You could see who owned what code at the top of every file, and go right to them for questions or feedback.

Every 4th sprint was two weeks where owners got to work on what they thought their systems needed the most. They'd often have a huge backlog of things they wanted to do, but that the schedule would constantly deprioritize for reasons. But these two weeks were always some of the most productive for the quality of our tools, and looked forward to by every developer working on them. Scheduling and committees have a way of deprioritizing things which are not QUANTIFIABLY important, but can have massive ramifications on the actual quality of the experience. It's important that a company, as it attempts to quantify and control everything it's doing, leave significant time and space for the smart people they hire to operate on their instincts as craftsmen, and to pour their passions into what they are doing, and have ownership over the features they work on.

As for Better Shaders, I think it very much represents my approach on solving the problems Aras tried to solve a while back, with newer problems like the integrations between different code assets, and more modular/testable reuse of code, both solved via stackables. However, I'd like to see a similar leap for lighting customization, which is fundamentally ham stringed by having to reverse engineer all the shader code every time Unity changes anything in each SRP.

Done properly, users should be able to write shader bits in whatever interface they want (graph, text, both), easily integrate code from 3rd party assets (Bakery, Curved World, etc), add common effects to existing shaders (snow, weather, etc), and apply custom lighting models, all without changing a line of source. Better Shaders has proven that all of that is possible, and the production gains from it, and new markets it would create, are quite promising.

 

Since Aras stepped down, there has been no progress on my games which are typically custom graphics heavy. In fact I had a lot of pain and it burned me out completely. That's probably because of the bizarre monolithic shaders in URP and HDRP with a shadergraph that was so crippled, I had to write a utility just to patch the resulting shader time and time again. Then Better Shaders arrived and I realised then the question wasn't about me buying Better Shaders vs waiting for Unity to do their version but simply just leaving Unity for something that just stopped just lying to me and drip feeding me even more problems.

Why do we have to fight and chase Unity? Why do we have to go to a special to-do list for product managers? Why aren't the product managers intelligent enough to know regardless, what we need? It's not rocket science for any senior game dev in this industry to quickly pull apart how bad the whole thing is and figure the right moves.

If politics are causing that then it's time for me to look for a new government.

 

I relate COMPLETELY.
I have been a paying "pro" user since Unity 3.
Now, I'll finish my current game with Unity and once it's done, I'll have to very seriously check other engines.
If Unity can't provide me with the user-friendliness and retro-compatibility it once had, I have no incentives to stay.
If I have to rewrite all my shaders, maybe I'll do it in Unreal instead.

 

@jbooth Would you even work at Unity overhauling how rendering is done as the new Aras, or would this just bore you? Do you have a price? Did Unity even contact you for it?

If not then it's Unity's loss since I don't see many others doing anywhere close to the problem solving you are.

 

Its always the problem with publicly traded companies, they get so eager to show a profit and impress shareholders that profit becomes Everything at the expense of anything else.

Its just a phase.

Mind you, does any publicly traded company ever come out of that phase.....

 

It was more or less the same situation before they became a public company.

 

Agreed but I suspect a part of that was anticipating going public and getting ready for it.

It explains a lot of their decisions for me, like posting about the user stories about how Unity is so great its being used in the robotics industry. What the hell has that got to do with making games?

Or buying a company that owns software around importing very high detail auto CAD models....

Or their posting about how Unity now has new tools to support cinematics and films industry....

Seems like a manager decision of "find new ways to sell licenses wherever you can" without realising that you have now become a jack of all trades and a master of none at all.

 

Unity are *literally telling people not to do this* - I received an email this week from Unity support telling me that Unity will "only" accept input via the forums, nowhere else.

Why aren't people using that feedback place? BECAUSE UNITY SAYS NOT TO.

... it's very very hard to take seriously a team at Unity who isn't even trusted by their own colleagues, and who apparently is so ignored in the org that their preferred feedback system is forbidden.

 

By Occam's Razor: they don't want feedback from users, they want to build whatever it is *they* want to build, whatever is convenient/interesting/fun for them, not for the users.

That can work great if you're personally the target audience. Are they, though? They don't seem to be.

 

What the hell??

 

Unity has not reached out for any such things - only a meeting well over a year ago (when I started #SRPLife) where they made lots of hand wavy gestures about why Surface Shaders 2.0 was impossible ("Mesh Shaders!", "We might change shading languages!"), made promises about getting me some info/plan and responding to any questions I had, then proceeded to do absolutely none of it and ignore my follow up emails. It didn't exactly fill me with confidence that the person I met with was being genuine.

I'm pretty sure they view me as an adversary instead of someone trying to fix the issues. I've basically designed the system for them, provided a sample implementation, then published a sample implementation as an asset (which will never recoup the time I've spent on it). But you know, shoot the messenger is easier and all.

 

Sounds strange and almost unbelievable that a leading tech company would behave such toward someone who has contributed years to Unity including onstage at Unite with a rock solid history of industry experience.

 

I'm loud and direct, say exactly what I think, and will go around normal procedure if those procedures do not work. That can be extremely unpopular, especially if people are more concerned with image or management than the actual output. I've been the loudest voice on this issue for the longest, so it's not surprising to me at all- I've played this role more than a few times before, and I'm happy to play it if it gets attention brought to the issue.

 

That's a very poor support experience --if this is what you are being told, the support representative *literally* has incorrect and/or outdated information. Can you post an excerpt or screenshot of the email with the support rep's name so we can make sure they are retrained? Which is a problem in itself if even one support rep is not properly trained or informed. In any case, Product managers like @ali_mohebali have made it very clear on the Unity platform roadmap portal itself (that officially launched a few months ago) and on the forums that the roadmap is to be used as an official channel for submitting feedback (in addition to the forums, unity pulse surveys, bug reports, etc) This is coming directly from the product teams that are asking for the feedback, not some support representative that by my estimation is feeding you some line (that may have been true in the past, before these new platform roadmaps either existed or officially launched.)

This is the current and official word from "Unity" on providing feedback via the Unity Platform Roadmap: https://unity.com/roadmap/unity-platform/rendering-visual-effects

 

That has been the official word from Unity (paraphrased in various ways over the years) for a while now and it has always been just words.

 

Have any of you been in regular, direct contact with Unity's Success team?

Just wondering, as their role seem to be to bypass the usual channels and report directly to higher-management client issues and priorities or go straight to the development team. I might be mistaken and naïve but it sound like the best channel for all your grievances.

Only had positives experiences so far with them but I am not dealing with such high level of problems as you are.

 

I'm just clarifying what the official word is on providing feedback via the unity platform roadmap is --that specifically and not anything else. So that people will channel their critical feedback into this as well and make it louder from all angles until it gets through to more people at Unity that can do something about it. So do you wanna help amplify the message?

Anyone that wants to help and is able to, please articulate critical feedback and send it en mass and in force to Unity Tech via every channel available to us.

That's my intention here in this post and then in the most recent post i'm just correcting bad info that came indirectly from some Unity support rep that might keep people from actually using the feedback feature of the Roadmap, and limit the message from being as loud as it can possibly be.

 

I do not wish to amplify any message and I do not wish to give Unity any more feedback on anything.

There has been enough feedback and they haven't been acting on it, I'm not sure why you think this time it will be different.

We are discussing this in a thread that is all about them telling us they have heard our feedback and they are going to act on it... a year ago. But we should give them more feedback now, because... What?

 

Because It's very helpful to provide feedback to shape development while it's happening to make sure we get something that works for all of us. Maybe in an ideal world that would be nice for it magically to just materialize and Unity goes out of their way to practically read our minds and deliver exactly what we all want, but most of us know by now Unity can't develop something in a vacuum without constant community feedback and reminders throughout the entire process. There's never enough feedback. There should be, but there isn't. That's why people like @jbooth and @rz_0lento, and many, many others are continually providing feedback as features develop and the teams and priorities at Unity move and shift around like they do at organizations of this scale. But if you don't want to be involved in the development process or don't have time or energy to do so, and just wanna take whatever Unity comes up with according to feedback from a year ago and whoever else is available to provide feedback, that's understandable, too. Or if you have some exit strategy after your current project to use some other engine or framework instead that targets your use case better, by all means go and do that. whatever works.

We're all frustrated about this and know it's taking painfully long, we've been promised things and haven't gotten them from Unity or not in the timeframe we were told, I feel you. I just assume if you are still posting here in the forums now and again you actually care and have some stake in Unity being better and probably have a bit spare time to articulate some feedback and post it in the Roadmap portal about how the Shader API should be, even if that's just saying "make it like BetterShaders, hire / consult @jbooth because I can actually trust his judgement on this" or similar. if you didn't have valuable feedback to contribute and something to gain from your workflows with Unity improving, you (likely) wouldn't be here. But idk, maybe you just need to vent. I don't blame anyone for being at that point about it after the ride we've been taken for here.

 

Since feedback is so important, here’s my feedback to you: I do find your posts patronizing.

 

Here's the problem really. In Unreal and other tech places I've been privy to, they assign one dev per feature. That's immutable. That's the sole job of that dev, to keep that feature pristine, optimised and up to date with current needs.

That just does not happen at Unity for a lot of Unity's features. You just said that yourself. "...teams and priorities at Unity move and shift..."

But see, that doesn't happen like you claim it happens elsewhere. That's at Unity and maybe wherever you worked before. But for big engines, no. That doesn't happen. With very, very good reason.

And then you call for us to keep on shouting until we are blue? Really, because you're not managed well enough to hear the first, second, third time?

 

I did, and said to them everything I said here. For more than 2 years in a row.
If other paying users also complained to them, maybe that's why they changed their mind on Surface Shaders 2.0 or maybe not. I'll never know.

We are all very much burnt out about all of this.
We have said this things so many times that it's literally impossible for Unity managers to don't know already.

I'll say it loud and clear: all we are experiencing right now started when Unity hired John Riccitiello as their CEO.
- The shift to a subscription model instead of permanent licenses
- The loss of focus on game development
- The default features converted to packages
- The half baked scriptable render pipelines
- The push to abandon the built-in pipeline

Sadly, all of this is not happening by accident. All of this is the result of a specific type of management that puts in charge people that don't understand what it's actually to make a game.
I don't think this will change in the future while the directive board stays the same.
Sadly, again, I don't think it will happen, since Unity Tech has an incentive to don't fix things, given it pushes its users to buy add-ons in the asset store for features that should be there by default, making great profits of otherwise non-paying customers.
This is the same technique as free-to-play games: give the game for free and monetize the hell out of it.
This is giving Unity Tech great profits... IN THE SHORT TO MEDIUM TERM.
In the long term, the company will lose it's relevance and just disappear.
But the directive board doesn't care about that, they only care for profit right now.
So, I'm afraid this will never get really fixed. They will only apply band-aids to a bleeding wound.

Unity is now the shell of the company that it once was.
The corporate suits have taken control. And as usual, they rot everything.

 

That's a good point! There's certainly been ownership and accountability issues at Unity, above and beyond the "average" fluctuations at other similarly positioned companies. In the last ~3 years especially. Of course another huge factor (many of us are aware of EDIT: and @atomicjoe just broke down with extra detail above) to highlight is since that Unity doesn't have any real "dog-fooding" like say Epic and other companies who makes the engine for themselves to be able actually use it and deliver a project in-house using their own tools. So that also contributes to why Unity requires an unreasonable amount of community feedback. I don't like it, I don't wish it was this way, and I'm not defending Unity's numerous failures and shortcomings. And for the above reasons you and I have both shared, I don't think we should have to be providing this steady and strong stream of feedback to make sure things stay on track. But here we are.

So I'm just doing the best I can to assess the situation and highlight happenings and pathways that could help it, such as the the roadmap portal that didn't even exist a few months ago. And while I'm strongly suggesting to anyone that reads this to provide critical feedback about the Shader API, and SRP improvements by clicking the "Critical" button on the Roadmap item and optionally sharing details. In addition to being vocal in the forums, and other feedback channels, officially or not. Let me clarify I'm not saying anyone *should* be or needs to be giving regular feedback, we all have different situations, roles, responsibilities, and I cast no judgement on whether anyone does or not, nor do I have a bias against anyone that wants or chooses to "jump ship" and use another engine or framework that is able to better fit their needs. I just want to support projects and teams to succeed by whatever means possible. So here I am posting links and screenshots for the Roadmap items for the SRP "Surface Shaders" API to try and help spread awareness for anyone in the greater community that might come across this thread and didn't know what the latest was or what they could do to show support beyond posting in this thread. That's all folks.

 

It's quite a state of affairs when even the moderator is laying into Unity here.

I don't have anything substantial to say that's not already been said, other than to register my frustration along with everyone else. I've spent the day seriously evaluating moving my project over to Unreal and I think that there's a good chance that I'm going to do it. Both here in the graphics department and throughout the engine Unity is just a constant battle against letdown after letdown. I'm actively looking for ways to get it out of my life.

This is a shame, because Unity is the engine that most closely matches my ideal design philosophy of being relatively low level and unopinionated, and I believe that there's a version of the Unity idea that has the potential for excellence. Sadly the execution leaves everything to be desired.

 

built-in has been completely stagnant, with the aim on URP to match it's set of features.
Yet, years later, only in 2021 does it even get _close_ while all other engines have been making major improvements and updates in said years. To top it off, many of it's features are lower or matched in quality.

It blows my mind. If i can get a position at management, with any big company -- man will I have a good time. It seems that quality doesn't matter, hit the right shoulders and you'll be just fine.

 

Either take me back to 2015 or forward to 2025.
I can't take this no more

 

Hahaha!

It's not all doom and gloom. Much good is coming from the louder and clearer feedback people are giving. Unity takes none of this personally (if they do, then they need to stop, take time off and eat a banana).

 

Having been here since the start of Unity 5 .. when an " exciting " feature is announced then takes Years .. to get done, and even then, not right, it takes a toll. ALL of those built up " letdown after letdown " a person can only take so much. Some are more forgiving then others. I can see someone who is/ was 18 .and has years ahead of their life, then yeah, broken /delayed promises do not hit as hard. I agree one person should be responsible for a feature. How does swapping out people help development at all.. " to many cooks " are never helpful.

The last kick in the gut for me was BOLT. I still can not see how a company could buy that and NOT realize that out putting C#, with what they had in mind would not be possible. IF* they did not know, then WOW... IF* they did know then WOW. One good asset that could have helped > 60% ? of users to do better and actually make a game.. all lost. Bolt was made by one person ..and look how fast development went.

Talking about Unity listening to its user base... how many people remember how many forums and places people were complaining / begging for the Dark theme in part because of eye strain. How many years....?
Even browsers now have the capability to use a dark theme if you want.

My only other real point is if things have not changed with UE4, if you make a game with UE4 and do publish it.. you have to get permission to use their splash screen ...think about what that means and means to UE4 company, and why they would want it that way. I don't think I have to write about all the implications about Unity making you pay to take out the splash screen... res ipsa loquitur

EDIT: " This is a shame, because Unity is the engine that most closely matches my ideal design philosophy of being relatively low level and unopinionated, and I believe that there's a version of the Unity idea that has the potential for excellence. Sadly the execution leaves everything to be desired"

I feel the exact same way. That is why I felt to reply to this post.

 

When I first started working in game development, I used inhouse tools for years, and not the nice ones, the S***ty ones where you had to rebuild the game to see some color changes for particles, it took probably 10 steps to import a model in the game, if all vertices had the same vertex color the model failed to import so one of them must be a different color, all environments made in 3dsmax exclusively, the editor was only usable to look at it, everything was broken all over the place, hot reload? what's that? you had to rebuild the game for any change.

At that point, I discovered a nice tool called Unity, v4.6 I think, and I was mesmerized. You could drag and drop assets, I discovered the Strumpy shader editor, particle systems, I could see the lighting in real-time in the editor, wow! In the meantime, I discovered the asset store and I said, hmm, I could try to make a water shader. I started using Amplify and I made it. Guess what, it works to this day!

Then SRPs came, I said OMG this is huge, you will be able to extend the rendering, incredible. Since most features used to be good in unity, I was expecting the same quality for SRPs. Hmm, something small that bothered me from the beginning, was the inconsistency for the most basic shader features: _BaseMap in URP / _BaseColorMap in HDRP, Metallic Smoothness Ao as separate maps in URP / Mask Map in HDRP. So how am I supposed to move from one to another when (if) they add cross pipeline compatibility? Reauthor all my materials and textures? Shaders are one part of SRPs, what about the other features?

So now that we have them and I'm trying to support them for my assets, I work in Birp 2019.4 / URP 2019.4 / HDRP 2019.4 / URP 2020.2 / HDRP 2020.2 and 2 more are coming, luckily they are in beta so I have some time to breathe. Guess what can go wrong? I posted this before, but I'll put it again, for effect:



Going back to the first paragraph, I just wrote it to tell how I feel working with unity today. Could be me, but I guess having a tread with 8 pages of mostly complaints, says a lot :| I wonder how the SRPs devs feel about working in 3 versions of SRPs (bug fixes for 7x, bug fixes for 10x, new features for 12x).

That being said, I'll have a beer and continue the journey, what doesn't kill you, makes you stronger. My fingers have muscles already from clicking 5 (7 with SRP 12) times more thanks to SRPs.

Buy my assets, they are good, and have SRP support ))

 

All I wanna do is screw in this goddamned lightbulb but the person holding the ladder keeps moving it around.

Is it at all possible to render a shader replacement pass in URP? Yes yes they have that Render Objects thingy but it replaces the whole damned material, textures and all. How much boilerplate code is it going to take to remake the equivalent of Camera.RenderWithShader?

Followup: I managed it by using my own custom LightMode Tag with a Render Object, and then I had to add that custom LightMode tag to the shader and since the Shader Graph doesn't allow for editing Lightmode tags, I had to do that via Amplify or within a handwritten custom shader. So yes, Camera.RenderWithShader replacement is semi-possible via LightMode-tagged with a RenderObjects pass in conjunction with custom lightmode tags.

This is like taking a sip of water from a fire hydrant.

 

I’m not convinced this isn’t the start of the end for Unity.