Current 2D Renderer's missing core features

Well, I thought 2D Team would be focusing on much needed features, instead they released another demo for marketing. URP 12 is almost there and it seems there aren't many difference between v11 which to my versioning isn't even v1.1.
Almost all Unity products lately is prone to waiting for developers. There is no room for anything.

Edit: And it seems almost only renderer developer out there is Chris_Chu.

 

It has been almost 5 months... 5 MONTHS! since I started this thread, but still no news about Emissive channel.
I was talking about the emissive channel long before this, so it is probably being closed to 1 year for asking about emissive channel.

I mean, come on... how can you possibly make any game practically without emissive channel for any rendering pipeline.. Not sure how difficult this could possibly be, but it can't be that difficult can it? We have secondary texture feature already, we should be able to just insert emissive add operation after rendering is done.

1. Gather emissive rendering texture by pretty much doing the same thing like how mask or normal map buffer is prepared.

2. Just blit that over after the lighting phase.?

 

For me it's clear Unity company is not dedicating enough resources to the team dedicated to this part of the engine, they just don't care (I'm trying not to be foulmouthed). It's like there are just 2 or 3 devs in the department, doing as much as they can, unable to attend to all the people, like us, that ask for features / fixing bugs.

 

Indeed, feels like butter has been spread too thin. We are not even asking for little, small, optional stuff. Emissive Channel is ... like wow, it has to be there even before "Shadow" in my opinion...

 

It's because they have too much extra "bloat" in the app and it all has to be maintained. Unity over the years has really gotten absurd with giving people 20 different ways to do one thing. All in the hopes of appealing to EVERY potential customer in the market and bringing in more revenue. There really is no reason for something like this (or the new input system for that matter) to be in the state they are in and be considered v1.0 releases. SMH. It is really frustrating.

 

@Chris_Chu , _CameraSortingLayerTexture isn't quite working. The result is the capture of the Normal map buffer.
Can you confirm please? Using Unity 2021.1.7f1

 

Well, I've spent a pair of days on diving into the URP source code and implementing some 2D Renderer features I've asked for so far:

• Emissive color.
• Blending mode.
• Falloff offset (it was removed in v2021, I brought it back).
• Custom post-process FX.
• Limiting the Bloom FX to emissive / unlit pixels only.
  
• Multiple render targets.

I feel relieved to not depend on Unity's development schedule anymore, putting bugs aside.
If any of you is as tired of waiting as me or need any of these features urgently, I can lend you a hand.

Btw I'm not using Shadergraph anymore either.

 

@ThundThund How do you get Unity to not overwrite your changes to a package?

Also, is not using Shadergraph a consequence of your changes? Or merely incidental?

 

This is exactly why I am not very comfortable modifying the package source. If things are more like "plugin" then I would attempt it myself as well. Having said above, ThundThund managed to add all that in just few days of work... And it beats me to understand why Unity takes so long.

 

You need to download the source code from Github and add the folders of the packages as a disk directory in the Package Manager.

I could keep using Shadergraph for shaders that don't need the new features, but I don't want to, unless I need to do quick tests when developing new shaders.

 

Looks like VFX Graph support has been partially added in 2021.2 beta: https://unity3d.com/unity/beta/2021.2.0b1

And...

Seems like that means 2D Renderer is now no longer experimental.

 

I still say without proper support for Emissive channel and tilemap shadow support, 2DRenderer is still not production ready.

 

I'm a bit confused on the current state of `_CameraSortingLayerTexture`, if anyone could clear things up.

1. My use case is a post processing effect, while using the 2d renderer in URP.
2. Specifically, I want to create a shader that gets the current color of the screen
3. In some of the posts on the 1st page you guys said you got it to work, but for me I seem to always get "white".
4. I created a shader graph with alpha always 1, color = SampleTexture2D
4.1. Texture: parameter of type 'Texture2D', name '_CameraSortingLayerTexture', Reference '_CameraSortingLayerTexture'
4.2. UV: Scren Position with mode default, direct connection
5. In Project Settings, Quality -> I have my Scriptable Render Pipeline Settings referencing a custom 2d renderer.
6. The custom 2d renderer references custom 2d renderer data
7. the 2d renderer data has:
7.1. Post-Processing enabled: checked
7.2 Post-Processing Data: built-in data that was automatically assigned after enabled
7.3 Camera Sorting Layers Texture Bound: Default
7.4 Camera Sorting Layers Texture Downsampling Method: None

I'm currently on 2021.1.16f1.

am I missing something, or is it just not supported at the moment?

p.s.
I'm also a bit confused on the sorting layers bound. specifically, how do I specify which layer is set to a specific sprite/mesh/material ?

 

Is the input texture exposed (checkbox)?

 

I tried with and without, doesn't matter.

 

2DRenderer need a RoadMap!!!

 

Well, we do have this thing which Unity has referred to as its 'roadmap'.

Feel free to submit your 2D Renderer ideas!

 

I submitted 5 or 6 ideas moths ago, but I don't know where they are, it's not displayed anywhere and I didn't receive any notification.

 

A submitted idea doesn't go anywhere on the roadmap until the unity team has put the idea 'under consideration'.

Which, frankly, isn't great design.

It would be nice if submitting an idea put it into a "submitted ideas section" where any user could vote on the importance of said ideas so that way each user didn't need to manually submit the same idea over and over — but I suppose doing it the way unity did could cause an increase in idea quality and may provide a better gauge in how wanted the feature is (internally speaking).

But yeah... still not great design.

Regardless —

Hmm, that's odd. Did you receive a confirmation email when you submitted your idea?

After you receive a confirmation email (and confirm your idea via that email) you can view that (and your other) submitted ideas by clicking on the "submit idea" button again.

Edit: Actually that last part doesn't seem to be the case — I don't seem to be able to see my submitted ideas anymore.

 

Was not the purpose of using packages was to avoid releasing at the same time with the editor release? And now the editor release is only twice a year, with one or two engineered, no wonder why the development is so slow. I actually suspected this was happening... when I saw release of URP12 2Drenderer with features almost the same as the URP 11 version.

Being able to extend the 2DRenderer is a great idea, and it will certainly help, but for things like emission channel, I suspect it has to change the core rendering code. I feel your pain, it is a lot of work to handle all by yourself. But I also just feel how much Unity cares about 2DRenderer. It is currently not a priority for Unity.

Anyway, thanks for the insight. Unfortunately, if this is the pace, then I think I will have to move away from using 2DRenderer for our current project. I don't think emission channel will be added to URP 12 2DRenderer, not at this pace, when it has not made to it to the current beta.

I think given the circumstances, there is nothing we can do about it, but still knowing what's going on helps the developers to make decision to use the 2DRenderer or not. For example, if Emission channel is not going to make it to the 2021.2 then we need to know. If tilemap shadow is not going to make to 2021.2 then we need to know. This is much more important than actually getting the feature implemented. If the feature will take one more year to make, so be it. We just need to know so we can make decision of using the 2DRenderer or not. Bottom line is that as long as we know what/when for any feature or bug fix, we can go with it.

 

Thanks for the info. Where can we find example of emissive shadergraph?

 

Also the shader grpah 2DRenderer lit / custom lit does not allow control over depth writing... any reason for this?

 

Thank you Chris for your honesty and transparency. I agree with everything said by @castor76 , as we supposed it was not a problem of incompetence but a problem of size of the department, only 2 engineers... I understand the slowness of the development process as there are many versions of Unity out there and the maintenance cost must be huge, that's ok. But mixing a tedious process with a very small team is the perfect recipe for frustration, yours and ours. I hope you are able to convince your manager to hire another engineer so you are at least 3 and you're not left alone if one gets sick, or you don't lose a lot of codebase knowledge in case one of them quits.
I'm a bit skeptic about the possibility of making URP so flexible that everybody can add features like which I added without touching the code of the URP, but let's see the magic you can do!

 

2 engineers working on the 2D Renderer, doesnt mean they're working on soley that either.
I believe Chris is one of the main Shadergraph guys aswell.

 

Excuse my naivety:

Do URP and HDRP shaders have to be made in the ShaderGraph, by Unity?

Couldn't Unity just make Materials with Inspectors and coded shaders that provide for most common production needs, and are absolutely optimised for performance and/or appearance as and where needed?

Especially when the ShaderGraph is seemingly a blocking issue with regards getting at expected and commonly desired functionality in shaders.

 

I agree. We should be able to do everything without shader graph..

 

Where are the shader graph samples to look put for?

 

You can see the shader code produced by ShaderGraph, then you can copy it and use it in a material. Or you can dig in the repository of URP and take the shader https://github.com/Unity-Technologi...m.unity.render-pipelines.universal/Shaders/2D

 

That's not the point of what I'm suggesting, at all.

Perhaps this will make it clearer:

How about Unity hand write some custom, hand coded shaders that fill all the normal, expected and most commonly desired needs, with appropriate materials/inspectors, whilst we wait for them to figure out how they're going to make the ShaderGraph more capable and fully functional?

Especially in areas where a shader's development is being hamstrung by limitations in the ShaderGraph - as per the quoted statement.

 

Whilst the example given is the current shader, I'm asking the question in general:

Why not make a full range of custom, hand coded shaders that do all that's generally, commonly expected and desired for game development, as fast as possible, without any holdups, so users can crack on and get games done and out the door as fast as possible.

Because... it seems that ShaderGraph is massively slowing down URP and HDRP shader development, and because it's been years, and because Unity keeps pushing URP and HDRP as production ready. etc etc.

Right from the get go, really... I'm curious as to why a full range of the most common, general, essential and commonly desired shaders weren't provided.

Everyone, by now, understands that ShaderGraph shaders can be as performant as hand coded shaders. That's got nothing to do with the point I'm making and the question I'm asking.

Which, to sum up, is:

Why let ShaderGraph hold back providing best-in-class hand coded shaders for URP and HDRP - especially when there's been (for years) huge gaps in what can be done with ShaderGraph, but (presumably) can still be done with hand coded shaders in both these SRPs.

 

It feels like the boat sank on love for 2D in "new" Unity when it first became apparent that LWRP wasn't particularly light weight (more just Limited than Light) and wasn't designed with any significant consideration for how to be a performant and empowering 2D pipeline for mobile, where "LightWeight" would have been most welcome.

At the same time, Project Tiny was still 2D only, and looked like it might not only fill this gap, but be exceptional in both size and performance.

Of course, that wasn't to be. Project Tiny went quiet for too long, and then became 3D focused. And now looks like it's a ghost town.

To me, the missing core features are foundational trust in Unity's understanding of the 2D gaming spaces that are not high powered Steam gaming PCs and, increasingly for years now, its lack of care for anything not AAA 3D.

I understand you're in a 2D focused team, but it's clear you've not had the remit to make a great 2D pipeline, which really should have been one of the pipelines. A 2D, ultra performant, ultra lightweight, ideally stripped version of an SRP really should have been one of the pipelines created with SRPs, by Unity, from the beginning.

That it wasn't speaks volumes about Unity's lack of insight into the differences between the needs of those making 3D "light weight" games versus those making 2D games for mobile.

I'd still choose Builtin, if starting a 2D project, because I can't trust URP's performance, can't find how best to optimise it and can't be sure that anything about it won't change significantly, nor that I won't simply find whole chunks of things missing, broken or that can be done in Builtin better, or simply can't be done at all in URP.

All the light casting stuff, the bending stuff, it all looks great, until trying to use it, when performance and usability problems immediately scream at me to go running back to something that's been tried and tested for a decade on lesser gaming devices, and was honed back in the days when Unity was more responsive to its communities and their needs and issues.

It feels, sadly, like 2D Renderers in URP are tacked on to satiate a demand, not conceived to be good, let alone great.

So I'd say it's love and understanding of 2D game making, that's missing, at a management level.

 

Well, like I have already mentioned, we need full control over depth writing, blend mode, render queue etc, just like any other URP lit shader. Not sure how difficult this would be.. The reason why I am not using ShaderGraph is because of this. Instead, I am using Amplify Shader Editor which does support all of above, except it is not working with current 2021.2 beta and they are refusing to support beta Unity releases. So I am "stuck" in between not able to try Unity 2012.2 beta and the latest 2DRenderer and not being able to use ShaderGrpah neither.

 

In my opinion, 2DRenderer does have so much potential. It just lacks that last 20% to become close to "production ready". All it needs is to fill the last critically needed features. But like Chris explained, things like tilemap shadow will only be implemented in 2022.1 (and this is hopefully) which is kinda like still a year from now on.

So, how do I feel about it? Back in Unity 3 or so maybe even 5, the tech development and response from Unity was ok, but I feel like sometime around switching to subscription, something changed. Now Unity feels ... more or less like Adobe Photoshop, where innovation and update pace is a lot slower than before.

I do actually feel that Unity's current focus is quality than innovation and features. Which is fine in its own. Except that for 2DRenderer, it was caught in the wrong cycle where its development was not "finished" yet to be production ready. And for some reason, engineering power level does not seem to be what we normally be expecting from technology company such as Unity. One or two engineer for 2DRenderer? Wow.. that is... really sad.

 

This is one of the problems of no management oversight and consideration.

Why "compete" with the provisions and capabilities of Amplify and other store assets, when Unity can do something nobody else can - optimise 2D performance and low level functionality that's in the engine bay of C++ etc, and create optimised access to these optimisations surfaced through C#, Burst, Jobs and DOTS - to folks like Amplify and their market competitors?

 

Can we expect this kind of trivial update to happen in 2021.1.x or 2021.2.x? Or because this is "new" feature, we have to wait for the next Unity 2022.1 ? (another half year wait?)

 

Also, things like... 2DLight missing some properties setters. There are many properties in the 2DLight component that has only getter, but no setter with no obvious reasons.

 

I can’t believe you only have 2 engineers.In my opinion, the development efficiency of URP2D is extremely low.

Can't believe it anymore, I decided to change the source code through my own study to achieve my goal.

 

Yeah, maybe the best way is to have better document or examples on how to dig into the URP source code and customizing it could be the only way forward until maybe 2022.2 arrives..

 

I think for something like 2DRenderer, the "number" should be at least 5. Not everyone should develop a new feature, but some does and some take care of the bugs, minor needs (such as property setter for example). Once the renderer becomes much more stable and feature complete, then the I guess the number could get down to 3 or so?

As for the customization, I want to see the shadergraph emissive example first. Then maybe I can work around it using that. Or the custom lit mode. I have tried the custom lit mode, but I could not see any difference between the normal lit as far as the structure goes, and have no idea how to actually make use of it.

 

Whatever number would have been needed to finish and polish a class leading, all purpose, amazeballs 2D renderer by 2019.4LTS, and it not be reliant on ShaderGraph, DOTS, ECS or any other aspect of Unity's growing list of foibles, flails and fails.

How important is modern 2D Rendering to Unity?

3500/2

 

I daren't speak for everyone, except to say this:

I'm reasonably sure EVERYONE is incredibly grateful that YOU, specifically, are working on 2D rendering, and that you take the time to post and read and think about 2D and Unity and it's users.

You think clearly, respond well, seem to genuinely care for the user experience, and have some significant political chops when it comes to dealing with the systems you're within versus the needs/desires of users.

I'd guess you're probably also a gamer, too.

Please don't take any perceived attitudes or otherwise as anything other than a desire for better tools to make better games.

It is wonderful, truly, to interface with a blacksmith about the swords we will wield in the future. It is a great shame that Unity hasn't seen fit to empower their blacksmiths to build a better armoury ahead of time, for 2D, given that it's the (dare I say) best use of Unity3D.

Well, perhaps 2.5D is, I think, the best use of Unity3D.

Whatever Ori is... that kind of thing seems to be Unity3D's sweet spot.

And the issue is that Unity as a company doesn't get this, instead thinking they can be some kind of Unreal competitor, and seem to neglect (at best) the opportunity to be the absolute foundry of new and modern 2.5D FUN!

Cause, let's face it... 2D and 2.5D games are much more gamey, much more fun and much quicker to pick up and enjoy with friends.

Short aside... how much time do you spend talking to the UGUI folks, and how many are they?