Standard Assets 2018 - let us know what you think!

I think that Unity will always fall short of asset store or bespoke controllers / vehicles etc. That's a given. You can't please everyone and you would probably need AAA experienced staff to get a car (for example) that is robust and fully featured so the standard assets should just be depreciated for the following reason:

The standard assets are IMHO in a difficult place. Templates are pretty much 9 times out of 10 going to be more appealing, more useful and more obviously useful across a wider range of developers. When devs see tower defence template, they think about what they can rip from it or build from it.

When devs see standard assets they're like "whuh, whut" and wonder why they need it or why isn't "official". In short I recommend the immediate depreciation of standard assets and instead shift focus and effort on more Templates. Templates are exciting, visible in what they can provide (much more useful than standard assets) and also encourage building or modding a project instead of being lost with a bit of ill-fitting lego.

TLDR: Templates are more important to make noise about for user quality of life as standard assets will never be battle tested, focused or usable enough.

 

I agree. The only stuff I've found standard assets useful for are shaders and character controllers (latter of which would be better served with a template). There's a lot of stuff in there that I don't think will ever be particularly useful to anyone, because it's so basic and generic that it's not really good enough for anything.

 

Yeah I really only use Standard assets mainly just so i can get the particle systems out for quick prototyping particles I want lol.

Now if the new post effects were in it, I'd gladly use them, hate having to go search for them. In fact I think the new post effects should just be in unity from the start. I think it's been in beta long enough, considering V2 is being worked on, why not go ahead and put V1 inside Unity?

 

Standard asset are not useful for game, but they are for experimentation and demonstration, I would expend them to go more into whiteboxing level of usefulness, instead of random toys. But we can do that with template too anyway, so template might share a reusable "standard" among themselves?

 

Standard Shader that supports vertex colors.
Considering how many millions get spent I would think this would be easy, but they still don't do it.

 

I filled out the survey but I just want to publicly echo what everyone else has said here in that the standard assets would better be served by templates than the current setup we have now.

 

Can you expand on this hippo?
Are you (and it seems everyone else) saying instead of a standard asset pack Unity should create templates that would be starting points for certain types/genres of games - ie template fps controller, 3rd person controller, mobile UI ??. template top down shooter, template (basic) inventory system? I don't know - I think I can see the benefit from of course character controllers, since those are the examples I have, and mostly the only reason I use the standard asset pack (for prototyping). But I can also foresee a headache for Unity when they start getting user help requests on incorporating all the templates into one system (help me make my game with all the templates combined) type of requests.

With or without the current standard assets - in conjunction with the handy training material available - what purpose would templates serve?

 

I've always thought of myself as moderately skilled with Unity..... Thanks for the ego boost as I had to select Advanced according to the descriptions

I'd like to see complete extendible systems added to the standard assets, rather than just some helper scripts and starter models.

Something along these lines:
* FPS system with UNET integration
* RTS system with UNET integration
* Day/Night cycle system
* RPG Inventory system with UI integration

As it is I find the standard assets to be useful when learning the engine, as it gives you something to play with and some example code to look at, but their usefulness in a shippable game is dubious. Turnkey extendible systems, where a game type is basically ready to go out of the box, and can be adjusted to fit the flavor of the game, would be really awesome. Allowing these to be extended with additional game specific mechanics, with commented code well enough for easy customization, would really raise the bar for standard assets.

You do that and you'd see standard assets appearing in a good number of shipping games. The key is to design them with customization in mind so everyone's game doesn't look like everyone else's. Include horribly terrible models with these systems, so devs aren't tempted to use the same default models as everyone else

 

They might just call them "low-poly" and ship them anyway... D:

 

I think I understand what he's saying. The templates that are broken out into their own packages are very robust and nearly fully featured. For example two really good ones are:

Endless Runner - Sample Game https://www.assetstore.unity3d.com/en/#!/content/87901
Adventure - Sample Game https://www.assetstore.unity3d.com/en/#!/content/76216

Now that I think about it Angry Bots was another one that was massively popular and re-used by tons of developers to showcase their own products.

As for Standard Assets, I don't see these sample games replacing that or even a need for that. One of my suggestions in the survey was to break out the Standard Assets into smaller packages that keep stuff in a single theme. For example I don't want to import a bunch of vehicle assets if I want to strip out the Particle FX. I can't speak for other developers, but I use the Standard Assets all of the time to strip out the stuff I need to prototype with. I know I can manually un-checkbox each asset I don't want...but it's a time dump doing that. For example, if I need a camera rig or cool camera prototyping stuff I should be downloading a Camera Standards Assets Pack. If I need particle FX I should be downloading the Particle FX Standard Assets Pack.

 

I'm surprised there wouldn't be any analytics sent to Unity that you guys can use to determine what Standard assets are being used.

 

As a few others have said, the standard assets are good for prototyping. Because of this however I don't think they should all be shoved into templates. I don't want to have to break open the RPG game template or whatever to pull out a generic grass texture to use.

 

It's a bit hard to fill out a survey saying you haven't used Standard Assets when you keep forcing me to say I've used them when I haven't.

Did that not occur to the person who made this survey? Do you think that the Standard Assets are so amazing that we would be forced to use them?

I work mostly making 2D games. I generally avoid using Unity Physics so everything you have to offer is largely useless. I also work using pixel art - so I'm mostly writing my own scripts to force Unity to not compress the crap out of all the graphics.

 

What if the standard assets also came with artist friendly formats e.g. 3d models in Blender/Maya file versions, Photoshop textures, Substances, Sound files, Music files.

Or Unity asked some of the 3d modelling/art/sound/music software providers if they could help make standard assets that can then be stepping stones for budding artists to work with.

Then you can do art/sound challenges and learning courses that take the standard assets and make them something amazing, whilst learning and improving their skills.

So the standard assets are more than replaceable prototypes they are stepping stones for artists and developers to improve their skills with.

 

On the programming side what if standard assets were based on the single responsibility principle and used interfaces so even if the developers upgrade to asset store code it can conform to a good Unity game SRP interface pattern that would allow enhanced systems to be drag and drop replacements for standard asset code (SAC).

Also SRP code could allow the learning section to build a framework that works in many game formats and can be built upon. Or with this new SAC it could work like a meta template that other templates extend for their game genre and style.

 

I agree with the artist-centric focus & formats + template scripts that are well-developed.


(Full-disclosure: I'm an artist/animator/gamedesigner/toolsmith -- but mostly an animator.)


I feel like if you guys offered some "template-style" art assets -- such as various heads/body shapes for multi-part 3d characters, or hair styles that could be manipulated in a .blend file or something, that could be plugged-in to, say, a character customization template script "rigged" to a 2d body with some basic walking/running/etc. animations (to give people some toys to play with while simultaneously encouraging beginner artists/animators to author their own content by, for example, providing really low-quality textures with the geometry that is nothing more than a test-bed asset to let the budding gamedevs author their own content.

This could also be useful to experienced gamedevs equally by providing them with some simple gray-boxing assets they could either heavily modify in their 3d tools, or toss entirely when they create better assets.

Providing some script templates that are easily modifiable (to extend or replace portions of them) allows an experienced dev to easily, say, swap gray-boxed models/sprites/levels/geometry with more-final assets, but also giving it a general "use" to a beginner as well (by offering them an example of a "character customization" system created via this script.) That script has a larger purpose of being useful to anyone to, say, swap graybox assets with more final assets, and would essentially serve multiple purposes (with a little modification to any extender scripts and such) and would ultimately be a huge benefit a larger amount of people. Something like "auto-graybox-swapping" would be immensely useful to someone like me, who isn't interested in programming such a tool from scratch, nor buying one from the Asset Store, but would totally extend it to suit my own purposes -- if I had it available to me in some form (in a very-usable state) to easily do so.

These types of scripts are the golden-nuggets of an experienced Unity-dev to keep them from having to reinvent the wheel on everything as most of us pretty much tend to to do.



Only later do we find out that there was a free asset or script out there that did (almost) the exact thing that WE were trying to do~!

How many here have had this happen to you?

Wouldn't it be great to have a common set of "standard" script and art templates that genuinely serve some use to /everyone/ for each of their own different/unique applications and purposes? -- i.e. scripts and assets that were genuinely meant to be modified?


Also -- I think lots of "buyer's remorse" comes from the Asset Store because of things like this too. There are plenty of asset devs out there who like to capitalize on stuff that already exists in a lesser-known (but free) form.

Just food for thought.

 

Unfortunately, I can't honestly complete the survey. One of the first questions are "have you used standard assets", which I haven't is my answer. But the rest of the survey only allows answers for people who have and are mandatory to answer. I can't answer 'what packs have I used' because I haven't used any. Your data will not be accurately represented due to the people who haven't used it forcing to pick answers that aren't true.

 

Exactly this.

To add my 2 cents : I think to prototype a game, people needs very basic assets and very basic features (few script very useful like character controller to simply move on the scene to show level design for exemple, but not that complex (sounds of walk, etc), because it's not useful).

Plus, it's a bit harsh and it's a very personnal opinion but you should not gives that much scripting features like this, because every amateur indie use them without modify it or understand it, splash the "made with unity" logo because they of course use Unity Personnal and here it is, a bunch of Unity games contains those noticable standard behavior everywhere.

I know this is another subject but I think the moutain of script in the Standard Assets participate to the "Unity games are S*** games" speech we still read on forums and reviews nowdays.

My recommendation is, like hippocoder, depreciate it, and focus on templates. And for me, in addition to that, just provides standard ASSETS (basic textures, basic sounds, basic shaders, mesh (and I'm not talking about a car, i'm talking about simple shapes, like cones, pyramid, torus..)) and not "Standard behaviours" already full coded and full featured, or move them onto another package, or just spread them onto the asset store, for limiting their use and push people to build or mod their behaviour themselves. Or do templates

I know this is opposite to the Unity mindset "Even my mom can do a game with Unity(tm)" but maybe this punchline is actually more harmful than helpful for the sake of this marvellous engine.

 

Greetings my learned gentlemen.

Some things should definitely be standard. Last time I tried to use the Unity Character Controller it was a disaster due to rigid body and other things not matching up. I had to learn UE4 just to know what the standard was for a biped.

I agree with the comment that Standard Assets should be split up into more useful specific things.

Also slightly on topic-
* can we please have a filter on polycount in the Asset Store. I don't use the asset store because its impossible to know which assets are suitable for PC, Console or mobile. Possibly assets made by professionals vs those made for fun would be helpful too.

* can we have some standard asset shaders for mobile that work and aren't treated as Legacy? I'm betting that the gamma inflicted fresnel effect is still a problem on mobile standard shader. The idea of bundling it all into one shader has proven to be a disaster for mobile.

Our latest Android demo in Unity has only been possible by throwing most things away that were supposedly standard.

 

This.

At the very least, if you do nothing else, listen to THIS ^^^^^^^^^^^^^ Unity Team.


Also -- please make it easier to do this with other Assets too -- not /just/ the "Standard Assets". I can't tell you the number of times I want to grab just a script (or some decorative models / shaders) from another project or asset I have already imported somewhere once, yet I get forced to always have to import those same selections manually by lots of check/uncheck options for stuff that will always be the same. For example, I don't want to import the "demo" scene and its models/materials for an Asset I've already bought/downloaded from the store and have been using for years after I learned how to use the tool -- What is the point of continuously having to uncheck it when importing the Asset?

Perhaps go one step farther with importing assets and, instead of forcing us to continuously click all those stupid checkboxes each and every time, why not give us the ability to create "profiles" for those asset packages? -- i.e. "particles" -- "models" -- "scripts" -- "etc." -- "all" -- that a user can later select from a dropdown upon importing the asset package (or a similarly-named .unitypackage) to ensure that everytime he wants to import some assets from that package, he/she can "pick" them all simultaneously by simply selecting a particular "profile" that contains all of their desired "selections" that correspond to the asset he wants to import. Maybe let Asset Developers store this data in the .unitypackage file itself though -- that way, if you download a model pack, and only want, say, rocks, the creator of the asset can simply make a "profile" that is labeled "rocks" so that the user can always import /just/ the rocks when he wants to.

"Import Assets 2.0" right there in that second big paragraph -- ^^^^^^^^^^ -- Make our lives easier please?

 

Instead of standard assets, one can introduce the concept of the Standard Library, a searchable place for individual scripts, possibly even user submitted instead of a blob of assets from asset store, most of which languish without updates nor purpose.

It's a fallacy to assume that standard assets would solve artist code problems. That's what visual scripting is for because a block of code can't please everyone and yet another component can't possibly cut it, and hasn't so far since Unity was launched over a decade ago.

If you can't manage it with how easy native Unity is, you're going to struggle exactly the same with whatever other component is used. Non programmers can spend weeks looking for the perfect script. I'm pretty sure an updated standard assets would not contain the perfect script.

Visual scripting solves this, especially if there's a new concept of a standard library - a database of visual script components, prototype assets etc... Anyway, plenty of food for thought, I'll be off now.

 

+1 for templates rather than standard assets.

But perhaps the most important thing that Unity is missing imo is an actual good character controller. The ones from the Standard Assets are quite simply abysmal. They are not deterministic for networking (they're based on rigidbody physics), deal with slopes/ledges/bumps/angled-corners very poorly, and stutter all over the place

Not only is the character controller terrible, but the character model is awful too. A standard assets character should look like a generic dummy; not a specific character like 'Ethan'. UE4's got the right idea for this:

 

Maybe standard assets need a new improved prefab system to be more modular in nature.

Imagine being able to build up prefab assets and filter their content e.g. FPS human player with weapon animations and controller that can be trimmed down to an NPC style asset by choosing only the layers or elements needed and/or swapping out components like controllers/sfx/materials/sub-objects.

Or a house asset that can work in an FPS with doors/windows/stairs/interior/furniture or trimmed down to just a shell for an RTS just with LOD/LOC (Level of Complexity) settings or layers.

Now if Unity could do that it would really empower creators and developers IMHO.

 

You just articulated one of my biggest issues with the Standard Assets. We need characters to be proxy artwork...not finished character models. A cubeman or anything very generic and obviously placeholder. The same applies to vehicles, environments and so on. All proxy and modular and not locked into a "theme" like the existing characters have been.

 

Whenever I'm starting out on a 3D game of some sort, I always grab the humanoid animations from the characters package. It's really, really convenient to have a set of good-enough-for-a-long-while animations to start working with. Of course, that would be much better handled with a 3rd person controller template project. In general, templates in the hub is just a better way to deliver this kind of pre-made content.

Unreal has this stuff down - you just boot the thing and click a button and get a dude with a gun in a room that works reasonably well.

 

Drops bomb and runs.... te he he ...LOL.

After learning Blenders material nodes and compositing nodes I am a total convert to visual scripting methods.

throws bomb to next in line ....

 

Yes, Yes, Yes amen

 

this would be really nice. That and the ability to collaborate, submit bugfixes/feature updates to existing content

 

I think Templates and such are something that could help but a lot of these things simply serve as a crutch rather than a tool for learning and a lot are already available on the app store, however to keep pace with Unreal and also for rapid prototyping, official, free templates may be a nice addition.

Not sure it would fit the Standard assets paradigm but I'd like to see more basic assets to help beginners get a grip with the Unity way of doing things as well as scripting.

The current FPS Controller, for example, has a lot of stuff that clutters the script. While headbob, step sounds and CrossPlatformInput might be wanted in an asset you intend to just drop in and use, they're just complicating things for those trying to learn from the ground up.

TL;DR More basic assets for noobs to learn from and build off of rather than things that can be just dropped in and used.

 

Thanks Will.

Input and CCs +1. Generic characters look good too.

I seem to remember the elastic man having some animations. Would it be possible to have a set of standard professional animations correctly tailored to them? ie run jog walk etc.

bonus points if a Blender file for generic characters is available so Rigify could be used with it ;-)

 

It's nice having a male and female variant, although robots do solve both and would have been less needless work (not sure if this is even a worthy criticism). Loving where it's all going with Package Manager. Regarding controllers, I'm not sure you should be doing a thing more than Epic's example because essentially - they get real custom real fast, so just the basic one is fine I suspect. So that's not a huge problem for Unity.

I'd even bet just using built in Character Controller (Physx) with decent input and a decent 3D model would silence 99% of people. I bet most of the complaints are just people trying to get things moving and not realising where these things are in Unity (Physx controller + Animator + Mesh + Input). I'm sure some will disagree because they want it to support their use scenario out of the box

 

https://forum.unity.com/threads/free-opensource-aura-volumetric-lighting.441387/

This is the kind of asset I would like to see standard in unity

 

Is the CharacterController component a native PhysX thing? I'd always assumed it was Unity-made.

 

Will...dude, you are awesome my man. These look great. I for one love the idea of having a male and female generic prototype characters.

 

Although all of this is great news, I wanted to respond to just a couple of things:

• 1. This is great news! -- a package-based import system is amazingly appreciated. It sounds like a higher-level version of the "using" command in C# -- but Unity-specific, which is actually a great concept imo. So great job to everyone involved in making this happen!

• 5. As an artist who might consider making use of a heavily-modified version of these professionally-developed models (alongside parts of the future cross-platform input-based character controllers), would it be possible to separate any clothing and accessories from the models into a form that would let them be easily-skinned (dynamically, but not necessarily at runtime) alongside the character customization system?
  
  Perhaps if we provided weight-correction data to help with this process? -- I feel like too many of us are reinventing the wheel on such things and it'd be nice to have a Unity-supported solution to the constant issue of things like character "customization" efforts. (No offense to the developers of UMA, but its approach is much too convoluted and something (native to Unity) that would simply take the technical stress off of those of us who are purely artists to let us define our own skinning/weighting for characters and their clothing slots would be a great "Standard" asset/template to include.)

Thoughts?

 

Wrapper for http://docs.nvidia.com/gameworks/co.../physx/guide/Manual/CharacterControllers.html - in general it's an excellent start point for any developer and Unity does not really veer far from Physx API according to @yant and I think it's a good decision.

 

And while we're mentioning PhysX stuff alongside Characters -- has anyone at Unity looked into the PhysX problem with Cloth?

Here is what I mean (an interesting watch -- but it proves the math at Nvidia is actually /wrong/ for cloth simulation):

 

Ha! Yes, I remember seeing that too when I studied it. At line 1397 here:
https://github.com/NVIDIAGameWorks/...acterKinematic/src/CctCharacterController.cpp
(you have to be registered with Nvidia to access it though.... otherwise you'll get a 404 error)

It looks like it was put together rather quickly, almost as a kind of little hack week project by someone at Nvidia, but it ended up staying there forever. The code is full of somewhat funny little comments like:

• "why did we do this ?"
• "A bit slow, but everything else I tried was less convincing..."
• "this kind of works"
• "AAARRRGGH"

Overall, it seems pretty solid for basic stuff, but using it in your project is a risky gamble in my opinion. The big problem is that since this is in c++ land, we'll never get access to the sources, and getting access to sources is extremely important for character controllers. There will almost always be a situation where you want to do special case handling in the movement/collision code, and won't be able to since you don't have the sources

 

Not entirely. I just mean it might be helpful to include a barebones "slot" system for those with low art skills to be able to use and, say, subdivide the face, sculpt one onto it in a sculpting program, and eventually make their own clothes/hair for the character (since a naked "humanoid" model probably wouldn't be fitting/ideal to use in any game), but since the clothing slots are included in the Unity project itself, the dev could replace them with their own clothes (and rigging information to let Unity do this either dynamically in-editor or at runtime), thus causing a lot less of a "generic" human look like you'd get from something like iClone, and more of a "hand-modeled" approach to help with the pipeline of making clothing/accessories/props for more detailed characters down the line.

I agree with this.

That said, I think what I'm trying to get at is more along the lines of the fact that "programatically" rigging in the Unity editor is both unsupported and also beyond the technical capabilities of most artists who want to create their own content (or who want to use another's free content as a base for their own content.) Additionally, the solution to this problem could be combined with a "slot editor" template that lets one do this very thing.

The issue of "standard" assets not being developed to be customizable (whether they're crappy low-poly or nice high-quality ones like in your screenshot) further lends itself to the deeper problem of character customization support -- or rather the lack thereof for that kind of thing (except in *very* limited circumstances that, for example, /dont/ involve potentially-dynamic rigging scenarios, which have been common since the days of The Sims) Yet, almost everyone, whether they're making 2d or 3d games these days has to support some form of character customization -- and then, that also means the more dynamic the better.

Just some food for thought -- I feel like Unity really needs to consider this side of the fence too when standardizing anything to do with characters.

 

Sound like Unity needs to really get a move on with the old visual scripting lark. Gonna be big business then.

 

When it comes to standard stuff for characters, what I'm missing is examples of correctly set up Humanoid characters for all of the 3D programs.

A fbx file with a male and female character is well and good, but what I'd really like would be a .blend file (and the same for Maya/3DMax) that contains a rigged humanoid character that will import correctly for Humanoid with all bones and whatnot set up correctly. We've got that in-house now, but it took a lot of experimentation and failure to get it just right.

Unity does the correct thing with regards to Blender and recommends the Rigify plugin, but that plugin has different ideas than Humanoid about what a biped rig should look like, so custom work is required to get it right. For an animator that's new to the engine, having a rig they can import to their program of choice that shows "this is how to do things" would save a lot of headache.



Another thing; examples of correctly set up Unity components are sorely missing. As I think I've mentioned upthread, the default 3D character controller is super usefull because it shows how to set up a character with root motion and a 2D blend tree. That helped a ton when I needed to figure out root motion-based movement. As a contrast, I've never managed to set up cloth physics correctly. If there was an official cloak I could download that looked good, that'd go a far way towards understanding how the system's supposed to work.

 

I mostly use Utility scripts, would be nice to concentrate on boilerplate functionality by category like input, controllers, AI techniques, physics (drag rigidbody), animation (waypoint circut, smooth follow), optimization (object pool).

A Standard Packages installer would be nice too, to include essential Unity-made packages that should've been integrated: postprocessing, cinemachine, Navmesh components, 2d extras, textmeshpro, render-pipelines etc.

 

It would be great to have a flexible (rigidbody based) character controller.

 

Here's the problem for us. I'll open up a bit on how things work at the moment. Perhaps it's out of turn but we need to move on and time is short.

Today our character system is based on root motion. This was due to the party line from Unity that RM was the new thing.

For touchscreen input is based on an old angrybots script from ancient times. In part this due to the new UI system being oriented around building blocks that didn't scale for this task.

UE4 I note has very little support for root motion.

Their character control is extremely responsive on touch screen.

Both engines use physX.

I have watched videos where Unreal were actively working with the Blender community using Rigify to animate characters.

Our market is constantly telling us our controls are not responsive enough.

Is there a clear website or tutorial which states what Unity's current position is regarding root motion, character controllers and input please?

Its heartening that Unity have a plan to address input. It's not so helpful thats its been so slow. I have a game ready to go as you can see but its controls are 70%.

I live 10 minutes from the Unity office in Brighton. We have a 750k download game so we have a reasonable idea of what we are doing. I would like to be involved in anything that could guide a path through this.

Are they any local events? I have to say I've never heard of one. I note the SF are doing a week long thing with Devs who can just drop in around GDC.

We had a subscription for a year but still couldn't get any precise information about how this was supposed to work.

I will be spending the next week on this (probably wasting a great deal of time). It would be very helpful if Unity could make a statement if they don't want to engage directly.

Currently we stick with Unity mainly because of 5yrs built up knowledge and the forward rendering being faster still than the new UE4 system. That probably won't be the case once they start optimising. I think they still do some per pixel shader stuff which is slow for mobile.

Years ago we used NeoAxis. We ditched it due to Unity's wonderful mobile support. I would be very sorry to have to ditch Unity because it can't do top line 3d animation. It would be even more unfortunate if it could but no one wanted to explain how to do it.

Note I would prefer not to move to UE4. If we can't fix this then I will have no choice to move. I made the support staff aware of this and they just tried to tell me about other things I didn't need. Perhaps they were doing as they were told but it wasn't enough.

From the sound of things Hippocoder is also thinking along the same lines which should really worry you (I too am very tempted by visual scripting). I think a community event would be a good idea. London/Brighton pref

I very much appreciate Will reaching out to us on this. Sorry if this seems like a mega-winge.

My comments here are meant to give a real world developer situation as an example of why there is a problem.

My hope it's that it might illuminate the darkness so that someone can say 'You idiot, you should be using this, so shut-up already' )