Weekly Topic GDC 2017 Discussion!

In SF, our team is kicking off Game Developers Conference 2017 this week! While some of you might be at the event in person, we would love to start talking speculation for the week and what you’re expecting to see!

There are a great deal of events going on throughout the week, but if you’re interested in our specific talking points for the week make sure you tune in to our GDC 17 Keynote tomorrow morning, 8:30am PT. Our event page will be live with the stream embedded right before the show is live. Tune in and see what is coming up with Unity!

So what do you think we will be talking about? What news are you expecting to come out of the event as a whole? GDC is a great week to be a developer and we hope we can share some of the experience with you.

Once the events have kicked off feel free to discuss both our own, and industry announcements here with us!

-----

Follow our Topic Index to be notified of each weeks new topic!

 

So far, over an hour in, the keynote is a huge disappointment to me. I don't understand why presenters (or should I blame investors?) still think, to this day, that devs care about generic statistics, buzzwords, more pointless platforms that will drain development resources, run-off-the-mill mobile games, and celebrities being paraded on stage. The audience is so uninterested in these things that it feels painful to watch, and you can feel their lifeforce being drained out of them every time they reluctantly have to clap.

I just don't understand. This presentation is so disconnected and bizzare.... Who does this appeal to?

Any talk about "enabling success" is completely pointless in the first place, because if you enable success for everyone, you'll be making it harder for everyone to be successful at the same time. So nothing will change in the end. Success will always be relative to the success of others.

 

So, where was Octane?

 

I gotta say, though, that the last half hour was pretty good. The first hour and a half of the presentation should've been compressed into no more than 5 minutes at the start (all things related to statistics, finances, analytics, success stories, "enabling developers", "democratizing stuff", etc...), and the rest of the presentation should've been all the Unity2017 stuff

 

Visual Scripting in Unity 2017 got a mention. Ah if I had a pound for every time someone asked for that in the forum over the years I would have quite a lot of money. End of an era

 

What happened in the last few minutes?

 

They talked about all the good stuff:

• new High-end PC/Consoles renderer
• C# safe multithreaded code generation API that will supposedly "out-perform c++"
• Auto-LOD generation
• More commitment to artist tools (including visual scripting and neural net content generation stuff which seems pretty vague/risky to me)
• Timeline presentation
• Big changes being made to transforms, internal callbacks, old obsolete parts of the engine, etc....

And also they will have actual demos for some of these things at Unite Amsterdam in June

 

I'm really excited about Timeline and Cinemachine!

 

As an update, if you missed anything you can check out the whole live-stream at this link. We've also just released a blog with a quicker glance at everything that was covered in the keynote that you can check out here.

While Otoy / Octane Render was not in the keynote, they have a presence at our booth during the conference.

 

I saw the 2D tilemapper listed. any news on that?

 

That bit about using deep neural networks for content generation... the future sure looks exciting.

 

Is anyone else worried it would just end up resulting in super stale and generic worlds, though?

 

I had a question regarding the new Visual Scripting additions - will this be compiled or interpreted?

I would strongly prefer having a interpreted option if possible (maybe building on-top of Playables/AnimatorBehaviours?); it would be excellent if they could be loaded at runtime along with art assets - not needing player recompilation to embed the designed scripts.

I realise this is slower than compiled, however:
- Artists can work with this without needing developer involvement and compile times (which can creep into the minutes on larger projects)
- They can be loaded remotely as part of bundles.
- Debugging is presumably easier to implement
- Most visual scripting tends to involve the more complicated functionalities being passed into precompiled node blocks

Secondarily: is there a planned ETA on this feature? Thanks!

 

i'm totally confused with the presentation tho (maybe because i'm just too technical), so what do Unity have for 2017.1(ish) release?
i do hope Scriptable renderLoop and this part are included tho

 

They said they'll go into more detail at Unite Amsterdam, but they did seem to say Unity 2017 will have the HDRenderLoop, Timeline, and the c# jobs system.

Maybe not the node-based material editor that you pointed out though, but you never know

I guess we'll have a better idea when 5.6 is released on march 31 and the roadmap page gets updated

 

Aimed for 2017.1! And you can already download an Experimental Preview Build of it here: https://forum.unity3d.com/threads/2d-experimental-preview-3-update.454536/

 

It does sound like it is not worth spending 2 hours on watching the keynote, but the bit I have put in bold about changes to transforms could be great, could be very bad. I cannot find any information about what the changes are. Can anyone help?

 

All we know is that transforms were handled in a non-optimal way, and the they are rewritten to be more optimal. I don't think this could end up being a bad thing, unless you mean it could break compatibility... but personally I hope Unity breaks absolutely all the compatibility they want to break for Unity 2017. It has to happen at some point, and it'll benefit the engine immensely

 

There was no real detail, it just came up as part of a brief discussion about some of the changes under the bonnet they are doing to bring Unity up to scratch/fit for the future.

Personally I will be waiting till after 5.6 is released before expecting much detail about Unity 2017.

I would also suggest that not all that was mentioned will necessarily make it into the first version of Unity 2017 and I dont really see yesterdays keynote as being a proper intro to 2017 - I'd expect far more at Unite than we got at GDC.

 

I'm sure they do not intend to break compatibility freely when moving on to 2017. Maybe some stuff will have to be changed at various points but I expect them to keep breaking changes to a minimum, have the script updater handle a lot of the most common scenarios, etc.

 

I wouldn't worry about that:

https://www.ostagram.ru/

 

Great keynote, nice presentations.
Here is I found more or less interesting, plus some random thoughts:

Indie success stories blah blah - power rangers playing unity games under the strict eye of a tough Asian:
Not so interesting for me, might be interesting for others.

Unity 5.6 notes:
Vulkan is coming, Metal works. Lighting improvements, progressive lightmap debuts, improved navmesh.
Services - they @willFinallyWorkAtSomePoint, but not in the 5.x lifetime. Hmmm.

"Daydream and Cardboard are on board, for Android as well as IOS" Well, Cardboard for IOS makes sense, Daydream for IOS doesn't - so far, only a controller hack works and there is no intention from Google to expand this platform on IOS unless I missed something.

Congrats to Stephan B for the TextMesh Pro acquisition by Unity.

5.x lifecycle ends soon, so does my Editor dark skin. Hex Editor to the rescue.
Ads - IAP - China market penetration via Xiaomi - Facebook Gameroom. Money opportunities.

VR - so interesting, so fragmented, Owlchemy rocks.
A stamps collection seems a more worthy investment than VR for indies right now, due to fragmentation - high development costs - low platform retention - small market size. I hope this changes soon (tetherless Vive, more affordable headsets from all key players, Apple AR, affordable Daydream phones, Hololens for consumers, huge investments from all key players) but chances are that 2017 will not be the year for VR. Perhaps in 2018 or 2019 will be a better time to enter this market.

Unity 2017 promo time: Graphics improvements - C# code acceleration - faster build code updates - visual scripting.
Seeing is believing.

Tatarchuk is onboard!
Unity graphics team has some of the most talented people in graphics industry already and Tatatrchuk has a great record in delivering results - https://unity3d.com/company/public-relations/news/unity-technologies-hires-natalya-tatarchuk-graphics-director.

Unity 2017 promo time: Scalability.
Multithreading might finally deliver this time, seeing is believing.

Unity 2017 promo time: Greater visual fidelity - assisted asset creation - deep learning for content generation - composition.
I do believe that deep learning will become usable at some point, but not in 2017, perhaps around 2018-19, demand for VR content creation is the force that could drive this workflow, it could surface as a service front-end with an Amazon Cloud or some other cloud service back-end providing the processing muscle. See Renaldas Zioma early results on this field here https://bitbucket.org/Unity-Technologies/ml-imagesynthesis

Unity 2017 promo time: Cutscene/timeline editor.
Apropos the Cutscene/Timeline Editor experimental preview: Experimental previews practically means that you download many GBs that sit in your SSD just to see some features and try to implement stuff that might or might not work when these features come out. A modular approach, where the user could download a few MBs module that would work with the current Editor would be much preferred. I have unlimited bandwidth and enough SSD space, yet, I am not tempted to use the experimental previews, there are many developers who do not have either enough SSD space or enough bandwidth available and will not use these previews ever. It is a shame because a lot of effort is put in these previews and a lot of customer feedback is lost.

No mention about Octane integration progress, faster asset importing, nested prefabs, but what counts is to see those features implemented, not mentioned.

It is great to see Asteroids!, Ghost of a Tale plus others being featured in GDC Unity booth.

 

You clearly haven't been through the 3.0, 4.0 and 5.0 upgrades. They broke plenty of compatibility on those.

 

Not enough!

 

Haha, the component quick references depreciation in 5.0 gave me PTSD. The automatic script updater wasn't working - so we updated a good 2,500 entries by hand over the course of two weeks. :'(

 

Great to hear! And thanks for the link.

 

No idea (maybe some of the experimentals will be in it).. even the roadmap (I still say the page is worse than it was before its redesign) hasn't been updated, it would have been showing 5.7 stuff by now, no?, 2017 is a mystery

I was excited to hear artists and designers in the presentation...but then its mainly visual scripting , short of shaders/materials it has little interest to me, unless yknow, its designed and fully fleshed out so damn well in all those other areas I'm drawn to using it but yknow so is blueprints in UE and I don't think its all that great.. so artists and designers being catered to with better tools like through terrain improvements and other related areas or designer tools.. still laying out the ugui in the scene editor complete with all the rest of non ugui objects in the scene or doing other workarounds to avoid that. Still making scene levels out of cubes I mean having to use 3rd party modelling tools to design levels etc.

So more tools and features are needed for that, even timeline and cinematics doesn't really help artists and designers until much later in the project development like once you've got the stage set and assets in your scene and all the 100 and 1 other things done...but its not something you start with. I obviously given up on thinking untiy will have some amazing new terrain, scalable largeworld support from the start, water shader base line improvements, vegetation, time day sky cloud type of improvements for building inspiring worlds in anytime soon.

Animation FX IK improvements?...I know asset store.. what about Unet phase2? improvement on example assets to have unet supported examples, that fps controller needs an update imo. I feel like Unity is in the position to provide best practices for helping developers get multiplayer done right, I thought it was phase2 authoritive server design but it seems to have fallen off the rader.

Nested prefabs is still a mystery, is that even going ok?

Anyway I hate having to have multiple projects for different platforms, worse I hate having to import the same damn texture/model assets in multiple projects because I might want to use something from them, and if I don't use all of it, its just bloat and clutter in the project (cus unity provides no project cleanup, archiving or exporting only what is used in the project type of tooling either) ..so why can we not have a global asset folder option where the library data for said textures for different platforms can be used from that central location in all other projects. And same with plugins, editor assets that aren't likely to change but are often wanted in other projects as they become core tools to fill in all the gaps in unity, again has to be imported.. with no in editor option or plugin ui to enable/disable these things per project being able to easily disable scripts/assets in the project would be easier than other workarounds.

Best thing to come out of this event was that ugui was single handedly made better for all unity users with TMP, so for that alone it was a success

The rest of the stuff mentioned in the talk to me is all just wait and see unite Amsterdam, seems to be where things will be shown off. That multithread c# jobs system, and the other stuff as seen in experimental previews I already look forward to actually being able to use, and the scripting c#4.6 stuff,

I just don't test experimental previews that often because well 1) they don't update them alongside the beta releases often enough for me. 2) what if you want to try out more than one experimental preview feature in the same build..3) what if you grow attached to using it. so yeah rarely test those out... when it comes to testing stuff in an existing project that may have a gigs worth of stuff to duplicate just to try out an experimental build or feature.. sometimes not worth the hassle. It is what it is though and can at least try it out.

 

Will we be able to see videos of the other talks from GDC 2017?
I am mostly interested in: 'Destiny' Shader Pipeline.

 

If anyone want to watch more bitesized parts of the keynote, we have broken it down into six parts here:
https://www.youtube.com/playlist?list=PLX2vGYjWbI0SB3FWOxQ-SnyMBAHQTJAw5

 

Thanks. I can actually skip the boring parts now

 

On the timeline demo, there's a component called interactive objects. is that from Ghost of tale or part of the timeline system/demo?

 

Sometimes the authors post them online, sometimes not.

 

You might be able to find it on GDC Vault later on, or on their youtube channel.

 

I really hope so