Be More Critical Accepting "Public" Worlds. Rigor Tests.
closed
Shanie
I went to 8 worlds last night and 7 of them didn't pass what I would consider "basic muster".
- Too many real-time lights
- Holding 45 frames barely on a 1080
- Collision meshes too complicated and drags the frame rates down
- Clouds aren't VR friendly
- Shaders dragging the frames down to sub-30
- Land clipping because it stretches too far away
- Grass dragging the frames down because it's not occluded properly
- Mirrors not hidden on launch
- Water isn't VR friendly
- And the list goes on.
All these awful worlds at the most basic levels being accepted as public brings down the entertainment value of the experience.
When you're entering a public world you go in with the assumption that it has passed BASIC rigor. Instead it feels like pre-NES days where the gaming market was saturated with bad games which caused people to effectively just stop buying. It wasn't worth it. It's not worth it in VRChat right now.
Make some basic rigor rules with some instructions and ban submissions from people who break them for a week.
Log In
Ruuubick - Designer
closed
Rokon
We should also take strides in advertising videos that have to do with optimization
xxx_red_xxx
Well not everyone knows how to do some of these things.
Also not everyone has a VR to test their world. I published my first world long before I got my VR.
owlboy
xxx_red_xxx: It's a VR application. You are creating content for a VR application.
If you are saying people do not know how to do these things, AND are unwilling to learn, then that is a problem. Otherwise, they just need to learn how to play in this sandbox.
Unity is a
game engine
there is no getting around the fact you are basically doing the work needed to make an optimized game.xxx_red_xxx
owlboy: I know how to do these things and make a decent world and have a VR now I'm referring to those who can't. Not everyone has the time/has a VR to test their world. How do they advance forward? Or do we deny them the right to use one of the game's biggest features which is content creation just because of it.
owlboy
xxx_red_xxx: I don't think so. But one can learn the metrics to follow. keep draw calls as low as possible, use VR safe shaders, don't turn on deferred rendering, don't use some post-processing effects, don't use 6 real-time lights all at once, etc etc.
All of those things don't need "testing" they just need to be done.
Of course, there is more testing to see where the exact line is. But they can ask friends or strangers to help and give them feedback on how it's working in VR.
xxx_red_xxx
owlboy: Just reading your post I can understand how many won't understand how to do this. I may know what a draw call is but I can guarantee a lot of people don't. And a lot of people don't understand the concept of baked lighting. Also not all shaders are clearly marked as VR safe. For example, I got my hands on a water shader for my one world, determined it looks nice, came back the next day on my VR and realized how awful it looks in VR. Now if I didn't even have a VR and decided to publish it... Also New Users can publish things and with VRChats current situation where all the trust ranks seem to be very segregated, a New User trying to publish a world and may only have 10 or so friends and with a good chance that none of them have VR. Lastly, not everyone knows which postprocess effects to not use. Now I know you're gonna say that this "isn't an excuse" and I'd probably agree. I'm not trying to make excused why crappy worlds should keep rolling out but rather inform people that those who are new to making worlds need a better tutorial video or something.
Tupper - VRChat Head of Community
xxx_red_xxx: If you create content for VRChat, you are in essence a developer. You must keep in mind all things that any developer must keep in mind-- optimization, proper processes for scenes/models, etc. It is up to the user to make a "good" world.
xxx_red_xxx
Tupper - VRChat Head of Community: Fair point.
Zennshi
Tupper - VRChat Head of Community: And what metrics are we going to use to achieve that "good" rating exactly?
In normal games - you balance the complexity and features with performance to a desired end with a hardware benchmark in mind.
Vrchat has nothing of the sort.
There is no real metric, just arbitrary standards that aren't entirely true or given proper context.
Vrchat is like a blank unity scene. A world can be anything. An avatar can be anything.
What if I make an avatar designed for 1 on 1 interactions?
What if I want a world for 3 people at the most?
Would you apply the potato pc 50 vr user small map metric to that, despite the obvious intent behind the actual DEVELOPER who made it?
Doesn't your front page say something like, create your own worlds, express yourself, invite friends, etc?
Because that is like seeing a sign say "come paint anything you want", bringing your own supplies, then getting in there and be told what you want us to paint and how.
You guys need to make a decision.
Are you a open platform or a curated art gallery?
Hagbard Celine
Zennshi: The past told us what happens when VRChat as a plattform is too "open". If it means at least some consistency in world and avatar quality/performance I'd go for the "curated art gallery".
Zennshi
Hagbard Celine: Yeah because things in the past were so much worse than now.
In the past it was a handful of modded clients and subemitter crashes.
Now everyone has a client and the mesh/shader crashes can actually cause damage to your pc.
The only thing the past tells me is that they need to seriously consider real and actual security for their art gallery. Any rule makes no difference if anyone can open the door and bypass any/all of them.
So far as "consistency", what are you on about. There is no consistency with anything in vrchat besides how absolutely full of holes its security is.
You can't judge a world or avatar on the metric of other worlds of avatars. Much less assign what you call quality.
A knuckles is very performant. Does that make it a good avatar?
I don't think so, it has a green dot though. Do you really think things like that should be the benchmark? Is that what you mean by quality?
I've told lots of others who are on this vrc dev white knighting - I'd sooner see 2 feature rich complex avatars, than 30 boring featureless ones in vr. If I could turn off the IK and nametags of people I don't care about I would. If they start complaining I'd kill their audio too.
If I wanted an experience like Rec Room or High Fidelity, I'd go there.
So I don't know what you think you know about "quality" or "performance." But if you don't properly define what you mean when you use those blanket buzz words I'm going to assume you mean it in the way some of the actual "developers" mean it. Which is grossly uninformed at best, and flat out lies at worst.
I really don't think you want a curated art gallery. I really don't think you fully thought about what that means in terms of vrchat. Less options to customize your own personal vrc experience is not better for anyone besides the vrchat team because removing options is easier than adding new ones.
Shanie
Zennshi: Not exactly sure what you're rambling about, but VRChat has guidelines for avatars now as you know. Not rules per say, but guidelines, and I can tell you that over the past week I've heard "I'm so glad I get good frames now" or something alongside that at least a dozen times.
There's absolutely no reason to permit shit performance in a world.
BASIC
rigor for the majority of user created content, and I'm talking baking your lights, use performant shaders, and occluding appropriately, would bring framerates for most VR-operating machines to 60 or higher.You either have some basic rules like "your world needs to not be shit from a performance perspective" or you have a garbage experience in worlds that have been "approved". You can have your shitty private worlds, go for it. But not public.
Zennshi
Shanie: "Not exactly sure what you're rambling about, but VRChat has guidelines for avatars now as you know."
They always have, and the majority agreed they were bad.
"over the past week I've heard "I'm so glad I get good frames now" or something alongside that at least a dozen times."
The people I hang around understand that feature rich avatars cause performance issues in crowded lobbies. They also know of a block avatar button if they can't handle it. I run around in a 15k poly atlased avatar and I still hear people complain about frame loss, which depending on their own setup will heavily determine how much people can handle. Frame rate across people's platforms isn't constant, that literally means nothing.
"There's absolutely no reason to permit shit performance in a world."
What if the world isn't intended to be packed full of people :/....?
"You either have some basic rules like "your world needs to not be shit from a performance perspective" or you have a garbage experience in worlds that have been "approved". You can have your shitty private worlds, go for it. But not public."
I mean... I suppose... but some maps, even not very performant are still popular... Look at the void club.
L
Lamb_
Zennshi: It might be a bitter pill to swallow, but Vince has made some very valid points. VRC is advertised as more of an open platform, but recent restrictions seems to really contradict that. The team should have a nice long discussion on which direction to take VRC moving onward, assuming they haven't yet. The disconnect between the open platform promise and the actual restrictions are too jarring. The safety system is a brilliant move in allowing players to pick and choose to suit their own circumstances, yet the follow-up avatar ranking system seems to aim to take away choice from the players with the looming promise of upload restriction based on ranking.
So which is it? Do you want players to be able to choose for themselves or do you want to choose for your players?
Shanie
Lamb_: This was solved in the 80s. You have system requirements in VR, just like in VRChat, just like in every video game since they were sold. The rules are that a minimum-specced VR user needs to achieve at least 45 frames with nobody in the world to be hardly considered acceptable.
You audience is VR-geared people and you have to cater to your audience. Everyone wants to be able to choose for themselves what is acceptable in their world, but they need to accept realities of current processing equipment, just like every other computer-based experience.
If you don't want to play the game of basic optimization, you shouldn't be public, full stop.
L
Lamb_
Shanie: Of course, there should be a baseline performance requirement for public worlds. I have no qualms with that, and on that note, I've not had an issue with the world publishing criteria so far; all of the worlds I've been to have been acceptable in terms of performance. I don't check every single world that's in the new tab though, so take that as you will.
Sorry if I didn't make it clear in my comment, but I was referring to avatar ranking and the possibility of avatar uploads being restricted based on the ranking system.
Chdata
Tupper - VRChat Head of Community: While it's up to users to make good worlds, it's up to the platform to cull bad worlds properly, considering that not all users can or will make good worlds.
Rasmus Lindahl
Maybe have a requirement that 50 random people must give the thumbs up or down for the room and it will not go online unless at least X% think the world is acceptable?
Emma
Same issue with 2080 ti.
There's currently a werewolf map where i rarely go above 60fps. Even without any players in it.
owlboy
Emma: In these cases, I would definitely use the in-game report features to report bad perf.
Hagbard Celine
I just uploaded a world and must say 10 minutes from the request to public status seems a bit quick. I hope thats just because the world is prety simple. But thumbs up for the quick status change anyway....
W
Wulfhere Cyning
Is there no way for the game to recognise whether the player is using a VR headset? If yes, maybe when worlds are "accepted as public" they could be separated into two categories, "VR-friendly" and "Not-VR-Friendly". That way, when a VR user is going to public worlds, the game could lookup VR-friendly worlds and ignore the "not-VR-friendly" ones, while if the game sees its in desktop mode it can pull from both lists. The VR-Friendly worlds could be held to the standards you suggest, and the not-VR-Friendly worlds could be held to a lower "public acceptance" standard.
The problem goes two ways though. I was playing desktop mode recently, and found a few public worlds that I would call VR-functionality worlds, but had nothing (not even a signpost in-world) to say "Um yeah, this is for VR people".
Zarniwoop
Wulfhere Cyning: I think segregating worlds into "VR friendly" and "Not VR Friendly" categories would be awful. VRchat, as the name implies, is and should be about VR first and foremost. Even though desktop functionality should still be considered, it's a secondary feature in this app. The main appeal should always be to be in VR to drive it forward.
If a world is unbearable in VR but fine in desktop then it should simply not be made public at all. By that I mean for example non-stereo shaders which hurts your eyes to just look at stuff.
As for the game mechanics in the worlds themselves, if they want to cater to VR users or desktop users, that's up to the world author.
Shanie
Zarniwoop: I concur with Zarniwoop. VRChat is for VR, whereas keyboard and mouse users get an introduction to VRChat in a restricted sense, giving them an option to go all in or not. Priority is always VR.
xxx_red_xxx
Wulfhere Cyning: I agree with this and have even recently encountered worlds that were accepted as public and DIDNT WORK AT ALL IN VR (double-rendering). However, I like the world, returned on desktop mode and had fun. Should it be taken down? No. Too much work for creator. Should it be made clear it is not VR friendly? Yes. I don't want to waste time loading into it and then have to switch modes.
ivankazuya
Perhaps a rating system for worlds then, therefore, everyone can continue making their own worlds bad or good. I think we should consider the path to getting better is with practice. I don't have an issue when seeing basic worlds with avatars or a world someone new to blender made. If I don't like a world, I'll just leave. I would rather keep the terrifyingly awful worlds. I feel it gives new developers a reason to improve and to gain experience. But that's just me.
StormiTheAvali
ivankazuya: I'd love an "experimental public" tag that could be put on public worlds that didn't meet performance requirements, but that devs were actively working on. Like a time based thing, where each week or two weeks it's checked to see if that world has been updated at least once, and if it hasn't been, then it can be set back to private, but at least something like that would allow for creators new to Unity to be seen, and more than likely get decent feedback as a public world.
C
ContentCreator
Signed, please test the worlds also on gaming laptops or machines with less CPU
Hagbard Celine
ContentCreator: I dont care if a world looks underwhelming from an artistic point of view, we all had to start somwhere, but I'd prefer a good performance at least when the world is empty. If it is meant to be used on high end hardware there should be a warning at the spawn area.
And it should be tested on lowest spec VR hardware in VR. I don't care how well it performs in desktop mode.
Wich is the biggest problem I guess. People without VR can also upload worlds. They have now way to know how well or bad it performs and some of them don't seem to care. I got reactions like that several times now: "I don't care about VR performance..." or "Runs fine I have 60 fps...". In Desktop and in VR I had 30-45 with 4-7 people in the map.
Ron Millar (CCO and Design)
in progress
Load More
→