Guess I’d better get off my butt and finish the VRC music video I’ve been working on for the past 6 months.
Any plans for mobile support yet? like a mobile desktop/google cardboard mode?
also FOV slider?
I don’t know if it’s the same issue but when my cat is in my lap and blocks my hip tracker and I playspace move my hip is locked in place and ignores the place space move.
For me that’s a much more common issue than seeing someone with 2 trackers instead of 3
Usernames are dead
In a previous Dev Update we noted that the VRChat API will no longer provide the username of other users to you.
This has been deployed! Usernames are dead.
At some point in the future, they will be removed completely. We’ll post here and inform relevant parties when that occurs.
Any chance that when the username API drop happens, we also get some sort of function like Steam profiles have, at the same time? Where it lists “This user has also played as:” with a dropdown menu?
People change their names constantly.
Any news on VRS / Foveated rendering after that demo from a previous post? I recently aquired a pimax and have been desperately needing a little something something to help me get better frames lol
Has the team considered of integrating LODs (Level of distance) into the game? I feel like that’s what’s causing the performance to tank, especially when in large events with at least 15 other avatars. Has the issue with avatar/world/prop decimation come into mind? What about shader effects?
Yoooo I’m excited.
I suppose they start to approach that idea with the Avatar distance hiding system
There’s no button for creating a Group anywhere in the video. Is creating a group will be website/api-only?
Actually, the fastest way is to toggle OSC through your radial / action menu (Options>OSC>Enabled)
One thing that could be done to make it more obvious on how to block users is just making how to select users more obvious in the tutorial, the current tutorial world just picks an avatar for you, tells you how to walk and then spawns a bunch of portals to public worlds, which the experience of the new player might vary a lot from there. The only way you can really know without asking other people is either by dumb luck or taking the time to read all the posters in the default home, which not everyone does. It’s not really made obvious how to use it unless you know already
Also explaining what all the buttons do would be helpful too, because most times when there’s a new player I have to explain that they have to press the eye to show someone’s avatar, the microphone to mute someone, ect…
Please do!!!
Mayyyybe, the Steam implementation can also be cleared at-will by the user so that’s another function we’d have to add in (and UI for it and etc), dunno if I can sneak webdev/API time for that >_> they’re mostly busy on Groups atm
No new news, those projects have all stalled or stopped completely due to being unfeasible or being blocked behind things like “doesn’t work on this engine” and “everyone would have to output motion vectors from their shaders, which is not going to happen”
I could talk about it in a future dev update but it just tends to make people argue about it for pages, which isn’t what I want to do
LODs already exist in worlds, they don’t exist for avatars mostly because the people that have avatars performing poorly enough to need LODs are also the ones least likely to implement LODs
We have other solutions on the way but its a big, long term project
nah aside from just really badly made avatars, its mostly animators and setting remote poses last I checked. mostly animators. wowee there’s really bad animators out there. wowowowo
- no plans but the feature request is tracked
- filed, tracked, but no movement, as no resources are available at the moment
- wait don’t we already cull things outside of view
I think so, yes
This is good feedback, thank you!
Still hoping you can change your mind about this. It was expressed previously that, as far as I am aware, I can:
- Install VCC
- Update/migrate a project to a new SDK version
- close VCC and never open it again
If that’s the case, then why even use VCC? As a creator who knows how to manage my packages and read error message, why can’t I just continue to download the unity package? This would also help ensure that users on MacOS and Linux can develop with no issues, since you are not developing or doing QA for those platforms.
Lastly: Another request to make the clipping point option persist between restarts, at least behind a warning like horizon adjust.
I often find myself getting frustrated when I need to open up the menu because the setting did not persist, especially if my game crashes, freezes, or otherwise screws up. It’s getting to the point where it is a habit, so the concerns of users making false reports that are due to having the feature enabled are probably gonna keep happening due to it being a habit of many users when they start up the game.
It appears that a majority of what I wrote was deleted because I either wrote too much or pressed the Quote button too many times. Awesome!
I wrote out a lot for this, but it was deleted. To shorthand: It is sad to me that the underlying data structure for VRC requires that avatars be handled in these groups.
I would like to understand why there is not a data structure change to just make avatars and worlds a long list of ID references, probably in its own table, where you can just have a “worlds” and “avatar” list that is locked under a specific number, and can be paginated through. Then from there, you can create groups like you can with friends, and just have groups limited and total avatars limited without VRC+.
This one I wrote the most for, so it’s sad that it was deleted.
Basically: While I appreciate the current optimization tips on the docs, it would probably be a good idea for VRChat to team with content creators and/or people that are good at optimization, to create an optimization guide. I personally have spent a bit of time in Blender, and while I haven’t made many avatars or worlds, I have spend countless time learning Blender and Unity from the ground up, rather than from the top down by hacking existing objects and avatars.
Easiest quick solution: Make a giant red popup in the SDK showing the avatar optimization guide and avatar ranking doc if someone’s avatar is medium or lower.
Many methods of optimizations are somewhat inherent in my workflow now, and I really do think these things are easier to learn with a good guide. I do plan to eventually try to make my own version of this as I learn more, but having proper guides to teach people to how optimize would go much farther.
The other request with this was to allow more granular avatar blocking, especially as it concerns hiding the armature. Being able to block an avatar with too many materials + meshs, or too many physbone components, would go a long way in forcing users to optimize their avatars. To be even better, while not realistic probably, it would be nice to see active stats, where a user can view the performance of their avatar based on currently activated meshes, materials, etc.
I made a general note while at an even with 30+ at a mirror once: all of their armatures and physbones were being calculated, despite their avatars being hidden. I do understand there was a mention that there may be hitching if you did disable the armatures, which is why I think you didn’t do that, but it would be nice if this was an option hidden behind a warning. I would prefer the option to disable armatures for avatars that are hidden automatically. This would be huge for large events, especially for worlds that have teleport rooms, where the rooms are physically separate as to not render many world components already.
Another general idea: While not LOD levels (since I guess this was a bad idea for some reason?), it would be cool to have a fallback version of the avatar that is not strictly quest-compatible. This would allow creators to make very basic versions of their avatars that are very performant, which can be chosen by the other players if they are at an event or something.
It is pretty frustrating going to an event and seeing people that have countless skinned meshes, few dozen materials, and dozens of physbone components. I’d prefer to have options in general to dynamically block avatars like that without needing to manually hide and unhide specific people all the time.
I think the easiest and quickest solution to this would be to have a more transparent view of user reports. Try to collect some data:
- How many reports does a user get in a single instance?
- How many reports does a user get across a certain number of instances (or also depending on instance type)?
- How many reports does a user get in a period of time (day, week, month, etc.)
You can then use this data to, while not mark someone as nuisance specifically, but to mark their account and show all the instances of reports and what they were for. This sort of reporting system would allow toxic users to be marked faster and easier. It would allow people to also set their own thresholds on auto-hiding or blocking users, as if it were another field on the safety settings (still waiting for custom presets, btw).
At least those are my thoughts. The chances of this being abused are slim, assuming the numbers are tuned correctly.
Edit: Speaking of which, by the way, it was noticed that on ToxMod’s website, VRChat was listed as a customer back in April of 2022. It apparently was removed shortly thereafter.
Can VRChat confirm if it had at any time, currently does, or plans in the future to use this software?
I just want to say: I’m sure y’all can understand how much distrust you would sow in the community if you used this, right? I would hope that if you do, it is only in public instances.
Oh yeah! I see there was already a canny made in this same vein so I will link it (Canny).
However, closer to my idea: Can we get the option to toggle chat boxes like we already can with name overlays?
I see people using a Spotify integration that someone made to be able to show the current song they are playing. While it is a cool idea, I do not appreciate seeing a bunch of text boxes, constantly activated and never vanishing, of what someone is listening to, even if it’s paused. This is distracting and clutters the screen too much. The initial tests to prevent chatbox abuse were good, but I would qualify this as a abuse as well and I would prefer to have the option to not see it.
And before you say, “We can’t do that because we want to allow users to have their messages seen”, the current default option is friends-only, so all you see as a non-friend is a red typing indicator, which is even more annoying imo.
I would prefer if you at least implemented the “hacky” solution as an option hidden behind a warning message. I personally am getting a little tired of seeing you guys simply not implement any features unless they are absolutely perfect (in your eyes). Or even added texture memory to the avatar rating, with something like:
- 50 MB - excellent
- 100 MB - good
- 150 MB - medium
- 200 MB - poor
- 300 MB - very poor
Also, what does it matter if it’s already gobbled it up? Would it not be possible to trigger garbage collection for that avatar’s data shortly after their number is calculated?
It is a bit frustrating to see VRChat max out my 8 GB of VRAM and then use anywhere from 10 to 20 GB of my system RAM while at an event with 25-35+ people. At least try to punish users that make their avatars waste so much space and are terribly performant.
Really, considering the hard cap of worlds is about 80 players (kinda), I think it stands to reason that more than 100 MB of texture memory should not be necessary (assuming 8 GB of VRAM). I know that’s now how VRAM works, but it’s at least a start.
Can’t agree more. There are some SDK2 worlds that I know that have not been updated in 2 or 3 years that still work perfectly fine. Heck, some of them look way more graphically intensive than many worlds that exist today, but are actually more performant because they were built by people that understood how Unity works.
While true, I feel like this is a silly reason. Though it goes well with the theme of “We aren’t going to implement X feature because of [arbitrary reason].”
Does it have nothing to do with the fact that armatures and physbones are being rendered even when the avatar is hidden? Either way, I still feel like VRC could implement a slew of great tools to further block poorly performing avatars (especially to prevent them from being loaded), especially if it means it is a hassle for users to show those poor avatars and encourage users to make their avatars more optimized.
You can’t keep pointing the finger at users that need to learn months worth of Unity and Blender to do good optimization for their complex avatars while doing little to actually incentivize them to do so.
Are there any plans on making the knee angle setting separate for each leg? It would be nice as i have a different skew on my left vs right foot.
google stoped support for cardboard same with samsung and there phone vr its higly unlikely and we will see vrchat support a unsupported platform
You can use Google Cardboard to talk most people out of the idea of VR.
I suspect its easier to detect a world crashing players than specific avatars. Just sayin, if you’re thinking about it.