20 Players on Quest

If players in area exceeds 8, they stay as the green unloaded character, this loading would be done locally, and by order of player loaded, if so say a group of 10 people are loading at the same time, it would be chosen randomly who gets loaded in, or potentially a priority load, with friends being the first to load.

note
I don’t actually know what’s holding back the developers from having 20 players on at a time on quest but I’m guessing it would be a performance probably. And I think this would help the game stay running well meanwhile having more then just 8 players on a server, that is if the unloaded characters don’t have much of a performance hit.

The other problem may be having the players item either not loading, or appearing and potentially causing lag

Personal Twist
When players in local area exceeds 8, everyone else’s model and items are switched to a lower level of detail to stay performing well. It might look ugly but it may just work. This would also probably need a lot of work for the devs to make lower poly models unless they already have them.

  • Addition to this idea would be to have the closest player have a high quality model, meanwhile the rest or a lot of the further ones are lower quality.

I think the quest is missing out on the community feel when there’s little people on, so I’d like to see some increase in player count, if or how the devs decide to accomplish that.

So you don’t know how code works, but you think it could work…? :thinking:
First, read this comment by SeanBannister (self-proclaimed game dev) on Reddit, regarding this exact topic of server expansion.

quote

Game developer here. There’s no hard limit on the number of players a quest game can have in a multiplayer game but there is a performance limit. For example take a look at the Performance Targets from Oculus: https://developer.oculus.com/documentation/unity/unity-perf/

let’s use triangles per frame as an example, if you were trying to hit the target of 300,000 triangles. Your scene (environment) might take up 200,000 triangles leaving you with 100,000 triangles. If each of your player models is 10,000 triangles you could have 10 players on the server at once.

This is the reason vrchat on quest displays a default avatar for most players. They don’t have enough performance to display all the complex avatars people choose.

I think it’s OK for you to let the devs solve code problems on their own schedule & brainpower. They may have OTHER reasons, such as OTHER game features that aren’t ready for massive player counts (moderation, climbing, AOE combat, for example). The public roadmap doesn’t even have a planned ‘to-do’ card regarding this subject, so they may not even be concerned with increasing the count, at least not to the level of 20 people. 16 maybe.

Not to mention this subject was addressed in 2019 in the topic Server Optimization.


This just won't work, sadly. Explanation hidden here.

Let’s pretend they put this suggested system in. They take the cap off servers, and suddenly they get SWAMPED with players. The blacksmith or carpenter is almost ALWAYS packed.
Let’s say for our example 10 people are in the carpenter building. 1 is working at the station, 1 is storing items. If I’m un-loaded player 10, I’m likely not receiving the updates made by the other 8 players, meaning our stored items could collide and crash the server, we could merge crafting station components, I could load into the space they’re standing or hand-held items could merge, fire a projectile(kunai/arrow) that doesn’t show for others, certain features might not behave correctly, trolls could abuse this in SOME fashion, so much potential for chaos.

Further, what happens for the unloaded player then, with your desire of community? Everyone else is off interacting, chatting, having a good time, and I’m a shapeless blob, possibly not even projecting audio, depending on the connection strain? At least, until 2 people leave, and my player is permitted to load. What if it’s a town event with free items, music, and a party? Does everyone after the first 8 just have to suck it up and not get to participate?? That’s not OK as a feature based on equal accessibility for all.

I think the largest issue with this concept, however, is the game will need to do ANOTHER calculation constantly: “Are there 9 > x player objects in 1 chunk?” That logic would require a good amount of code to detect 1 chunk, count players, compare vs logic, transform some players to low-quality, basically a bunch of extra work. Assuming you NEVER have more than 8 people in 1 region, you’ve now taxed on MORE code to those that gain nothing from it. That isn’t a solution, that’s a band-aid at best.

I think I have an idea of how coding works, since I’ve made like 4 unity games? But I haven’t made any multiplayer games or really optimized my games much so yeah I don’t really know, just kind of what I think it works like. I probably would like to get deep into how the devs of this game make their stuff but well I don’t have access to the project.

Responding to the part where you said it could break things heavily, regardless if it even would work correctly to how I’m thinking it, is that, actually well shit you might be kind of right. I was thinking that the items would work independently of the player, or at least show up as much to other players when unloaded. Like a a weapon that is being held, but appears floating next to the unloaded player.

Like this - image

Though I think this may cause issues for backpacks on peoples backs. I believe this would still allow the other players to receive updates from the other players, it may just look not right.

Oh and also, I think with what your saying that you’d just stay a shapeless blob until people leave, I was thinking that it would be different for everyone who sees who, that it would be calculated locally on your own game rather then server side, stopping it once it hits the cap in whatever order the players are being loaded first. And no I don’t know how the audio would strain with this change, it would really be up to how well it could handle it, and I’ve been on a pc server where the audio sucked so it may turn out similar to that. And I don’t think that people wouldn’t be able to participate in events like that, it may just turned out as a limited experience, which I wouldn’t want either.

Honestly, you make a lot of well thought out points with that, it seems like a hard concept to iron out, but that’s okay, I don’t think this concept really needs to be exactly right but you pointed out a lot of negatives that I hadn’t thought of, and I don’t like that. I feel that it’s true that this is kind of a band-aid fix. EXCEPT FOR MY LOD IDEA. I have a lot more belief in that idea working then the other one.

It would take up a lot of processing power for it to constantly check over and over again if there are 9 players in a certain chunk, which is why I think it should be placed on a when if statement, whenever a new player is loading in for a player it would check, is this over 8 people? If so, change level of quality based on such and such parameters. And if I’m correct that this is done locally to the player rather then chunk by chunk on the server side, I don’t think this would require a whole lot of code, if I’m wrong, that might end up being taxing on the server.

And I wonder, if it would be possibly to progressively increase or decrease a persons level of detail based on the performance of the game at that current moment, that might be another level of optimization regardless of expanding player counts on servers, cause damn does it still get laggy in the mines with a lot of people. Unless, they already have something like this implemented in the game? Besides world detail. They may even be able to use some of the code that changes the world detail based on distance to have this work without creating too much more code.

I understand your desire for an improved game and more players per server. I want the same thing. But you’re posting feedback on a forum, so there’s going to be discussion of some level, and some of us (hi, that’s me) are critical of ideas rather than blindly agreeing “Ya that’s awesome just do it!!”, personally because I’m familiar with the process involved of group code work and adoption of new ideas in business, because the COMPANY is going to have to do the same thing, picking it apart for any flaws or vulnerabilities, potentially for HOURS or DAYS.
So knowing that time sink and work expected, wouldn’t you want to lighten that workload any way you can? Even if it means you have to take off the rose colored glasses and instead put your trash goggles on. If an idea can’t stand up to that harshness, it’s going to be REALLY challenging to implement for the devs. And you clearly are trying to make things EASIER with a suggestion.

I would assume a company that is dedicated to VR games would likely have tested your theory of LODs on their own, during the process of trying to increase player count back in 2019, when you COULD have more than 8 people according to the post history in the linked topic. Now the game has grown MUCH more complex, and they haven’t gone back to that project of more players. There’s probably a SIGNIFICANT reason why.

I want to embrace your idea of the individual headset balancing the workload, but there are 2 implications here:

  1. The individual oculus will have to work harder, it already gets pretty warm and works hard.
  2. In order for the when if block to function, the code has now gone from individual computation per headset being casted in real time to multiple connections, to now also include each of those headsets being registered for a full-visibility queue. That’s going to require the server to communicate individually with each headset constantly, to be able to decide when the players are in 1 shared zone.

Depending on how hosting works, even if the people past 8 players aren’t in the zone, the added server stress could lag people without even getting to play around each other.

I don’t think Oculus Quest 2 has the hardware necessary. OQ3 hopefully, or maybe with a PC Tether. But even then, if someone is still on OQ2, what happens for them? They simply can’t be on a server with more than 8 people? They can’t play cross-platform? There are some hard questions here that are a bit more complex than your standard PC requirements or games released to specific consoles, given that its already out and we want to increase player count.

It would be neat to hear from the devs on the subject, especially on the topic of whether we’ll ever hit that sweet 16 cap when they focus on server optimizations. Buuuut it sounds like they have MUCH more on their plate with the other priorities

You make some fair points there, there are a lot of unknowns with how this could and would work with its flaws and pros, I would like to hear from one of the devs too on this subject. And I feel a lot more optimizations may need to be made in the game if this were to go forward. And I know that there will be criticism with whatever ideas go on here, I just think we have conflicting beliefs about it. If I were to be in the devs shoes, I think I’d want to get through the work already in front of me first, with all the bugs and such. And then consider what’s important next. And no I don’t think I’d want to make their workload light or heavier, they may work differently or even enjoy when a workload is heavier, I have no fucking clue. So I don’t really care much, the reason for me giving the idea was to give the devs an idea of how they could implement this feature if they wanted to, rather then how they should implement it.

I don’t know what you mean by a 1 shared zone

And yeah that’d feel sucky if people couldn’t play with each other with the added server stress, the quest servers may work differently or less powerful compared to the pc servers, but that’s what I was going off of.

And I thought it was the oculus quest 1 that was holding that back, since that had considerably less power compared to the 2. But I may be wrong, I do think too that if the quest 3 comes out and it’s way more powerful, it could probably handle 16-20 players, without optimization. From what I’ve heard, quest 2 can still be pushed in the graphics department, so there may be some room for things like that. I think there’d just need to be some bigger graphical changes between consoles. And some general performance improvements.

And oooh looking at that list makes me feel excited for the future, like dragon ore? That sounds awesome. I want the housing update to come out too, but with quest features missing that may be far out. It seems as though they’re able to fit smaller features into quest updates occasionally, like the more recent one, I really liked it but I didn’t like all the crashes. So depending on how much work and time this would take, this may be able to fit into their workload.

And I’ll say, with your analogy on glasses, I feel it’s much harder to look at the crappy sides of an idea, I believe it’s important to look at the full truth of things, rather then just the exciting cool stuff.

1 Like

From what i heard was that if there were more then 10 players on a quest server (just logging in, not doing anything), it would hard crash and it was toned down to 8 player, because they got to play and will cause some lag.

2 Likes

Huh, I did some research on that on discord and seems right, that the quest 1 was system that crashed hard when it goes to 10. I also read that it was a RAM limit that caused the crash.

yeah, as long as the games running on quest 1, even quest 2 to some extent, 8 players is probably the best amount of players

1 Like

Do note that they’re working with 4GB of RAM for the Oculus Quest, not to mention its CPU which is a mobile processor designed for phones in 2019.

While it would be nice if they could just double the player count of the quest platform, there are a lot more things that contribute to RAM gobbling than just player models.
On top of each player consuming a fair bit of memory by themselves, there’s also other data stuff going on like syncing their character and literally everything that’s typically on a player (backpack, tools, weapons, etc). This could include a variety of things, such as models, physics interactions, position/save data nonsense I forget the names of, updating info based on the server, etc.

It’s a miracle as-is that they’ve managed to get this game on the original Quest, but we can always hope.

2 Likes

I hate myself for even suggesting it but…

Township Tale 2…?
I hate to chop a community in half, but if there was a separate download version for OQ3, perhaps we could get past 8 people

Either that or people will just need to accept that PC VR > Oculus, on a technical level.

This idea is kind of trashed for adding more players then. Seems only useful for increasing performance, which might not even work as additional low LOD players and items may increase RAM usage. Ah well