Figured I should mention that Roblox had a Oculus Rift demo at RDC 2016, and the laptops were only GTX 980Ms. Considering the other amount of optimizations that have happened since then, my guess is this is less of a concern, in terms of multiple perspectives. Obvious problem is still mobile, which sometimes have problems with the single perspective like the iPhone 4S.
RDC is not a good sample. The main attendees were developers – the average audience isn’t going to have decent hardware, even on PC. Even with optimizations this feature would only be performant enough on a low proportion of user hardware, and even then its use would be limited (e.g. a $2000 PC can’t comfortably run 4 Vives at once)
1 Like
There could be some rendering limits which would depend on device performance. Oh and I am sure too that it has been improved / optimalized to the present.
You are correct about that. VR’s two smaller-than-PC-monitor views that don’t depend on each other is a different beast from what the OP is asking for. Things like mirrors, security cameras on GUIs, etc. imply first rendering views to textures, which are then used in the final main scene render, much like any other texture or decal. Totally different problem, as expensive as it sounds.
2 Likes
Support!
I am talking about the splitted screen too… The Camera Instance idea is just a improved and way more complicated version of that. I know it would be very difficult to optimize. (editted the post to be more clear)
So just to be clear, we should never expect to see split-screen capability and RT cameras & in-game screens? Even if Roblox was able to make poor-quality versions that don’t eat up too much power?
Split screen might be feasible – ROBLOX already supports multiple controllers, so this might even be on the roadmap.
This, on the other hand, is more complicated. Other games have similar features: Bethesda games, for instance, uses portals to connect two interior logical rooms which show the other side even though it’s not truly loaded, so that’s kind of like what’s being requested.
These generally come with constraints though, and are specific to the game. In Bethesda’s case, they suffer massive performance losses when you see a long tunnel of portals. The reason why you see constant stairs/turns right after rooms in Skyrim is partially for performance reasons. It prevents multiple portals from being viewed at once.
We have some really talented engineers at ROBLOX, so I wouldn’t say the answer is “never”. I’m sure they could implement something that would address user needs, but the major challenges are priority and design.
Design:
It’s simply not feasible to render many portals at once, so we have to constrain it somehow. The question is determining how we can constraint it that inhibits generic use as little as possible. We can’t do what Bethesda does because that isn’t applicable to other use cases. Determining how else we can constrain portals would require a lot of time decide.
Priority:
Split screen is a major use case, but it can be implemented in a unique way that doesn’t have the same problems a generic solution would. Past that, most of the use cases that have been presented for this feature are "nice-to-have"s, rather than critical needs. There are other features with higher priority, so this can’t happen until engineers have free time (never going to happen except during Hack Week) or we see top games finding critical uses for it.
1 Like
I would definitely like to have a GUI camera option. It would make security cameras so much better in-game.
2 Likes
I commented this exact thing, then saw your comment. xD
I am saying mainly about splitted screen feature, but with ability to set the position and size of “second view”.
As has already been described multiple times, the same arguments apply
- Takes up too much processing power
- ^ (Hence not worth adding - if 90% of users can’t utilise it then the feature may actually be a hindrance considering it may mean 90% of players can’t play certain games)
- Isn’t in a great demand
- Isn’t exactly possible with the current rendering engine - at least, not at levels efficient enough to be considered.
It was done in the hack week so I’m surprised that Roblox thought of this actually. 
1 Like
Hack week is hack week. Don’t take it as a promise for the future or a promise of a feature being feasible on millions of different devices across multiple platforms.
3 Likes
My exact arguments still stand.
I know about still. Still glad about it’s in hack week at least! 
Update on thread:
Released as viewportframes (third idea of mine).
Yeah but, this isn’t the same at all.
- It is localised to single models
- It is heavily drawn back rendering (for the reasons I described above)
1 Like
I’m going to clear the solution on this thread because a feature request is for a feature that all developers will use, not necessarily just the poster, and I don’t think ViewportFrame satisfies the split screen use cases properly.
ViewportFrame might be a workaround for now, but I don’t think most developers will be satisfied with this solution since it results in lower rendering quality, is inconvenient to use for this purpose, and is not really a performance-friendly way to pull this off, since you still draw a scene behind the two ViewportFrames.
5 Likes
Personally, I am not 100% satisfied with viewportframes for this purpose so I appreciate the change.
5 Likes