[mncharity]

> walking a pretty thin line just trying to not make people sick

Yes and no. There seems a lot of confusion about that. About which design constraints are coming from which goals. Yes, for the common style of immersive gaming. But when I'm working, I'm usually down around 30 fps with high variance, running on integrated graphics.

What constitutes a "horrible" "immersion-breaking" "visual artifact"? And how much will you pay to avoid it?

I'm looking at a laptop's desktop. It's obviously a display panel. And not my wooden desk. And that's fine. When looking at my HMD "desktop", it's also obviously a panel. And not my office. And that's also fine. Using emacs isn't fighting slimey zombies. Usually.

Paper novels can be immersive. Even if you sometimes notice turning a page.

The design goal "avoid reminding the user they're wearing an HMD" is a very challenging one. It prunes the design space, discouraging many things. Camera passthrough AR. Lag. Different objects having different lag. Visible boundaries in display space. And so on.

For example. If you mostly care about text, then you want to see unblurred pixels, which means only the center half of the lens-blurred display is useful. Passthrough AR can provide balance even at low fps. So it can be shown beyond the center. Which creates a visible boundary in display space. And that's fine. And if when you turn your head, some graphical elements slowly chase others across the screen, that's fine too. So now you can run on old Intel integrated graphics. It's a different point in design space from fighting zombies in a warehouse. And it has different design constraints.

So when people say "VR requires X", that's worth translating as "SteamVR games require X". And it can be fun to consider other things you can do with a tracked head mounted display.

[ss248]

>The design goal "avoid reminding the user they're wearing an HMD"

There is a difference between design decisions based on "striving to achieve presence" and design decisions based on "not make people vomit". And high fps with good tracking is closer to "not make people vomit" group.

I agree that it's okay to break presence to do something (i would trade presence for fun for example) and let user push the limit if he can handle it, but you have to understand that VR is already a niche market, so devs usually are just playing it safe.

[mncharity]

> There is a difference between design decisions based on "striving to achieve presence" and design decisions based on "not make people vomit". And high fps with good tracking is closer to "not make people vomit" group.

A goal of avoiding visible lag, requires a mechanism of predictive tracking, which has an undesired side-effect of judder, which can be far more sickening than the original lag, which motivates a mitigation of higher fps and lower variance, which requires a stronger GPU.

A goal of wide fov (absence of tunnel-vision "comfort mode"), increases user sensitivity, especially during rapid head and spatial motion, increasing needed performance and user discomfort risk.

A goal of avoiding visual artifacts, discourages introducing visual artifacts which increase user comfort. A goal of visual seamlessness, discourages segmenting the display to permit separately optimizing for task and comfort, compromising both, motivating mitigation.

With current hardware, "presence" and comfort are conflicting objectives. Optimizing for presence, is why we're pushing the envelope on comfort. But there's been a lack of community awareness of just what tradeoffs are being made, and why, and of alternatives and opportunities. In part because of the niche economics you mentioned - if SteamVR doesn't support it, why think about it? But happily, improving hardware will make the issue mostly go away.