Hacker News Reader

Comment by nobody271 | Hacker News Reader

The main problem with web browsing in VR is it's way too low resolution to read text and you can't see your hands so you have to type everything in with that remote character by bloody freaking character.

But I guess, if you want to browse the web in VR you might as well do it in Firefox.

View on HN

Sorry to be dense, but why would one want do to web browsing in VR at all? It's all built on a metaphor of flat sheets of paper, so it's paradigmatically 2D.

Privacy.

Or simply use a Laptop Sock:

https://www.flickr.com/photos/bekathwia/2462986853

This was a genuinely interesting rabbit hole. Never seen this before. Thanks.

For the same reason I alt-tab out of a game to check something.

Searching for VR porn?

I want to be working in VR. Once the resolution gets decent enough I can just get 20 screens wherever I look. Plus I can sit, stand, whatever. Sounds fantastic. Coffee might get a bit cumbersome though.

A straw, maybe?

Yeah me too, but I guess it will take a long time before we have headsets that can reproduce the same experience as high resolution screens.

We just need one 8k screen per eye, which will be doable as soon as 8k smartphones are on the market.

Sadly having an 8k screen on a smartphone seems mostly pointless, and achieving the necessary economies of scale is hard with VR headsets alone.

My biggest hope is a paradigm shift in UI design. Everything we have is designed to accommodate low resolution screens. With ubiquitous high resolution screens UI patterns from print like "lean in to read this detail" become viable, and if popular would justify much higher display resolutions, directly benefiting VR.

Maybe vector-based CRTs like in oscilloscopes can be made small enough to use in headsets (if they haven't already). Either project floating text on transparent AR glasses or project a text layer on the lower resolution graphics display. I'd prefer the AR approach if it means I won't need a $500 GPU to generate floating text.

Put trackers on your hands and coffee mug and you should be able to drink coffee normally with a bit of practice.

I'm mostly worried about the goggles being in the way

This is my ideal for VR. I'm really hanging out for the day I can create infinite amount of virtual screens, and reconfigure screen layouts on the fly.

Throw in a camera so you can toggle AR mode when needed (coffee), and you have my dream workstation.

> Once the resolution gets decent enough I can just get 20 screens wherever I look.

Isn't that stressful on your neck? In the current environment you can also have 20 screens and switch via key-combinations or using the mouse.

I think this problem is solved already, see this video about TrackIR from 1:00 https://www.youtube.com/watch?v=_AO0F5sLdVM (it is called motion scaling). Of course people may still get injuries from this if they do it for 40 hours a week...

In VR that would likely make you violently ill

I think my ideal VR workspace would be one screen directly in front of me, and then as many screens as I want off in the distance.

You'd be able to easily see the screens in the distance and look for updates/etc and just click on it or whatever to pull that screen closer to you if you want to actually read on that screen.

Maybe it could improve ergonomics? A lot of people suffer injury from the posture required to sit at a desk looking at a monitor all day. A sufficiently high resolution VR display could be combined with a reclined posture that supports the head, neck, and back. Of course there's the whole input method problem to solve too. :)

Here's a VR text-input concept by Elevr. Using three dimensions you only need one hop from the center of a 3x3x3 cube to hold 26 letters, plus a space in the middle. http://elevr.com/keycube-3-torus-based-typing-in-vr/#h.a9zrf... (This shows the setup but it's not the final version, because eventually they realized that returning your cursor to the center after each character works better with muscle memory.) So use various controller axes to indicate positive-neutral-negative for three axes, and you have a letter. If you use chording you can type reasonably fast. There is quite a learning curve though.

I'm open to new ideas but I can't imagine this being a very efficient way to input words and characters into a computer. I have worked with older game consoles and remote controls that tried to do this stuff and I'm not sure anyone likes it.

Why not simply speech to text? I don't think a lot of people are doing VR in a room with other people.

The article states the firefox reality actually supports tts.

Just because people don't like it initially, doesn't mean it can't become a fairly fast input method given enough practice. Here is another example of a keyboard in much the same vein as using a controller, for people with carpal tunnel syndrome: https://orbitouch.com

> you can't see your hands so you have to type everything in with that remote character by bloody freaking character.

Learn to touch type! (Though even then, you're stuck with the problem of finding your keyboard.)

Logitech was working on a keyboard with a steamvr tracking puck

https://uploadvr.com/vr-keyboard-logitech-vive-tacker/

That announcement was nearly a year ago and nothing but crickets since. It doesn't seem that technically challenging. I wonder what the hold up is?

Driving something like Dasher (http://www.inference.org.uk/dasher/) with the remote would be easy enough.

(Typing on a keyboard would still work better, if you can do so without looking.)

The issues with text input are a large part of why we added voice input even in our very first release! w.r.t. text rendering, we're doing a lot of fundamental work in rendering higher-quality text (you can't fix resolution fundamentals, but you can take into account the optics of the hardware better than we all do today), but nothing to announce there yet.

What is doing the speech-to-text processing? Is this part of the Common Voice project? Is the code available to use outside of Firefox Reality?

I checked and it's using Mozilla's DeepSpeech library: https://github.com/MozillaReality/FirefoxReality/blob/dee6f4...

which is here: https://github.com/mozilla/DeepSpeech

DeepSpeech is the actual voice recognition library, CommonVoice is a project to improve its accuracy by gathering more voice data.

Interesting. The DeepSpeech github indicates that it's not capable of realtime transcription, even using a GPU. How is that fast enough for this use case?

I guess with some kind of haptic or kinesthetic gloves you will not need to see your hands to type anyway.

Steam's VR home or whatever they call it made great use of the touchpad by treating them as two fingers on a virtual keyboard. It was effortless to learn and rather quick.

Too bad Steam VR's desktop is awful. I have 3 monitors side by side, and Steam shows your desktop at a constant width, which means my display is super super tiny. Can't even zoom in, and to left click, you have to keep the controller super constant or it registers as a drag.

> way too low resolution to read text

Latest Vive should be decent.

Reply on news.ycombinator.com