Friday, January 11, 2013

- augmented reality -

After playing with the 3D TV in our manager's office, I realized I'm not a big fan of 3D. Actually, I realized that after watching a few 3D movies and playing with the 3DS. 3D stuff doesn't make me feel like I'm immersed in some virtual world - it makes me feel queasy.

On the other hand, I found augmented reality (AR) quite awesome. The 3DS comes with these augmented reality cards. You could place one on the table and it would turn into a box. And a dragon would come out and you'd have to shoot it down. Of course none of this actually happens - you see it through the 3DS.

So when my rotation with my current R&D-style team started, I made a brief mention of my interest in AR to my mentor. He mentioned while the technology is interesting, there's not much there for telcos to look into.

In the meantime, more and more stores are using AR for marketing - like American Apparel. I really wanted to check it out today, but (1) there isn't an American Apparel store nearby and (2) I don't have an iPhone.

I'm sure AR will take off in the near future (especially with Google Glass!) I can already imagine a bunch of really cool uses like

  • Being able to draw up someone's profile just by looking at them
  • RPGs where you defeat monsters in your neighbourhood
  • Blue lines leading you to your destination

To be honest, once you start merging the digital right into your vision, the possibilities are almost endless. Vision is the one sense we rely on the most, and the one sense that this world is designed for.

Wednesday, January 09, 2013

- Vocal Familiarity -

In just the last few years, you can verbally shut down Xbox Kinects and Samsung Smart TVs; you can ask your smartphone to point you to the nearest gas station is; you can compose text messages by talking it out.

The point is speech recognition is becoming more and more common. But I've found that the accuracy is still pretty bad. By "pretty bad", I mean not 99% accurate. If it's not at least that accurate, I feel that I might as well type it out. As a sidenote, I have the same issue with predictive keyboards.

Considering there are a very limited number of pronounceable sounds (phonemes) in English, shouldn't we be able to vastly improve speech recognition by asking the user to speak out every phoneme as set-up?