4 ms·

I'm thinking that Hololens rev. 2 will be out before Magic Leap rev. 1 is out and will be cheaper. So unless Magic Leap is an order of magnitude better than Ho

I'm thinking that Hololens rev. 2 will be out before Magic Leap rev. 1 is out and will be cheaper. So unless Magic Leap is an order of magnitude better than Hololens, they have failed. Say what you will about Microsoft, at least they already have a platform on which this "new era of computing" can be based: Windows. That is to say, it makes tremendous more sense to evolve from what we have now into the "new era of computing", than to write an entire platform from scratch, which is what I assume Magic Leap is trying to do.

SparkyMcUnicorn 10y ago
Hololense = 1268 * 720 pixels per eye Magicleap = supposedly infinite (or they don't exist at all?) Hololens says they have 2.3 million “light points”. I don't know much about light points other than a few paragraphs, but I wonder how Magicleap actually measures light density and how/if it compares or correlates.
- bobsil1 10y ago
  MLeap is scanned, but they'll have a limit on the switching speed of the laser + color components.
  - smaddox 10y ago
    Diode lasers can easily amplitude modulate at 10s of MHz, which can provide more effective pixels than 4k displays. If they're still using the piezoelectric beam steering they demo'd a while back, then that will likely be their scanning speed limiter, and thus the pixel density limiter. Their most recent patents suggest they've moved on to something more advanced, though, so it's tough to say for sure. Regardless, GPUs will likely limit the render resolution for the first one or two products. Upscaling might be possible, though.
- white-flame 10y ago
  1920x1080 ~= 2 million pixels. Assuming an optical scanning system, you could give them the benefit of the doubt that they can route R, G, and B through a single light point during scanning. I presume a "light point" is some sort of shaped structure on their glass which allows the scanned light to escape to the eye when requested. And in theory, might be a place where they can do interference cancellation of incoming light to darken points. All their hype probably points to metamaterials-style shaping of light. Remember that pixels aren't rectangles (hence why scaling images larger always looks more blocky or blurry than native res), they're point samples. So a "light point" should be equivalent to a pixel in the simple case. Whether or not they can bend light between points to make it "infinite", who knows. Any of that is always going to be limited by the resolution that the GPU can crunch in time.
carterehsmith 10y ago
So it does seem that investors are impressed with the Magic Leap. Just how impressed are they? Let's see. Magic Leap = 4.5bn valuation Snapchat = 20bn valuation So if, as the article states, ML could disrupt the $160bn LED screen market (among other markets), the valuation would reflect that. Presently, it does not.
nirav72 10y ago
I think Microsoft just needs to make the FoV in rev. 2 better than right now and they'll have a decent product.

SparkyMcUnicorn 10y ago

Hololense = 1268 * 720 pixels per eye Magicleap = supposedly infinite (or they don't exist at all?) Hololens says they have 2.3 million “light points”. I don't know much about light points other than a few paragraphs, but I wonder how Magicleap actually measures light density and how/if it compares or correlates.

bobsil1 10y ago
MLeap is scanned, but they'll have a limit on the switching speed of the laser + color components.
- smaddox 10y ago
  Diode lasers can easily amplitude modulate at 10s of MHz, which can provide more effective pixels than 4k displays. If they're still using the piezoelectric beam steering they demo'd a while back, then that will likely be their scanning speed limiter, and thus the pixel density limiter. Their most recent patents suggest they've moved on to something more advanced, though, so it's tough to say for sure. Regardless, GPUs will likely limit the render resolution for the first one or two products. Upscaling might be possible, though.
white-flame 10y ago
1920x1080 ~= 2 million pixels. Assuming an optical scanning system, you could give them the benefit of the doubt that they can route R, G, and B through a single light point during scanning. I presume a "light point" is some sort of shaped structure on their glass which allows the scanned light to escape to the eye when requested. And in theory, might be a place where they can do interference cancellation of incoming light to darken points. All their hype probably points to metamaterials-style shaping of light. Remember that pixels aren't rectangles (hence why scaling images larger always looks more blocky or blurry than native res), they're point samples. So a "light point" should be equivalent to a pixel in the simple case. Whether or not they can bend light between points to make it "infinite", who knows. Any of that is always going to be limited by the resolution that the GPU can crunch in time.

carterehsmith 10y ago

So it does seem that investors are impressed with the Magic Leap. Just how impressed are they? Let's see. Magic Leap = 4.5bn valuation Snapchat = 20bn valuation So if, as the article states, ML could disrupt the $160bn LED screen market (among other markets), the valuation would reflect that. Presently, it does not.

nirav72 10y ago

I think Microsoft just needs to make the FoV in rev. 2 better than right now and they'll have a decent product.