spacediver wrote:Light23 wrote:
2nd Modified Sony laser projector: Specific Neutral Density filters are placed over the light sensors so that both projectors are 'tricked' into increasing the brightness of all lasers
heh, nice idea
Indeed that's a clever hack.
It could be of immediate benefit to anyone who owns one of these pico projectors and is willing to do even the slightest bit of modding, plus ND2 / ND4s are really cheap.
Might want to read the breakdown from Karl Guttag about the Celluon / Sony pico projector:
http://www.kguttag.com/2015/07/13/cellu ... ical-path/
I was all interested in buying it to play around but the resolution is low, and it's because the inertia of their mirror forces them to scan back and forth and it gives a kind of wavey raster pattern instead of straight lines. The effective resolution suffers because of the way the image has to be resampled due to its bowtie scanning pattern.
I was also considering buying a Sony projector and replacing the lasers with more powerful ones, but I have to ask, why stop with just changing the wavelength of the lasers to expand the gamut? Or even using ND filters to short circuit their detectors?
Just use a step up transformer and higher powered lasers with the same output signal and you're done.
Also what I wrote before is wrong apparently. The green lasers in the Sony are probably at least 50mw according to Karl Guttag. And there is a 50% lumens loss (at least) due to all the de-speckling being done optically. I believe Light123's screen could result in a double of the lumens just by skipping the despeckling steps in the design. Of course then that begs the question, why start with the Sony or Celluon to begin with, they already "solve" the speckle issue internally. Lasers that are scanned i.e. not stationary have a different regulatory power limit than stationary ones, since the light is spread out it is less dangerous and therefore higher mw lasers are allowed in consumer grade devices (hence why 50mw green lasers are used in this projector legally and still achieve class 3R certification). Karl posted a link to the reasoning behind this in the comments of his article.
But the hack to improve the color gamut, or lumens, by either replacing the lasers or increasing the power going to the current ones indirectly through the ND filters is something which many enthusiasts / tinkerers could immediately benefit from.
I encourage you, Light123, to post picks of your homebrew laser hacks on these devices and share the results here. Personally I think your screen solving the speckle issue is solving an already more or less solved problem and loses out on one of the best benefits of this projector : a focus free image allows incredible placement flexibility. And also curved screens, like curved black screens for gaming with those projection mapping middlewares would be super cool, if the laser power were increased enough.
There are just a couple aspects of the design of these devices that I fundamentally don't like, such as the inability to resolve horizontal / vertical lines without blur. Of course those issues could be fixed by increasing the internal resolution to 2x 1080p but that is such a colossal waste of pixels. Call me old fashioned but horizontal lines in a raster scanner should be perfectly straight and fully resolved. This is a fundamental design limitation of the way they move their mirrors. They would have been better trying to achieve perfect 720p resolution instead of trying to fake 1080p and not even reaching 720p. That's right, these projectors are lower than 720p in terms of resolvable detail. Still very innovative and fun toys, but not (yet) serious competition for the low end projector market.
If you boost the laser power you could have a 100-120 inch low persistence display. Boost the gamut and it could reach DCI P3 and maybe even do HDR with the right tone mapping curve.
You could do that with some kind of variable transformer with a built in curve to convert from Rec 709 to PQ gamma. HDR decoding support needs to be done at the projector side, not at the PC side. Well, actually both. You would tunnel PQ gamma encoded video over an HDMI connection (it's just bits after all), but then you need to properly map those values to voltages inside the projector.
And then there's the issue of what kinds of bit depths are supported. I wonder what these pico devices look like with Deep Color (30 bit) and showing a banding test pattern:
http://www.bealecorner.org/red/test-pat ... -16bit.png
Can you tell us? The lasers are analog so the only limit there is the bit depth of the DACs they are using to drive them (and any other signal processing, plus limitations of HDMI input bandwidth. This is why, for example, you need to drop to 422 chroma to send 10- or 12-bit at 1080p / 60 over HDMI 1.3-1.4 devices like this one probably is). But we won't know until we test.
I suspect it's 8 bit all the way because otherwise they would announce support for 10-bit or 12-bit support, wouldn't they? They did some silly mistakes in the design of these projectors, like for example the charger not charging fast enough to use the projector continuously forever without running out of power, even while plugged in. Total facepalm that one. That is probably my biggest deal breaker for these projectors. Actually, even worse is you still can't even buy the Sony here in Canada, it's so absurd / annoying. I know, I know, that's like going to a restaurant and saying the food sucks and the portions were way too small.
But I'd like to do some laser hacking too. Let us know if you can pump up the lumens from the paltry / neigh worthless 32 lumens. I have a 2000 lumen projector and it's still not bright enough for my wall sized display. I suspect the biggest problems you'll run into are power supply issues and heat dissipation.
Just sticking an ND4 might overload the power circuits, so I'm curious how much you can boost the lumens using that hack.
