We should use the appropriate terminology: light fields. A traditional camera captures a point sample of a light field: you see a correct image from 1 point of view. A hologram captures a 2D planar sample of a light field: you see a correct image from any perspective that looks through the film (as if it's a window). To capture a volume sample of a lightfield is not really possible (at least, not at the same instant in time), since that requires having a sensor be placed everywhere in the volume, and of course the sensor itself interferes with the light samples that it's taking.
They just need to figure out a way such that only properly GM-licensed car mats, seat covers, etc. will work on the Volt.:-)
Cars actually offer tons of revenue streams (accessories, service/maintenance, financing, insurance, gasoline, etc.). The problem is that most of these have numerous providers. They're doing all they can to make service/maintenance first-party only, though...
The reason the US doesn't have a system like Isreal's is because they've taken a systematic look at the problem and have implemented a comprehensive, multilayered, efficient solution. In the US, we prefer one-step, silver-bullet type "fixes". Anything more complex would be argued out of existence.
Because news sites *want* search engines like Google to see their content, so that people searching for stuff will be directed to them. And they want the people following links from Google to come back. So they try follow the drug dealer's model: we'll give you a bit for free, so that you'll come back and pay for more later. Of course, smart people figure out how to not pay ever, but that's only a small percentage of viewers.
Old plaster walls often use wire mesh as a support material. Some reinforced concrete does also.
Look at the bright side: it keeps your wifi secure from your neighbors! (Or: it encourages you to use a fast, reliable wired network instead of a slower, flakier wireless one.)
The zBoost repeaters work pretty well. I got an YX-500 from Ebay for $100, and it works nicely for T-mobile. (Be sure to check that the frequencies supported by a given model match the ones used by your service provider.) Proper setup is important: the antenna to the cell tower must get an adequate signal, and it must be a minimum distance away from the repeater. The zBoost is nice in that you can use regular TV coax cable to connect the external antenna to the repeater.
In the building where I am, the cement is reinforced with wire mesh, and thus I could only get a signal standing near a window. With the repeater, I don't have that problem anymore.
It's explained to some extent in PrimeSense's patent: WIPO Patent WO/2007/043036.
The processing chip knows the position of each dot as projected onto a plane.
Where the IR camera sees each dot depends upon what it (the dot) projects onto (due to the offset from the IR laser). For any given dot, it's just a matter of triangulation to find the depth of it.
The trick, however, is in identifying the dots it's looking at. For this, the processor has to look at the relative dot positions and try to match areas of dots to areas within its stored pattern.
This is also why the resolution is limited: it needs more than a couple dots projected onto the same surface in order to be able to recognize the pattern.
Indeed I have. So have many others. You can search for "ascension flock of birds" and see lots of related research. For a limited volume (that's relatively free of ferrous objects), it's okay. It tends to jitter & distort towards the limits of its range (at least the model I used). I imagine that optical trackers are better for bigger volumes, assuming you can string up an appropriate constellation of LEDs (for outward looking) or cameras (for inward looking). You could probably do something quite decent with enough Wii remotes and LEDs, if cost is a concern.
It depends upon what head-tracking you're talking about. The Polhemus and Ascension magnetic trackers have allowed this kind of thing for ages, and are available for under $10K (sometimes you see them on Ebay for a pittance). In addition, various optical (OptiTrack) and ultrasonic (Intersense) trackers have also had this capability.
And let's not forget mechanical tracking, as pioneered by Ivan Sutherland back in 1963. True, range was an issue, but the main idea was already present.
Way back when, there were these very exclusive devices known as video terminals. There were many different kinds of different capabilities, but the basic idea is that they could interpret commands from a serial port (running at up to 9600 baud if you're lucky) and draw points and lines on the screen. Some early models used a video storage tube that didn't need to be continually refreshed, but it had to be erased by erasing the entire screen at once. They were monochrome (green). Eventually, models incorporating huge amounts of RAM (like several K-bytes) came out and supported raster-scan graphics, with 16 or more colors. Eventually, they got really fancy and could draw using a palette of 256 colors out of a possible 16 million. Someone figured out you could hook three of these together to achieve 24-bit full color output, but this cost tens of thousands of dollars, typically (heck, the big color CRT alone would go for about $10K).
Of course, prior to such advanced technology, people were still doing computer graphics. For instance, you could hook up a D/A controller to an oscilloscope and control a dot of light on a small screen. By putting up a color filter and a camera back on the screen, you could draw a fancy image, one color layer at a time. You'd have to do that and then develop the film, of course, before you could see anything.
In such times, you would write your code carefully (if you were lucky, you had a printing terminal with a line editor; not so lucky meant using a card puncher to code a card at a time). You'd then run it and wait to see what the final output would be: no preview of any kind, except what you drew on paper.
is what Simul-cam sounds like. However, the technique from the article is a bit different, in that it lets a hand-held device to act as a camera controller within a virtual environment. Nothing particularly new there either. Previously, the tracking device on the user's head performed such a function.
This is why the device is "optimized" to work within a given depth range. Note that the depth cam looks to be fixed focus, not auto-focus as the summary says.
I'd like to see the results of someone operating the device in a dark room and looking at the results through a high-speed camera with good IR sensitivity.
Though I must admit: the rate of improvement of many technologies does seem to have increased much in recent years. Even gas engines, which have been around since long before the Model T, seem to be getting much more powerful (per liter) recently compared to when I started driving. Still, I wonder where we'd be now if we'd started with electric instead of gas. Heck, where would we be if GM hadn't killed the EV1, for that matter?
Dual boots? What a pain. Even two VM's would be unwieldy. Let me know when there's an OS that gives full API support to either set of apps (like Linux with Wine, but better).
It maintained its altitude until they ran out of field. Not bad for a first flight. Give them time to refine it further.
I was really impressed by how graceful the wing flapping looked. It's not just a simple up-down motion of the ends, but a more complex wave-like motion over most of the wingspan.
The Move has 4 sensors: 1) Eye camera sensing the glowing ball. 2) accelerometer 3) gyroscope 4) magnetometer (3D compass)
It is the combination of all 4 of these that allows accurate tracking. There is no single sensor answer that gives a decent solution to the general tracking problem.
Should be possible using an off-the-shelf HDMI receiver chip, an off-the-shelf DVI transmitter chip, and a microcontroller. Just reroute the incoming DDC line from the HDMI port to the microcontroller and let it perform the HDCP authentication (instead of having the receiver chip do it). Then just hook up the data lines from the receiver to the transmitter, and let the microcontroller coordinate all the necessary setup.
I believe that all you need is a Black Magic Intensity capture device, a small microcontroller to intercept the DDC (I2C bus of HDMI), and a PC. The microcontroller would perform the necessary authentication with the source device, which would then pass the encrypted bitstream over HDMI. The PC would receive the decryption keys from the microcontroller and decrypt the captured bitstream.
Slashdot Mirror
We should use the appropriate terminology: light fields. A traditional camera captures a point sample of a light field: you see a correct image from 1 point of view. A hologram captures a 2D planar sample of a light field: you see a correct image from any perspective that looks through the film (as if it's a window). To capture a volume sample of a lightfield is not really possible (at least, not at the same instant in time), since that requires having a sensor be placed everywhere in the volume, and of course the sensor itself interferes with the light samples that it's taking.
Just don't fly into any nano-particle fields, or your ship may be destroyed too!
They just need to figure out a way such that only properly GM-licensed car mats, seat covers, etc. will work on the Volt. :-)
Cars actually offer tons of revenue streams (accessories, service/maintenance, financing, insurance, gasoline, etc.). The problem is that most of these have numerous providers. They're doing all they can to make service/maintenance first-party only, though...
Article that describes some of the Israeli methods: http://www.thestar.com/news/world/article/744199---israelification-high-security-little-bother
The reason the US doesn't have a system like Isreal's is because they've taken a systematic look at the problem and have implemented a comprehensive, multilayered, efficient solution. In the US, we prefer one-step, silver-bullet type "fixes". Anything more complex would be argued out of existence.
Isn't an antenna itself an amplifier of sorts? After all, what does "gain" mean re: high-gain antenna?
The idea is to have a high-gain antenna directed towards a cell tower connected to a moderate antenna that covers your living space.
Because news sites *want* search engines like Google to see their content, so that people searching for stuff will be directed to them. And they want the people following links from Google to come back. So they try follow the drug dealer's model: we'll give you a bit for free, so that you'll come back and pay for more later. Of course, smart people figure out how to not pay ever, but that's only a small percentage of viewers.
Old plaster walls often use wire mesh as a support material. Some reinforced concrete does also.
Look at the bright side: it keeps your wifi secure from your neighbors! (Or: it encourages you to use a fast, reliable wired network instead of a slower, flakier wireless one.)
The zBoost repeaters work pretty well. I got an YX-500 from Ebay for $100, and it works nicely for T-mobile. (Be sure to check that the frequencies supported by a given model match the ones used by your service provider.) Proper setup is important: the antenna to the cell tower must get an adequate signal, and it must be a minimum distance away from the repeater. The zBoost is nice in that you can use regular TV coax cable to connect the external antenna to the repeater.
In the building where I am, the cement is reinforced with wire mesh, and thus I could only get a signal standing near a window. With the repeater, I don't have that problem anymore.
It's explained to some extent in PrimeSense's patent: WIPO Patent WO/2007/043036.
The processing chip knows the position of each dot as projected onto a plane.
Where the IR camera sees each dot depends upon what it (the dot) projects onto (due to the offset from the IR laser). For any given dot, it's just a matter of triangulation to find the depth of it.
The trick, however, is in identifying the dots it's looking at. For this, the processor has to look at the relative dot positions and try to match areas of dots to areas within its stored pattern.
This is also why the resolution is limited: it needs more than a couple dots projected onto the same surface in order to be able to recognize the pattern.
Indeed I have. So have many others. You can search for "ascension flock of birds" and see lots of related research.
For a limited volume (that's relatively free of ferrous objects), it's okay. It tends to jitter & distort towards the limits of its range
(at least the model I used). I imagine that optical trackers are better for bigger volumes, assuming you can string up an
appropriate constellation of LEDs (for outward looking) or cameras (for inward looking). You could probably do something
quite decent with enough Wii remotes and LEDs, if cost is a concern.
It depends upon what head-tracking you're talking about. The Polhemus and Ascension magnetic trackers have allowed this kind of thing for ages, and are available for under $10K (sometimes you see them on Ebay for a pittance). In addition, various optical (OptiTrack) and ultrasonic (Intersense) trackers have also had this capability.
And let's not forget mechanical tracking, as pioneered by Ivan Sutherland back in 1963. True, range was an issue, but the main idea was already present.
Way back when, there were these very exclusive devices known as video terminals. There were many different kinds of different capabilities, but the basic idea is that they could interpret commands from a serial port (running at up to 9600 baud if you're lucky) and draw points and lines on the screen. Some early models used a video storage tube that didn't need to be continually refreshed, but it had to be erased by erasing the entire screen at once. They were monochrome (green). Eventually, models incorporating huge amounts of RAM (like several K-bytes) came out and supported raster-scan graphics, with 16 or more colors. Eventually, they got really fancy and could draw using a palette of 256 colors out of a possible 16 million. Someone figured out you could hook three of these together to achieve 24-bit full color output, but this cost tens of thousands of dollars, typically (heck, the big color CRT alone would go for about $10K).
Of course, prior to such advanced technology, people were still doing computer graphics. For instance, you could hook up a D/A controller to an oscilloscope and control a dot of light on a small screen. By putting up a color filter and a camera back on the screen, you could draw a fancy image, one color layer at a time. You'd have to do that and then develop the film, of course, before you could see anything.
In such times, you would write your code carefully (if you were lucky, you had a printing terminal with a line editor; not so lucky meant using a card puncher to code a card at a time). You'd then run it and wait to see what the final output would be: no preview of any kind, except what you drew on paper.
is what Simul-cam sounds like. However, the technique from the article is a bit different, in that it lets a hand-held device to act as a camera controller within a virtual environment. Nothing particularly new there either. Previously, the tracking device on the user's head performed such a function.
Love to. How about Global Cyber Internet War?
There are multiple ways to use cameras to detect depth. Both your summary and the above summary are possible, as well as other methods.
This is why the device is "optimized" to work within a given depth range. Note that the depth cam looks to be fixed focus, not auto-focus as the summary says.
I'd like to see the results of someone operating the device in a dark room and looking at the results through a high-speed camera with good IR sensitivity.
the Model T were electric.
Though I must admit: the rate of improvement of many technologies does seem to have increased much in recent years. Even gas engines, which have been around since long before the Model T, seem to be getting much more powerful (per liter) recently compared to when I started driving. Still, I wonder where we'd be now if we'd started with electric instead of gas. Heck, where would we be if GM hadn't killed the EV1, for that matter?
The guy who sells their advertising?
Dual boots? What a pain. Even two VM's would be unwieldy. Let me know when there's an OS that gives full API support to either set of apps (like Linux with Wine, but better).
"It's got a more powerful battery. This new model will *fly* !!!"
It maintained its altitude until they ran out of field. Not bad for a first flight. Give them time to refine it further.
I was really impressed by how graceful the wing flapping looked. It's not just a simple up-down motion of the ends, but a more complex wave-like motion over most of the wingspan.
The Move has 4 sensors:
1) Eye camera sensing the glowing ball.
2) accelerometer
3) gyroscope
4) magnetometer (3D compass)
It is the combination of all 4 of these that allows accurate tracking. There is no single sensor answer that gives a decent solution to the general tracking problem.
Should be possible using an off-the-shelf HDMI receiver chip, an off-the-shelf DVI transmitter chip, and a microcontroller. Just reroute the incoming DDC line from the HDMI port to the microcontroller and let it perform the HDCP authentication (instead of having the receiver chip do it). Then just hook up the data lines from the receiver to the transmitter, and let the microcontroller coordinate all the necessary setup.
I believe that all you need is a Black Magic Intensity capture device, a small microcontroller to intercept the DDC (I2C bus of HDMI), and a PC. The microcontroller would perform the necessary authentication with the source device, which would then pass the encrypted bitstream over HDMI. The PC would receive the decryption keys from the microcontroller and decrypt the captured bitstream.