MIT Shows Trillion-Frames-per-Second Video
Researchers at the Massachusetts Institute of Technology (MIT) have developed a fascinating streak camera system that can take picture of light particles moving through space and create and "ultra slow motion movie" from those images.
The novel camera takes pictures at a rate of one trillion frames per second. In an example, the MIT researchers showed light moving through a soda bottle.
To create the movie, the $250,000 camera system as well as the laser emitting the photons produce "hundreds of thousands" of data sets that provide information about the positions of photons as well as their times of arrival. The data is then stitched together to create a movie that expands a process that takes only about 1 nanosecond in real time, but is stretched to about six seconds in the slow-mo. The effort of repetitive picture taking and combining them into a video takes about one hour, the researchers said.
The intriguing part of the invention may be two-fold: Scientists working with light could get a much better opportunity to monitor photons. However, the MIT researchers also said that the ultra-slow motion technology will make its way to the consumer at some point.
- Kindle Fire Tablet Getting Major UI Update Soon
- Verizon May Simply Purchase Netflix Rather Than Compete
- FBI May Be Using Carrier IQ Too; Carrier IQ Issues Report
- Costco and Verizon Point to Dec. 15 Arrival for Galaxy Nexus
- Verizon May Soon Get a 4G Xperia PLAY Too
- Modern Warfare 3 Beats Avatar: Passes $1 Billion in 16 Days
- Kindle Touch Jailbroken With Just an MP3
- Online Holiday Spending in U.S. is Nearing $25 Billion
- Unmanned Drones Are Already Being Used in the U.S.
- Facebook is Unequivocally the Most Popular Android App
- Nokia Says That Young People Are Bored of the iPhone
- Security Firm Questions Google Wallet Data Storage
- Cequint Sues Apple Over Advanced Caller ID
- Google's Answer to Siri Codenamed 'Majel' After Star Trek?
- Facebook Launches Tool for Reporting Suicidal Behavior
- Safety Board Recommends Total Cell Phone Ban for Drivers
- Nokia's First U.S. Windows Phone Lumia 710 for T-Mobile
- MS: Free Windows Phones for Disgruntled Android Users
- Sony, Universal, Fox Caught Pirating via BitTorrent
interesting, even if same method is actually used in DAQ when sampling fast signals for quite some time. many digital oscilloscopes do the same (sample same signal using different ADCs that are sampled with slight delay, then data is stitched together and displayed).
Wonder what they are using to render the 6 second movie.
Life is good...
The applications of this may be unlimited, but the price is way beyond.
This is useless in case of making movies because our eyes can't detect the difference after 60fps or so. But it may be used to make the coolest super ultra slow videos of the fastest processes in the world.
This is useless in case of making movies because our eyes can't detect the difference after 60fps or so. But it may be used to make the coolest super ultra slow videos of the fastest processes in the world.
In general I agree with your comment. But for very fast moving objects or cameras on video even 240 fps could make a enjoyable difference we can perceive. However it will pale in comparison to a trillion fps, so your comment is still good by a solid margin. After 240 fps i am not sure if we could have another useless "fps" marketing war from monitor and tv makers similar to the dynamic contrast ratio.
As for slow motion I think Hollywood could still find ways to produce some effects with a camera like this, even if this technology is used as a way to sale movies.
@theuniquegamer, it is made for analysis of physical phenomenon and volumes, etc. It isn't made for slow motion neo effects, but the sky is the limit using 1T scans that are broken up into scanline compositions. Imagine something an pop, there's a new idea. The scientist was just throwing ideas out there, some pretty good ones too especially for 3-8 seconds snippets of an interview.
Boy do those two sound clever!
This is seriously awesome work, MIT does have some of the brightest minds on Earth!
I can see where this will go, especially in the Medical World as he says, I think this will be the first point of call.
This technology is not about gaming / fps. Geez! Its the ability to ACTUALLY see light **move**. Light is not instant, but its very very fast - faster than the fastest rocket will ever get.
This could open to the door to other types of tech that could use light. Think back to the the 1970s with the first LED dispay used on huge calculators (way bigger than our phones) - then about 1980, the LCD display which was only black on a silver surface (I actually have mine handy - its quite dead) for history reasons. Fast forward 30 years and we have TVs that combine LCD and LED technology to make killer 62" 3D TVs.
For scientist, this is about being able to observe - and from there who knows. Maybe we'll have our light-sabers for cutting butter.
Since light behaves as a wave and particle and creates interference patterns when shot through two parallel pinholes with some odd quantum behavior thrown in, could this then be observed with this method of video capture? It would be interesting to see.
The only problem with this is that they appear to have repeated the pulse multiple times to generate the video for multiple scan lines. Their camera takes pictures of a line. So to make an image, they had to use that slow-moving mirror to gather multiple lines like a raster-scan, sending out a duplicate pulse for each scan line. That won't work if you can't generate multiple equal events, such as trying to record the effects of an explosion for a film. So they need to find a way to put these cameras into a 2D (or 3D, if it's Hollywood) array instead of a line.
Also, it isn't taking pictures of photons moving through space. It's taking pictures of photons getting reflected towards the cameras from surfaces. That makes me wonder what would happen as the photographed area got deeper; the photons bouncing off the farther-away surfaces would take longer to reach the camera, and so would appear to trail the photons bouncing off surfaces closer to the camera.
The applications of this may be unlimited, but the price is way beyond.
Actuially I am impressed that the camera costs so little for what it is and does. No doubt a calibrated camera that takes less 'streakey' pictures will cost much more, but this is a very impressive prototype/proof of concept!
Very cool MIT, Very cool.
*hear the sirens calling* My eyes are sensitive enough to tell the difference between 120hz and 60hz. Anyone saying they can't probably has never seen higher than 60hz.
The applications of this may be unlimited, but the price is way beyond.
250K is actually very cheap in the world of university projects. Hell mine spent dozens of times more than that on that putting marble benches all around campus. Why? I dunno. So they can claim they have no money and need to raise tuition 15%?
Interesting, but does it require repeating the same event to take multiple pictures? There are already cameras being manufactured that take 1 billion fps of a single event, but I'm sure this opens a whole new level of problems with the physical limits of the universe. Maybe they can partner with the photonic chip people.
I am glad you think so.
How will this be used to further advance pr0n?
@del35, uangrybro?
This is useless in case of making movies because our eyes can't detect the difference after 60fps or so. But it may be used to make the coolest super ultra slow videos of the fastest processes in the world.
Movies are not presently the intended application of this technology. 1-trillion frames, i.e., 1,000,000,000,000 frames played at 60 frames per second would last over 11.574 million hours. The most practical application of this technology are processes that can be repeated and have extremely short durations like chemical reactions.
Here's a much better version of the article.
1 Trillion frames played at 30 frames per second would last over 1050 years. that's a lot of harddrive space.
1 Trillion frames played at 30 frames per second would last over 1050 years. that's a lot of harddrive space.
I completely agree....doing some math:
1920x1080x12(average of 12 bpp in 1080p video) / 8 (byte) / 1024 (kb) / 1024 (mb) = 2.966 mbytes per frame of uncompressed data.
Now multiple by 1 trillion and divide by 1024 to get terabytes = 2,896,785,736 gb PER SECOND of video....need a new hard drive...
where's the damb edit option!
I meant divide by 1024 twice to get tera bytes....I did list it as close to 3 billion gb but wrote terabytes before...in tb 2,828,892 terabytes per second of video.
This will be great for viewing on my 60 hertz monitor...
In general I agree with your comment. But for very fast moving objects or cameras on video even 240 fps could make a enjoyable difference we can perceive. However it will pale in comparison to a trillion fps, so your comment is still good by a solid margin. After 240 fps i am not sure if we could have another useless "fps" marketing war from monitor and tv makers similar to the dynamic contrast ratio.As for slow motion I think Hollywood could still find ways to produce some effects with a camera like this, even if this technology is used as a way to sale movies.
I wonder if animals with faster optic nerves are fooled with a screen that utilizes a much higher refresh rate. So for instance, would a hawk be fooled into believing there is a real mouse on a screen?
*hear the sirens calling* My eyes are sensitive enough to tell the difference between 120hz and 60hz. Anyone saying they can't probably has never seen higher than 60hz.
Technically, your eyes can throughput that. Its our brains that are too slow to interpret that data I think. Someone correct me if I'm wrong?
Too bad they can't film at 1000 trillion frames per second. If they actually could they might see that some of the frames show nothing.
Now we just need a camera that can do full frame 4K video at 1 trillion FPS.
Imagine getting cool cat videos at 1 trillion frames per second. You would be able to do awesome slow motion shots as well as great high speed shots with no motion blur.
Since light behaves as a wave and particle and creates interference patterns when shot through two parallel pinholes with some odd quantum behavior thrown in, could this then be observed with this method of video capture? It would be interesting to see.
From the video, it did not see actual photons (probably need a significantly faster frame rate, but the movement of it seems to be a motion blur of photons, or not a high enough resolution since the photons will be in the atom range)
so far this is mainly a good proof of concept of what high speed cameras can do but has no true practical uses as of yep. All in all it is still a great discovery from a technology standpoint and will most likely make it's way into many aspects of society and technology.
Basically you see time slowed down, light/time in realtivistic sense.
what?
Watched the bottle video as well. It's stunning.