Trillion FPS camera developed at MIT camera can 'watch' the movement of light

mesa

This is nothing short of amazing. MIT figured out a way back in February to view how light 'moves' on, about and through objects.

Here is a link:
http://www.youtube.com/watch?v=snSIR...eature=related

They apparently use multiple staggered high speed cameras to capture the images and then turn it into a continuous film; very, very clever.

Interesting to see how the light moves and how normally we basically 'see' everything as a constant even though the photons are in continuous motion.

I wish they had videos showing more accurate trajectories and different materials as opposed to this showboating publicity stuff, but still cool none the less.

dipole

They should do some diffraction/interference experiments, but I don't think it will work because they're not filming a single wave front, they're filming thousands of different pulses of light and staggering the shots to make a video.

mesa

Well the technology will improve with time, it took only 60 something years for us to go from Kitty Hawk to the moon; rather staggering if you really think about it.

I love the time in which we live; so much left to discover and with the technology, ever increasing acess to knowledge, and colleges/universities at our disposal anyone who actively chooses can be a part of it.

Science and Education rock! :)

Cthugha

I considered this article pretty misleading when it came out. Streak cameras like the one used here with a temporal resolution of roughly 2 picoseconds are known and have been used for way more than 10 years now .

However, typically you direct some emission into the streak camera to check the dynamics of the emission instead of waiting for stray light of light passing through air as it is performed here. The only "innovation" here is that they add a second standard camera covering the whole scene, so they can merge a still showing the scene and the streak camera info showing the dynamics of the light moving.

davenn

yeah saw that video a few weeks ago
there seems to be lots of discussion elsewhere that the camera ISNT doing a trillion fps, nowhere near it. its the way they are manipulating the imaging that gives the appearance/impression of an extreme frame rate

Dave

Cthugha

Well, the camera has a time resolution that would allow to get a trillion fps if one could operate it continuously, but one cannot. You have some photocathode from which electrons are emitted as light strikes it. Then you apply a sinusoidal deflection voltage synchronized with the excitation laser frequency to deflect these photoelectrons. These hit a phosphorus screen and the position where they hit gives you information about the arrival time. However, you only record a very small fraction of the sinusoidal cycle, typically the linear part.

If you really want a time resolution of 2 picoseconds, you can only record a length of approximately 140 picoseconds. If you accept slightly worse time resolution, you can record 700 picoseconds. As the time delay between consecutive laser pulses is typically 13 nanoseconds, you miss a lot of the frames in between. Taking into account that the CCD which takes a picture of the phosphorus screen operates typically at 100 Hz, you lose even more frames (or integrate over them).

Andy Resnick

Streak cameras have been around for a while, and can resolve down to 1 ps:

http://sales.hamamatsu.com/en/produc...art-c10910.php

http://learn.hamamatsu.com/tutorials/java/streakcamera/

For some reason I'm having trouble seeing videos, but based on the posted comments MIT essentially used heterodyne mixing to slow down a periodic process; similar imaging techniques have been performed of a sonoluminescing bubble.

mesa

I certainly understand the argument that the camera's do not in fact show 1 trillion frames per second. Before this video the best I had seen was million FPS, link below:

http://www.youtube.com/watch?v=QfDoQwIAaXg

MIT said they used 500 high speed cameras, assuming they were as good as these MFPS camera's that would give 5x10^8 frames per second so they only capture one frame per 2000 of the claimed 1X10^12 FPS and that is assuming the miliion FPS camera's don't 'cheat' the viewer as well.

However, maybe we shouldn't lose sight of what can be done with the technology as it sits. Even though this camera only takes 'snapshots', as far as our eyes can detect the imagery is remarkable and now brings the motion of light into the real world of observation and study of it's physical characteristics.

There are formula's to describe the motion of photons but reality trumps pen and paper for visualization.

I am grateful that technological developments such as this happen and we should exploit it to it's fullest advantage; we can get lost in a conversation about "well it's not really a 'TFPS' or instead think about how this can be used to gain new knowledge and help us better visualize and understand what we already know.

Cthugha

Basically they do not really use 500 cameras, but a CCD consisting of 640x480 pixels and each of the 480 rows is used as a row camera for a different frame.
However, my main point was a different one:

My main point was something different. This simply is no new development. The exactly same technique has been known and is routinely used for more than ten years now. The technique has already been used to its fullest advantage. Using it to watch stray light from a passing laser pulse is nice and gets media attraction, but it is nowhere near as new or groundbreaking as it sounds.

mesa

I understand, although I didn't realize 10 years was such a long time lol

"The technique has already been used to its fullest advantage." - That is a rather big claim, so you believe everything that can be done experimentally has been accomplished with these cameras?

I would like to see what has been done; I would imagine there has been testing of setups with specific trajectories and materials but where can we see it?

Cthugha

I understand, but in scientific terms time passes quickly. I mean, 17 years look like a short time, too. 17 years ago the first BEC was created and by now study of BECs has become absolutely standard and a huge field. Progress sometimes happens at an insane pace.

Well, let me put it this way: at least more sensible things than watching photons fly by have been done.

Typically, such cameras are not used to check spatial movement, but the emission is directly passed on to the camera to monitor emission dynamics or similar things - often together with a monochromator to get spatially and spectrally resolved information. So most setups will focus on photoluminescence or similar things. However, other measurements are possible, too. I have written some papers on experiments using streak cameras myself, however they are somewhat aimed at a rather specific audience. If you are interested I can give you some links per pm.

mesa

I could imagine, I didn't even know we had the ability to observe groups of photons until two days ago lol!


I am sure there are many wonderful things even more interesting than light but the idea of actually watching groups of photons in motion is nothing short of astonishing to me. To actually watch them move through air and interact with objects is, well, great!


Very interested!

Andy Resnick

Except the video showed none of this- the light field was very incoherent (broad spectrum and no speckle), and the only interaction shown was macroscopic.

The video is a stunt, nothing more.

mesa

I agree, the video is in many regards 'just a stunt' which is frustrating but seeing how a group of photons move and interact with objects is still fantastic. If these cameras were used experimentally they could create some wonderful visualizations and teaching materials and perhaps shed some light (no pun intended lol) on behavior as well even if only macroscopically for the time being.

Technology needs time to develop and I suspect the science will get better, at least so long as we encourage this development as opposed to just offering harsh criticism.

Physics started out by studying what we can easily see and relate to, this is the first time I have seen, in the real world, how light interacts with physical objects and as I understand it (thank you Cthugha) this technology has only been around for 10 years. If you think this is unimportant then I am sorry.

sophiecentaur

I wonder whether people are getting a bit too excited about this idea of watching the progress of an EM signal through a medium. It's something that has been done for decades with RF signals. You can see and measure the change of phase (progress) along a transmission line with some very humble equipment.
Yes, of course it's good that they managed it with visible light but I think it should be viewed in context. Is there really a difference, in principle?

voko

Their statement that they can see individual photons fly is simply false.

sophiecentaur

Precisely. All they are doing is to use the same techniques as Sampling Oscilloscopes and Time Domain Reflectometers have used, which involves repeated fast pulses (of light, in this case) to build up a picture over a length of time / space. Very pretty though, but he could have cleaned up that mirror for the demo! I would never take a regular snap with that amount of muck on my lens.