World's Fastest Camera Shoots 10 Trillion Frames a Second

bbsguru shares a report from New Atlas: Slow-motion video has always been fun to watch, with the best rigs usually shooting on the scale of thousands of frames per second. But now the world's fastest camera, developed by researchers at Caltech and INRS, blows them out of the water, capturing the world at a mind-boggling 10 trillion frames per second -- fast enough to probe the nanoscale interactions between light and matter. For the new imaging technique, the team started with compressed ultrafast photography (CUP), a method that it is capable of 100 billion fps. That's nothing to scoff at by itself, but it's still not fast enough to really capture what's going on with ultrafast laser pulses, which occur on the scale of femtoseconds. A femtosecond, for reference, is one quadrillionth of a second.

So the team built on that technology by combining a femtosecond streak camera and a static camera, and running it through a data acquisition technique known as Radon transformation. This advanced system was dubbed T-CUP. For the first test, the camera proved its worth by capturing a single femtosecond pulse of laser light, recording 25 images that were each 400 femtoseconds apart. Through this process, the team could see the changes in the light pulse's shape, intensity and angle of inclination, in much slower motion than ever before.

Finally

[bobstreo]

I can prove my Canadian girlfriend is real. /s

double slit experiment

[Wycliffe]

This would be interesting to use to record the double slit experiment and find out what is really going on.

Re:

[alvinrod]

Wouldn't the camera count as observation and result in the same pattern that's seen with the detectors that are currently used? It's an interesting idea, but I suspect that it would just capture the electrons behaving as expected without being able to provide any additional insight into what's really going on.

Re:double slit experiment

[Anonymous Coward]

If you're thinking of using the camera to record the state of the photons or electrons as they leave the slits, all you'd end up with is two distinct bars recorded on the photographic plate. If you're thinking of recording the target media itself, all you'd see is small dots accumulating into a series of bars making up an interference pattern.

Re:

[rajkiran_g]

Depends on what we mean by reality!

In the classical sense, reality is independent of observation. Not so in the realm of quantum mechanics, where observation is not possible without interaction which in turn effects the state being observed/measured.

Re:double slit experiment

[schweini]

CHeck out PBS SpaceTime's videos on the double slit experiment! Simply recording quantum things makes for very very strange outcomes:
lhttps://www.youtube.com/watch?v=8ORLN_KwAgs [youtube.com]

Re:

[grep -v '.*' *]

record the double slit experiment and find out what is really going on.

All of the electrons have a happy face with a tongue sicking out as they pass thru either/both/neither slits.

Link [google.com] I like #2 personally.

Meh..

[Anonymous Coward]

It doesn't do 4k.

What's the resolution?

[UnknownSoldier]

A trillion frames at 1x1 is useless.

That said a super high frame rate like this is awesome.

What's the resolution though? I couldn't find it mentioned in the article.

Re:

[Anonymous Coward]

This isn't definitive, but one of their previous papers shows images at 150 x 150.
(See bottom of page 3: https://authors.library.caltech.edu/67908/11/nihms636729.pdf)

I'm guessing that since they are looking at frickin lasers, they aren't too concerned about getting much better than that.

150x150 isn't enough for lasers

[rsilvergun]

you can't even get the angry sea bass in the shot, let alone a shark.

The reason we use exponents

[petes_PoV]

A femtosecond, for reference, is one quadrillionth of a second.

Probably one of the world's least useful explanations.

Re:

[CaptainDork]

LOL

They should have used a car analogy.

Re:

[GrahamJ]

How many parsecs is that?

Re:

[epine]

Google "6 inches in attoparsecs" and it replies "4.93895". In other words, not a flattering unit for the male anatomy, but useful if you want a close approximate to a decimal foot.

Re:

[epine]

Hmm, had I been awake, I would have described that as cosmology's decimal inch.

10 attoparsecs = 1.01236 feet

Re:

[wyoung76]

https://www.britannica.com/top... [britannica.com] Yes, quite a useless explanation.

Re:The reason we use exponents

[Solandri]

Some more intuitive descriptions:

• In one femtosecond, light will travel 0.3 micrometers (0.0003 mm).
• In the 400 femtoseconds between the successive image frames shot with this camera, light travels only 0.1 mm.
• If you shot one second at 10 trillion fps, and played them back at 60 fps of YouTube video, it would take over 5000 years to watch that one second.

Re: The reason we use exponents

[troon]

The UK hasn't used long scale numbers in quite a while. I remember this being an issue as a kid in the early 1980s, but not now.

Re:

[TeknoHog]

The UK definition is 10 ^26

I think you mean 10^24. Also, I'd rather call it the European definition (or more precisely the long scale), because I've seen both definitions used in the UK, probably via American influence.

The European/long scale is easier to remember, because the number of zeros is 6 times the number from the name. Here, quad = 4 so we get 24. In the US scale, the logic goes as 3 + 3*quad = 15.

Re:

[TeknoHog]

Is that an American or a European quadrillion?

As it passed, ...

[CaptainDork]

... the pulse of laser light said, "I've been framed."

Interleaved?

[Jfetjunky]

Is it real time, interleaved, or synchronous sub sampling (aka aliasing aka stroboscopic effect)?

Re:

[account_deleted]

Comment removed based on user account deletion

Re:

[account_deleted]

Comment removed based on user account deletion

Re:

[Anonymous Coward]

Synchronous, if I understand correctly. Physically, they capture two 2D images simultaneously - one "straight" and one "sheared" - then use them to digitally reconstruct a 3D spatio-temporal image.

It's a bit like how your eyes, observing the same scene from two different angles, let you construct a three-dimensional mental image - but in this case, the third dimension is time.

Frame rate is high ... but

[140Mandak262Jamuna]

It took only 25 frames in total. And looks like each frame is a low res pic of about 100 x 100 pixels, as far as I can tell.

For those who are curious

[sconeu]

1 ten trillionth of a second (100 fs or 0.1ps) is 29 microns at light speed.

Re:

[hcs_$reboot]

So what's the distance between the cam and the laser beam? I mean, a cam reacts to light, and here it shows a laser pulse advancing within a few ps, but that advancing light needs also to reach the sensor (or whatever it is)...

LiFi would be great

[mtaht]

if we could only get LEDs to switch even 1/100th this fast. I think it's about 2ms now.

Re: LiFi would be great

[troon]

It's a bit quicker than 2ms: that's only 500Hz, which would preclude SPDIF.

transform, !ation

[cellocgw]

Just for those who don't mess with image analysis, it's the "Radon Transform," not a 'transformation'. The wikipedia article is long on equations and short on simple application, which is that it's a way to find critical image parameters like lines or edges.

First Real-World Application will be...

[BeerMilkshake]

We all know it will be porn somehow.

Don't use misleading scales!

[MS]

By 10 trillion, you mean 10^13 aka 10 billion in the rest (=non-us-english) of the world, right?!?

Where is my 100 yotta bytes SDCARD ?

[Andre Dias]

I want one, so I can store all 10 seconds of my daughter birthday. Can someone point me to a reseller of a 100 yotta byte SDCARD?