A team of Caltech researchers has built (again) a camera that captures more than a trillion photographs per second. The old “fastest camera in the world” could fire 10 trillion dpi. Up to 70 trillion dpi will be shot by the current camera.
The new breakthrough is an evolution of previous Caltech “trillion+ fps” cameras, each of which has a “Compressed Ultrafast Imaging” or (CUP) shape. CUSP, which reflects compact Ultrafast Spectral Photography, is the newest version from Caltech that was able to record 1 trillion fps of translucent events, including neuron bursts and shock waves.
CUSP, designed by Lihong Wang and his team, integrates a short pulsed laser light with optical systems which split each individual puls into a train of even shorter light pulses, each of which is able to create an image in the camera. The laser is only one quadrillionth of a second long. Yet those are just words … just a few example images.
In the GIF below, access the full-resolved video here, a burst of light was spread over a group of letters (which, naturally, spell Caltech) and shot with the T-CUP camera on 10 trillion fps (left) and the current CUSP camera on 70 trillion fps (right). Each pulse is just 1.5 picoseconds, or 1.5 in 10 ^ -12 seconds, or 0.00000000015 seconds, respectively.
For those who love scientific contexts, the speed of light is 300 billion millimeters per second (just below). Now, take 70 trillion frames and break it by 300 trillion mm / s and get ~233.3 frames per millimeter of light. Put another way: in order to transfer a millimeter, this sensor will take roughly 230 frames.
It may have been suspected by you now, but the CUSP device does not take photos at any moment … or ever. The technology would be used to record extremely short-lived processes such as nuclear fusion, fluorescence of molecules or light wave activity itself. Although some of this was feasible with T-CUP, this 7x speed upgrade aims to expose previously unexplained natural phenomena.
Find out the Caltech website to read more about this new advance in femtosecond visualization. If you want to delve into the whole study article, you will find it here or in the Nature Communications issue of April 29th.