neuroanatomist said:
kennephoto said:
Pretty sure you can use any light, since its quite literally capturing a photo of light. You just adjust the exposure to compensate for the amount of light like we already do.
Right, so at 1 billionth of the current max shutter speed of 1/8000 s, to 'just adjust the exposure to compensate for the amount of light like we already do,' you'd need to go 36 stops higher than ISO 100. Know any cameras that can set ISO 6,871,947,673,600? :
Actually you both are wrong. You cant use any light. You need femtosecond pulsed laser source to illuminate the scene. Then streak camera, PMT or other device is used to capture reflected pulses to create a probability distribution by acquiring millions of scans and calculate arrival time of photon. Image is integrated for whole scene by integrating using piezo-controlled mirrors. In short not non-technical background, repeated measurement scene illuminated by femtosecond spaced pulses combined with maths is used to calculate the time information at a point/line. This is then calculated for all points/lines in scene to construct image. So its takes quite a while to construct one super-fast timelapse (counterintuitive isnt it).
Small issues getting it into Canon 1DXXX:
1) You need at least two people to hold even simplest of femtosecond lasers
2) Its a two piece assembly with micrometer beam aliment on optical table; good luck with handholding
3) You need the scene to repeat millions of times. Like if you want a video of kid running, you will need the kid to run same path millions of times (this is only for time lapse imaging; does not apply to their around the corner imaging)
and many more but I am bored.
I am sure they will overcome these problems sometime in future!!!!
Source: I use the same camera at nanometer scale. So I collect images with nanometer spatial resolution and picosecond time resolution.