Nvidia is back at it again with another awesome demo of applied machine learning: artificially transforming standard video into slow motion – they’re so good at showing off what AI can do that anyone would think they were trying to sell hardware for it.
Though most modern phones and cameras have an option to record in slow motion, it often comes at the expense of resolution, and always at the expense of storage space. For really high frame rates you’ll need a specialist camera, and you often don’t know that you should be filming in slow motion until after an event has occurred. Wouldn’t it be nice if we could just convert standard video to slow motion after it was recorded?
That’s just what Nvidia has done, all nicely documented in a paper. At its heart, the algorithm must take two frames, and artificially create one or more frames in between. This is not a manual algorithm that interpolates frames, this is a fully fledged deep-learning system. The Convolutional Neural Network (CNN) was trained on over a thousand videos – roughly 300k individual frames.
Since none of the parameters of the CNN are time-dependent, it’s possible to generate as many intermediate frames as required, something which sets this solution apart from previous approaches. In some of the shots in their demo video, 30fps video is converted to 240fps; this requires the creation of 7 additional frames for every pair of consecutive frames.
The video after the break is seriously impressive, though if you look carefully you can see the odd imperfection, like the hockey player’s skate or dancer’s arm. Deep learning is as much an art as a science, and if you understood all of the research paper then you’re doing pretty darn well. For the rest of us, get up to speed by wrapping your head around neural networks, and trying out the simplest Tensorflow example.
Continue reading “Nvidia Transforms Standard Video Into Slow Motion Using AI”
[Paul] created a frame that uses an Arduino and LEDs to create a slow motion illusion of a delicate item (like a flower or a feather). The effect is striking as you can see in the video below.
[Paul] had seen similar projects (both one-offs and sold as a product), but wanted to do his own take on it. The principle is simple: The device vibrates the objects at one frequency and strobes LEDs at a slightly different frequency (80 and 79.5 Hz, in this case). The difference between the frequencies (the beat frequency) is what your eye perceives as a very slow (0.5 Hz, here) motion.
Continue reading “Slow Motion Frame Will Be the New Magic Mirror”
[hw97karbine] has made a pretty cool tennis ball cannon. While making a cannon of this sort is nothing new to us, we were impressed by the effort taken to get a stoichiometrically ideal mixture of 3.2% butane and air in the combustion chamber.
[hw97karbine] filled a syringe with butane and then dosed exact amounts into the chamber using a hole in the back. To control the air mixture he marked lines on the outside of the cylinder with magic marker. Simple but effective.
More rewarding than the methods was the cool slow-mo videography of the explosions in the chamber. You really have to check it out. [hw97karbine] shows clearly the difference between a well-balanced fuel to air mixture and a poorly balanced one. It’s one thing to say that more fuel does not mean better combustion, as we all remember from our personal potato cannon experiences, but it’s another thing entirely to see it.
The Raspberry Pi Zero was announced this week, so you know what that means: someone is going to destroy a Game Boy Micro. If you’re interested in putting the Zero in a tiny handheld of your own design, here are the dimensions, courtesy of [Bert].
[Ahmed] – the kid with the clock – and his family are suing his school district and city for $15 Million. The family is also seeking written apologies from the city’s mayor and police chief.
There are a lot — a lot — of ‘intro to FPGA’ boards out there, and the huge variety is an example of how the ‘educational FPGA’ is a hard nut to crack. Here’s the latest one from a Kickstarter. It uses an ICE40, so an open source toolchain is available, and at only $50, it’s cheap enough to start digging around with LUTs and gates.
Over on Hackaday.io, [Joseph] is building a YAG laser. This laser will require a parabolic mirror with the YAG rod at the focus. There’s an interesting way to make one of these: cut out some acrylic and beat a copper pipe against a form. A little polish and nickel plating and you have a custom mirror for a laser.
You know those machines with wooden gears, tracks, and dozens of ball bearings? Cool, huh? Tiny magnetic balls exist, and the obvious extension to this line of thought is amazing.
[David Windestål] is awesome. Completely and totally awesome. Usually, he’s behind the controls of an RC plane or tricopter, but this time he’s behind a slo-mo camera, an RC heli, and a watermelon. That’s a 550-sized heli with carbon fiber blades spinning at 2500 RPM, shot at 1000 FPS.
How do you label your cables? Apparently, you can use a label printer with heat shrink tubing. Nothing else, even: just put heat shrink through a label maker.
[Matt Kane] works at a really cool company in the UK where he recently finished working on the Triggertrap Ada — the highest-performance, most feature packed camera trigger out there. So just for fun, he decided to challenge himself again — could he make a super basic, super fast, bare-bones camera trigger for $2 instead?
At the most basic level this is just a laser pointer and a light sensor. When the object your photographing breaks the light path, the flash triggers. Typically this is done with an IR laser, but since he’s going for a low-cost system, he’ll use a basic 1mw red laser pointer — the only downfall is you might see it in the picture.
Next up is the sensor. Ideally we’d use a photodiode which is very fast, but also expensive. A photoresistor is cheap, but not fast enough. A nice medium between the two is a phototransistor, which is relatively fast, and cheap. Finally, we need a minimum trigger period to offset the flash. [Matt] thought about using a 555 timer but instead decided to just generate a pulse with an Attiny45. Continue reading “High Speed Laser Based Camera Flash Trigger for Only $2”
Earlier this week I saw a video that was showing how some guys made some really cool bullet time effects with 15 tiny rugged awesome cameras called “GoPros”(that wasn’t a paid endorsement, they’re awesome). For those unfamiliar, the bullet time effect gained popularity from the first matrix movie. The footage slowed down to slow motion while [Neo] dodged some bullets, but the camera was still able to move around. To do this, they built a massive circular rig and mounted tons of cameras all around. Using these multiple angles, they were able to stitch together the scene in slow motion and “move” the camera.
Continue reading “Bullet time with a single GoPro”
Though we can’t really tell you how this is hacking related, we’ve always shared random videos of stuff getting destroyed with you. Invariably someone in the comments goes on a rant about how wasteful and/or dangerous it is. This clip, from a Danish TV show called Stupid and Dangerous, fits that description quite well. It is also freaking awesome.
We’re really not sure what draws us to these videos exactly. Is it vicarious destruction or possibly our natural affinity for slow motion?