About this project
More example videos:
What the people are saying:
"A product for the professional market, but it's well below the cost of the Phantom" - Norman Chan, Tested
"A high-end "microscope for time", finally within relative reach." - Norman Chan, Tested
"If you’re a sucker for super slow-mo, this could be the best game in town" - Darrell Etherington, TechCrunch
"The quality truly is professional grade." - DL Cade, PetaPixel.
"Edgertronic camera shoots Hollywood-style high-res, slo-mo videos, minus the big budget" - Les Shu, Digital Trends
A high-speed camera that belongs in your bag:
What frame rates and resolutions are currently supported?
The edgertronic supports resolutions from 192x96 up to 1280x1024 in multiples of 16 horizontal and vertical. That's 4,130 possible resolutions.
Are you frustrated that access to professional grade, high-speed video cameras is limited to big budget shows on the Discovery Channel? Do you dream of having the frame rate and resolution of a truly professional camera sitting on your tripod? We’ve created an easy-to-use system that rivals the quality of high-priced cameras at a fraction of the cost.
edgertronic is disrupting the high-speed video camera market.
We started this project for one simple reason: There are no low-cost high-speed video cameras that offer the video quality that professional photographers demand. Since Mike first modified a low-cost strobe flash in 1975, he's been pursuing his passion for low-cost, high-speed photography. He's been frustrated by ridiculously expensive rentals (thousands of dollars per day) of medium quality, high-speed cameras that are bulky and complex to use. In 2011, we set out to solve these problems, and now we're ready to show it to the world.
We've been hard at work for almost 2 years, and with your help, we're ready to take the edgertronic into full-fledged production. By supporting this effort, you’ll be the first to get your hands on an amazing high-speed video camera AND help usher in a new generation of high-speed photography.
Tech specs for the revolution:
We've gone through exhaustive lengths to design a new type of high-speed camera:
- Engineered from the ground up to produce high-quality, high-speed video
- Intuitive and simple to use software so you can spend less time setting up the camera and more time shooting
- Uses standard video codecs, and video is stored on a local SD card so it can easily be imported into the video editing suite of your choice
- Standard Nikon F mount allows use of fast primes, zooms, and other lenses
Read on for a more detailed description of why this camera is revolutionary and how you’ll help our efforts if you support this project.
How does the edgertronic work?
The edgertronic contains a specialized CMOS image sensor, ultra high-speed electronics, memory, and image processing electronics in a compact form factor.
The camera runs a web server and connects to a computer/laptop or to a LAN over Ethernet. The user directs a standard web browser to the camera’s IP address and controls the camera via a user interface (UI) appearing on the web browser. The camera’s UI allows the user to set exposure, frame rate, preview composition, adjust focus, and finally trigger the camera to take a high-speed video.
While running, the edgertronic is constantly capturing frames of high-speed video into an internal buffer. Depending on frame size and frame rate, this buffer contains, at a minimum, the last 8 seconds of video. When a trigger occurs, video before and/or after the trigger is captured and compressed into H.264 video and saved to a removable SD card.
Continuous capture into a this large buffer allows the user to trigger the edgertronic even after an event has occurred. Unpredictable events, like a lightning bolt, are captured with ease.
Once the video is saved, it can be downloaded to the computer, or replayed in the web browser. Alternately, you can remove the SD card and download the videos to their computer or laptop.
High-speed videography has unique challenges in how frame rate, shutter speed, aperture, and resolution settings interact. We completely rethought the control interface to reduce complexity while allowing full creative override. You can specify as many or as few of the settings as you want, and the camera will fill in the remaining ones. This allows you to optimally concentrate on your creative process.
Clicking outside of the control panel puts the camera into live preview mode, allowing the user to focus, compose, capture, and save a high-speed video.
Revision A was a hardware proof of concept and a software development platform. Design began in December 2011, using complete chip and board level simulation (no rework wires!). Revision A powered up in March 2012 and took the first demo video in May 2012. A great start!
Revision B was a repackaging of the Revision A into a more compact design. Improvements included thermal, image quality, and manufacturability. Good but still not perfect.
Revision C is ready for production. Changes made the camera smaller, lighter, easier to manufacture, and more elegant. The last image quality issues were solved. Revision C successfully passed FCC Class B emissions testing, allowing the camera to legally be offered for sale. Along the way, we wrote code. A lot of code.
Excellent design inside:
- Custom camera hardware and software designed for high-speed video
- 1280 x 1024 color sensor (high-speed monochrome sensor limited production)
- Proprietary Global Shutter technology allows crisp exposures without rolling shutter jello effect
- Exposures down to 1/200,000 seconds
- Resolution settable from 192x96 to 1280x1024
- ISO 100-400 sensitivity (color), ISO 400-1600 (monochrome)
- Frame rates up to 17,791 frames/sec. depending on resolution
- Stores captured videos in H.264 format on a removable SD card
- Web-browser based control panel
- Compact 111x108x79 mm form factor (roughly the size of a D-SLR) allows it to fit into spaces previously too small for high-speed cameras
- 2 USB ports
- 10/100 Ethernet
- 12 VDC Power, 1.5A with 2.5/5.5mm locking connector
- Hard Anodize Type III Aluminum enclosure
- Audio Input port (not supported in ver. 1.0 SW)
- Accessory expansion I/O port (external trigger, etc.)
- Manual trigger button
- 4 status indicator LEDs
- Accepts Nikon F-mount lenses (manual and D series)
- Built-in fan
- Size: 111x108x79 mm exclusive of lens
- Weight: ~862 gr exclusive of lens
- Field upgradeable software/firmware
What accessories come with the edgertronic?:
- Ethernet Cable
- AC Power Adapter
- Nikon 50mm F1.8 D Lens
- Airline Grade Fitted Case
- 12V Car Adapter
- Wired Remote Trigger
Where can I learn more about the edgertronic?
So where are we in this whole process? We’ve:
- Formed an awesome team in the San Francisco Bay Area
- Built and tested 3 generations of hardware prototypes
- Engineered the system to be mass-produced
- Written and optimized tens of thousands of lines of code
- Tested a production run of alpha units
We’ve spent the last year designing, prototyping and identifying the best manufacturing processes and partners for local production here in the San Francisco Bay Area. We need your help to complete the final phases that will put our first cameras in the hands of artists, directors, scientists, experimenters and creators just like you. This will position us to ramp up production so we can insure that everyone has access to professional quality high-speed cameras at prices amateurs can afford.
This final phase includes:
- Purchasing components and raw materials
- PCB assembly and testing
- Machining and anodizing the camera’s aluminum housing
- Final camera assembly and testing
- Local San Francisco bay area manufacturing
- Software improvements and testing
- Developing community site for video sharing, best practice suggestions, and posting ideas for improving the camera software
We’re nearly ready to set up full-scale manufacturing to get your edgertronic delivered as soon as possible. We take this last crucial step seriously, and we’ll need a lot of support to do the job right. That’s why we are turning to Kickstarter to get the production process up and running!
If you decide to pitch in, you’ll get a discount and be one of the first to have our product in your camera bag. More importantly, you’ll play a huge part in starting this revolution in high-speed photography by providing your feedback and creating amazing images and videos with the edgertronic. We can’t wait to see what your imagination will capture!
The edgertronic team
I have extensive experience in the architecture and design of high performance digital systems. I use a detailed knowledge of analog and digital design techniques to create state-of-the-art systems that improve performance, lower cost, and work the first time. Products include work stations, super-computers, computer graphics, digital audio/video, and consumer electronics. Notables include the Stellar GS Series and the Apple Powerbook 540c.
I am interested in building all things digital. I have broad knowledge and experience in computer/digital software/systems at all levels from web/desktop apps to DOM to operating systems to digital architecture to firmware to RTL to analog circuits and even some device physics.
Mike's been designing and building products in Silicon Valley since 1990. Over the years he's dealt with many suppliers, consultants, and assembly shops … some good and some bad. He's created a shortlist of firms that we trust for this most important project.
For example, the company Mike's used for years to do PCB layout also assembled the boards for our Rev A, B and C prototypes. These boards had zero defects and worked the first time. We'll use the same company to assemble our production boards as well.
Risks and challenges
The biggest challenge we face is managing the timing of our supply chain beyond development of the first 60 cameras.
The majority of the components in the design are commonly stocked. A few specialized components have long lead times, and we have held these in reserve for the initial limited edition run of 60 cameras. After that, we’ll be batching up production once per quarter to manage any supply chain timing issues. This timing is reflected in the tiers that are available in the campaign. One advantage of moving to a batch production model is that we’ll be able to produce a higher number of cameras with each batch. This increase in production is also reflected in the tiers.
At this point, the hardware (the enclosure and the electronics) are very solid. Our first prototypes worked and were used to take the first demo videos. Rev. B made the camera more compact, and improved video quality. Rev C incorporated changes to refine the mechanical design and perfect the video quality. This is now ready to go into production.
We have built a total of 19 test cameras from Rev A through Rev C. All 19 cameras powered up and worked on the first try. Production will use the exact same suppliers for components, fabrication, machining and assembly as were used for prototypes. Based on the previous turnaround times of our suppliers, we are very confident of meeting our production schedule.
Once the campaign closes, we’ll keep you posted on the status of production and inform you should any issues arise.Learn about accountability on Kickstarter
I have more products in my development pipeline related to high speed photography and backing this project will help me continue to disrupt aspects of high speed digital photography. Also, each camera helps keep manufacturing jobs here in America as each camera is fabricated in the San Francisco Bay Area.
We did a lot of research and the Nikon F Mount is the only logical choice. If you bear with me, I'll explain why:
First, high speed cameras need fast lenses because at high frame rates and/or short exposures, you can never have too much light.
Second, auto focus and auto exposure just don't work when you're doing high speed video. Imagine photographing a car crash test. Even if auto-focus was fast enough, what if it focused on the flying hubcap and blurred out the passenger compartment? Fail!
So we need fast manual focus/aperture lenses, but which mount to pick?
Over the years companies have developed a lot of 35mm lens mounts. Here're the ones that have a long enough flange focal distance to work with our camera:
Of these, only the Nikon F and Leica R are in current production. The rest are only discontinued and only available on the used market.
What about Canon EF/EF-S (EOS) mount? Except for the $5K and up Cine lenses, Canon has never made a fully manual lens in this mount. None. Never.
We want to sell a lot of cameras. What happens when all the quality used lenses are gone? Sure there millions out there, but most are sitting forgotten in some basement with fungus, sticky diaphragms or some other problem that makes them worthless. We need to design the camera around a manual lens mount that is supported in the foreseeable future
So our choices are down to two: Nikon F or Leica R. Most Leica R lenses cost more than the camera and are overkill for this application, so by process of elimination we're down to one lens mount:
Take a look at BH Photo and you'll see that this mount is still supported by Nikon, Zeiss, Samyang, Bower and others. It's popular in industrial machine vision, scientific and medical markets, and it's going to be around for a while.
Even Olympus uses the Nikon F mount on its I-Speed brand high speed video cameras.
This is an excellent question.
First, let's talk about light. In high speed video, you can never have too much light. I've taken videos at 4658 fps, 10us exposure, full sun and could have used more light for better depth of focus. Repeat … you can't have too much light.
The monochrome sensor lacks the color sensor's Bayer filter and is 2 f-stops faster. In some applications this can make all the difference.
If you're a dog and can't see color anyway, get the monochrome. You can never have too much light.
If you're an artist or DP you'll want color. You'll need to buy more lights.
If you are a scientist, experimenter etc. The choice is up to you. Color vs light sensitivity.
Monochrome sensors are sold in lower volumes and thus cost more than color. Simple as that.
If you understand basic photograpic principals like ISO, shutter speed, aperture, and image resolution you can operate the edgertronic. The default settings get you started quickly, from there the user can then adjust the settings and immediately preview changes on their browser. Between video captures, the camera is in a real-time/viewfinder mode that lets the user setup the shot.
Around 1975, I saw some high speed photographs taken by MIT’s Doc Edgerton in a magazine. This combined my interests in electronics and photography. I wanted to know how Doc took the photos, and how I could take my own as well.
The electronic strobes Doc used were quite specialized and expensive. As a kid in high school, there was no way I could afford one, but I thought I might be able to modify a consumer grade electronic flash to do nearly the same thing. After some trial and error, I had a high speed electronic strobe and an acoustic trigger to fire the strobe at the precise instant.
My father was an avid golfer, and I had the high speed equipment, so let’s go hit some balls:
I sent the photo to Doc Edgerton, along with a letter saying I'm applying to MIT and I hoped I'd get to meet him.
A few years later, I'm taking Doc's Strobe Lab course, and I wanted to see what happens when you put a thousand amps through a hair thin piece of wire. Hint: the wire goes away real quick. Doc dusted off an old TRW Image Converter camera and said "It hasn't worked since that H-bomb test, but if you can fix it you can use it.” I got it working and took some interesting sub-microsecond pictures with it.
in short, it comes down to the sensor. Like many things, the design of an image sensor involves many tradeoffs: light sensitivity, noise, size and readout speed are the main ones. It is not possible to take great slow motion videos with a sensor optimized for still photography. Similarly, you can’t take great stills with a sensor optimized for slow motion video. All of the existing consumer and prosumer cameras use sensors optimized for the still end of the spectrum and it is a good choice for that market.
Readout speed is the primary difference between slow motion sensors and other sensors. Our sensor reads data out at ~1 GByte/sec. It’s necessary for slow motion, but comes at a price: higher power, more image noise, complicated interfaces, and hard to design an image processor to handle all that data so quickly.
A global shutter is the second difference. A global shutter exposes all the pixels simultaneously, just like an intra-element leaf shutter in a still camera. All consumer and prosumer cameras with a CMOS sensor have a rolling shutter. Just like a focal plane shutter, different pixels in the image are exposed at different times. This is what causes the jello effect or the bizarre distortion is the image below. So, why don’t all cameras have a global shutter? It’s a tradeoff: global shutters are more expensive, higher noise, less sensitive to light and lower resolution. Rolling shutters work OK as long as there isn’t a lot of movement, but we’re making a slow motion camera. We want a lot of movement.
The link below demonstrates why you never want a rolling shutter in a real high-speed camera: http://en.wikipedia.org/wiki/File:Turboprop_Rolling_Shutter.jpg
Sorry for the link. Kickstarter doesn't have a way to insert a picture in a FAQ.
Not really. to understand why we need to understand what exactly is happening in the D7 + Twixtor case:
The Canon D7 is a great camera and shoots beautiful video, but it only takes 60 frames per second, or one frame every 16.66 ms. Let’s call these keyframes.
Twixtor is program that uses computer interpolation to create new frames based on a guess of what happened between the keyframes. It’s no different than image morphing software that can turn a butterfly into a dog. In the middle of the morph, the half butterfly, half dog it’s just the program’s best guess of what a butterfly-dog would look like. It’s not real, but it looks convincing. Similarly, when creating a 960 frame per second slow motion effect, the 15 new frames that Twixtor creates between each keyframe aren’t real, they’re just a guess as to what happened between the keyframes.
Sometimes Twixtor works OK. Consider a bicycle stunt, where the bike is in the air for about a second. In this case, you’ll get about 60 keyframes and the bike’s motion is rather fluid. Thats enough keyframes, combined with the predictable motion of the bike, for Twixtor to create a rather convincing slow motion video.
Sometimes Twixtor doesn’t work. Take the case of a firecracker. In one keyframe it’s intact and in the next keyframe, it’s entirely gone. What happened in between is lost information and it impossible for Twixtor to create the frames showing how the firecracker blows up.
This is an example of an important concept in information theory. Once data is lost due to sampling or quantization, it can’t be recovered. Cliff Notes work by throwing away data. This is great if you haven’t read the book and the test is tomorrow, but you’ll never know what exactly happened on page 52 of Hamlet.
Slow Motion is used in industry to design and maintain high speed production equipment. Jams, stoppages and other hiccups can be clearly seen and corrected when the motion is slowed down.
Slow motion is an important tool for coaches and athletes alike. Slow motion can help perfect a tennis serve, a high dive or even a skateboard stunt.
Slow Motion is educational and insightful. In slow motion, a slinky behaves in an unexpected fashion and provides a lesson in physics and wave propagation.
Slow motion is used to make products safer. In a crash test, slow motion allows safety features, like airbags, to be optimized for maximum effectiveness.
Slow Motion is essential in the entertainment arts. Discovery Channel can't go for more than a few minutes without cutting to a slow motion clip.
Slow Motion makes everyday occurrences visible. Seeing how water drops splash isn't very practical, but it's both cool and beautiful.
- Attach lens and put camera on a tripod.
- Configure computer or LAN network (instructions provided with camera)
- Connect camera to computer or LAN
- Connect power to camera
- Browse to the camera’s IP address
- You’re ready to start shooting
- A desktop or laptop computer with a Google Chrome
- An ethernet port or a LAN router running a DHCP server
- A tripod
- Good natural light or studio lighting as needed
When the camera is in preview mode, RAW video frames are continuously stored to the 8GB ring buffer. This buffer will hold at least 8 seconds of video (longer if the frame rate is set to less than the maximum for that resolution). The pretrigger and duration settings in the UI are used to calculate how many frames are to be stored after the trigger event. When a trigger occurs, the camera stops after this number of frames has been stored. This is how the camera can capture video before and after the trigger event.
For example, if the user sets the duration to 8 seconds and the pretrigger to 25%, there will be 2 seconds of video before the trigger and 6 seconds after.
The edgertronic design goals were simple: design the smallest, lightest, lowest cost, easiest to use, straight-up high-speed video camera, and bring it to market quickly.
The interfaces listed above are typical of the status quo in high speed video cameras. We're not making a $100K complex beast, bristling with all sorts of controls and interfaces, that's intended to plug into a huge TV studio suite. We're changing the rules.
We could have added some or all of the above interfaces, but it wouldn't have made the camera faster or better, or able to do anything new. It just would be bigger, heavier, cost more and taken longer to develop.
The edgertronic is designed so that future SW releases can be downloaded and easily installed. That said, what's on the list following SW version 1.0?
96KHz Stereo audio sync'd to the captured video.
Save h.264 video to external USB drives, in addition to SD Card.
UI enhancements (advanced controls etc)
Image Quality enhancements
Possible depending on customer demand:
Save RAW to SD Card/USB Drive
SDK for customer specific applications
Support this project
- (31 days)