Edgertronic - The first affordable high speed video camera | Hacker News Reader

Edgertronic - The first affordable high speed video camera

kickstarter.com

60 pointscocoflunchy21 comments

The edgertronic camera was designed to fill a huge hole in the high speed video camera market.

Currently, at one end of the market are consumer cameras where high speed is basically an afterthought (there's only so much you can do with a rolling shutter). At the other end are the ultra high end cameras used for big buck productions that cost $250K and up.

The edgertronic has a global shutter and all the other specialized design features that make it a straight up high speed video camera.

We could have come out with a camera that's faster/higher resolution but it would cost more. Just like with cars: "Speed costs money. How fast do you want to go?" This camera strikes a nice balance between price and performance.

https://www.kickstarter.com/projects/1521768569/edgertronict...

Mike

Can you guys tell us more hardware specs like the processor/fpga (I would assume fpga?), sensor (pixel count, pixel pitch), ram size and speed, etc?

Wondering if it would be possible to cool the sensor to lower the noise floor and thus require less light. Will you guys also allow programming of the fpga for image processing and other uses?

Appreciate it!

Edit: Also, why 100Mbps Ethernet?? Almost all networks are GigE and with the frame rate you guys are talking about it could use the faster speed!

  Also, why 100Mbps Ethernet??

1.) It records to SD card. Ethernet is just used for live preview, not production footage.

2.) The bottleneck is the internal buffer. (and H.264 encoding) Gigabit ethernet wouldn't be fast enough to stream uncompressed high-speed video. Upgrading to either 10 gigabit ethernet (expensive!) or realtime 1000FPS H.264 encoding (really expensive!) would break the $5,000 budget.

  cool the sensor to lower the noise floor and thus 
  require less light

Difficult...

A high speed camera obviously has to have very short exposure times in order to work at all. (Can't get faster than 2FPS with a 1/2s exposure time)

The general rule of thumb is that you can trade off one "stop" in either direction-- 1/100, f/4, ISO200 is the same as 1/50 f/5.6 ISO200 is the same as 1/100, f/5.6, ISO400.

My guess is that even with heroic cooling (liquid nitrogen) you'd only get from the usual wildly excessive lighting needed for high-speed cameras, down to merely "really bright." You'd never get to a point where you can light a scene with just indoor lighting.

sbierwagon,

Thanks for the excellent response to reportingjr. You beat me to it.

I spent a lot of time making the edgertronic a balanced design. All the parts work well together. On the computers we've tested with, the speed of the web browser and cpu/graphics processor is the bottleneck, not the ethernet.

I did look at GigE but it would have increased cost, size, and power without adding any real benefit.

With this type of sensor, cooling won't measurably improve noise.

We'll design other cameras in the future, but we had to start somewhere, and this is it.

Curious if this will find support with the media industry. The first thing that jumps out is that it doesn't use the standard TV and film resolutions 1080/720/480. From talking to film makers I know about digital photography I think they would find that list of resolutions quite alienating and possibly assume it wasn't suitable for their needs.

The quality of the images and the price look very good however.

[edit] I see it does have 720 and 480, but not the important 1080, which would make it "HD". Having had a brief look at some of the competitors though, most of them seem to operate on weird resolutions like 1024x1024[0], so I guess it must be expected from a slo-mo camera.

[0] http://www.photron.com/?cmd=comparison

Phantom has HD, 2K, and 4K high-speed cameras; they're crazy expensive compared to this. I don't see any evidence that it's possible to provide high-speed full HD cheap.

It does do 1280×720 which is 720p, most people would call that HD.

True its not 1080P but at 700 frames/s at 720p you are going to get incredible pictures.

Minor note, but the p (or i) is not important here, so it would just be HD720 or HD1080. The p/i denotes a method of displaying the image, either through interlacing[0] or progressive scan[1].

[0] http://en.wikipedia.org/wiki/Interlaced_video [1] http://en.wikipedia.org/wiki/Progressive_scan

However, its highest (17k fps) speeds are at much smaller resolutions, and it's more like 1000fps and under at the higher ones, which isn't quite as impressive, although still fast.

Indie movie guy here. It will be fine. Existing pro and semi pro cameras that offer slo-mo functionality as a fringe benefit typically do so with compromised resolution (eg with a RED Scarlet you have o drop drom 4k to 2k to shhot at 60FPS and down to 1k for 120FPS, IIRC). It's a fact of life and editors are used to dealing with it, you just have to be selective about how much you use, your composition and so on. There are also ways to address shortcomings in post-processing with super-resolution plugins and so forth. 700fps at 720k is more than acceptable for many purposes (especially when you lay a breathless narrator saying 'shot with a special camera...', because this will get used more in documentary and newsgathering than in narrative.

The more advanced frame rates will only be useful in science labs due to the increasingly poor resolution, but they'll still be useful - there was a consumer camera from Casio that did something similar at much lower resolutions but it had a fixed lens which severely limited its application. However, it became quite popular with amateur filmmakers and in high-school science labs.

I like that this is just a head that goes out to a computer. I don't think every director of photography will have one, but that some people will specialize in setting up and shooting this and then rent themselves and their gear out to productions for special effects shots etc. They'll make their target comfortably.

Backed. I've used a few different high speed video cameras for walking robot research, and this seems like exactly what I want.

Awesome. There's been a lot of neat optics work in the last few years driving costs down. I can't wait for every camera to be able to do this and this: https://www.lytro.com/

I also have hopes for http://www.muoptics.com/ too!

With respect to Mu, check out this EEVBlog thread: http://www.eevblog.com/forum/crowd-funded-projects/m-thermal... .

As a disclosure I backed just for the fun of it and haven't asked for a refund yet but with each new "update" their "CEO" releases I grow increasingly skeptical that Mu Optics are ever releasing a real product. Not only did their pricing seem too good to be true to begin with but their updates are needlessly vague even with IP concerns in mind.

I really hope they get funded, but for a hobbyist the camera is still pretty expensive. I suspect there are many "weekend" GoPro users who would love to have a supplemental camera that shoots at 240 or 480 fps instead of of the GoPro Hero3's max of 120 (ideally also getting rid of the rolling shutter). But, I think a camera like that would only really sell if you can get the price down to below $500. That way you can use the camera in interesting, somewhat risky locations, but while it still hurts, it's not the end of the world if you lose it or break it.

the max framerate for the go pro is 240fps at WVGA resolution.

Yes, I forgot about that, but it has a max of 120fps in HD quality (720p) / Protune video.

Can someone explain to me some things that I'm too lazy to Wikipedia?

How long can this camera shoot continuously for? And what frame rate can it transfer image over its I/o ports?

Nearly $5k is 'affordable'?

Maybe they're trying to be modest by not mentioning the prices of their competitors which are $100K or more AFAIK.

In many parts of the country, a nice house costs less than the high speed camera used in the last World Series.

Reply on news.ycombinator.com