I’d like to create 'The Matrix' bullet time style flipbooks on the cheap
January 13, 2013 8:09 PM
Remember the time freeze sequences introduced in the first Matrix movie? I’d like to quickly film, edit and print a flipbook as a giveaway.
The self-funded project doesn’t need to be pretty but must be functional in ~6 months... I think plenty of time if I don’t hit dead ends. Unfortunately, to get this done I’ve have a lot of learn’n to do so I’m hoping I can get some project kick-off pointers here.
The plan:
1. A circular array of ~36 shutter synchronized video cameras film a ~5 second clip of the actor.
2. Film clips are collected on one computer
3. Actor selects the primary camera for the normal speed lead in
4. Actor selects the freeze point for the viewpoint pirouette
5. Movie frames are extracted and assembled into a final clip
6. Movie frames are printed ~12 per letter size sheet
Camera specs:
1. External shutter sync.
2. Color
3. 640x480 is probably plenty, more is better
4. ~24 fps is OK, ~48 would be nice for a slow motion lead in
5. Cheap or moderately priced if re-sellable
My guess is that the camera will need local memory to buffer the video instead of streaming to the computer.
My questions:
1. Camera source? I’m hoping for DIY sources of a cell phone caliber camera. Something fancy like GoPro probably would be easy, but I’d have to float big bucks to buy-use-sell them.
2. Camera interface? I’m guessing Ethernet would be preferred.
3. Program language to edit the video in a Windows environment? I’m sure lots can do it, is there a favorite flavor? Again, cost is an issue. (I do have barely used copy of Visual Studio.net Professional v2003 in a box). I could probably justify more investment if I ended up learning some IEC 1131-3.
4. Other wisdom?
5. Am I crazy?
Thanks.
The self-funded project doesn’t need to be pretty but must be functional in ~6 months... I think plenty of time if I don’t hit dead ends. Unfortunately, to get this done I’ve have a lot of learn’n to do so I’m hoping I can get some project kick-off pointers here.
The plan:
1. A circular array of ~36 shutter synchronized video cameras film a ~5 second clip of the actor.
2. Film clips are collected on one computer
3. Actor selects the primary camera for the normal speed lead in
4. Actor selects the freeze point for the viewpoint pirouette
5. Movie frames are extracted and assembled into a final clip
6. Movie frames are printed ~12 per letter size sheet
Camera specs:
1. External shutter sync.
2. Color
3. 640x480 is probably plenty, more is better
4. ~24 fps is OK, ~48 would be nice for a slow motion lead in
5. Cheap or moderately priced if re-sellable
My guess is that the camera will need local memory to buffer the video instead of streaming to the computer.
My questions:
1. Camera source? I’m hoping for DIY sources of a cell phone caliber camera. Something fancy like GoPro probably would be easy, but I’d have to float big bucks to buy-use-sell them.
2. Camera interface? I’m guessing Ethernet would be preferred.
3. Program language to edit the video in a Windows environment? I’m sure lots can do it, is there a favorite flavor? Again, cost is an issue. (I do have barely used copy of Visual Studio.net Professional v2003 in a box). I could probably justify more investment if I ended up learning some IEC 1131-3.
4. Other wisdom?
5. Am I crazy?
Thanks.
Enter the Ghetto Matrix is probably not exactly what you want, but maybe useful background? The individual cameras just fire once to do the rotation effect, and don't take video.
I wonder if a Raspberry Pi + the coming-soon Camera module would work. I'm not sure if you'd be able to sync the shutters though.
posted by BungaDunga at 8:36 PM on January 13, 2013
I wonder if a Raspberry Pi + the coming-soon Camera module would work. I'm not sure if you'd be able to sync the shutters though.
posted by BungaDunga at 8:36 PM on January 13, 2013
Would it maybe be more manageable to do a computer animation? There are software packages like Poser and DAZ Studio—according to Wikipedia the latter is free with registration—that are designed specifically for setting up scenes with human and human-like figures.
You could even do a "dry run" with a computer animation to test out all of the steps of producing the flip book first, before investing the money and effort to do it with cameras and actors.
posted by XMLicious at 9:42 PM on January 13, 2013
You could even do a "dry run" with a computer animation to test out all of the steps of producing the flip book first, before investing the money and effort to do it with cameras and actors.
posted by XMLicious at 9:42 PM on January 13, 2013
Programming language? processing could certainly do the job, and it's free. It can read video files and output PDFs for printing. There's an active community of processing users who might be able to help you get the code together.
For cheap cameras, check dealextreme.com, all kinds of electronic junk shipped impossibly cheaply from Hong Kong. They have big bulk discounts if you buy a lot of any one thing. You could get a big pile of flimsy tripods there too.
I think you could go a lot cheaper if you use cameras that record onto SD cards rather than those that transmit over a network. You could hack up some sort of shutter release that shorts out the "GO" button on all the cameras at once, or just go push the triggers on each one by hand. It would be a chore to fetch and copy all the SD cards, though.
If you did find networkable cameras, or just plugged 36 USB webcams into a bunch of USB hubs, I'm not sure any computer could handle grabbing 36 live video streams simultaneously, even at low resolution. Ideally, you'd have a setup where the computer could tell all the cameras to record to their cards simultaneously, and then automatically download from the cards one at a time.
posted by moonmilk at 7:59 AM on January 14, 2013
For cheap cameras, check dealextreme.com, all kinds of electronic junk shipped impossibly cheaply from Hong Kong. They have big bulk discounts if you buy a lot of any one thing. You could get a big pile of flimsy tripods there too.
I think you could go a lot cheaper if you use cameras that record onto SD cards rather than those that transmit over a network. You could hack up some sort of shutter release that shorts out the "GO" button on all the cameras at once, or just go push the triggers on each one by hand. It would be a chore to fetch and copy all the SD cards, though.
If you did find networkable cameras, or just plugged 36 USB webcams into a bunch of USB hubs, I'm not sure any computer could handle grabbing 36 live video streams simultaneously, even at low resolution. Ideally, you'd have a setup where the computer could tell all the cameras to record to their cards simultaneously, and then automatically download from the cards one at a time.
posted by moonmilk at 7:59 AM on January 14, 2013
Don't forget the green screen aspect. In the Ghetto Matrix example linked above, you can see the other cameras behind the actor when the shot gets to the far ends of the arc.
posted by scottatdrake at 11:57 AM on January 14, 2013
posted by scottatdrake at 11:57 AM on January 14, 2013
Thanks all, you’ve got me thinking...
ottereroticist: Yes, Fliptography is very close to what I envision, I guess I can’t claim novelty. I’ll have to ask if they’re bringing their skills to ttitd... I’d rather not be compared to competence.
BungaDunga: I’ve really got my fingers crossed for the Raspberry Pi + the coming-soon Camera module. I’m guess’n the camera is powered up and waiting for an instruction, I worry the camera ‘shutter’ has a heartbeat during the wait. I found Allied Electronics has stock on the Raspberry Pi, I’ll order a couple tomorrow to play with (the SD card pricing is in flux, no online orders for the $7 bit).
moonmilk: I’ve ordered a MAKE: book on processing, it’ll arrive Tues.
An ex co-worker had the same though as you on the cameras with SD cards, he came up with a Wi-Fi SD card and a HackHD Camera . HackHD says:
they should sync up just fine if you ground the buttons at the same time. The clocks should match. To test the accuracy I might have to do some testing but I am pretty confident they should both sync up at the same speed if using the same microSD card in both of them.
Having 36 devices simultaneously phone home sounds scarey though.
I think networking will be essential, I expect at least 100 flipbooks/day for the week.
posted by tinker at 8:13 PM on January 20, 2013
ottereroticist: Yes, Fliptography is very close to what I envision, I guess I can’t claim novelty. I’ll have to ask if they’re bringing their skills to ttitd... I’d rather not be compared to competence.
BungaDunga: I’ve really got my fingers crossed for the Raspberry Pi + the coming-soon Camera module. I’m guess’n the camera is powered up and waiting for an instruction, I worry the camera ‘shutter’ has a heartbeat during the wait. I found Allied Electronics has stock on the Raspberry Pi, I’ll order a couple tomorrow to play with (the SD card pricing is in flux, no online orders for the $7 bit).
moonmilk: I’ve ordered a MAKE: book on processing, it’ll arrive Tues.
An ex co-worker had the same though as you on the cameras with SD cards, he came up with a Wi-Fi SD card and a HackHD Camera . HackHD says:
they should sync up just fine if you ground the buttons at the same time. The clocks should match. To test the accuracy I might have to do some testing but I am pretty confident they should both sync up at the same speed if using the same microSD card in both of them.
Having 36 devices simultaneously phone home sounds scarey though.
I think networking will be essential, I expect at least 100 flipbooks/day for the week.
posted by tinker at 8:13 PM on January 20, 2013
Yikes! I just checked Allied stock, down to 6 from ~380 on Friday. I made sure their stock is down to 4.
posted by tinker at 8:22 PM on January 20, 2013
posted by tinker at 8:22 PM on January 20, 2013
« Older Looking for a therapist in Los Angeles. | "We hired someone whose skills more closely match... Newer »
This thread is closed to new comments.
posted by mmascolino at 8:29 PM on January 13, 2013