Benchmarks

1- Processing time

I’m really obsessed with the performances of our app, and I keep pushing my guys to improve the software as much as possible. There are ways to speed up things by having proper hardware, network configuration and proper settings. Once everything is sorted out, you should get results as fast as what we’ve been measuring here at the studio

Bullet-time

(time required to download all files, apply digital calibration and render the full mp4 at 1080p, ready to replay and share)

Number of cameras File type Dimension Hardware Duration
16JPG1080p4 Raspberry Pis3s
16RAW6k/24MP4 Raspberry Pis17s
16RAW6k/24MP8 Raspberry Pis15s
16JPG1080pusb hubs + one computer6s
42JPG1080p14 Raspberry Pis6s
42RAW5k/18MP14 Raspberry Pis18s
16RAW6k/24MPusb hubs + one computer23s
16RAW6k/24MPusb hubs + two computers20s
154JPG1080p22 Raspberry Pis19s
154RAW5k/18MP22 Raspberry Pis58s

*** Raspberry Pi’s architecture @ 42 cameras: Canon SL2, One Gigabit Switch (no daisy chaining), Raspberry Pi 3b+, Gaming i7 computer
*** Raspberry Pi’s architecture @ 154 cameras: Canon SL1, Gigabit Switches with daisy chaining, Raspberry Pi 3b and 3b+ mixed, Gaming i7 computer
*** USB hubs architecture: Canon SL2, 7 ports powered usb-hubs, Gaming i7 computer.



2- Trigger precision (usb triggering)

Starting from version 1.11.0+, USB triggering through Raspberry Pi' computers OR using usb-hubs on a single Windows computer is about the same as using an analog trigger (using the trigger connection port). And that is true even when setting 4 cameras on a single Pi. From our experience, the only precision problem left is about the cameras themselves (mechanical shutter latency at more or less 1ms). Our latest tests have been made using Canon SL1/100d and SL2/200d, and we’re working on various test scenarios to provide you guys with more details.

20x SL2/200D with continuous light (USB-Hub architecture)

18x SL2/200D with continuous light (Raspberry Pi architecture)

154x SL1/100D with continuous light (Raspberry Pi architecture)