There is much discussion here regarding quantum efficiency (QE). Keep in mind that figures for sensors are generally _peak_ QE for a given colour filter array e

There is much discussion here regarding quantum efficiency (QE). Keep in mind that figures for sensors are generally _peak_ QE for a given colour filter array element. These can be quite high like 60-70%. But - this is an 'area under the graph' issue. While it may peak at 60%, it can also fall off quickly and be much less efficient as the wavelength moves away from the peak for say red/green/blue. From what I can tell from the tacky promo videos, the sensor is very sensitive for each colour over a wide range of wavelengths, probably from ultraviolet right up to 1200nm. That's a lot more photons being measured in any case, but especially at night. Their use of the word 'broadband' sums it up. It's more sensitive over a much larger range of frequencies. I also wouldn't be surprised if they are using a colour filter array with not only R/G/B but perhaps R/G/B/none or even R/IR/G/B/none. The no filter bit bringing in the high broadband sensitivity with the other pixels providing colour - don't need nearly as many of those. Edit - one remarkable thing for me is based on the rough size of the sensor and the depth of field in the videos, this isn't using a lens much more than about f/2.4. You'd think it would be f/1.4 or thereabouts to get way more light but there is far too much DoF for that.