Their documentation and video targets this as an astronomy camera pitted up against EMCCD cameras. See video:
That looks cool. This is the kind of thing I'd expect from Canon; their DSLRs have long be favorites for low-light photography - they've even produced variations of some of their cameras (the older EOS-20Da and more recently, the 60Da) optimized for astrophotography.
Does this have any potential for being a surveillance camera? Seems super niche, even far more niche than Avigilon/Canon's 29MP CCD cameras.
Seeing as low light is probably THE holy grail for surveillance - think about it, better low-light sensors means you can use higher shutter speeds and/or smaller apertures for sharper images and better DOF in more situations, even under "okay" lighting, basically solving a lot of other resultant problems - I think this is a technology that, if it's really THAT GOOD, could easily make its way into surveillance cameras.
I would guess Canon is building it into a 35mm sensor initially because the full-frame-DSLR market is probably the most likely to pay the initial adoption costs (full-frame models already being the "cream of the crop" systems), but should the technology prove itself, I don't see why it couldn't be used in smaller sensors.
And that's why EMCCD has been so widely adopted. <grin>
Low light is super important but can Canon get this to a price point and size that works in normal surveillance applications? And will its performance make that much of a difference against mainstream IP camera offerings that are clearly improving themselves.
I imagine it will be like any other technology, filtering down through the various markets, price dropping the more it's implemented. I don't think the question is WHETHER it will reach that price point, but how soon. Get a couple of higher-volume CCTV manufacturers who see a benefit in it and start it out in some higher-end cameras, I expect you'll see it come down fairly quickly.
I didn't notice whether the article said what the cost of this technology is vs. existing sensors - not all major leaps forward come with major price points :)
Yes but most do come with major price points in this space, and you have no evidence to show that this is to the contrary. Not all technologies 'filter down', see EMCCD as an example.
My guess is that a derivative could make its way to CCTV. Canon already makes IP cameras (albeit they're really not a dominant player), making a 5MP 1/2" sensor version of this would work well.
Of course I have no evidence, especially since this is still in development. It's hard to know without knowing how much of a technology shift this really is - is it a major new branch that's required a lot of R&D, or one of those "fortunate mistakes" that's shown how a slight change to way things are already done can make a major difference?
Granted, technologies don't always "filter down"... but hey, how long did it take for CMOS itself to become common in CCTV sensors?
Ok, let's table the speculation - both of ours. Unless someone can share additional facts about this product offering, it is not productive.