Snap! Nokia's gyro stabilised camera tech now on open market The tech that makes Nokia's smartphone cameras so awesomely good has become available on the open market for the first time. Gyroscopic stabilisation, which means cameras' shutters can stay open for longer without blurring the final image, has been around for a while in digital cameras. Cramming the technology into the minute …
Language isn't maths and doesn't work like maths. If you insist on seeing the word "pretty" as having no purpose other than to describe the word "unique", you get it wrong. That's not how people actually use words. That everyone knows perfectly well what "pretty unique" means, even if they object to the usage, demonstrates that the phrase sends its intended signal perfectly effectively.
"Language isn't maths and doesn't work like maths".
Rubbish. That's like saying that two wrongs don't make a right, when they clearly do.
Given that pedantry is high art round these parts, I think the linguistic pragmatists are on a hiding to nothing, and should accept that "unique" is a concept that is absolute.
Well, if I may out-pedanticize your pedantry, I agree that "unique" is an absolute concept and didn't say otherwise. What I said was that the word "pretty" conveys a linguistic signal beyond and apart from mere modification of the word "unique". So nerr.
> That's like saying that two wrongs don't make a right, when they clearly do.
Upvoted, sir.
well, if thats the case what do you make of 'remarkably remarkable' (architecture professor on the telly) either it is, then it needs not the emphasis or it's simply ahem, advanced normal. The current fashion for (seemingly) anything above the ordinary to be described as 'unbelievable' despite being there, in front of eyes and therefore very much believable etc. As for pretty unique, well sorry to pop your bubble mate but this IS crappy English, it should be 'pretty much unique' or 'pretty well unique' or better still 'damn near unique'. see how much more precise it is. Using your approach would leave us with 'like unique' being like acceptable these days...
You might just as well object to "damn" being used to describe things that haven't actually been and are never likely to be damned by God or "sinister" being used to describe people who aren't left-handed or "terrific" being used to describe things that do not cause anyone to experience any terror or "nondescript" to describe something that is actually known to the world of science or "fast" to describe something that can move or "fucking" to describe something that isn't having sex. And maybe you will.
This is not replacing anything, it is a part that will move the optical component of the camera (not the sensor) contrary to the movement of the camera on a number of axis' - there is not a part that does anything of this sort in any non-Lumia smartphone today.
Surely this is only a minor part of Nokia's special sauce. They're using image stabilisation, but also over-sampling on bigger sensors and then a bunch of processing in the camera, before the image is passed onto the phone's chip. I thought the size of the raw images was too big for either Android or Win Phone to process, as the OSes hadn't been written with this in mind.
I'm not sure Nokia had any realistic chance of a hardware advantage. It's just that they did the work to integrate it all, and get it small enough not to make a huge bulge in the phone.
Replicating it was always going to be research budget + time.
The actual algorithm that chooses which pixels are good to keep and which are to be thrown away is also part of the sauce. They baked into into a separate GPU just for the camera. If it's that good on a mobile with a crappy lens, I'd like to see what it could do in a compact camera. Nokia could do well for themselves by licensing all the sauce, but no doubt MS will have a clause in the contract to prevent them doing it any time soon.
No, other companies sell phones that do the same, they just don't have Microsoft's marketing budget to allow them to shout about it.
The Nokia pictures are nothing to write home about. The latest high end Xperias take better pictures in daylight and low light...
Not sure what you mean - my Panasonic video camera outputs 1080p/50 video to an SD card without losing any frames & that's ~~3MB/sec coming off the camera admittedly already compressed to .MTS format (Codec: H264 - MPEG-4 AVC (part 10) (h264))
"You're still going to get jerky video recording due to frame dropping because of SSD card latency, if you use that to store your media."
Sorry to be blunt but you are talking out of your backside
1 - Frames would not be dropped in a video pipeline because of a slow write speed unless you were recording in a non stream format (i.e. CinemaDNG) - doubtful that we'll see any smartphones doing that soon.
2 - I assume you meant SD Card, not SSD - there are cards that are capable of recording raw video streams, for example the BMD Pocket Cinema Camera writes to an SD card and is capable of putting out 30 frames per second in DNG format, quite a few SD cards are capable of that.
3 - Latency has nothing to do with this, write speed is what is important - you need to read up on the difference before posting and trying to educate the world on your thoughts
It's a static ring of material they get to resonate in a particular plane. Due to Coriolis force, any rotation in the axis perpendicular to the ring causes the vibration to move in sympathy which is detected capacitively as with most mems devices. 3D sensing means making rings in three different planes which is difficult to do on a flat silicon surface.
Nokia 1020 camera has movable optics which takes up the bulk of the volume. The mems sensors are only one small component.
Foucault (of pendulum fame) observed that if he placed a metal rod in the chuck of a lathe and hit it from above, it would vibrate up and down - as one would expect. What he found notable, however, was that if he then rotated the chuck of the lathe, the plane of vibration didn't change, i,e, he didn't observe the rod wobbling from side to side, it was still up and down.
MEMs gyros are based upon this same principle, but are constructed using techniques developed for silicon chips.
You can add all the fancy stuff you want, all the megapixels, all the processing.
It comes back to the problem that cell phones have really tiny CCDs. Until they can work out a way to make them much larger without having a big thick protrusion all these improvements will only be at the margins. Someday we'll see a phone with a lens made from a metamaterial so they can use a 1" CCD but still have a thin body without the lens protruding and the pictures it'll take will make everything else look look a decade behind.
> so they can use a 1" CCD but still have a thin body without the lens protruding
You are unaware of the limitations that affect cameras, or believe in magic.
You are forgetting the effect of 'focal length'. A 'Four-Thirds' camera has about a 1" CCD (22.5mm). Having the 'thin body' of 8mm or so will give a focal length of maybe 4-5mm. On a 1" CCD (crop factor 2.0) this is an 35mm equivalent of 8-10mm - super-wide-angle or fish eye. It would have to have a small aperture to get a reasonable focus across the width of the CCD.
Another issue is that on a flat CCD the light collection is done in stacks and these are most effective when the light arrives perpendicular to the surface. Quality lenses for digital cameras produce the image so that it is as perpendicular over the whole surface as possible.
In phones or compact cameras the easiest way to do that is to have a tiny sensor - which also makes the equivalent focal length more useful.
> and the pictures it'll take will make everything else look look a decade behind.
No, it would just look like crap.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
