Samsung has announced the availability of two image sensors designed for smartphone cameras, one cramming 48Mpixels on the sensors, the other 64Mpixels. Both have a pixel size of 0.8µm, hence the 64Mpixels sensor is a bit larger in size than its sibling.
The small size of the pixel is bound to make noise noticeable in low light (having a small size to capture incoming photons means that in low light very few photons will end up on the pixel hence the signal to noise ratio will go up). This would result in poor images. To avoid this Samsung engineers have designed the sensor in such a way that when the light is below a certain thresholds the electronic circuit managing the pixel will use 4 pixels and merge them into one. This will reduce the resolution by 4 (from 48 to 12 Mpixels and from 64 to 16Mpixels respectively). This will bring the sensor resolution au pair with the sensors in today’s top of the line smartphones.
At the same time, when photographing in daylight the circuit will pick up data from each individual pixel, thus delivering the maximum resolution.
Ingenious, isn’t it?
The question is: what can I do with so much resolution? For one you can have a zooming capability that can transform your lens into a tele, still keeping a sufficiently high resolution to make photos that can be printed on a big sheet. For another, applications can do tricks on the data captured by the sensor, like improving the image stability or extending its exposure latitude (HDR) by using nearby pixels in different ways. This results in lower resolution but since you are starting with a very big resolution the smaller one will still be quite good.
Nice to see the ingenuity of researchers, finding ways to go around physical limitation that seemed unsurmountable just few years ago!