Anonymous, 18 Dec 2019Try making one 12mp sensor with 2.4 micron pixel size and fit that into a smartphone That sensor would be the same size as the one inside S11+.
Sam, 18 Dec 2019Exactly, this has always been my point, why make 108mp sensor, and then need to reduce it to ... moreTry making one 12mp sensor with 2.4 micron pixel size and fit that into a smartphone
Anonymous, 18 Dec 2019For this size of sensor , f1.7 is more than enough.
People that think bigger aperture is bett... moreOf coure f/1.7 and OIS would be extreme for this sensor size. It would be equivalent to an iPhone 11 sensor with f/1.0 lens. An aperture of f/2.55 would capture the same amount of light as the Galaxy S10 at f/1.5.
Therefore I bet that Samsung will use a variable aperture. In order to get the previous Galaxy S10 f/2.4 depth of field, their variable aperture needs to have f/4.1.
Whackcar, 18 Dec 2019I think this is the wrong approach. Going for 100+MP cameras, only to capture 12MP image in th... moreYou mean, Sony did.
Htc bought the sensor. Did not made.
One of the older iphones had 8MP 1,5 micron.
Performs worse than iphones 12MP with 1,12 micron.
Whackcar, 18 Dec 2019I think this is the wrong approach. Going for 100+MP cameras, only to capture 12MP image in th... moreMany people don't know that a high megapixel count makes the image look less noisy, if pixel size is constant. So a phone with large pixels, but with a very low megapixel count, usually produces noisier results than a phone with slightly smaller pixels, but with a much higher megapixel count. That's a reason why HTC's approach wasn't so smart.
Low light performance is approximately proportional to
(Pixel size)² x (number of megapixels) x exposure time / (f-number)²
Only when the noise gets irregular, the megapixel count gets less important. Though even then a higher megapixel count will still make the image look less noisy.
Whackcar, 18 Dec 2019Nona is a multiplier value for 9 times (same as octa is for 8 & deca is for 10).
See th... moreSo then its dicta cell not Nona cell.
Thats also multiplier.
And much more effectuive in this word to make a sense.
I think this is the wrong approach. Going for 100+MP cameras, only to capture 12MP image in the end is a waste. It takes more processing power, and also doesn't provide any major improvements to daylight images (and even low-light result might not be impressive as we expect). If they just wanted to increase the pixel size, then they should've developed a 12 MP sensor with 2.0-2.4um Pixel size. HTC was able to do it way back in 2013 (HTC One M7 had a 4MP camera with 2um native Pixel size), and the only reason it wasn't well-recieved is because people weren't willing to settle for 4MP when the competition had started offering 13MP. But now, 6 years later, Samsung could easily create a 12+MP camera with native large pixel size without the need for pixel binning.
If they want to go for massive MP count, then at least provide a legitimate benefit in max attainable resolved detail (as you get with Mi Note 9, Mi Mix Alpha & so on with 27MP images).
CptPower, 18 Dec 2019Its nanocell not nonacell.
They have a mistake there.
Because if its a non-a-cell then its... moreNona is a multiplier value for 9 times (same as octa is for 8 & deca is for 10).
See this: https://en.wikipedia.org/wiki/IUPAC_numerical_multiplier
Sam, 18 Dec 2019Exactly, this has always been my point, why make 108mp sensor, and then need to reduce it to ... moreBinning reduces noise and each pixel captures a photo with different exposure so you have HDR photo without your phone taking multiple pictures and combining them this also reduces motion and image in 9-1 binning each of the 9 pixels will capture an image at different exposure so you can have 9 different exposure levels which will greatly improve HDR pictures
Anonymous, 18 Dec 2019Wouldnt a large 12 mpix sensor without binning have larger pixels, and thereby better low ligh... more12MP sensor of same size will have 2.4nm pixel size same as 108MP with 9-1 binning so low light should be pretty similar
from smallest sensor in market but still top tier camera (cause good software?)
to largest sensor in market (plus good software?)
Next year, Flagship camera phone war gonna be crazy.
Dude , 18 Dec 2019From nano to nona. From micro to mocri.Nona means nine, tetra-penta-hexa-septa-octa-nona. I was suspecting this would elude some people.
"that's significantly larger than the current quad-bayer implementations having 1.6µm effective pixels. This would ensure stellar low-light performance and dynamic range in all scenarios."
Pixel size doesn't affect image quality much in low-dynamic range scenes or average light conditions, when sensor size, etc are constant. Pixel size only gets really important when the image quality starts to fall apart. When the image quality of a certain area is "good", then larger pixels won't make it "very good". It will be still just "good".
Anonymous, 18 Dec 2019If the f-number and shutter speed are not high, then compared to 27 megapixels this likely won... moreFor this size of sensor , f1.7 is more than enough.
People that think bigger aperture is better (ignoring many other factors) will complain.
They think phones need f1.0, f1.1 .....
Anonymous, 18 Dec 2019Wouldnt a large 12 mpix sensor without binning have larger pixels, and thereby better low ligh... morePixel size refers to the pixel pitch, which won't change. It would be still 2.4 micrometers. But yes, it might be a bit more efficient. The main advantage of Quad Bayer sensors is a low resolution camera and a high resolution camera (but with terrible demosaicing artifacts) without the costs of two cameras. Some people talk about Quad Bayer HDR modes, but no company has really confirmed that they use this technique and it's not clear whether Quad Bayer HDR works better, Apple and Google don't use it. In fact, Samsung's article even doesn't mention HDR. https://m.gsmarena.com/samsung_details_the_features_of_its_108mp_isocell_bright_hmx_sensor-news-40631.php
If the f-number and shutter speed are not high, then compared to 27 megapixels this likely won't really improve the noise of normal low-light photos because you don't capture more light. Only in complete darkness or very dark shadows there will be an advantage because read noise, which is dominant in complete darkness, is less visible with larger pixels. Also, a high shutter speed and high f number (more depth of field) will be more usuable in low-light.
But normal low-light conditions with a low shutter speed and low f number won't benefit much.
Therefore I think a 16 or 17 megapixel sensor of the same size could have been better because usually the high resolution modes produce terrible demosaicing artifacts.
Why not just make a proper 12mp sensor, with 2.4um photosites? It will perform better at high ISO and will not require different pixels to be exposed for different intervals just to get noise reduction and dynamic range up to par with a normal bayer array sensor.
The Last Oracle, 18 Dec 2019Is this the first flagship (Apart from Huawei), that is using the Quad Pixel and larger cluste... morethey won’t do that all at once. It’s a gradual process, because they would be making use of the same sensor for consecutive generations just like what they did with s7
The Last Oracle, 18 Dec 2019Is this the first flagship (Apart from Huawei), that is using the Quad Pixel and larger cluste... moreLittle Problem:
Dynamic range of small cellphones sensors are pure software.
Not natural from the high amount light caught by big sensors.
FF has more DR than Apsc.
Apsc has more DR than 4/3.
4/3 has more DR than 1".
They boost HDR to prevent exposure issues, but end up with artificial photos.
Wrong WB, brighter areas than they are.