In reality, the light-gathering surface of a 8x6mm size 12MP sensor with 2 micron pixels is equal-to-or-less than a 8x6mm 48MP sensor with 1 micron pixels, because there is always some small space between the pixels. Also. In daylight, the 48MP sensor in unbinned mode is still useless, because the diffraction limit is given by the lens, in smartphone the lenses are so small that the effective resolutions for tipical sensor sizes is limited to about 16-20MP at best. It is physically impossible to resolve details beyond 16-20MP resolution on such small sensors (smaller than 1'') with lenses so small. It is also impossible to resolve details fully on pixel-sizes less than 0.8 microns, since red light as a wavelength of 0.8 microns at max.
Anecdotally, a 50MP image in daylight on a Galaxy S23 Ultra is far better than a 12MP image. While the 200MP image resolution difference might be unnoticeable, the increased file size IS noticeable, with far more detail being preserved in lower contrast areas. The biggest advantage to 200MP is definitely the increased detail from file size, not from resolution.
@@vivek_vfrom what you’ve said it sounds like just allowing photos to be compressed less will result in equally good photos as a huge mp count sensor ? No?
Cameras that use pixel binning almost always don't produce more details when using full-resolution mode. The reason is the quad-Bayer filter. Even though the sensor resolution is 48 MPx, it sits behind a 12 MPx color filter. Therefore, the only benefit of high-resolution cameras with pixel binning is bragging rights. Smartphone manufacturers can market their cameras as high-resolution without producing terrible images in lower light.
You have got this all wrong from the beginning... More pixels is more light? Absolutely not true. More pixels pro ide more details. What is really crucial for providing light is the size of the sensor. You can put 108 millions of pixels to a small sensor: it will not help a bit with a low light photography. But hey, now you are asking how is it possible to produce a decent night photo with a small sensor? Here is when artificial intelligence comes in hand. The low light photos are artificially enlightened. So the best sensors out there are big in size (1:1) and DO NOT have huge number of pixels. The number of pixels in best cameras is somewhere between 12 and 36 million pixels.
@@Fee.1 It was existing for 12-15 years but in slightly different way. In smartphone you have blocks with 4 pixels of same colour in one chunk binned to one. When in old cameras it was grabbing neighbor pixels across Bayer pattern in order to get next pixel of seme colour which is not next to ech other so it wasnt 2-to 1 it was 3 to 1 binning so it was slightly blurry but does the same thing without quad pattern. Fun fact that this kind of binning appeared first on large sensors for preserving full sensor data in small file instead of better low light performance, it was just side effect.
This is an old footage when he used to work at android authority . If you observe closely you can see this footage was used in multiple android authority videos.
12MP certainly isn't enough. Unlike a professional camera, you have no zoom options between 1x and 2-3x. 2.5x cropped into a 12MP image is completely unusable, let alone an 8MP image. It seems like some people are quite out of touch with phone cameras.
I think the explanation of Pixel binning could be better if you try and show an example. Having a hard time trying to explain this myself to another person but they still are kind of glazed over when I explain it or shared this video (watching at the same time) Maybe a picture example and then a close-up. I tried using this verbally but I think they cannot see the math so to speak about how this is done
One thing that is not clear to me is if pixel binning impacts on the size of the image.without pixel binning the same camera (same sensor same lens) is supposed to produce different image size if the picture is shot at 4k resolution or 1920x1080 for example.lower res image is supposed to have smaller horizontal and vertical angle of view compared to 4k res.i Guess this is not happening with pixel binning that actually uses the full size of the sensor at every shooting resolution or am i wrong ?
i still dont understand how it works, combining pixel, into 1 pixel, the 4 pixel has its own filter but combining them how do they identify rgb light since it act as one
this makes no sense. the question to answer is why splitting the available sensor surface into 50 or 200 pixels is advantageous over just using larger pixels if you pixel bin and just provide 12MP pics at the end. My canon r5 has 45MP, my Google 6 claims to have 50. loading the raws into Lightroom you see the real quality. the ultra high MP of phones are just marketing bs to win the spec sheet war.
The real question is "how smartphones got so many megapixels and still do shitty photos?" Even the top iPhones and Sammys still struggle with focus and noise.
I still do not understand why the influencers are still not calling out bullsh*t on quad bayer and pixel bining...! I am a professional photographer and I can explain why pixel bining and quad bayers are crap and do not provide ANY advantage over just having a 12MP sensor directly. You want high resolution in day time? then over sampling is the work around which still is crap but some how you have been trained to expect your tiny 1 inch smartphone camera to capture too much detail for posting it on social media. I do not have a UA-cam channel so if you wanna know just email me.
Pixel binning and oversampling are two totally different things. Pixel binning is the analog addition of signals from multiple pixels of the same color (because Quad-Bayer filter) before that signal gets converted to digital. I think it was first used by Huawei P20 pro Pixel oversampling is the digital addition of multiple pixels of a different color (regular Bayer filter). It was used by Nokia 808 PureView and later in Lumia 1020
Oh, so like the iPhone 5 in low light. It would be neat if the iPhone 14 had it on the 10th anniversary of the feature’s introduction to the iPhone. (Phil Schiller only talked about it for a few seconds on stage so most people missed it)
Depends on the size of the sensor. Sharpness/Detail can be capped/bottle necked by the lens, in which case making the pixels bigger will actually give you a sharper image. If the limit of the lens isn't reached, a higher megapixel count will improve detail. It's kinda funny how 108MP images from the S22 Ultra and 50MP from the S22 have pretty much the same sharpness, because the lens is the bottleneck xD
@@Elusive_Chicken I've seen samples of the 108mp and a 12mp from ip13 pro and there's definitely a lot more detail when pixel peeping especially in raws
can anyone suggest maybe a "class of smartphones" that aimed mainly to photography, and not for videogames at all (but for videocalls and messaging)? My target is to minimize non-practical things in order to get the nicest camera I can afford to myself but probably by means of worse processor or less memory...
Samsung S22 Ultra or iPhone 14 Pro are probably your best choices. In my opinion, the whole photography experience from shooting pics to editing them and even importing photos from DSLR is more seamless experience on iPhone. Everything feels more fluid and polished. Samsung on the other hand gives you more freedom. It has 10x zoom lens for extra tight shots. Also the main lens has more megapixels. Android overall is less limited but there isn't same kind of integration to what apple offers. On iPhone, the whole experience of shooting a 4k Dolby Vision video, editing it, then AirPlaying it to Apple TV, all feel like one connected experience. Or shooting ProRaw images, doing some quick edits and then later adjusting the same edits on Mac and copying the same look to all images, all without having to buy Lightroom.
@@eclipticpath Have you even used the S22 Ultra? You seem completely out of touch with an amateur/professional photographer's needs. Suggesting that the iPhone, a phone that doesn't even offer manual photo or video controls in the native camera app, is even a somewhat viable choice for photography is almost criminal. And have you even tried color grading Dolby Vision HDR video from an iPhone? They sure as heck make it nearly impossible to color grade... and that also brings up the question of what serious photographer color grades video on a phone. And what do you mean "without having to buy Lightroom"? Snapseed is a great free photo editing app on Android and iOS that allows for RAW processing and basic photo editing tools. Though the lack of HSL curves is a huge pain that prevents the app from being usable for any type of serious work.
So….in laments terms pixel binning gets worse the higher the mp count goes so computational software has to fill in where the sensor is lacking? PASS. just increase the sensor size and bump the camera up to 16-18mp and get a true 16 or 18mp shot. All pixel binning is doing is creating extra work for the processor that could have been achieved by simply increasing the sensor and pixel size.
They know that but can't do it, it'll be too much work for a thick phone that 1% of the phone market will buy, with this trick they can still keep phones really thin so the majority of consumers woukd be appealed by it and they can make there investment back and more, that's why even the sony phone's isn't going all out and making it super thick.
There is actually NO evidence to prove that pixel-binning actually reduces noise or improved low-light performance compared to just using an equivalent lower-resolution sensor with the same sensor size. If you compare a sensor with 2 micrometer pixel pitch in 12MP mode to another sensor of the same size but with 1 micrometer pixels in 4X binning mode, even though the "maximum" resolution is 48MP, the shots taken in the 4xbinned 12MP mode will produce worse results because of the sub-pixel spacing needed for binning. All you're doing with pixel binning is diving the 2 micrometer pixels into 2x 1 micrometer pixels, but keeping the same resolution. There is NO benefit whatsoever to this technique, the light-gathering is actually worse in the 2xbinned sensor because the space between pixels is wasted, and there is no better noise reduction, the same noise is present regardless of wether your pixels are "normal" or divided into 4 chunks operating together in a binned pixel. What pixel-binning actually does is tricks the consumer into believing the sensor is a higher megapixel count, when in reality the image quality is almost the same in the binned mode.
That's what i think too! Would love to see what a 16 or 20MP 1/1.5" or 1/1.3" sensors in a smartphone without pixel binning can do against current sensors with pixel binning.
"compared to just using an equivalent lower resolution sensor with the same sensor size" of course not... the comparison should be done with a traditional bayer sensor of the same resolution and size... so, 48 MP QB at 1μm vs 48 MP traditional bayer at 1μm. the quad bayer can pixel bin down to 12 MP. the 48 MP can't.
Larger sensor is far better than pixel-binning pixel-binning leads to blurry images Higher pixels like 200mp on small sensors don't improve image quality It's funny how stupid people think pixel-binning is better but in reality u are getting worse images 😂🤣
Literally didnt explain anything. "By combining 9 pixels into 1, smartphones are able to gather more light" WHY? For anyone actually interested why they do this: a diode (a physical pixel) will output a voltage level when exposed to light. Green/Blue/Red diodes can only sample light intensity, so no matter the color of the light, they output a high voltage if its high intensity, low voltage for low intensity. Having a larger diode, means it can catch more light => higher dynamic range, but because smartphone sensors are so small, you can only go so large, so if our dynamic range is 0-255, a large diode can read values from 50-150, but you can split that diode into 4 smaller ones, each tuned with a specific sensitivity (0-50,50-100,100-150,150-200). This way, instead of exposing an image for a longer time period to gather light in a big diode, you expose it for a small period, but each diode reads a different area of the wavelength. That's why manufacturers use pixel binning, so you don't need a longer exposure time to get enough light to saturate a single diode (longer exposure time means more motion blur and you'll have to wait for your phone to take a picture), rather you use smaller ones each sampling a different intensity. Also, keep in mind those diodes have some space between them that's just wasted space as its not used for light gathering.
where can i read about this? ive been trying to educate myself on quad bayers but this is my first time reading this. is this what samsung meant when they said that some pixels take a low exposure image and others a high exposure one for single frame HDR?
In reality, the light-gathering surface of a 8x6mm size 12MP sensor with 2 micron pixels is equal-to-or-less than a 8x6mm 48MP sensor with 1 micron pixels, because there is always some small space between the pixels.
Also. In daylight, the 48MP sensor in unbinned mode is still useless, because the diffraction limit is given by the lens, in smartphone the lenses are so small that the effective resolutions for tipical sensor sizes is limited to about 16-20MP at best.
It is physically impossible to resolve details beyond 16-20MP resolution on such small sensors (smaller than 1'') with lenses so small.
It is also impossible to resolve details fully on pixel-sizes less than 0.8 microns, since red light as a wavelength of 0.8 microns at max.
what if they can put longer focal lengths with larger apperture, is it gonna help with the diffraction issue?
Anecdotally, a 50MP image in daylight on a Galaxy S23 Ultra is far better than a 12MP image. While the 200MP image resolution difference might be unnoticeable, the increased file size IS noticeable, with far more detail being preserved in lower contrast areas. The biggest advantage to 200MP is definitely the increased detail from file size, not from resolution.
@@vivek_v is that because the 50mp image is 1/4 of the 200mp image sensor though??? surely you are still limited by the actual lens?
Can you point me to any idiots guide videos that will help me understand what you’re showing here ?
@@vivek_vfrom what you’ve said it sounds like just allowing photos to be compressed less will result in equally good photos as a huge mp count sensor ? No?
Cameras that use pixel binning almost always don't produce more details when using full-resolution mode. The reason is the quad-Bayer filter. Even though the sensor resolution is 48 MPx, it sits behind a 12 MPx color filter.
Therefore, the only benefit of high-resolution cameras with pixel binning is bragging rights. Smartphone manufacturers can market their cameras as high-resolution without producing terrible images in lower light.
Very informative thanks!
I like how you dismantle these complex concepts keep it up man💪💪
You have got this all wrong from the beginning... More pixels is more light? Absolutely not true. More pixels pro ide more details. What is really crucial for providing light is the size of the sensor. You can put 108 millions of pixels to a small sensor: it will not help a bit with a low light photography. But hey, now you are asking how is it possible to produce a decent night photo with a small sensor? Here is when artificial intelligence comes in hand. The low light photos are artificially enlightened. So the best sensors out there are big in size (1:1) and DO NOT have huge number of pixels. The number of pixels in best cameras is somewhere between 12 and 36 million pixels.
But will it help a lot to have 108mp in sunlight pictures?
Love how pixel binning is more recognizable now than what it was almost 10 years ago
What do you mean? Was it coming back then? Please explain
@@Fee.1 It was existing for 12-15 years but in slightly different way. In smartphone you have blocks with 4 pixels of same colour in one chunk binned to one. When in old cameras it was grabbing neighbor pixels across Bayer pattern in order to get next pixel of seme colour which is not next to ech other so it wasnt 2-to 1 it was 3 to 1 binning so it was slightly blurry but does the same thing without quad pattern. Fun fact that this kind of binning appeared first on large sensors for preserving full sensor data in small file instead of better low light performance, it was just side effect.
What-
Does Brandon Harvard from MKBHD work here at Android Autority? 👀
I noticed that too!
This is an old footage when he used to work at android authority . If you observe closely you can see this footage was used in multiple android authority videos.
@@ayanmajumder4144 ohh okay. Thanks for telling bro
@@dylanmoreira7697 🤣😂
the solution is simple, lower the resolution ( 8 or 12MP is more than enough) and increase the sensor size.
It means more expensive to make
12MP certainly isn't enough. Unlike a professional camera, you have no zoom options between 1x and 2-3x. 2.5x cropped into a 12MP image is completely unusable, let alone an 8MP image. It seems like some people are quite out of touch with phone cameras.
I think the explanation of Pixel binning could be better if you try and show an example.
Having a hard time trying to explain this myself to another person but they still are kind of glazed over when I explain it or shared this video (watching at the same time)
Maybe a picture example and then a close-up. I tried using this verbally but I think they cannot see the math so to speak about how this is done
I agree I’m still confused on how this works.
this video is kinda bad
One thing that is not clear to me is if pixel binning impacts on the size of the image.without pixel binning the same camera (same sensor same lens) is supposed to produce different image size if the picture is shot at 4k resolution or 1920x1080 for example.lower res image is supposed to have smaller horizontal and vertical angle of view compared to 4k res.i Guess this is not happening with pixel binning that actually uses the full size of the sensor at every shooting resolution or am i wrong ?
Thanks for this video! Now I know when to use pixel binning and when not.
When will you use it?
@@MuntahaZad Depends on use case, low light and action shot enable pixel binning to use fast shutter speed and for stills or good light disable it.
i still dont understand how it works, combining pixel, into 1 pixel, the 4 pixel has its own filter but combining them how do they identify rgb light since it act as one
this makes no sense. the question to answer is why splitting the available sensor surface into 50 or 200 pixels is advantageous over just using larger pixels if you pixel bin and just provide 12MP pics at the end. My canon r5 has 45MP, my Google 6 claims to have 50. loading the raws into Lightroom you see the real quality. the ultra high MP of phones are just marketing bs to win the spec sheet war.
The real question is "how smartphones got so many megapixels and still do shitty photos?" Even the top iPhones and Sammys still struggle with focus and noise.
Lol sammys
Sensor size, aperture size, lens size, etc
@@kenmastersmaster ye
Good info
I didn t knew that until now. Thx:)
Nice video! FYI - Bayer is pronounced "Buyer"
I still do not understand why the influencers are still not calling out bullsh*t on quad bayer and pixel bining...!
I am a professional photographer and I can explain why pixel bining and quad bayers are crap and do not provide ANY advantage over just having a 12MP sensor directly. You want high resolution in day time? then over sampling is the work around which still is crap but some how you have been trained to expect your tiny 1 inch smartphone camera to capture too much detail for posting it on social media. I do not have a UA-cam channel so if you wanna know just email me.
Great information.. thanks..
1:52 1:53 these people are mkbhd's editors
Elaborate explanation. Thank you
Can you tell from which smartphone pixel binning or oversampling method introduced ?????
Pixel binning and oversampling are two totally different things.
Pixel binning is the analog addition of signals from multiple pixels of the same color (because Quad-Bayer filter) before that signal gets converted to digital. I think it was first used by Huawei P20 pro
Pixel oversampling is the digital addition of multiple pixels of a different color (regular Bayer filter). It was used by Nokia 808 PureView and later in Lumia 1020
Oh, so like the iPhone 5 in low light. It would be neat if the iPhone 14 had it on the 10th anniversary of the feature’s introduction to the iPhone. (Phil Schiller only talked about it for a few seconds on stage so most people missed it)
it has it now.
Can PIXEL BINNING works on Video?
yes
I wish to see a video from you regarding custom roms
So... More light in night view and more detail in daylight? Wouldn't bigger pixels still have more detail?
Depends on the size of the sensor. Sharpness/Detail can be capped/bottle necked by the lens, in which case making the pixels bigger will actually give you a sharper image. If the limit of the lens isn't reached, a higher megapixel count will improve detail.
It's kinda funny how 108MP images from the S22 Ultra and 50MP from the S22 have pretty much the same sharpness, because the lens is the bottleneck xD
@@Elusive_Chicken I've seen samples of the 108mp and a 12mp from ip13 pro and there's definitely a lot more detail when pixel peeping especially in raws
So the best would be to take a high res pic and a pixel binned image and fuse into one image.
Well done young Harley
can anyone suggest maybe a "class of smartphones" that aimed mainly to photography, and not for videogames at all (but for videocalls and messaging)? My target is to minimize non-practical things in order to get the nicest camera I can afford to myself but probably by means of worse processor or less memory...
Samsung S22 Ultra would be a great choice.
Other than you could consider the Fan Edition version.
Honor Magic 4 Pro might be good for you too.
Samsung S22 Ultra or iPhone 14 Pro are probably your best choices. In my opinion, the whole photography experience from shooting pics to editing them and even importing photos from DSLR is more seamless experience on iPhone. Everything feels more fluid and polished.
Samsung on the other hand gives you more freedom. It has 10x zoom lens for extra tight shots. Also the main lens has more megapixels. Android overall is less limited but there isn't same kind of integration to what apple offers. On iPhone, the whole experience of shooting a 4k Dolby Vision video, editing it, then AirPlaying it to Apple TV, all feel like one connected experience. Or shooting ProRaw images, doing some quick edits and then later adjusting the same edits on Mac and copying the same look to all images, all without having to buy Lightroom.
@@eclipticpath Have you even used the S22 Ultra? You seem completely out of touch with an amateur/professional photographer's needs.
Suggesting that the iPhone, a phone that doesn't even offer manual photo or video controls in the native camera app, is even a somewhat viable choice for photography is almost criminal. And have you even tried color grading Dolby Vision HDR video from an iPhone? They sure as heck make it nearly impossible to color grade... and that also brings up the question of what serious photographer color grades video on a phone.
And what do you mean "without having to buy Lightroom"? Snapseed is a great free photo editing app on Android and iOS that allows for RAW processing and basic photo editing tools. Though the lack of HSL curves is a huge pain that prevents the app from being usable for any type of serious work.
i think xperias are good for that. i barely remember though.
So it's a 12MP sensor.
I’m here before it’s blow because of new iPhone
14 Pro is it! So yeah
i dont understand why smartphones company and sensors company don't use oversampling method to capture photo or video
Sony Xperia Pro I main camera uses only 12Mp and is best camera. Because pixel size is more important
Nice...
Manny from Modern family???
You contradicted what you said in the first half of your video.
So how do we disable pixel binning?
You don't because you can't
you can. If you take pics with the full resolution option, your phone won't use pixel binning.
So….in laments terms pixel binning gets worse the higher the mp count goes so computational software has to fill in where the sensor is lacking? PASS. just increase the sensor size and bump the camera up to 16-18mp and get a true 16 or 18mp shot. All pixel binning is doing is creating extra work for the processor that could have been achieved by simply increasing the sensor and pixel size.
They know that but can't do it, it'll be too much work for a thick phone that 1% of the phone market will buy, with this trick they can still keep phones really thin so the majority of consumers woukd be appealed by it and they can make there investment back and more, that's why even the sony phone's isn't going all out and making it super thick.
Same with Nokia 1020.
What was the same?
There is actually NO evidence to prove that pixel-binning actually reduces noise or improved low-light performance compared to just using an equivalent lower-resolution sensor with the same sensor size.
If you compare a sensor with 2 micrometer pixel pitch in 12MP mode to another sensor of the same size but with 1 micrometer pixels in 4X binning mode, even though the "maximum" resolution is 48MP, the shots taken in the 4xbinned 12MP mode will produce worse results because of the sub-pixel spacing needed for binning.
All you're doing with pixel binning is diving the 2 micrometer pixels into 2x 1 micrometer pixels, but keeping the same resolution. There is NO benefit whatsoever to this technique, the light-gathering is actually worse in the 2xbinned sensor because the space between pixels is wasted, and there is no better noise reduction, the same noise is present regardless of wether your pixels are "normal" or divided into 4 chunks operating together in a binned pixel.
What pixel-binning actually does is tricks the consumer into believing the sensor is a higher megapixel count, when in reality the image quality is almost the same in the binned mode.
True
That's what i think too! Would love to see what a 16 or 20MP 1/1.5" or 1/1.3" sensors in a smartphone without pixel binning can do against current sensors with pixel binning.
@@dushyantsharma4912 you would have to assume it would do better?
but surely that is why they use a bigger sensor? otherwise as you said you get worse results?
"compared to just using an equivalent lower resolution sensor with the same sensor size" of course not...
the comparison should be done with a traditional bayer sensor of the same resolution and size...
so, 48 MP QB at 1μm vs 48 MP traditional bayer at 1μm. the quad bayer can pixel bin down to 12 MP. the 48 MP can't.
I didn't know bending my P6P makes it take better photo 😱
Larger sensor is far better than pixel-binning
pixel-binning leads to blurry images
Higher pixels like 200mp on small sensors don't improve image quality
It's funny how stupid people think pixel-binning is better but in reality u are getting worse images 😂🤣
Yeaaaaah
pixel 6 is better
I think it’s called pixel bidding. I’ve never seen it written out until now. I’m usually never wrong. I’m so smart and great.
Narcissist spotted
@@ShamnathShaji lol fr 😂
Literally didnt explain anything. "By combining 9 pixels into 1, smartphones are able to gather more light" WHY? For anyone actually interested why they do this: a diode (a physical pixel) will output a voltage level when exposed to light. Green/Blue/Red diodes can only sample light intensity, so no matter the color of the light, they output a high voltage if its high intensity, low voltage for low intensity. Having a larger diode, means it can catch more light => higher dynamic range, but because smartphone sensors are so small, you can only go so large, so if our dynamic range is 0-255, a large diode can read values from 50-150, but you can split that diode into 4 smaller ones, each tuned with a specific sensitivity (0-50,50-100,100-150,150-200). This way, instead of exposing an image for a longer time period to gather light in a big diode, you expose it for a small period, but each diode reads a different area of the wavelength.
That's why manufacturers use pixel binning, so you don't need a longer exposure time to get enough light to saturate a single diode (longer exposure time means more motion blur and you'll have to wait for your phone to take a picture), rather you use smaller ones each sampling a different intensity. Also, keep in mind those diodes have some space between them that's just wasted space as its not used for light gathering.
where can i read about this? ive been trying to educate myself on quad bayers but this is my first time reading this. is this what samsung meant when they said that some pixels take a low exposure image and others a high exposure one for single frame HDR?
216mpx
Xiaomi
I guess Apple just jumped into the Pixel Binning bandwagon... ¯\_(ツ)_/¯
mmmmmmmkaaaay.....
🤔🌐🤯🤳🌍🌈👽🦾🍀😘.