A phone built around a 200-megapixel sensor produces 12-megapixel photos by default. Samsung's ISOCELL HP2 groups sixteen 0.6μm pixels into one, behaving as though it had 12.5 million pixels of 2.4μm each. The pixel count drops by a factor of sixteen, while the light-collecting area behind each output pixel grows by the same factor.
The technique is pixel binning, marketed by Samsung as Tetra2pixel. According to the HP2 specifications summarised by GSMArena, binning four pixels yields 50MP at 1.2μm and binning sixteen yields 12.5MP at 2.4μm. The published specs alone are enough to work out exactly what the trade buys and what it gives up.
Sixteen 0.6μm pixels make one 2.4μm pixel
The arithmetic is straightforward. Group an n-by-n block and the effective pixel pitch grows n times while the collecting area grows n squared times, because a pixel gathers light by area rather than by length. A 0.6μm pixel covers 0.36μm²; a binned 2.4μm pixel covers 5.76μm². Exactly sixteen times.
| Mode | Pixels binned | Output resolution (MP) | Effective pitch (μm) | Area per pixel (μm²) | Area multiple | Long edge (px) |
|---|---|---|---|---|---|---|
| Full resolution | 1 | 200 | 0.6 | 0.36 | 1x | approx. 16,330 |
| 4-in-1 | 4 | 50 | 1.2 | 1.44 | 4x | approx. 8,160 |
| 16-in-1 | 16 | 12.5 | 2.4 | 5.76 | 16x | approx. 4,080 |
Long-edge figures assume a 4:3 frame. Two hundred megapixels at 4:3 gives a long edge of roughly 16,330 pixels, falling to about 4,080 after 16-in-1 binning — still comfortably above the 3,840 pixels a 4K frame needs.
The sensor size can be reverse-engineered from the same numbers: 200 million times 0.36μm² is 72mm², which at 4:3 works out to about 9.8mm by 7.35mm, a 12.25mm diagonal. The optical format convention puts 1/1.3-inch at roughly 12.3mm diagonal, so the published spec checks out.
Binning does not add light to the sensor
This is where the common misreading starts. "Binning captures more light" is true only per output pixel. The total light reaching the sensor is fixed by that 72mm² of area and by the lens aperture; binning only changes how that light is divided. It is the same bucket of water poured into 200 million cups or into 12.5 million.
What binning produces is not light but signal-to-noise ratio. The amount of light entering the sensor is set by aperture and sensor area, and nothing downstream changes it.
So why does binning improve image quality? Every pixel carries read noise that has nothing to do with incoming light, and that noise is random across pixels. Sum sixteen pixels and the signal grows sixteenfold while random noise grows only by the square root of sixteen, or fourfold — a theoretical four-times improvement in signal-to-noise ratio. That is why a full-resolution file looks grainy in dim light while the binned output looks clean.
Samsung added circuit-level compensation as well. The HP2 spec sheet relayed by GSMArena lists a Dual Vertical Transfer Gate that raises full well capacity by more than 33%, addressing the tendency of very small pixels to saturate early in bright areas. Judged on published specs, a 0.6μm pixel is hard to justify without that kind of support.
When full resolution actually earns its keep
The value of full resolution shows up in digital cropping. Crop by a factor of n and you keep one n-squared-th of the pixels.
| Digital crop factor | Pixels left in 200MP mode (MP) | Pixels left in 12.5MP mode (MP) | Meets 4K (8.3MP) |
|---|---|---|---|
| 1x | 200 | 12.5 | Both |
| 2x | 50 | 3.1 | 200MP mode only |
| 4x | 12.5 | 0.78 | 200MP mode only |
| 8x | 3.1 | 0.20 | Neither |
Crop the 200MP file by four and 12.5 megapixels remain — the same resolution as the default mode, obtained without a telephoto lens. That is the one unambiguous use for full resolution. The caveat is that the cropped pixels are un-binned 0.6μm originals, so anything short of bright daylight fills the extra resolution with noise.
File size is the other constraint. Sixteen times the pixels means correspondingly more storage and processing time, and in burst shooting the bottleneck is usually the sequential write speed of the storage medium.
The number to check before megapixels
Pixel count is printed largest on the spec sheet and matters least. Two sensors can both claim 200MP while differing by more than 10% in collecting area depending on whether the pixels are 0.6μm or 0.56μm. Pixel count is simply sensor area divided by pixel size, so the starting point for any judgement should be sensor area and pixel pitch.
Manufacturer labelling deserves the same scepticism. Samsung's official specifications list a 200MP wide camera on the Galaxy S25 Ultra, yet the default output is a binned 12MP file. Korean outlet Asia Times made the same observation back in the 108MP generation. The 200 on the sheet and the 12 in your gallery are two sides of one design, not a contradiction.
Who this suits
- Shooting distant subjects in bright daylight — a 2x to 4x crop from full resolution is a genuine substitute for a telephoto lens, and on published specs this is the only place a high-resolution sensor pays off.
- Mostly shooting indoors or at night — leave the default 12MP mode alone. Raising resolution adds no light, only noise.
- Choosing between phones — compare optical format and pixel pitch rather than megapixels. The gap between 1/1.3-inch and 1/1.5-inch matters more than a doubling of pixel count.
- Short on storage — shooting full resolution all the time is hard to justify; switching only when the scene calls for it costs less in both space and processing time.