Samsung’s new 200MP HP1 sensor: sensible or marketing?


This week, Samsung LSI announced a new camera sensor that apparently pushes the limits of a mobile phone’s resolution. The new sensor S5KHP1, or simply HP1, pushes the resolution above 200 megapixels, almost doubling that of what is currently deployed in contemporary hardware in today’s phones.

The new sensor is interesting because it marks an implementation of a new binning mechanism, beyond the Quad-Bayer currently deployed (4: 1 pixel binning) or “Nonapixel” (9: 1 binning), but a new mechanism of binning “ChameleonCell” which is able to use both 4: 1 binning in a 2×2 structure, as well as 16: 1 binning in a 4×4 structure.

We’ve known about Samsung’s 108MP sensors for quite some time now, as they’ve been adopted for over two years in both Samsung mobile phones and Xiaomi devices, albeit in slightly different sensor configurations. Probably the most familiar implementation is the HM1 and HM3 modules of the S20 Ultra and S21 Ultra series, which implement the 9: 1 “Nonapixel” pixel clustering technique to aggregate 9 pixels into 1 for regular 12MP captures in most cameras. scenarios.

One problem with the 9: 1 scheme was that you actually had to magnify 3x compared to the two resolution modes, and for devices like the S21 Ultra this was mostly a redundant mode of operation because the phone had a 3x dedicated. telephoto module to achieve a similar spatial resolution of pixels, at a larger pixel size.

8K video recording was a use case when the native resolution of 108MP made sense, but even here the problem is that the native resolution is way above the 33MP required for 8K video, which means the phone has had to undergo a very large field of vision. because it did not support 108MP to 33MP resolution super sampling.

Sensor solution comparisons
Optical Sensor
35 mm
eq. Florida
(H / V / D)

Ventilated disc
Resolution Pixels


24.17 71.2 °
56.5 °
83.7 °
~ f / 1.9

1.15 µm
201.3 million natives
(16384 x 12288)

2×2 50.3M bin
(8192 x 6144)

4×4 trash can 12.6M
(4096 x 3072)

0.64 µm

1.28 µm

2.56 µm

15.2 ″

30.4 “

60.9 ″

1 / 1.22 “
10.48mm x 7.86mm
82.46 mm²

(S21 Ultra)

24.17 71.2 °
56.5 °
83.7 °
f / 1.8

1.09 µm
108.0 million natives
(12000 x 9000)

Tray 12.0M 3×3

0.8 µm

2.4 µm


64.1 ″

1 / 1.33 “
9.60 mm x 7.20 mm

(Mi 11 Ultra)

23.01 73.9 °
58.9 °
86.5 °
f / 1.95

1.18 µm
49.9 million natives

12.5MP 2×2 bin
(4080 x 3060)

1.4 µm

2.8 µm

32.6 ″

65.2 ″

1 / 1.12 “
11.42 mm x 8.56 mm
97.88 mm²
3x telephoto

(4: 3)

27.77 °
21.01 °
34.34 °

(4: 3)

f / 2.4

1.46 µm
10.87M native
(3976 x 2736)

9.99 M 4: 3 crop
(3648 x 2736)

12M to scale

1.22 µm 27.4 ″ 1 / 2.72 “
4.85mm x 3.33mm
16.19 mm²

(4: 3)

8.31 °
6.24 °
10.38 °

(4: 3)

f / 4.9

5.97 µm
10.87M native
(3976 x 2736)

9.99 M 4: 3 crop
(3648 x 2736)

12M to scale

1.22 µm 8.21 ″ 1 / 2.72 “
4.85mm x 3.33mm
16.19 mm²

The new HP1 sensor now has two binning modes: 4: 1 and 16: 1. The 4: 1 mode effectively turns the native resolution of 201 MP into 50 MP captures, and when cropping in a 12.5 MP frame of view, it would result in 2x magnification that would be more in line with what we are. accustomed to Quad-Bayer sensors. In fact, the results here would be roughly in line with Quad-Bayer sensors as the HP1’s native color filter is still only 12.5 MP, which means only one R / G / B filter site. covers 16 native pixels.

A 4: 1 / 2×2 binning mode is more useful, as generally the quality here is still excellent and allows mobile providers to support high quality 2x magnification capture modes without the need for an additional camera module, this is something that is used by a lot of devices, but was missing in Samsung’s own 108MP 3×3 binning sensors due to the structural tradeoff. Samsung LSI even states here that the HP1 would be able to achieve 8K video recording with much less field of view loss due to the less cropping requirements.

200MP – Potentially unnecessary?

This brings us to the actual native resolution of the sensor, the 201MP mode; here, the sensor’s native pixel pitch is only 0.64 µm, which is tiny and the smallest we’ve seen in the industry. What’s also quite strange here is that the color resolution is spatially 4x lower due to the 2.56 µm color filter, so the demosaicing algorithm has to do more work than Quad implementations. -Bayer or Nonapixel usual that we have seen so far.

Source: JEOL

At such small pixel pitches, we run into different problems, and that’s the diffraction limit. Usually these large sensors are used as a “wide angle” module, so they usually have apertures between f / 1.6 and f / 1.9 – the HP1 is a 1 / 1.22 “optical format sensor which is 19% larger than the HM3 in the S21 Ultra, so maybe the f / 1.9 aperture is more realistic. The maximum intensity center diameter of the ventilated disc at f / 1.9 would be 1.15 µm, and generally we tend to say that the diffraction limit where the spatial resolution noticeably degrades is twice as large as this – about 2.3 µm, that’s well beyond the Pixel Size of 0, 64 µm from the sensor.

The 200MP mode might have slight advantages and be able to solve problems better than 50MP, but I very much doubt that we will see much advantage over the current 108MP sensors. In that sense, it seems to me to be a pretty useless mode.

High: 3.76 µm native pixels at f / 6.4 (7.8 µm ventilated disc – 2.07x ratio) – Original
Bottom: 1.22 µm native pixels at f / 4.9 (5.9 µm ventilated disc – 4.83x ratio) – Original

That being said, we are actually seeing camera implementations on the market that are well above the diffraction limit in terms of sensor pixel size and optics. For example, above I removed two 1: 1 pixel crops – one from an actual camera and lens, and the other from the periscope module of the Galaxy S21 Ultra. In theory we should see somewhat Similar resolution, glass optical quality aside, but it is obvious that the S21U has a much lower actual spatial resolution. A big reason here is again the diffraction limit, where the S21U’s periscope at 1.22 µm pixels and an f / 4.9 aperture has a ventilated disk of 5.9 µm, or 4.82 times the pixel size. , which means that physically incoming light cannot be resolved more than a quarter of the resolution of the actual sensor.

For the HP1 sensor to take advantage of its 200MP mode, it would have to have a very large aperture optics to avoid diffraction, a very high quality glass to really solve even the details, and not to be in a very demanding high dynamic range scenario, due to the extremely low full well capacity of small pixels.

To alleviate dynamic range issues, Samsung says the sensor incorporates the latest technology: better deep trench isolation (“ISOCELL 3.0”) should increase the full pixel capacity, while also supporting dual converters. gain (Smart-ISO Pro) as well as offset HDR capture.

The only thing the sensor would lack in terms of more modern features is full-sensor dual-pixel autofocus, with Samsung noting that it uses “Double Super PD” instead, with double the dedicated PD sites like the Super implementations. Existing PDs such as on the HM3. .

Overall the HP1 looks interesting, however I can’t help but say that the sensor’s 200MP mode will have very little practical benefit. In theory, a device would be able to cover the magnification range of 1x to ~ 5x with quite reasonable quality and perhaps avoid having a dedicated mode in that focal length, but we’ll have to see how the vendors design their. camera systems around the sensor. The new 2×2 binning mode is welcome, however, simply because it is much more versatile than the 3×3 mode in current 108MP sensors, and should provide great real-world benefits to the camera experience, although the native 200MP mode does not. proceed as advertised.

Leave A Reply

Your email address will not be published.