HARDWARE

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


This week, Samsung LSI announced a new camera sensor that seemingly pushes the boundaries of resolution within the mobile phone. The new S5KHP1, or simply HP1 sensor, raises the resolution above 200 megapixels, almost doubling that of what is currently used in modern hardware in today’s phones.

The new sensor is interesting because it indicates the implementation of a new binning mechanism, in addition to the currently implemented Quad-Bayer (binning 4: 1 pixels) or “Nonapixel” (binning 9: 1), but also a new binning mechanism “ChameleonCell” which can both use edging 4: 1 in 2×2 structure, as well as 16: 1 edging in 4×4 structure.

For some time now, we have been familiar with Samsung’s 108-megapixel sensors, which have been adopted for more than two years, both in Samsung’s mobile phones and in Xiaomi devices, although in slightly different sensor configurations. The most well-known implementations are probably the HM1 and HM3 modules in the S20 Ultra and S21 Ultra series, which implement the “Nonapixel” 9: 1 pixel binning technique to aggregate 9 pixels into 1 for common 12MP shots in most scenarios.

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

Recording 8K video was the case when the original resolution of 108 megapixels made sense, however even here the problem is that the original resolution is far above the required 33 megapixels for 8K video, which means that the phone had to suffer a very large field of view because did not support super sampling resolution from 108MP to 33MP.

Comparisons of sensor solutions
Optics Sensor
35mm
eq. FL
FoV
(H / V / D)
Aperture

Airy Disk
Resolution Pixel
Pitch
Pixel
Res.
Sensor
Size
HP1

(Theoretical)

24.17 71.2 °
56.5 °
83.7 °
~ f / 1.9

1.15µm
201.3 million natives
(16384 x 12288)

50.3M bucket 2×2
(8192 x 6144)

12.6M 4×4 trash can
(4096 x 3072)

0.64µm

1.28µm

2.56µm

15.2 ″

30.4 “

60.9 ″

1 / 1.22 “
10.48 mm x 7.86 mm
82.46mm²
HM3

(S21 Ultra)

24.17 71.2 °
56.5 °
83.7 °
f / 1.8

1.09µм
108.0M originally
(12000 x 9000)

12.0M 3×3 trash can
(4000 x 3000)

0.8µm

2.4µm

21.4 ″

64.1 ″

1 / 1.33 “
9.60 mm x 7.20 mm
69.12mm²
GN2

(Mi 11 Ultra)

23.01 73.9 °
58.9 °
86.5 °
f / 1.95

1.18m
49.9 million natives
(8160 x 6120)

12.5MP 2×2 basket
(4080 x 3060)

1.4µm

2.8µm

32.6 ″

65.2 ″

1 / 1.12 “
11.42 mm x 8.56 mm
97.88mm²
S21U
3x telephoto
70.04

(4: 3)

27.77 °
21.01 °
34.34 °

(4: 3)

f / 2.4

1.46µм
10.87M original
(3976 x 2736)

9.99M crop 4: 3
(3648 x 2736)

12M scaled
(4000 x 3000)

1.22µм 27.4 ″ 1 / 2.72 “
4.85 mm x 3.33 mm
16.19mm²
S21U
10x
Telephoto
238.16

(4: 3)

8.31 °
6.24 °
10.38 °

(4: 3)

f / 4.9

5.97µм
10.87M original
(3976 x 2736)

9.99M crop 4: 3
(3648 x 2736)

12M scaled
(4000 x 3000)

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

The new HP1 sensor now has two binning modes: 4: 1 and 16: 1. 4: 1 mode effectively converts the original 201 MP resolution into 50 MP shots, and when cut into a 12.5 MP frame, it would result in a 2x magnification which would be more in line with what we are used to in Quad-Bayer sensors. In fact, the results here would be fairly consistent with Quad-Bayer sensors as the original color filter for the HP1 is still only 12.5 megapixels, meaning that one location of the R / G / B filter covers 16 original pixels.

The 4: 1 / 2×2 binning mode is more useful, as overall the quality here is still excellent and allows mobile vendors to support high quality 2x zoom shooting modes without the need for an additional camera module, something many use devices but lacked Samsung’s own 108MP 3×3 binning sensors due to compromise in structure. Samsung LSI even states here that the HP1 could achieve 8K video with much less field of view loss due to lower cropping requirements.

200MP – Potentially pointless?

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


Source: JEOL

At such small pixel heights we encounter various problems, and that is the diffraction limit. Typically, such large sensors are used as “wide-angle” modules, so they usually have openings between f / 1.6 and / 1.9 – HP1 is a 1 / 1.22 “optical format sensor that is 19% larger than the HM3 in the S21 Ultra, so perhaps the aperture of af / 1.9 is more realistic.The maximum mean diameter of the air disk intensity at f / 1.9 would be 1.15µm, and we are generally inclined to say that the diffraction limit at which the spatial resolution noticeably degrades is twice that of – about 2.3 μm, which is well above the 0.64µm sensor pixel size.

The 200 MP mode could have some slight advantages and be able to handle things better than 50 MP, but I doubt we will see many advantages over the current 108 MP sensors. In that sense, it seems to me to be a pretty pointless way.


Top: 3.76 μm original pixels at f / 6.4 (7.8 μm breathable – 2.07x aspect ratio) – original
Bottom: 1.22µm of original pixels at f / 4.9 (5.9µm breathable – 4.83x aspect ratio) – Original

Given this, we actually see camera implementations on the market that are far above the diffraction limit in terms of pixel size of sensors and optics. For example, above I took out two 1: 1 pixel clips — one from the actual camera and lens, and one from the Galaxy S21 Ultra periscope module. In theory we should see to some extent similar resolution, the quality of the glass optics aside, however it is evident that the S21U shows a far lower actual spatial resolution. One big reason here is again the diffraction limit, where the S21U periscope at 1.22 µm pixels and the aperture f / 4.9 has an airy disc of 5.9 µm, or 4.82 times larger than the pixels, which means that the physically incoming the light cannot solve more than a quarter of the resolution of the actual sensor.

In order for the HP1 sensor to take advantage of its 200 megapixel mode, it is necessary to have optics with a large aperture to avoid diffraction, a very high quality glass that even solves details, and not to be in very demanding dynamic range scenarios, due to extremely low full capacity small pixels.

To alleviate concerns about dynamic range, Samsung says the sensor incorporates the latest technologies: better deep trench insulation (“ISOCELL 3.0”) should increase the full capacity of pixel wells, and support dual-gain converters (Smart-ISO Pro) as well as gradual HDR recording.

The only thing the sensor would lack in terms of more modern features is dual-pixel autofocus with the entire sensor, with Samsung noticing that it uses “Double Super PD” instead, with dual-purpose PD locations as existing Super PD implementations, like the HM3.

All in all, the HP1 looks interesting, however I can’t get rid of the fact that the 200 MP sensor mode will have very few practical advantages. In theory, the device could cover a magnification range of 1x to ~ 5x with quite reasonable quality and perhaps avoid a dedicated way of working at that focal length, however we will have to see how vendors design their camera systems around sensors. The new 2×2 binning mode is still welcome, simply due to the fact that it is much more versatile than the 3×3 mode in current 108 MP sensors, and should provide real big benefits to the camera experience, even if you don’t use the original 200 MP mode as advertised.



Source link

Naveen Kumar

Friendly communicator. Music maven. Explorer. Pop culture trailblazer. Social media practitioner.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button