Author | Anonymous
With the rapid development of smartphones, the camera function of mobile phones seems to have become a competitive hotspot and marketing point for major smartphone manufacturers.
The main reason is that apart from daily functions like voice and scanning, many mobile phone users habitually use their phone cameras to record life. With their compact size and online platforms, the convenience of smartphone photography is unparalleled by professional cameras.
The main camera of flagship smartphones now offers very good shooting effects, but there is still a significant gap compared to professional cameras. This is mainly due to the limited size of the mini camera module, which makes it nearly impossible to achieve physical changes in aperture and focal length with current technology.
In recent years, smartphone manufacturers worldwide have invested considerable effort in photography, achieving many notable innovations. Excluding some fringe applications, these innovations are generally aimed at one direction: to minimize the gap between smartphone photography and DSLR cameras (even though there is still a long way to go). Currently, manufacturers have broadly divided this goal into three smaller objectives:
➢ Basic imaging quality
➢ Aperture blurring
➢ Zoom
To achieve these three small goals, the main solutions currently are sensor + photography algorithms, AI image segmentation algorithms, and cameras. Among these three goals, the first two have generally met the photography needs of ordinary smartphone users, but zoom still has some distance to cover. Speaking of zoom, this has long been a challenging problem in smartphone photography. Zoom can be divided into optical zoom and digital zoom.
Optical zoom relies on the structure of optical lenses to achieve zoom, which is done by moving the lens to enlarge or reduce the size of the subject being photographed. It occurs through changes in the positions of the lens, object, and focal point. The longer the lens focal length, the larger the image of the subject, meaning it can capture subjects that are further away, thus achieving higher zoom ratios. Naturally, this requires a larger movement space for the internal lens and light-sensitive components. Therefore, the greater the zoom ratio, the larger the lens size, making it very difficult to install large lenses in smartphones, which is why optical zoom is rarely used in mobile phones.
In smartphones, digital zoom is commonly used. Although it is called “zoom,” the focal length does not actually change. Instead, it enlarges the distance between two pixels in the original image using the phone’s processor, and then determines the color surrounding the existing pixels to interpolate and enlarge part of the pixels on the sensor to fill the entire image. This technique is similar to using image editing software to increase the area of an image, which comes at the cost of lossy cropping of pixels. Although digital zoom attempts to improve image quality using interpolation and other methods, the color and quality of the image significantly decrease.Therefore, digital zoom has a significant gap in photo details compared to optical zoom, especially for distant shots, where this disadvantage becomes very pronounced.
Comparison of Optical Zoom and Digital Zoom Imaging
The emergence of dual and even multi-camera smartphones has brought a new direction for smartphone zoom and is currently the mainstream zoom solution. This solution essentially uses cameras with different focal lengths, switching cameras when the zoom reaches that focal length, while other processes still rely on digital zoom. This is a hybrid zoom solution, and not true optical zoom, with clear pros and cons. The advantage is that when reaching a fixed zoom ratio, it is indeed lossless; the downside is that aside from the fixed ratio, other zooms remain lossy.
We take the Huawei Mate20 Pro as an example to explain the working principle of “smartphone optical zoom”.
The Mate20 Pro is equipped with three rear cameras in a “pillow” style:
1) 40-megapixel main camera, 27mm wide angle
2) 20-megapixel ultra-wide angle camera, 16mm ultra-wide angle
3) 8-megapixel telephoto camera, 80mm telephoto
Huawei claims that the Mate20 Pro has 3× optical zoom (80/27=2.96≈3), rather than 5× optical zoom (80/16=5), so it can be inferred that the ultra-wide angle camera’s involvement in the zoom process is not very high. Further research into the Mate20 Pro’s specific “optical zoom” principle:
1) When the image is at 0.6× to 1×, the 20-megapixel ultra-wide angle achieves digital zoom;
2) When the image is at 1×, the 40-megapixel wide-angle main camera is used to capture, producing a lossless photo;
3) When the image is at 1× to 3×, digital zoom is achieved using the main camera;
4) When the image reaches 3×, both the main camera and telephoto camera are used to produce a 3× lossless photo;
5) When the image is at 3× to 5×, digital zoom is implemented using the main camera and telephoto camera;
6) When the image is at 5×, the 8-megapixel telephoto camera operates independently to capture a lossless photo;
7) When the image is at 5× to 10×, the 8-megapixel telephoto camera still operates independently to achieve digital zoom.
Although, by definition, dual/multi-focus cameras achieving “optical zoom” are not true optical zoom, the results achieve the same effect as optical zoom in terms of output. Limited by the characteristics of smartphones, in the short to medium term, smartphone optical zoom will still innovate on dual/multi-camera “optical zoom”. To output higher magnification lossless photos, it is inevitable to add longer focal length telephoto cameras in the future and achieve zoom quality equivalent to optical zoom through algorithms.
This April, Huawei’s flagship P30 Pro was launched with a periscope telephoto camera, achieving an astonishing 10× hybrid zoom and up to 50× digital zoom. Whether it is the moon hanging over the treetops or the sunset on the horizon, previously unreachable beautiful scenery is now within reach. We predict that in the future, “ultra-wide angle + wide angle or standard focal length + TOF + multiple fixed focal length telephoto cameras” will become a trend.
As mentioned earlier, when zooming in a smartphone, switching between several cameras is required to cover focal lengths from ultra-wide angle to telephoto, meeting the user’s need for a single device to do it all. This switching of cameras needs to be accomplished by the MIPI switch. Generally, the CSI (Camera Serial Interface) interface of mobile phone platforms is limited, and it is impossible to assign a CSI interface to each camera. The MIPI switch allows sharing of the platform’s CSI interface.
Below is a typical application of the MIPI switch under the D-PHY protocol:
The selection of the bandwidth of the MIPI switch needs to be determined based on the actual pixels of the camera used, with a general formula as follows (taking a 20-megapixel camera as an example):
Each pixel: 10bit (common), the output of 20 million pixels uses D-PHY with 4 data channels, and the output frame rate is 30fps, allowing us to calculate the transmission rate of each data channel:
The industry typically requires the bandwidth (-3dB) of the MIPI switch on a channel to be greater than the data transmission rate. Of course, the higher the bandwidth of the MIPI switch, the less signal attenuation occurs. Similarly, a comparison table of camera pixels and MIPI switch bandwidth requirements can be obtained:
AW35646 has a bandwidth of up to 4.5GHz, perfectly supporting 60 million ultra-high pixels.
Generally, 30fps is acceptable, and the picture feels smooth. However, if the frame rate is increased to 60fps, a smoother picture can be obtained, significantly enhancing interactivity and realism; hence, current cameras offer several different frame rates.
For a 20-megapixel camera, using the formula above, the relationship between frame rate and bandwidth can be calculated:
For a 20-megapixel camera, AW35646 can perfectly support a frame rate of 90fps.
In the transmission line interconnection structure (TLIS) between the camera and the AP image signal processing unit (ISP), the inserted analog switch can be seen as a media channel, as part of the transmission line interconnection structure. As one implementation of the transmission line interconnection structure – printed circuit board (PCB), it is a crucial and challenging aspect of high-speed data transmission integrity design, requiring minimization of discontinuities to achieve impedance matching.
If the maximum allowable attenuation for the transmission line interconnection structure is -6dB, using a switch with 3.0GHz bandwidth means that at 3.0GHz frequency, the attenuation of the MIPI switch is -3dB, leaving a maximum allowable attenuation of only -3dB for the remaining part of the transmission line interconnection structure. However, if a switch like AW35646 with a bandwidth of 4.5GHz is used, the attenuation of the MIPI switch at the same 3GHz frequency is only -2dB, allowing a maximum allowable attenuation of -4dB for the remaining part of the transmission line interconnection structure, providing sufficient margin for the PCB’s high-speed transmission design to meet the signal integrity needs of the entire channel.
By comparing the eye diagrams of switches with 3.0GHz and 4.5GHz bandwidth at data rates of 1Gbps, 2Gbps, and 2.5Gbps, and measuring rise times, the advantages of high bandwidth switches in eye diagrams are illustrated.
The simulation schematic is as follows: (The rise/fall time of the original signal is 0.2*UI)
Below are specific test eye diagrams
1Gbps data rate eye diagram
2Gbps data rate eye diagram
2.5Gbps data rate eye diagram
|
Data Rate |
4.5GHz Bandwidth Rise Time |
3.0GHz Bandwidth Rise Time |
|
1Gbps |
175ps |
198ps |
|
2Gbps |
121ps |
144ps |
|
3Gbps |
102ps |
135ps |
➢ At a 1Gbps data rate, limited by the bandwidth of the signal itself, the difference in eye diagram open size is not significant; the 4.5GHz bandwidth is only faster in rise time compared to the 3.0GHz bandwidth.
➢ At a 2Gbps data rate, the difference in eye diagram open size becomes apparent, with the signal quality of the 4.5GHz bandwidth being better.
➢ At a 2.5Gbps data rate, the difference in eye diagram open size becomes even more pronounced, highlighting the advantages of the 4.5GHz high bandwidth with shorter rise times.
Thanks to the high bandwidth switch of 4.5GHz, the rise/fall times of the corresponding output signals are shorter, and the eye diagram opens wider.
Awinic, through rigorous market research and a high starting point, has launched the high-speed four data channel MIPI switch AW35646, which features industry-leading data bandwidth of 4.5GHz, meeting the demand for higher pixel cameras and higher data transmission rates in the future smartphone market.
This chip integrates 10 single-pole double-throw switches (5 differential channels) optimized for high-speed transmission, making it very suitable for MIPI data transmission and switching for two sets of 4 lanes or fewer, with a static power consumption as low as 25μA to meet the low power requirements of electronic products like smartphones.The AW35646 chip provides wafer-level packaging WLCSP-36B.
For more product information
Please call
021-54271166
Or send an email to
for inquiries
Click the mini program below toview all job details!