Lei feng’s network (search for “Lei feng’s network” public attention): author of 7Tens, app public: ICshare sponsors.
Dual camera composition:
Highlights of the high-end smartphones less Lei Tutu-run has been unable to show the advantages of mobile phones. More aggressive twin-camera debut. It says dual cameras, is not like the old smart phone, a webcam before and after, but altogether after using two cameras, simulating human eyes to realize more photographic capabilities and better photographic effect.
In General, the current smart-phone camera interface is MIPI (Mobile Industry Processor Interface) interface. Mobile platforms before only 2 MIPI interfaces, respectively, to the front camera and camera. Dual cameras, requires the platform to support at least three MIPI interface. In the previous high-end platforms, in order to achieve higher pixels, has dual ISP (Image Signal Processing, image signal processing device), such as to support the 16M camera, the ISP will use 2-way 8M. This type of platform is likely to only 2 MIPI, but this could not prevent engineers from going to do before and after camera + dual cameras.
Dual camera application:
The problem is gone, two cameras can do?
1, dual camera range, distance-related applications
Human eye is easy to locate an object’s distance, but when people close one eye, positioning capability will drop a lot. Application of dual camera is to simulate the human eye. Simply put, distance measurement, is calculated by algorithm, was recorded with the camera left/right angle θ 1 and θ 2, coupled with the fixed value of y (that is, two Center distance of the camera), it is very easy to calculate the z value (the distance from the Camera).
But it’s also easy to calculate, if two cameras Center distance is too small, the distance will be calculated. If you want to calculate distance, the distance must be left and right camera remote.
As shown above, due to the dual cameras through the algorithm, can determine the subject distance, so this feature, it’s easy to make some special effects, such as:
A, background blur
SLR cameras are one of the most outstanding feature is a large aperture. Due to the dual cameras can detect objects photographed different distances, the need for large aperture of the object into other objects at different distances blur, you can easily achieve large aperture effects.
(Belle-centric focus, blur the background, and mushrooms on hand)
(A mushroom-centric focus, blur the background and beauty)
B, by replacing the
Because you can measure distance, can be extracted photographed objects in the main, the replacement of background, it can be simpler than the PS, cutout.
C, background effects
Since you can cutout, just for background processing, is very easy. For example, the figure with the sketch effects.
D, measure distance
This figure is obviously identify different object distances, this distance information is marked with a different color. When the AP gained a different distance information, you can achieve the above functions.
2, dual cameras have optical zoom
Optical zoom is mainly about the camera uses a different FOV (view), two cameras for the scene is different. When they need a wide angle photo, use perspectives to 85 degrees and the left camera shooting, get a wide-angle effect. When you need a telephoto pictures, with the right camera angle is 45 degrees for the scene, get a telephoto effect.
In order to make the camera objects overlap and high optical zoom dual camera module not distance cameras of modules that distance is too large, but putting cameras around the closer, the better.
If the two Camera’s FOV is not the same, a large FOV, and a smaller FOV, and algorithm between two optical lens effects, making it easy to achieve optical zoom.
Without dual cameras, after you zoom in the picture, the text is not clear
If using two cameras, after you zoom in the picture, the text remains unclear
This figure is the mix of wide angle and telephoto, through the algorithm to calculate the Middle photo, detail without distortion
3, dark light effect
In General, low light enhancement is in two cameras: RGBG standard for a camera, a black and white camera get rid of RGBG filters. RGBG used to get the object’s color, and black and white cameras to get better light, judged to be objects of light intensity. Then two images merge in order to obtain better low light enhancement.
There are currently two methods of integration:
Picture in black and white as the main body, gets the color of each pixel on a color picture affixed to a black and white picture and two pictures together.
To body with color pictures, black and white pictures get each image’s intensity compensation to color photos, two pictures together.
As to which method is the more appropriate integration may be benevolent, it would not be in this discussion.
Similarly, did the enhancement, in order to make the camera objects overlap and high, such double camera module is the closer, the better.
Need to explain things, and Huawei P9 is the way the selection of modules
Of course, some industry insiders say this algorithm is currently doing the effect is not obvious. Low light compensation to the user really helps, especially at night time. But some customers think that Sony and Samsung’s Dual PD technology very well, more willing to do with Dual PD camera low light compensation.
What is dual camera Dual PD good low light compensation effect, we can compare the Huawei P9 and Samsung Galaxy S7 edge, you will have the answer.
The general use of color + black and white cameras. Through black and white cameras to get pictures of the light intensity to low light compensation to the picture.
4, 3D shooting and 3D modeling
3D algorithm and distance shooting and 3D modeling application is similar, except that it needs higher precision, sometimes even need more accurate distance with infrared distance judgment.
Mentioned double camera of algorithm, had to mentioned ISP (Image Signal Processing image signal processing device), ISP main role is on front-end image sensor output of signal do late processing, main function has linear corrected, and noise removal, and bad points removal, and within plug, and white balance, and automatically exposure control,, rely on Yu ISP to in different of optical conditions Xia are can better of restore site details, ISP technology in is big degree Shang decided has phone of imaging quality.
Function machine era, ISP is doing on the cameras, megapixel cameras with different properties of different ISP. With the cellphone camera pixels and higher, increasing demand for the ISP if the ISP to be integrated on the camera Sensor, will inevitably cause the camera module is too large and even the photographic effect. So the age of intelligent machines, ISP is on a main chip SoC. Some brands in order to achieve better results, even at cost plus an ISP used to achieve better and more professional photographic effect.
Good algorithm needs to match for the camera ISP,ISP and algorithms are complementary and indispensable. Two cameras on the ISP performance requires more. First, in order to make left and right camera signal can be processed at one ISP has been unable to meet the needs of two cameras. This requires dual ISP to implement this functionality.
In low light enhancement, for example, color/black and white image into separate ISP and calibration channels and two pairs of pictures do match (for example, two pictures of the same part extract, remove only the parts of a camera), and then by blocking, detection, algorithms such as compensation-related pictures. Finally, two images blend together, color enhancement. ISP coordination algorithm for actually doing things, of course, far more than this picture says.
Of course, there is also a small episode in this. Are two ISP, how many there are two ISP processing speed, dealing with different issues. In order to ensure both the ISP can be sampled at the same time, you need two cameras shoot out of the picture is taken at the same time. One solution is to give the Sensor a synchronization signal pins. Two camera sync dock, always read a picture, play previous timestamp pictures, ISP by timestamp, ensure the camera around from photos taken at the same time, and eventually fusing.
Different from the 3D movie shoot. Two on the phone during the filming of the camera is not in the image there is sufficient visual difference, this is due to the distance between the two cameras and eyes are not the same. And in order to make it more obvious to get 3D visual effects. Algorithms often need to be enhanced.
Because you can measure distance, subsequent dual cameras not only 3D photography, 3D modeling can also, by this time, I think the importance of dual camera is more important.
Others are increased, such as HDR, increased resolution, these functions are in fact camera can achieve, just two cameras can be made better, does not list them.
For now, these functions are the most common function dual camera mobile phone. Background blur/replace Dim light effects give users a more photographic effect; optical zoom lets us feel the Zoom camera; but I think the future is the most exciting 3D features.
VR so popular this year, where VR material? Also have to rely on dual camera algorithm optimization. If the algorithm of 3D and modeling for the camera mature, will make two cameras are becoming more popular.
Dual camera chain:
Former US dual camera application and principle. Says dual cameras for now we will focus on the industrial chain.
Algorithm 1, dual camera suppliers
Algorithm is needed with ISP coordination, so the algorithm and ISP are complementary, the algorithm well, there must be a good ISP.
As the primary platform vendor, Qualcomm and mediatek have their own ISP, so their dual camera algorithm is developed. The algorithm for better or worse, remains the market testing. disney iPad cases
As a supplier of Sensor, Sony, Samsung, OV is also actively developing double perturbation algorithm, also did not see mass production of products for the time being. But in a functional era, there is no ISP. When the Camera sensor vendors do 2M/5M, have to match their ISP, so the experience of these suppliers is your ISP. Development algorithm of dual cameras, relatively speaking, has relevant experience.
Apple last year bought Linx, patents and algorithm also has a number of cameras, whether in their dual-camera models, just watch this year’s iPhone7. Theory, said Linx has plenty of algorithms, so Apple is not to large IP authorization to buy dual cameras.
In addition to these platforms, sensor suppliers and Smartphone brands outside the Chamber of Commerce’s own algorithm for dual cameras, other vendors we enumerate:
Founded in 1996 in Taiwan huajing, developing independent ISP chip. Some high-end cell phones, cameras and in-car in their ISP cases. Dual camera cell phone in their ISP. Distance applications, optical zoom and low light compensation is made.
An Israel company. Its main advantages on the optical zoom and low light compensation. Depth of field, there is some research. Judging from the media, hTC uses its algorithms.
Rainbow soft was established in 1994, the headquarters in the United States in Shanghai, Hangzhou and Nanjing Technology Center. Rainbow soft strength is the optical zoom and low light compensation. In previous second optical Jiao Tu figure came from Rainbow soft light and dark added. Disney case
Shanghai XING core is a company that was founded in 2011, image processor is mainly engaged in research and development, the current products are mainly used in the automotive market. Independent ISP do few research and development company, is currently developing two cameras algorithms and ISP. After the dual camera market, X-Chip will be a potential dark horse.
Dual camera is just out, so the algorithm, each has its own advantages and disadvantages. However have launched dual camera cell phone from the current results, the algorithm needs to be improved. After various algorithms such as will going to make dual camera mobile phone standard.
2, dual-camera sensor vendors
Due to the dual camera when taking pictures, you need two cameras synchronizing timestamps, which requires Camera Sensor sync signal. Currently, this synchronization signal sensor supplier Sony, Samsung, and OV, Geke micro. So dual-camera Sensor is mainly used this product.
3, dual camera module suppliers
Dual camera module can now do many, light Sapphire, sunny, truly, Namuga, O-Film, Samsung Electronics, the dome of titanium. But mass production experience is mainly Lite, sunny and Lee. Huawei models mainly used up the treasure He Shunyu modules. Electrical and mechanical O-Film and Samsung, with its powerful workstations and now headed off to double camera module field. Namuga is the company to maintain a good relationship with the algorithm, and Samsung mobile phone suppliers, also in the field with double camera module.
Don’t depend on modules but different functions. We continued taking above said four functions, for example:
Application of distance, general practice, large cameras, common camera specifications such as 13M+2M,13M+5M. The following figure:
Or just put some distance in two cameras, as follows:
This two ways, to better calculation are subject to lens distance.
Optical zoom dual camera modules, most notably two cameras require different FOV, similar to the image below:
Different FOV, taken out of focus is different, algorithm, optical zoom will be achieved.
Low light enhancement, through a full color RGB camera + a MONO black and white cameras for filming, as shown below:
Algorithm, primarily through black and white to the right of the camera reads the amount of light, to compensate for the left color RGB camera.
3D photography and modeling modules associated with distance in this way is somewhat similar. 3D distances require a high degree of precision. In that case, a distance between the two cameras placed farther apart, and even some will use additional infrared-assisted positioning to achieve the distance measurement, ultimate 3D shooting and role modeling.
Dual camera module made of difficulties:
Dual-camera approach, generally, there are two kinds: common base plate, or total support. The following figure:
If total substrate, is to put together two Camera Sensor in the same substrate, then a FPC substrate. If the total support, as shown above, through the bracket securing the Sensor, each Sensor has its own Board and its own FPC.
Advantages of common base plate: the two Sensor can sit on the same plate, and is falling.
Total substrate shortcomings: low yields, resulting in expensive.
The advantages of total support: high yield, price is good.
Total support weaknesses: because two independent Sensor module AA calibration is needed to make it in the same plane, difficulty, resistance to bad fall.
In summary both methods have their advantages and disadvantages, at present only by way of common substrates in Huawei is. But either way, the current yield is low, so the cost is very high.
Two cameras in the future:
If improved, modules yield increase, two cameras there are many advantages to what is worth doing.
But because of the different functional modules display requirements are different, two cameras now, may not meet the requirements of everyone on camera.
When Google launched Project Tango, wisely proposed a three-camera concept: when a distance measurement and 3D modeling when needed, can be represented by a travel distance of the camera optical zoom and low light compensation can use two close shots. Even this Depth Sense to combine infrared, more accurate measurement of distances.
In any case, with the evolution of algorithm for dual cameras, increased demand for VR, cameras play an increasing role. Dual camera, even after the first dual camera + three camera phone, the next two years are likely to become more common.