Yole Développement recently launched the 2015 market and technology report entitled ‘Status of the CMOS Image Sensor Industry’. This report analyses the recent trends that are reshaping this industry. Sony has excelled in mobile photography and therefore captured a quarter of the global image sensor market. Yole Développement had a great opportunity in interviewing Mr Yasuhiro Ueda, SVP Corporate Executive, Senior General Manager of the Image Sensor Business Division of Sony Corporation to find out more about the future of Sony and of CMOS Image Sensors in general.
| 1- What is Sony Device Division’s importance within Sony Corporation? Within our organizational structure we have different divisions responsible for their respective activities. The Device Solution Business Group (DSBG) is in charge of our device business. Our major component business group, including semiconductors and batteries, belongs to this DSBG. Our image sensor group activities are also included under the DSBG unit. 2- Sony was a leader in the charge-coupled device (CCD) legacy technology and succeeded in the transition to CMOS, becoming the undisputed leader in 2014. Could you explain how Sony, as an image sensor company, has succeeded in beating the odds of disruption and did not end up like Kodak, IBM, or Nokia? |
 Yasuhiro Ueda, SVP, Senior General Manager of Image Sensor Business Division, Device Solutions Business Group, Sony |
The book “Innovator’s Dilemma” describes how disruptive technologies impacted the disk drive industry and how existing players unable to properly adopt the transition fell behind. This may have been the case for the companies you mentioned.
In addition the book also describes how technologies can be among the extended lines of product evolution. We didn’t position CMOS image sensors as a revolutionary, disruptive, product totally different to CCDs but instead found them to be among the extended lines of our CCD evolution. Transition necessities in CCD were led by our customer’s needs for taking high definition videos. CCD works fine handling 480p SD resolution videos but higher speed HD is difficult. As a sensor division it was then just a matter of course to consider our transition to CMOS sensors, as they have much greater speed advantages. We didn’t position this move as anything innovative as the evolution was a process driven by our customers’ needs.
3- How important is Sony’s Integrated Device Manufacturer model? Do you think it helped for the transition CCD to CMOS? And what about the current capital investments?
We have in-house production and are also out-sourcing some production to external foundry companies as the fabless companies do. As you are aware, our main sensors have a stacked design. These stacked sensors are fully manufactured in-house but widely available logic or similar products are supplied from external parties. Throughout our CCD years we accumulated unique process knowledge and today this is one of our definite core competences. We consider our competitiveness is coming from the strategy to complete our in-house manufacturing among the essential areas with a fully controllable environment. We have no intention to change this strategy.
4- Sony invested 35B Yen during FY14 ~ FY15 for image sensors, and similar amounts in the three previous years. Which part of this investment is linked to CMOS image sensor technology and which part is linked to capital investment?
The 35 billion Yen investment is solely for the production capacity increase. For R&D activities and ordinary investment for maintaining image sensor manufacturing, there is a different budget.
5- Do you see a major technological breakthrough ahead of us in CMOS image sensors (CIS) at device level, optics level and software level?
I shall talk a little about modules and the image sensor’s role therein later but concerning technologies related to digital signal processors (DSPs) and optics, these are handled by our other business groups with a different R&D budget.
6- How do you see the evolutions after 1.1µm/0.9µm? We will then be very close to the wavelength limit. Are there trends in image sensors emerging? For example, decreasing pixel size was a big trend but do you think this is now reaching some limit?
The image sensor’s cell size is indeed reducing. Today our developments are around the 1 micron cell level and we shall continue working towards reducing the cell size further. Lately academic demonstrations have proven light can still be acquired even when the sensor’s aperture is extremely small. This kind of technology is applied in our bigger cell size α7S SLR digital camera, capable of taking pictures in a pitch dark environment. It really depends on our customers’ demand but we feel the cell size reductions are meaningful to improve light usage efficiency and therefore believe we should continue our developments in this direction.
7- Increasing image quality sometimes means bigger pixels, and bigger pixels mean higher prices for image sensor manufacturers. Using bigger pixels seems to be a really important trend and Sony’s development of 3D stacking technology has helped push in this direction for better quality while maintaining a compact footprint. Do you see this as the peak of the trend?
Today you find many digital still cameras achieving the resolution of 24 million pixel digital SLRs without low pass filters. Some people may want better images and higher resolution, other people may want good quality but with a low pixel count. The direction depends on the camera producer’s approach. Your opinion is reasonable but we feel there are two directions.
Some people want to take pictures of the Eiffel tower during daytime. You don’t need the smallest sensors since it’s bright enough and you can enjoy a good image with high resolution. If you want to do the same at night this is also possible with pixel summation in order to reduce the noise and in this case you don’t need too many pixels.
Some people ask for greater pixel count as the application range is higher. On the other hand some find it bothersome and just ask for larger sensors. The demand is coming from the customers and we are prepared to answer either way.
8- There are a few technologies inching closer to realization. One is computational imaging and dual cameras, another one is 3D cameras. What is Sony’s position on computational imaging?
Computational imaging shall contribute further value to our sensors and we are highly interested in that sense. Our sensors are using a stacked structure and it makes sense adding features to the logic side to create differentiated values in the sensor. In line with this approach we have a strong interest in image processing technologies and so forth.
We are and also have been interested in supporting the path of Dual cameras and likewise 3D camera to answer our customers’ needs, meaning the demand is coming from customers.
9- May I ask whether Sony is interested in infrared imaging or not?
Yes we are highly interested. Not only infrared but also ultraviolet and x-ray as well.
10- What do you think will be the next technical evolutions in mobile phone applications, as the pixel size decrease is now going to slow down and mobile phones are becoming thinner and thinner?
To meet the thinner, smaller mobile phone requirements we need to consider many enablers like pixel evolution and computational imaging but it’s also a fact that mobile phone cameras have less image quality compared to SLR cameras. To reach that SLR level work has to be achieved with enabling technologies such as optical, image processing and sensors. We are and will continue challenging the SLR camera quality. Mobile phone cameras also have a bad video quality compared to broadcasting gear and similar to the SLRs we shall be aiming to challenge these quality levels. 
11- Sony would like to take more share in module manufacturing. Do you think future innovations could come from the system level, from the software or at the optics level?
In the case of module manufacturers, sensors and modules are normally developed separately from each other. In our case we do both at the same time, meaning we have optimization advantages. We are dealing with very sensitive sensors easily damaged by dust and similar environment issues. Under this context an established knowledge of producing semiconductors is critical when the modules are made in the same manufacturing line. This is something only we can do. For instance an optical axis adjustment will need know-how about optimizing the sensor and lens axis in 3D. We have good knowledge in digital imaging and thanks to this we can perfectly handle a 6-axis adjustment. Players who don’t have this knowledge need to compromise by enlarging the optics’ dimensions. Our manufacturing line can perfectly handle small size, high image quality, and dust challenges. It makes us totally different to the players who are only making modules.
12- Sony pioneered automotive imaging with Nissan, but it seems this is not your priority right now. How important do you think is automotive for the future of image sensors and for Sony imaging activities?
In the beginning there was only the backside view requirement and no particular safety standards. This was when Sony entered the automotive business. We had quite an amount of business but then new concepts like ADAS (Advanced Driving Assistant System) came into the field, and along with them a number of safety standards were set that products needed to comply with. We were very much aware of the new standards but decided not to follow and started losing our share of the business. This doesn’t mean we didn’t have any interest in the automotive market, we knew it would grow. It was simply a decision and concentration of means. We had decided to focus and place our resources intensively on the mobile phone market. Thanks to this strategy we enjoy a great part of the mobile phone business today. Now we are really starting to aim at the automotive market. We are planning to be in top position around 2018 onwards when sensors for automated driving will be in demand. Our work on satisfying standards is steadily progressing and we are in discussion with the tier 1, OEM players to put our top position aim into practice.
13- How do you see the overall evolution of the industry in areas such as M&A and the business models of the players? How do you see the consolidation of the industry? 2014 has been a very important year with much consolidation. Is Sony interested in taking part in it?
When we see the players they all have different roles in the market and from this point of view we don’t think there will be a necessity for further consolidation. Having said we find there are too many module makers today and we feel this is where the need to consolidate will arise.
Source: Status of the CIS industry 2015 report, February 2015
14- What are the next three new applications for CIS in the next 5 years from your point of view?
First, it should be the automotive market. Then we believe the “new mobile imaging” market will grow. This is a field consisting of action cameras, drone cameras and new moving region concepts. The third is an area which we defined as a “sensing region”. Traditionally we defined sensing in respect to imaging to be an enabler in order to enjoy high quality images. Today images are by contrast starting to be used as data, like you see in the automated driving technologies. Image data can be used in various ways and the Internet of Things market is one area that will steadily grow. We are trying to draw our next growth strategy plan around these sensing opportunities and this will include a variety of products, in infrastructure, entertainment, agriculture and more.
15- The last question is related to the strategy and direction of Sony. Is it correct that you have an interest in various markets and technologies and it is your intention to move in all directions?
To seek our possibilities we shall continue studying the different markets but we can’t say we will be interested in “all” directions. There shall be areas to focus our resources on but there shall also be those where we shouldn’t. We are not interested in certain home applications where less valued VGA sensors can play a role. We neither would be entering the low pixel selfie market for smart phones, nor the 720P low resolution surveillance cameras. We shall apply our resources where they shall be valued best.
Source: www.yole.fr / http://www.sony.net/
Source: Status of the CMOS Image Sensor Industry’ report
留言列表
生技醫療產業 (1)