Imaging Products

  1. Home
  2. Products & Support
  3. Imaging Products
  4. History & Technology
  5. Behind the scenes
  6. Nikon FX-Format CMOS Sensor

Nikon FX-Format CMOS Sensor

In a digital camera, the light from the subject that comes in through the lens is captured using an image sensor and converted into electronic signals, which are recorded on a memory card. Nikon's originally developed large image sensor, the FX-format CMOS sensor, as used by the D3 digital single-lens reflex camera, yields a hitherto unseen level of image quality in the form of rich gradation expressions and high-definition images. The man responsible for its development introduces the appeal of this new technology and some unknown facts behind its creation.

Masato Takeishi
Junior Executive Staff, Development 1, 1st Development Department, Development Headquarters, Imaging Company
PROFILE:
After joining the company in 1988, Takeishi was initially involved in industrial high-definition cameras that use image pick-up tubes or CCDs, before moving on to the development of digital cameras for use in microscopes. Since 2003, he has been in charge of digital single-lens reflex cameras. With the D3 and the D700, he consistently demonstrated an ability for imaging, with which he has been involved since joining the company, and he was involved in the development of the FX-format CMOS sensor. His hobby is motorbikes, but recently he enjoys simply tinkering with bikes more than riding them, as he is an engineer at heart.

Driven by calls for a film sized image sensor

Please tell us about the principal features of the FX-format CMOS sensor, which has been talked about a great deal since its release.

Tekeishi holds the image sensor while he describes its development and application into the D3.

Takeishi:

In terms of appearance, the sensor is as large as 35mm silver-halide film. In terms of capability, it boasts the same high level of sensitivity as ISO 6400 film (the most highly sensitive film in common use) and has a high-speed performance of 9 frames per second. It also features a wide dynamic range that enables the depiction of smooth gradations from dark areas to light. These are its four principal features.

What were the circumstances leading to its creation? Please describe the background to its development.

Takeishi:

Since the release of the D1, Nikon digital single-lens reflex cameras have been equipped with DX-format sensors. However, we received requests from customers who wanted the same lens angle of view and bokeh as 35mm silver-halide film or who wanted to shoot pictures that evoked the same feel as film, so we began to consider a so-called full-size (FX-format) sensor.

How was the product commercialized?

Takeishi:

We first considered whether a camera equipped with a 35mm full-size sensor would be practical, if we could make such a sensor. Silver-halide film comes in the form of a thin sheet, whereas a sensor is a type of IC. The thickness, width and height of the IC package make the entire component larger than film. Moreover, there would be optical components such as an optical low-pass filter and an infra-red filter between the sensor and the shutter. These would all have to be precisely fitted into the camera body. After detailed discussions between those responsible for camera body design and those responsible for lens design on questions such as whether the light from the photographic lens would reach every area of the sensor, we came to the conclusion that it would be possible to build such a camera. At the same time, work on devising possible sensor specifications was progressing. The first thing to be considered was the size of the so-called “full-size” photosensitive surface. It was decided that the size of the recording pixel area would be virtually the same as that of 35mm film. Next, a full calculation of the imaging capability was carried out, and by opting for high sensitivity as the major advantage of a full-size sensor, the sensor was brought into line with the concept of the D3.

What was this concept?

Takeishi:

After considering what kind of equipment we should be offering our target D3 users-professional photographers and high-level amateurs-we believed that it is important to provide a single camera that could be used to shoot almost anywhere. With 12 million pixels, the camera would be able to handle virtually any scene. Furthermore, we placed emphasis on areas of photography that require high-sensitivity and rapid-rate continuous shooting, such as sports, news and wildlife. Taking all these factors into account, we decided on the FX format, 12 million pixels and ISO 6400.

Light condensing performance greatly improved by micro-lens modification

Did you have target figures in mind from the outset, as regards increasing the pixel size for the 12 million-pixel array?

Takeishi:

For both single-lens digital cameras and compact digital cameras, the question of how many pixels are enough is a common one, both by people within the company and by users. The higher the pixel count, the larger the image files generated, and the greater the processing load on the computer.
The question of how high pixel counts will go is also often asked. Nikon's answer is 12 million. This is good enough for printing on an A4 two-page spread.

In what ways do the DX format and FX format differ?

The FX format's larger size is obvious
when compared with the DX format
(right).

Takeishi:

There is no difference between the DX-format sensors and FX-format sensors in terms of basic structure or the read-out method that they use. However, due to the respective sizes of the sensors, there is a difference in the degree of freedom in camera design. If a DX sensor and an FX sensor have the same pixel count, the FX sensor's pixel size can be made larger. Larger pixel size is advantageous to increasing a camera's sensitivity and dynamic range, as more light can be collected and accumulated in the form of electronic signals. Alternatively, if a DX sensor and an FX sensor have the same pixel size, the FX sensor will have a higher pixel count and be able to obtain a higher resolution. All this might suggest that the FX is the more versatile of the two. However, the DX format also has its advantages in terms of cost-performance and compatible camera size.

So what you are saying is that the FX format allows more design freedom and offers a wider range of possibilities. It would seem that there must have been real technical challenges in this regard.

Takeishi:

The larger sensor made it difficult to effectively gather light from the photographic lens onto the pixels around the periphery of the image plane. Compared to the DX sensor, the light comes into the periphery of the sensor at a much more slanted angle. In order to collect this light without any loss, we tested many designs and carried out numerous simulations and experiments.

What specific technologies did you use to overcome this problem?

Takeishi:

We devised a configuration in which a micro-lens and a photodiode are optimally placed on each individual pixel, so that even light slanting in from the photographic lens can be condensed into the photodiode. Since the photodiode's function is to store light in the form of electronic signals, condensing light efficiently-even around the periphery of the sensor-is of key importance. Specifically, in addition to the optimal placement of the central position of the micro-lens and the photodiode, the use of thinner sensor chips also allowed components to be spaced at an optimum distance from one another. In addition, the adoption of a gapless micro-lens and an inner lens allowed all incident light (whether striking the sensor straight on or at an oblique angle) to be directed to the photodiodes with the minimum of loss. By means of this technology, we were able to eliminate the influence of the optical axis angle-even at the sensor's periphery.