Digital Camera Information

It looks like Canon will be continuing to invest heavily on building another CMOS sensor plant expanding on the growth of digital camera. By expanding in another plant, it helps to reduce costs whilst increasing profits.

Here’s what ruters has to say:

“TOKYO (Reuters) - Canon Inc. (7751.T: Quote, Profile, Research) will invest about 55 billion yen ($451 million) to build a new factory in Japan to double its production capacity of image sensors used in digital cameras.

Canon is aiming to produce 24 million digital cameras this year, including 3 million SLRs, which are high-end models that use interchangeable lenses.”

Source

It looks like Sony is investing heavily in its sensor production facility in Japan. According to the press release, Sony Corporation announced that it would invest approximately 60 billion JPY in Sony Semiconductor Kyushu Corporation’s Kumamoto Technology Centre (Kumamoto TEC) Fab 2 facility, to extend clean room facilities by 5,000m² and reinforce image sensor fabrication capacity. This investment will take place from fiscal year 2007 to fiscal 2009.

Sony is positioning image sensors as a key focus area of its semiconductor business. Through the reinforcement of this facility, Sony will strengthen supply and provide the platform for further image sensor business growth.

Kumamoto TEC has continued to expand its operations as Sony’s principle facility for the fabrication of imaging devices, such as the CCDs, CMOS sensors and micro-display devices that are Sony’s strengths. In particular, demand for CMOS sensors has demonstrated rapid growth in recent years. In view of this, Sony is enhancing production operations for this market, with mass production having already commenced at Fab 2’s existing 5,000m² production facility in Spring t

Over the next three years, Sony will continue to strengthen its CMOS sensor manufacturing operations to provide growth markets such as mobile phones and digital still cameras with CMOS sensors that combine high image quality with advanced processing speeds.

Source

According to Kammagamma, they say that Canon has developed a 52 megapixel CMOS sensor in APH-S size.

Here’s what they say in their abstract:

“We have developed a new CMOS image sensor having pixels of more than 52M in APS-H size. The CMOS image sensor has the most number of pixels known to date without stitching. The sensitivity of the monochromatic image sensor is 39000e-/lxï½¥s. The sensitivity of the colour image sensor (green pixel) is 16600e-/lxï½¥s. Pixel size is 3.2um x 3.2um. Random noise is 5.5e- with a saturation level of 24000e-. The CMOS image sensor has 5 x 5 random block readout mode and 4(2×2) adjacent pixels averaging mode. The reproduced image shows splendid high resolution.”

Below is a list of interesting questions and answers from Canon about the new sensor:

Question 1: What are the features of the new image sensor?

Answer 1: The image sensor features a newly developed pixel structure that is well suited to a
smaller pixel size, realizing a signal-to-noise ratio equivalent to that of SLR cameras. It is our understanding that 50 megapixels is the largest number of pixels ever to be fit on a sensor with these surface dimensions (31.6mm x 23.1 mm). The sensor enables high-resolution images to be extracted from the total image captured, making it well suited for the creation of new markets, such as inspection equipment. The sensor has a 5 x 5 random block readout mode for high-speed reading.

Question 2: What problems commonly arise when the pixel size is reduced? (Including such areas as circuit design, production, image quality, etc.)

Answer 2: Generally speaking, the following difficulties are encountered: sensitivity decreases,
the dynamic range decreases, and blending of colours increases. Also, because of the increase in circuit size, there is a tendency toward media delays and slower reading times. Furthermore, due to the further demands in the area of micro fabrication technology, there is also a tendency for yield rates to drop. To address these issues, we have carried out the efforts mentioned in A3 below.

Question 3: What have you done to address the problems that can occur when reducing the pixel size?

Answer 3: We reassessed the structure of the pixels and adopted a new structure that, even with a smaller pixel size, maintains sensitivity, dynamic range, and low colour blending. We
have also made progress in the area of micro fabrication. We also reassessed the circuit block, employing innovations in the area of circuit wiring and a high speed amp to secure a readout speed on par with current speeds.

Question 4: How much of an improvement was realized in terms of the accumulated charge per
pixel compared with conventional CMOS image sensors (taking into account the surface area of the new sensor)?

Answer 4: Compared with an existing product of the same image size, the new sensor achieves
an increase or 50 percent.

Question5: Compared with conventional CMOS image sensors, how much of an improvement in leakage to neighbouring pixels was achieved?

Answer 5: Leakage to neighbouring pixels was improved by about 10 percent.

Question 6: Why did you choose to make this sensor an APS-H size instead of a full-frame 35 mm size?

Answer 6: From a production standpoint, the APS-H size made it was easier to realize the
necessary miniaturization.

Question 7: Is this sensor also compatible with video?

Answer 7: While we have only incorporated limited functionality in this sensor at this stage, depending on the specifications we use, it would be possible to adapt this sensor for video.

Question 8: How far has development progressed? (Is there a prototype?)

Answer 8: We have created a prototype and have confirmed that it is capable of capturing
images.

Question 9: What challenges must be overcome before you will be able to mass produce this sensor?

Answer 9: We would need to increase the yield rate for the sensors, and make necessary
adjustments to match them to the characteristics of the lenses, which would be determined by the demands of whatever application the sensor would ultimately be used for.

Question 10: Compared with 10- and 16-megapixel sensors, how much more expensive would it be to manufacture this 50-megapixel sensor?

Answer 10: That would depend on the specifications and the yield rate.

Question 11: In what Canon products do you intend to use this sensor?

Answer 11: We have still yet to determine how this sensor might be used.

Question 12: What applications would this sensor be suited for in areas outside of Canon’s current product line-up?

Answer 12: Possible applications for this sensor include special surveillance cameras or
industrial-use inspection equipment.

Question 13: What merits would this sensor offer if used in a surveillance camera?

Answer 13: If used in surveillance cameras, the sensor would enable users to view an overall
scene while also enabling detailed close-ups from any given area within that scene.

Question 14: For use in surveillance cameras, what capabilities could such a sensor offer? (For
example, able to read newspaper text from a distance of XX meters.)

Answer 14: If equipped with a lens with sufficient resolving power, in principle, the sensor could
would make a car’s license plate number legible from a distance of 300 meters (approx. 330 yards).

Question 15: Are there any plans to market this sensor to third parties?

Answer 15: We have still yet to determine applications for this sensor. As such, that has still yet to be decided.

Source

Now here is some interesting piece of news from www.smh.com.au which claims that Kodak has developed sensors that can take photos in low light. To me this seems like they are increasing the ISO to about 2 -4 times the normal level, but what about noise. These are no mention of the noise or the image quality that comes out from these sensors and it will be interesting if this is just a marketing hype or a true breakthrough.

Unlike Fujifilm which has already proven their sensors can handle low light situations with the Fujifilm FinePix F series such as the award winning F30, Kodak yet to prove their claims.

Here are some of the comments:

“Eastman Kodak said it has developed a color-filter technology that at least doubles the sensitivity to light of the image sensor in every digital camera, enabling shutterbugs to take better pictures in poor light”.
“We’re talking about a two-to-four-times improvement in (light) sensitivity.”

“It’s often the simplest concepts that can have the most profound impact,” said Chute of IDC, a market research firm near Boston. “This could be revolutionary in terms of just changing that very simple filter on top of the sensor and basically allowing companies to use it in all different kinds of cameras.”

Source

Samsung has announced an 8.4 megapixel CMOS image sensor with a 1.4㎛pixel design at Samsung’s fourth annual Mobile Solution Forum. This is the worlds smallest CMOS sensor.

The new CMOS image sensor chip provides a very high signal to noise ratio which is a key measure of overall image quality.

This was achieved by implementing advanced light sensing features and minimized noise levels. Notably, an extended photo diode technology was implemented to achieve higher light sensitivity and saturation levels, resulting in an enhanced fill factor.

In addition, it provides the same image quality as the charge-coupled device (CCD) image sensors currently used in the majority of digital cameras and camcorders around the world. Since the new CMOS image sensor only uses one-tenth the power of a CCD image sensor, it should quickly replace CCDs in all three key applications-mobile phones, digital cameras and camcorders.

The CMOS image sensor product line of System LSI Division is one of the five major product areas that Samsung has been focusing on to enhance and balance the company’s overall competitiveness. Samsung’s present portfolio of CMOS image sensor technology spans across the range of 1.3 through 5 megapixel resolutions with the 8 megapixel CMOS image sensor expected to be available in the second half of this year.

The CMOS image sensor market is expected to show high demand for high-resolution devices. The current outlook toward 2009 shows a compound annual growth rate of over 90 percent for 3 megapixel and higher resolution devices over a four-year term from 2006 through 2009.

Source

It seems that Canon is the only camera manufacturer to really push the advantages of a full framed sensor. They have two DSLR, the EOS-1Ds Mark II and the EOS 5D which incorporates the full-frame CMOS (Complementary Metal Oxide Semiconductor) image sensors. As Nikon and other top camera manufacturers have yet to enter the full frame market, Canon has really started to push the advantages with the release of the 30 pages White Paper on their Full Frame CMOS sensor. Interestingly they also explain why it is more expensive which in their own terms, is their justification to charge more for these camera. Here what they say:

.. “For now, appreciate that a full-frame sensor costs not three or four times, but ten, twenty or more times as much as an APS-C sensor. Here, then, is the greatest disadvantage of full-frame sensors and the greatest advantage of small sensors. Regardless of future technological developments, cameras with full frame sensors will always cost much more than cameras with smaller sensors. That’s why the EOS Digital Rebel XT, EOS 20D and EOS 30D are such excellent values, and it is also why the EOS 5D and the EOS-1Ds Mark II must come with a substantial price differential. (Interestingly, the APS-H sensor of the EOS-1D Mark II N is the largest size that can be imaged in one shot onto a wafer. Extended through the whole sensor production process, the difference in price between the 1D Mark II N and the 1Ds Mark II can be readily understood.)”

The White Paper on Canon Full frame CMOS sensor can be download here.

According to an interview with the Assistant Manager, Product Planning and R&D Department for Nikon, Charles- André Wamberque, full frame sensor may become a reality when the cost decreases.

Here’s what he says:

“We never said we abandoned the full frame. But then again, we never said when we will introduce a camera with a full frame sensor. As you might have read in the Canon white paper – full frame sensor is still six times the price of an APS-C one. That has a big impact on the retail price. 5D costs 3500€ and it’s not easy to sell such an expensive product. We have to be careful not to make the full frame idea too popular. We’re definitely still thinking and working on it but I cannot tell you when it might happen.

DX lenses are only four years old and they are still very popular. The quality is still very good and the users know it. But sometimes the full frame is necessary.”

Click here for the full interview.

After a long 3 years, Sigma will be releasing a new DSLR called the SD14 on the 26th September 2006. The new Sigma SD14 replaces the Sigma SD10 which is the only DSLR in the world to incorporate the Foveon X3 CMOS sensor.

Rumors has it that the new Sigma SD14 will have a 14 megapixel sensor. I have always admired the Foveon X3 sensor and I’m a bit surprised it has not been used in other DSLR.

The Foveon X3 sensor is different in that it captures all RGB colours at each and every pixel. The sensor features three silicon-embedded layers of pixel sensors, stacked to take advantage of silicon’s ability to absorb red, green, and blue light at different respective depths. This results in more accurate colours, sharper resolution, pixel for pixel than any conventional image sensor without the need for interpolation.

Conventional image sensors feature just a single layer of pixel sensors in a tiled mosaic pattern. Each pixel detects only one colour of light, which has been filtered through a single red or green or blue colour filter. As a result, two thirds of the colour information at each pixel is blocked out. Therefore to compensate for this, colour interpolation is used. Unfortunately, this leads to colour errors, colour artifacts and loss of image detail.

It will be really interesting to see the new specification of the SD14 and whether the new sensor can improve on the poor night performance mentioned in reviews of the SD10.

OmniVision has introduced the world’s smallest NTSC camera with a sensor size of only 1/18 inch. The new sensor is called the OV6920 and will be use in disposable cameras for medical applications such as intubation and diagnostic systems. Measuring 2.1 mm by 2.3 mm, the CSP packaged chip will enable a 3.2 mm diameter microscopic camera module making many medical procedures even less invasive for the patient. It addition, it allows certain medical devices such as intubation systems and endoscopes to be made even smaller to allow for use on infants and small children.

Hasan Gadjali, OmniVision’s Vice President for Advanced Products says “We consider it a major accomplishment by our excellent engineering team to enable such a significant reduction in size over our previous smallest 1/7-inch camera chip. The closest thing in the market today is a 1/12-inch CCD,”.

“We are encouraged by the initial response to the OV6920. Several customers are already evaluating samples and are very excited about the possibilities this small camera can offer.”

The OV6920 is also an ideal solution for low-cost applications requiring both excellent color video performance and a very small footprint.

The ultra small OV6920 is a single-chip OmniPixel2 CameraChip system, which only needs clock and power to get the NTSC composite signal out for direct interface with a VCR/TV/monitor. The OV6920 is designed for very low power consumption,
requiring only a single 3.3-volt DC supply.

Source

Here is an interesting article by www.photoreporter.com that summarizes the differnces in digital camera formats between the different manufacturers. The formats includes Full Frame 24 x 36mm, Four Thirds 17.3 x 13mm and the common APS-C 22 x15mm.

Here’s what they say about each format.

Full-Frame Format

With a continuing strong commitment from the likes of Canon, Nikon, Pentax and Sony, APS-C in its slight variations certainly seems destined to be the dominant DSLR format for years to come. Though it is possible that full-frame DSLRs at one end and FT-format cameras on the other will nibble at its market share. But unless something dramatic and unexpected happens, like an entirely new DSLR form factor or a radical world-beating design, it sure looks like all three formats will be around for years to come. I’m not normally a wagering man, but I’d bet on it.

Source