Crave

(Credit: Fujifilm UK)

The next generation of Super CCD sensor that Fujifilm announced last September, the Super CCD EXR, makes its debut this month in the FinePix F200EXR compact point-and-shoot. The sensor technologically weds Fujifilm's older Super CCD HR (high resolution) and Super CCD SR (more sensitive) in a single chip with 12 million photosites. It can generate 12-megapixel images by using the data from each site individually, or 6-megapixel images by combining adjacent same-color sites. Although the latter technique, known as binning, isn't new, Fujifilm claims it does it better job because it doubles up the adjacent same-color pixels.

The F200EXR's implementation will offer shooters access to these capabilities via three different modes: a standard 12-megapixel mode; a 6-megapixel Pixel Fusion mode, which bins the pixels to increase light sensitivity for theoretically better low-light photos; and a 6-megapixel Dual Capture mode, which combines two simultaneous exposures where every other pixel captures either shadow or highlight detail. All in all, a pretty intriguing idea. The camera will also have an EXR Auto mode, which uses scene detection to automatically choose which of those three modes to use.

(Credit: Fujifilm USA)

The rest of the camera seems like pretty typical fare. It has a 5X 28-140mm lens and 3-inch LCD, and incorporates sensor-shift image stabilization. Though the company states a maximum sensitivity of ISO 12,800, that's at 3 megapixels; it does ISO 3200 at full resolution and ISO 6400 at 6 megapixels. It includes all the standard Fujifilm features, like film-style simulations, flash output compensation for shooting in macro, and dual-shot (with and without flash), plus category standards like enhanced face detection. And--yay!-- this model will take SD cards in addition to the company's albatross xD-Picture Cards.

It's slated to ship this month for $399.95.

Kodak 50-megapixel CCD

(Credit: Kodak)

update: 7/10/08: I made some errors in the original post, which I've corrected, plus added Kodak's comment.

You thought Sony's 24-megapixel CMOS was high res? Well, as Kodak's announcement on Tuesday of a 50-megapixel CCD shows, there's always room for more--pixels, that is.

Granted, that's a bit of an apples-to-oranges comparison. Sony's chip is designed for full-frame dSLR cameras, those with a sensor the size of a 35mm film frame (24x36mm), which generally go into pro-level handheld cameras. In contrast, Kodak's KAF-50100 CCD is 49.1x36.8mm, for medium-format digital photography,which tends to be used more by commercial and fine art photographers in studio settings. The KAF-50100 is only the latest in Kodak's line of high-resolution medium-format CCDs. For instance, it joins Kodak's 39-megapixel KAF-39000 in the lineup, which now becomes the second highest-resolution sensor for non-scientific imaging. Hasselblad recently announced the H3DII-50, a camera based on the 50-megapixel CCD, in addition to its older 39-megapixel H3DII-39. Professionals fork over upwards of $30,000 for models like these.

To fit more pixels on the same chip, Kodak had to perform some voodoo shrinkage on them; they're 6 microns, compared with 6.8 for the KAF-39000. Kodak claims that the chip has increased data throughput with a maximum of 18MHz per output vs. 24MHz for the 50-megapixel and 39-megapixel CCDs, respectively. Since the newer chip uses a 4-channel output, up from 2, that's an overall bandwidth increase. While the specs on the two cameras show the higher-resolution version to be faster by one measure--1.4 seconds per capture for the H3DII-39 over 1.1 seconds for the H3DII-50--the higher-resolution model has an overall slower capture rate of 33 captures per minute vs. 39. I'd attribute that last slowdown to the larger 50mp files the camera has to process.

Other performance specs seem to take a hit; quantum efficiency (how successful the electrons are at getting where they need to go), blooming protection (how well the chip handles the electron overflow caused by bright light) and dynamic range seem to drop as well, despite Kodak's claim that "key performance parameters" are retained compared to the current 6.8-micron designs. The company also says the new chip consumes less power, but there are no specs to judge by. (Check out the specs yourself.)

Kodak's comment:

While there are changes in other specifications for this device compared to the current sensor, our customers have indicated that these are minor and will not significantly impact overall camera performance in these markets. For example, while blooming protection is indeed slightly lower for the new device, it is still well above the level of protection needed for this application. And while dynamic range is also a little lower, it is still above 70 dB, which is a critical threshold for this market.

That sounds reasonable to me. Really, though, it does take some magic to cram that many more pixels in a small space without losing anything. We'd look forward to seeing what Kodak pulls out of this hat, but we probably won't be lucky enough to get our hands on the camera. Do you want us to try?

PhaseOne Chief Executive Henrik Hakonsson is bridging a vast digital photography divide.

A medium-format camera with a Phase One digital back.

His company makes top-end image sensor housings called digital backs, each costing tens of thousands of dollars and attaching to high-end medium-format cameras with similarly high price tags. But he just signed a partnership with Microsoft, which gears its products for the broadest possible audience.

The Phase One product that brings these two worlds together is Capture One, software that helped pioneer the area of processing "raw" images taken directly from image sensors without any in-camera processing. The software exists chiefly for Phase One's high-end customers, but it also supports many mainstream cameras.

Through the partnership, terms of which were not disclosed, Microsoft will help Phase One tackle technical challenges of improving that software, Hakonsson said. And according to Josh Weisberg, Microsoft's director of digital imaging evangelism, Capture One will be able to handle files encoded with Microsoft's HD Photo format, which the company is advocating as a higher-quality replacement for the ubiquitous JPEG and is standardizing as JPEG XR.

Phase One, based in Copenhagen, was founded in 1993 and is owned by its 130 employees. On the hardware side, its top-end P45+ back uses a 39-megapixel sensor from Kodak. It sells two versions of Capture One, the $499 Pro and the $99 LE, but with the upcoming version 4, the LE version will simply be named Capture One 4.

I chatted with Hakonsson about his company's software, hardware, and Microsoft alliance earlier this month. Here's an edited transcript.

Phase One CEO Henrik Hakonsson

(Credit: Phase One)

Q: Most people haven't heard of Phase One. Can you give us a thumbnail sketch?
Hakonsson: Phase One is the world's leading digital camera back manufacturer. If you take a copy of Vogue magazine and look at the first 50 pages, approximately 80 percent of the images are shot with Phase One digital back and Capture One software. Our position in the market is the very top maybe 1 percent of the photo segment--shooters who work with the biggest clients and the most demanding photographic applications.

What's your sales volume for digital backs?
The global market will exceed 10,000. Phase 1 has more than 50 percent of the market. Some of our digital back competitors are working to make less costly solutions. We try to target the most demanding photographers.

What will result from the Microsoft partnership?
For Phase One, the main reason for doing this was the ability to get access to some tools which will help us provide better services for the kind of photographers we're working with. We're getting into file sizes that may be two to three times what we have today, and the speed of being able to handle these files requires other tools than what we have in our portfolio.

For me, performance is No. 1. The parameters on which we position our product are speed, image quality, and ease of use. On the performance side, we needed a partner.

How big are your image files?
Typically 150MB. We expect larger file sizes for the next two to three years. The ability to make sure that people can browse and process images is important going forward. Microsoft has a range of tools for assuring that we can serve our high-end customers, who are the ones we are predominantly concerned about.

Eastman Kodak just sold its first CMOS image sensor for digital cameras. The customer? Eastman Kodak.

OK, that's being a little flippant. Kodak's camera division is separate from its sensor division, and the latter must compete with other suppliers for the camera business, so the deal is a significant achievement in the company's attempt to transform its sensor business.

Kodak will use its new KAC-05011 sensor in the new Easyshare C513, a $99 model with a 3X optical zoom lens and 2.4-inch LCD screen. It's due to ship this month, Kodak plans to announce Tuesday.

CMOS, which stands for complementary metal oxide semiconductor, is the ordinary process by which computer processors and memory are made. Most digital camera sensors today are built with the more specialized CCD, or charge-coupled device, technology.

Kodak's sensor group builds its own CCD products, but it's begun a parallel effort to design CMOS sensors built by IBM and Taiwan Semiconductor Manufacturing Corp. (TSMC). Selling the CMOS sensor--a 5-megapixel model the company promised earlier this year--is a milestone indicating Kodak has attained a certain level of expertise.

All things being equal, a CMOS sensor costs somewhere between 5 and 15 percent less to build than a CCD sensor, but Kodak is interested in the CMOS market more because it also can incorporate some processing tasks, said Michael DeLuca, Kodak's marketing manager for image sensors. For example, it can incorporate circuitry for basic sensor functions such as analog-to-digital conversion or chip timing, he said. And in the longer run, it could house circuitry for reconstructing full-color images from sensors using Kodak's new color filter patterns.

Ultimately, CMOS will likely replace CCD in some product categories, DeLuca predicted.

"For mass-market consumer products, it's probably a question of time," he said.

CMOS sensors are widely used in mobile phone cameras but are less common elsewhere in the digital camera market. The most notable example is Canon, which uses CMOS sensors in its SLR cameras--including some "full-frame" 36-by-24mm ones that are very large by digital camera standards.

The company also has said the sensor will be used in mobile phones.

Leave it to JVC to do something just a little bit different with its high-end consumer products. The company's Everio GZ-HD7, announced today at CES, has all the right creds: three progressive-scan 1/5-inch chips; a Fujinon HD broadcast lens; a hard disk capable of recording five hours of maximum-quality video; and a reasonable $1,799.95 price tag. JVC bills the GZ-HD7 as the first consumer model that records full HD video, which might be true; I could swear that Sony's AVCHD models record at the same 1920x1080 resolution, but I don't have time to check (and the Sony Web site is quite uninformative on this point). It's definitely the first 3CCD model. Regardless, first or not, it's a good thing.

But the company cut corners on the CCDs. First, they're really small. Second, each chip is a mere 976x548 pixels, using JVC's pixel shift technology to generate--dare I say "interpolate"?--it's vaunted "full" HD resolution. The jury on this one will have to remain sequestered until review units are available. The camcorder is slated to ship in April.