The story of cubesugar melting with a blacktea

Fujifilm EXR

Fujifilm's EXR technology was first introduced back in 2009 in the F200 EXR. The X10 uses a considerably larger sensor than the F200 did and one based on CMOS, rather than CCD technology, but the underlying idea is the same: to create a sensor that can be used in a series of ways to improve its resolution, dynamic range or noise performance.

There are two aspects to this design: a sensor where alternate rows of photosites can be read separately, giving you the option of treating the chip as if it were two interleaved sensors, and a color filter pattern designed to make the most of this capability.

The color filter array

Almost all digital cameras use what it called a Bayer color filter array, named after the Kodak engineer who developed the pattern. Over the years it's proved to be an excellent way of capturing both color and detail in a scene.

However, much of what the EXR system tries to achieve depends on being able to combine neighbouring lines of pixels, and the Bayer pattern isn't well suited to that. Instead, the EXR array can be thought of as two Bayer patterns slotted together, with one of them 1/2 a pixel offset, compared to the other.

The common 'Bayer' pattern used in most digital cameras	The color filter array pattern of Fujifilm's Super EXR sensor

The reasoning for this start to become apparent when you consider how close the nearest neighbouring photosite sensitive to the same color is. With the Bayer pattern, there's always an entire photosite gap between any two photosites of the same color whereas the EXR arrangement puts pairs of similarly colored photosites together. The benefits of this approach become clear when you see how this is used.

With the Bayer pattern, the nearest similarly colored photosite is always a pixel width away	The EXR arrangement means that there is always an immediately neighboring photosite of the same color

The EXR sensor can be used in three ways. You can either manually select which mode is used or you trust the camera to choose for you. In the X10, EXR Auto mode assesses the scene and compares it to the 54 scene types it recognises. It then checks whether there is motion in the scene, giving it a total of 99 different patterns it can respond to. It then chooses the appropriate white balance, exposure and EXR mode:

SN mode

In low light situations, the camera uses its SN mode which combines data from the pairs of similarly-colored photosite. Because noise occurs randomly, averaging out the noise between two photosites reduces its effect in the final image. This combination of pixels results in a 6MP image, of course.

DR mode

The DR (Dynamic range) mode also makes use of the proximety of similarly-colored photosites. In this mode the sensor starts reading from alternate lines of photosites (ending their exposure), part of the way through the total exposure time.

This means half the sensor stops capturing light early, reducing the chance of it overexposing and retaining more information about hightlight detail. This additional highlight detail is combined with the information from the neighboring fully-exposed photosites in the final images. Again the DR mode images are 6 megapixels.

	The EXR sensor is able to 'underexpose' half of its photosites by reading out data from them before the end of the full exposure. Again the pattern of the color filter array means pairs of pixels can be easily combined.

HR mode

The High Resolution, 'HR' mode treats the sensor just as any other camera would - producing one output pixel per photosite. The EXR pattern isn't quite as good at capturing resolution as a Bayer sensor with the same pixel count though - the downside of clumping together pairs of similar colors is that the gaps between one pair and the next is much greater than the spacing between similarly-colored photosites in a Bayer sensor.

The degree to which this is an issue will depend on the sophistication of the image processing, which we'll see when we conduct a full review.