Wide Dynamic Range Technology

Even for a scene that includes great differences in brightness, a natural photograph can be recorded by combining 2 images with different exposure times.

Both the brightest and the darkest areas of an image are reproduced smoothly

When shooting a scene that includes great differences in brightness with a conventional digital camera, as shown in Photo 1 (a) (b), detail is lost in bright areas because of saturation and in dark areas because of low light. Dynamic range of a digital camera sensor is typically 7~8EV, where EV (Exposure Value) is a unit of exposure.

When the dynamic range is insufficient, as shown in Photo 1 (a) (b), detail is lost in bright saturated areas and in dark areas with low light. The proposed technology makes a wide dynamic range image automatically by combining two images with different exposure times (long and short). Appropriate exposure areas of each image are combined to make a wide dynamic range image with a maximum range of 12EV.
Photo 1 (c) show a combined wide dynamic range image, without losing detail in bright and dark areas.


(a)Long exposure image
(a)Long exposure image

(b)Short exposure image
(b)Short exposure image

(c)Wide dynamic range image
(c)Wide dynamic range image

Photo1: Different exposure images and a wide dynamic range image

When combining two images shot at different times, displacement often occurs between the images. Direct combination of the two images results in a double-edge image because of the image displacement. The technology uses a high-speed image sensor to capture images so quickly that the ultra-short time difference virtually eliminates displacement.

Advanced technology realizes wide dynamic range of 12EV

This technology uses two images with different exposures to make a wide dynamic range image. Although brightness is different at the same place in the two images, the optimal weight function is designed to combine images naturally. In the dark areas of the combined image, the weight of the long exposure image is high and that of the short exposure image is low. In the bright areas of the combined image, the weight of the short exposure image is high and that of the long exposure image is low.

In the combined wide dynamic range image, one image pixel has more that 12 bits of information (in the case of CX1). But for the widely used JPEG image format, information of one pixel is compressed to 8 bits. It is necessary to compress 12 bits of information to 8 bits. This is done by a process called tone mapping.

Figure 1 shows an example of tone mapping. Horizontal axis is input value of 12-bit wide dynamic image signal. Vertical axis is output value of 8-bit compressed signal for JPEG image. The optimal tone mapping curve is designed to conform to continuous gray scales without any unnatural change in color. Consequently, a natural image is recorded, conveying the impression that the image is identical to the real object.


Figure 1: Tone mapping curve
Figure 1: Tone mapping curve

The technology is embedded in the Ricoh CX1 and GR DIGITAL III digital camera. Using this camera, even when shooting a scene containing great differences in brightness, both bright and dark areas are imaged properly and a natural, smooth grayscale image is recorded.

Page Top