The Science of Digitizing Art

The two main processes for converting your 2D art into a digital file are photographing and scanning. In reality, they’re just two sides of the same coin. A scanner is simply a large, slow, bulky camera. In the old days (5 years ago,) dedicated, special purpose scanners costing $10,000 or more were the norm. They were old and outdated even then but shops kept using them because they had so much invested. The best known of these systems was Better Light. Their claim to fame was a scanning back (digital camera) that didn’t require or use a Bayer Filter meaning each pixel represented an uninterpolated color. The trade-off was that each scan took much longer to complete and it had to be attached to a computer to check framing and focus. Bottom line, it was a tedious, expensive and labor intensive process to digitize each piece of art.

Today, digital camera technology has advanced by leaps and bounds. Yet, many cameras still use Bayer filters, producing interpolated (less accurate) colors and AA (anti-aliasing) filters that induce slight blurriness to suppress moire.

pixel shift

The Pentax K-1 used in our studio incorporates Pixel Shift Technology to improve the color fidelity and sharpness of images. First, it has no AA filter so there’s no induced softness. If an AA filter is needed, it can be switched on electronically. This might be desirable for art on canvas with a distinct weave pattern and a light paint load that results in moire.

Secondly, for improved color fidelity, Pixel Shift Technology shifts the sensor below the Bayer Filter so that each photosite (light sensor) captures four pure color samples per image. The illustration above shows the sensor (gray base ) covered by a Bayer Filter. In a conventional camera, each photosite captures the light passed through one filter lens, Red, Green or Blue. These samples are then interpolated (fancy term for “mixed”) with surrounding photosites to create a single pixel.

Unfortunately, in electronics, noise is a fact of life. Think of noise as bits of contaminants on your palette. This noise alters the final color and creates problems in recreating light, dark, shadows, highlights and colors. To defeat this problem, Pixel Shift Technology takes a red image (1 above) then shifts the sensor down (2 above) to capture a green image with a green lens over each photosite. Then it shifts the sensor to the right (3 above) for a blue image and finally up (4 above) to capture a second green. In the end, it has four frames of  pure red, green, blue and green color data. There are two green frames because human vision is optimized for green, probably to distinguish predators that might be lurking in green fields or forests. These four frames are then combined to create a much more pure color rendition of the original color.

As you might image, moving a sensor by the width of one photosite (~4.88 μm) is an incredible feat of engineering precision. To do so 4 times in less than a second is beyond incredible.

But you’re asking WIIFM (what’s in it for me.) Bottom line, it’s less expensive to convert your art into a digital file because the colors are cleaner and purer from the get-go which means less photo editing time. The final image is also sharper and has better shadow and light details which give it depth. It also means less time to set up the capture because the camera is small and easy to position, manage and focus. It’s a win-win-win for your budget, your clients’ pocketbook and my sanity.