Wednesday, May 30, 2007
The Bit-Depth Decision
8-bit versus 16-bit workflow is among the least understood aspects of photography for most professionals. This primer will get you up to speed quickly.
Is High-Bit Worth It?
So what about the bit-depth debate, you ask? There are those who would say that there's no need to work with more than 8-bits per color since all output devices will only accept this amount of data. First, this isn't necessarily true. While you probably won't encounter too many output devices that can accept this additional data, they do exist. In fact, if you were to use the ColorByte ImagePrint RIP driving many of the Epson printers, you can send and use the full 16-bit data path.
Will you see the difference? It depends on the image. In fact, many of the naysayers suggest high-bit workflows are a waste of time because, they say, the benefit of this additional data is never seen on the final output. This depends upon the final output source, though. Examine a halftone dot from any four-color reproduction device and compare that to the dither of a really good inkjet printer. Under a loupe, you can see how the halftone dot can obscure detail that the inkjet could easily resolve. Take an image that has very fine gradations of saturated color, such as a blue sky, and you may see banding from a better output device. Much of that banding could and often is the result of editing the file in 8-bits to the degree that, much like the black and white gradient discussed, the fine tones are discarded.
That's one of the problems with an 8-bit workflow. You never know for sure where any image may be reproduced nor do you know for sure how the file may be edited. Note that when printing your high-bit images out of Photoshop, there's no reason to convert the file to 8-bit because using the Print with Preview command will do this for you on the fly.
On the other hand, say you're shooting 500 widgets on a white seamless that will be printed on a 133-linescreen press for a parts catalog. The final size of all images is 2x2. Is it prudent to work in a high-bit workflow where each image is twice the normal file size and you'll never output the documents again? Probably not, but it's your call.
What if the image you captured is so stunning that you plan to include it in your portfolio or sell it as a large fine-art print? Discarding the additional data your camera was able to record would be ill advised. You don't know where you'll reproduce that image in the future nor on what output device. If you examine the recent history of desktop printers, in just a few years we've seen an amazing increase in image quality, color gamut and fine reproduction detail (dithering). Imagine the quality of digital printers we'll have at our disposal in just five years. Does it pay to discard data that may be useful in the foreseeable future?
Bit-Depth And Gamut
Another area where you want to implement a high-bit workflow is when you use very wide-gamut working spaces. For example, many photographers are moving from Adobe RGB (1998) to wider-gamut spaces like ProPhoto RGB. This is because as capture devices and printers improve, so does their ability to capture and reproduce wider-gamut colors. The new Epson K3 inkset is capable of exceeding the Adobe RGB (1998) color gamut.
If you utilize a wider-gamut working space, you want to use more bits and here's why: The wider the gamut of a color space, the farther apart each bit is from its neighbor. Imagine you have a file in sRGB, where the 16.7 million possible colors are represented as dots painted on a balloon. As the gamut of the working space increases, consider what happens as the balloon is inflated. The space between each dot also increases. When you attempt to define 16.7 million possible colors in a small color space like sRGB, the bits are close to each other. But as you work with progressively wider-gamut color spaces, the bits are spread farther apart, just as the dots move away from each other as the balloon is inflated.
Page 3 of 4