Posted by Ken Wyatt on January 12, 2010 at 11:03 PM
I know several of us do an occasional color >> B&W conversion. In Tim Grey's daily "Ask Tim Grey" email column today, he explains the importance of doing the processing in 16-bit mode, rather than 8-bit. I'm assuming he's talking about RAW processing here. Since I don't know how to link you to his email, I'll reproduce part of his explanation where he starts taking about bit-depth and tonal values:
From is column 1/12/10: "Things change dramatically when we're talking about black and white images. Even if you keep the image in the RGB color mode, you're only actually producing the equivalent of a single channel. What that means is that for a black and white image in 8-bit per channel mode, you only have 256 tonal values available. Compared to over 16.7 million possible values for an 8-bit color image, you can appreciate that an 8-bit black and white image simply doesn't contain very much information. By contrast, if you are working in 16-bit per channel mode for a black and white image, you have 65,536 possible tonal values to work with."
Wow! Something to consider...!
He says he'll be covering much more on B&W conversions in his next Digital Darkroom Quarterly (Winter 2010)
Posted by Todd Caudle on January 13, 2010 at 1:18 AM
Yes, 16-bit is very, very beneficial, both in color and b&w. It's sort of like going from absolute black to absolute white in 10 steps (8-bit analogy) or 1,000 steps (16-bit). the latter would produce much smoother tonal transitions. These numbers aren't exact numerical translations, but merely a demonstration of the concept.
Also, I suspect what he's suggesting is, when you convert a raw file, you have the option of converting to a jpeg (8-bit), or a tiff in either 8-bit or 16-bit. If you convert to 8-bit, you're starting with an image that has far less information to work with. Always convert to 16-bit.
Posted by Stephen Weaver on January 13, 2010 at 10:22 AM
I always work in 16 bit
Posted by Ken Wyatt on January 13, 2010 at 10:37 AM
I work in 16 bit for color as well, but It seems obvious now that 16 bit for B&W is even more important. < light bulb flashes on at this point! >
Discussion
3 repliesFrom is column 1/12/10: "Things change dramatically when we're talking about black and white images. Even if you keep the image in the RGB color mode, you're only actually producing the equivalent of a single channel. What that means is that for a black and white image in 8-bit per channel mode, you only have 256 tonal values available. Compared to over 16.7 million possible values for an 8-bit color image, you can appreciate that an 8-bit black and white image simply doesn't contain very much information. By contrast, if you are working in 16-bit per channel mode for a black and white image, you have 65,536 possible tonal values to work with."
Wow! Something to consider...!
He says he'll be covering much more on B&W conversions in his next Digital Darkroom Quarterly (Winter 2010)
Here's the link to his store: http://www.timgrey.com/store/index.php?app=ecom&ns=prodshow&ref=DigitalDarkroomQuarterly
Also, I suspect what he's suggesting is, when you convert a raw file, you have the option of converting to a jpeg (8-bit), or a tiff in either 8-bit or 16-bit. If you convert to 8-bit, you're starting with an image that has far less information to work with. Always convert to 16-bit.