For photographers of a certain age, the decision of whether to work in 8-bit or 16-bit is as easy as the decision between playing Nintendo and Super Nintendo in 1991. More colors led to more realistic graphics and better gameplay. The 16-bit Super Nintendo was king. The same principle holds true when working on your images in Photoshop.
You might think the difference is obvious. After all, 16 is twice as much as 8, so it should be twice as good, right? Well, as Unmesh Dinda of PiXimperfect points out, that's not exactly correct. Using 16-bit will, obviously, allow Photoshop to render more color variations than 8-bit. However, it's not twice as many, as the numbers might imply. Instead, you'll be working with exponentially more color variations.
The result, as Dinda demonstrates, is that when you are working on an image with multiple curve layers and many vibrant colors, 8-bit processing may cause color banding and artifacts in your image. Using 16-bit color, on the other hand, will render the colors more smoothly and give the image a much cleaner appearance. Dinda notes that there may be some instances where working in 8-bit is necessary, but by and large, 16-bit is preferable.
Once you're done working in 16-bit, you can later reduce the image to 8-bit for use on social media, other online platforms, printing, or other media that doesn't support 16-bit.
Do you usually work in 8-bit or 16-bit? Why do you prefer one setting over the other? Drop a comment below, and let us know.
I use 8 bit, some important actions are not available in 16 bit.
I wonder when 32-bit will be the norm.
When projects arise that need 16777216 times the information per pixel and colour channel (compared to 8bit)
My understanding (which could be wrong) is that the internal calculations accumulate less rounding errors even though you may eventually go back to 8 bit for display
That's true but wouldn't you prefer a solution that automatically uses more precision internally but still outputs 8bit? Could be just one slider in your global Photoshop settings and you'd never have to worry again.
but it's a different if I have 256 (8bit) or 65.536 (16) information...
16bit would be very beneficial.
The main advantage of 16bit and up is that it's enough information to make it a linear colour space. In 8bit, the brightness gets doubled each time you increase the value of a pixel by 1. We need to do this because our eyes are very good at making out small differences in shadow detail. 16bit provides little extra information in the darkest areas of an image but gives a huge boost to midtone and highlight details.
Agreed.
But your display have to work on a 16-bits basis.
Actually, most of the screen on the market displays 10-bits color depth.
16 bit is there to avoid cumulative mathematical rounding errors and the more work you do, the worse it gets. Importing raw files in 16 bit mode also gets the most out of the files. I always use 16 bit files for photoshop work but the files get chunky and my IQ150 files start at 900mb PSD files
Yea those 1GB files per PSD were reason for me to get back to 8-bit version. It's nice to know about this problem, and when you see posterization to know how to fix it, but in quite lot of cases it's an overkill for simple retouching tasks.