12 bit image on an 8 bit monitor

[dVonAllmen]

So whlle I was busy freaking out over the fact that the Red's native format will have over 4000 times the color depth of, say, a native Photoshop file, it occurred to me that my monitor will not. In fact, neither do most monitors, they're 8 bit.

Heard of some 10 bit monitors, but still 10 isn't 12, now is it?

I'm sure 8 bit monitors will be able to approximate, but one of the reasons that I'm so excited about the color depth is that I've got an upcoming project on which gradients could cause banding with less color depth. I wonder if I'm going to be seeing banding that wouldn't be there if I had a way to look at it 12 bit.

I'm opening this thread to see if anyone has thoughts on this or other potential problems caused by the bit depth difference between the Red format and monitors and / or possible sollutions.

[GlennChan]

Actually Red's 12-bit might be slightly not quite perfect.

Red's 12-bit is 12-bits LINEAR (as in linear to light). Suppose you want to make a DCI master from that. DCI calls for 12-bit with a gamma of 2.6. If you convert between the two, a whack of the linear values make to the same 2.6 gamma value. So you lose a bunch of values / have insufficient values/precision to begin with.

A similar effect happens with 12-bit linear light --> 8-bit sRGB/601/709.

2- In practice, the result of insufficient bit depth is banding artifacts. If there is noise in the image, that will dither the signal and hide the banding artifacts.

I wouldn't worry about it too much, as long as you are monitoring on a decent 8-bit monitor. If it's a 8-bit LCD, you'll lose bits to calibration (or your monitor is uncalibrated, which is also a problem).... that may or may not be something to watch out for.