Camera Bit Depth

[HurtinMinorKey]

So higher bit depth allows you to create contrast(in post) in a low contrast scene right? If my A/D conversion does not have enough bit depth, then i won't be able to detect subtle differences in exposure, right? Higher bit depth means a more precise A/D conversion.

 

Do i understand this correctly? Or am I completely full of sh**?

[maxotics]

Not sure you question, but in general bit-depth is a computer (bits) + photo (depbth) lingo way of saying how wide the range in shading of any given color/brightness sampled be the electronic equipment.  

 

You can store 256 shades (or depth values) in an 8-bit memory slot (or byte). 

 

When working with Canon RAW, the camera saves each sensel/pixel as a 14-bit value.  That would give you a range from 1 to 16,384.

 

Once you save those values in smaller bit values you have to reduce the precision.

 

I tried to explain the "precision" problem on the ML forum like so:

 

One can think of it this way.  You have a palette of 14 scales of gray.  You need to convert them into 8 scales for something else.  So 1-2 because 1, 3-4, become 2, 5-6 becomes 3, 7-8, becomes 4; 9-10, 5, 11-12, 6, 13-14, 7 (and we throw out the 8, for example)

Let's say you have two gray colors in what you shot, and they are 2 and 6.  You want to reduce the exposure by 1 (increase contrast).   

They were convered to 1 and 3, so now they become 0 and 2.   You went from a 300% difference (in 14bit) to 200% (in 8bit).  More importantly, you went from some gray 1, in 14bit, to no gray 0, in 8bit.

 

What people don't understand, especially about H.264, is you can't just take 8 bit values and put them in a 14-bit space, for example, and get that 14bit precision again.  

 

Does this make sense?  You seem to know your stuff, so maybe you're asking something else?