12 bit NEF's?

[ashleysmithd]

Hi.

I'm trying to work out if Nikon mean 12-bit compression, or 12 bit as in 4096 levels per RGB channel per pixel.

If they mean compression, then what is the bit depth of the Nikon D40?

And what do they actually mean behind the '12-bit compression'?

It's just that of course 3x12bit would be 36bit which I'm not sure is even fully useable in photoshop, last time I checked photoshop went up to 32 bit editing.

Thanks for any info.
Dan.

 

[Garbz]

Firstly 12-bit compression means the NEF RAW is compressed using a lossy algorithm (I haven't noticed the loss so I actually use compression all the time on my D200), but the compression does not reduce the bit depth of the image.

NEFs for All cameras up to the D200 are 12bits / channel so you are right, each colour has 4096 possible values between black and fully saturated. This is opposed to 8bits / channel which you normally get out as a processed JPEG giving you 256 possible values.

As for photoshop, it thinks in bits / channel too. It has no native support for a 12bit image but when you import a NEF using CameraRAW you have the option of spitting out the 8bit image, or upsampling it to 16bits / channel.

Photoshops 32bit editing is for HDR images and assumes 96bits per pixel that's 79*10^27 possible values and is in the realm of purely computational. No display currently has that dynamic range. And if you want to get really technical you don't need more than 8bits / channel for a final image. With 8bits/channel you get more than 16million individual colour values, more than can be displayed, printed, and resolved by the human eye.

What is important is for editing. If you have 12bits / channel to work with when you make a change like the colour temperature there is no guess work. You can accurately change the colour temperature without any loss in image quality which is one of the main benefits of working in RAW.