[Tig] When to use 444RGB or 444YUV for Dual link monitoring in a DI

Simon Burley simon at rpsfilmimaging.co.uk
Thu Aug 9 10:24:42 BST 2007


adrian hauser wrote:

> Hi Simon,
> Are you sure the RGB cube sits inside the YCbCr cube?
> I would assume it is the other way around. I thought YCbCr was a method
> for scaling rgb data and 'compressing' it into luminance and offset color
> components. As it uses (most fo the time) 16- 235 values, surely it would
> sit within the RGB cube...no?

Hi Adrian,
Every R'G'B' triplet transforms to a valid Y'CbCr value.
Only 2659110 Y'CbCr values transform to a valid R'G'B value. (~16%)
(Calculated in 8 bit precision with YCbCr black at 16 and white at 235)

Therefore the Y'CbCr gamut is larger than R'G'B' but codeword 
utilisation is poor, resulting in a lower signal to noise ratio for the 
same bit depth.

http://www.poynton.com/PDFs/Merging_RGB_and_422.pdf has a diagram on page 5.


Simon

-- 
Simon Burley
RPS Film Imaging Ltd
T: +44 (0) 208 652 5220
F: +44 (0) 208 770 7152
M: +44 (0) 7702 732 655
e: simon at rpsfilmimaging.co.uk





More information about the Tig mailing list