FW: [Tig] green in YUV

Mike Parsons mike.parsons
Mon Dec 3 09:03:52 GMT 2001


On 03/12/2001 4:33 PM, "Rob Lingelbach" <rob at emsh.calarts.edu> wrote:
> I guess I have a problem with the first line: Y isn't only the
> luminance, i.e. a monochrome signal?  Because how can I say that
> R+G+B isn't the whole video signal?
> 
> the rest of it only makes sense if I suspend my presumptions
> about Y.

Who is the target audience? This can get a bit complex!

At the end of the day Y is monochrome, it is only the luminance of the
scene. The high frequency chroma is split between U and V vectors and
superimposed on the video signal - using a low pass filter you get to see
just the luminance (mono) component of the signal. The chroma info is
isolated using a high pass filter leaving the luminance behind. Then a bit
of clever circuitry isolates the U and V signals from each other (they are
990 degrees out of phase with each other so its quite easy). The two signals
are processed a bit to get the actual B-Y and R-Y signals (multiplied by a
weighting factor and gamma corrected).
Then its all maths,
B = B-Y plus the Y signal separated earlier
R = R-Y plus the Y signal
G = the Y signal - B -R as just calculated,
If you want a monochrome signal you just take the Y signal, if you want
colour you send the RGB as calculated to each gun.

Your presumptions about Y are 100% correct, but just as light is waves and
particles so Y is the whole monochrome signal and RGB is the full colour
signal, YUV is a fudge to fit all the data into less space, remember back in
1953 monochrome compatibility was considered important!

Why U and V? The big deal is just not to send the green info, in TV land
white = 59% green 30% red and 11% blue (hence noisy blue channel). So
instead of sending RGB over the air we decided to only send R and B, at this
stage they could have stopped and sent YRB but as Y was already being sent
there is no need to send the luminance component again hence R-Y and B-Y. U
and V are essentially R-Y' and B-Y' (gamma corrected) multiplied by a
'weighting factor' which scales the volts to make the signal more
symmetrical.

-- 
Best Regards,

Mike Parsons

Visual Effects Supervision, Compositing, Editing
Telephone: 617 5529 4414  Mobile: 61 414 29 1936






More information about the Tig mailing list