On 2020-03-17 02:48:00 +0000, Jeff Urban said:
> From what I've read PAL is very similar to NTSC.
Colur encoding concept is the same.
> It has a scheme with a comb filter that makes a phase (tint) control
> unnecessary,
It inverts phase every other line so that phase errors cancel each
other out when summed.
> 50Hz frame rate rather than 60 and a higher horizontal rate resulting
> in more lines in the raster.
> (your spelling says you are across the pond)
:-)
> The burst phase is considered -I, where I is 180º out that is blue or
> actually B-Y, which is blue without the Y component which is what the
> black and white signal is called here. For red there is Q for
> quadrature which is 90º out of phase with the I and therefore the two
> do not interfere with each other. our vector algebra will tell you that
> a mixed 90º signal with the original phase does not affect the
> amplitude. Thus you get two discrete signals from one carrier. In the
> US it is 3.579545MHz which inherently causes interlace, I imagine the
> PAL carrier, which I think is 4.43MHz accomplishes the same in PAL.
Yes, the PM variant of QAM modulation is used in both. The differences
are in frequencies and PAL error cancellation.
> A vectorscope would give you the data. It is a bit easier to see when
> you have a phase control. You DO but it is not user adjustable because
> of the system making undesirable phase shifts cancel each other out.
> These color levels and phases are selected due to several factors
> having to do with visual perception. They also match the contour - for
> lack of a better word - to the makeup of the Y signal. It is not 33%,
> 33%, 33%, it is actually mostly green. Thus the -Y for the green is the
> least precedent in comparison.
> The I and Q do not exactly match red and blue for whatever reason. They
> are a hair off, so many manufacturers went with a different
> demodulation angle than 90º, usually wider like 105º or whatever.
> In the US the I signal is transmitted with a wider bandwidth and there
> are rare receivers that can use that. It requires an extra delay line.
> Also with that the only choice is true I and Q so whatever matrixing
> that has to be done falls on the engineer. Very very few units had
> that. I do not know if anything similar to that exists in the PAL
> system.
> The bottom line is that I is blue and Q is red. Green was derived from
> those usually here, not demodulated itself.
> I have worked (in the states) on some sets that had PAL capability but
> it was mostly all in chips so further detail would have to come from a
> manufacturer training courses or the standards committee of the
> government.
>
> So you may gloat that your analog color TV system was better than ours,
It was better in terms of colour fidelity and resolution. It was worse
in terms of flickering, due to lower refresh rate and some el cheapo
decoders being employed too often.
> however it is based on the same principles. Also your PAL, phase
> alternate by line system's one advantage depends on the comb filter,
> which I BELIEVE was invented or at least made practical by Sony for the
> color under home video recorder format.
> But we had it first. And from what I heard the Russian system was the worst.
> Now the French developed system called SECAM, I really have little to
> no idea how that works.
Instead of a variant of QAM modulation to encode U and V (I/Q) it sends
the two in sequence, one per each line.
> However the US was first and the government insisted on compatibility
> with the kagillions of black and white sets out there in use. I think
> PAL also has that capability but I doubt it with SECAM.
All three have. The requirement was the same everywhere.
[...]
> Enough. This is probably more information than you wanted. (I am an
> expert on the NTSC system) But you can prove it all with the
> vectorscope. All you need to do is to find a way to vary the phase of
> the subcarrier a bit to see what happens. It will come clear.
Yeah, thanks - the thing is though: I can measure this with somewhat
reasonable accuracy by sending pure R/G/B on to vectorscope and this
will tell me how _does_ the generator encode the signal. What I am
looking for though is the official/norm/standard specification telling
how it _should_ encode it.
If chroma phase in reference to burst phase is the hue, what is the
colour of 0 degrees, what angle should pure red, green, blue lie at in
reference to the burst. You say "it's a hair off". Why so? Etc.
Something like a real standard specification.
--
SD!