[opendtv] Re: Europe now debates i vs p
- From: jeroen.stessen@xxxxxxxxxxx
- To: opendtv@xxxxxxxxxxxxx
- Date: Mon, 13 Sep 2004 08:58:00 +0200
Hello,
Mark Schubin wrote:
>> Not even necessary. 50 Hz incandescent lights don't flicker,
>> only discharge lights do.
and:
>> If you watch a 10k incandescent bulb after the power to it is cut, t=
he
>> glow duration can be measured in seconds. When a tungsten filament =
is
>> heated to the point of incandescence, it will not extinguish in 1/20=
0 of
>> a second (the duration from a cycle peak to its zero point).
Doug McDonald:
> Sure they do (flicker).
I'd say that Doug is right. I remember long time ago a hobby
project (in a magazine) where a lamp was used to carry speech
over a distance, by modulating the filament current. While it
may take a long time for a filament to cool off completely,
initially the cooling goes very fast. It has to do with the
radiated power being proportional to the FOURTH power of the
temperature. Thin white hot metal will cool very quickly !
tjharvey@xxxxxxx wrote:
>> The baseband luma sampling is 74.25MHz and so anything above
>> 30MHz must be aliasing and bad filtering.
Doug McDonald:
> It's not aliasing ... aliasing only starts at 37 MHZ, and at the
> next "wrap around" the 18 MHZ analog cutoff would get rid of it.
> There was stuff in the first alias range above 37 MHZ, by the way.
> The stuff in the 30-33 MHz range was real. It may have been
> garbage, but it was not alias and not noise. In one case there
> was stuff at 30MHz that was clearly the 3rd harmonic of
> stuff at 10 MHz ... the stripes were visible on the screen.
I'd say that Doug is almost right, at the end. It IS aliasing.
Even if the transmission signal is bandwidth limited to 18 MHz,
there will still be higher harmonics that are generated by the
gamma correction function. The latter is needed because the signal
is transmitted "in the gamma domain", whereas the light is
generated "in the linear-light domain". Here lies the cause for
aliasing.
With a CRT, the signal is first made (horizontally) time-continuous,
with proper (electronic) anti-imaging filtering. Then there is the
non-linear conversion due to the inherent properties of a vacuum
tube cathode with voltage drive. There will be no aliasing (because
there is no sampled signal anymore), except maybe some shadow mask
Moir=E9 (because the shadow mask is a sampler). But that depends on
the spot size (which should be an effective anti-aliasing filter).
With a matrix display, the signal is usually gamma corrected while
still in the time-discrete (sampled) domain. This WILL cause
horizontal aliasing, when the higher harmonics fall above the
Nyquist limit. If the signal was free of (horizontal) aliasing in
the gamma domain then there will be aliasing in the linear-light
domain. This could be the case if the signal was designed for a
CRT. The other way around: if the signal was originally free of
aliasing in the linear-light domain and then converted to the gamma
domain, then there is aliasing in the gamma domain. Subsequent
inverse gamma correction will remove that aliasing, and you have a
perfect signal again. IMO, this applies to the vertical sampling,
even for old analog CRT applications. If everything is done right
then there should not be vertical aliasing due to the gamma.
Interesting i-versus-p debate, by the way. I had already seen it
coming because of the latest paper in the EBU technical review
(downloadable from: http://www.ebu.ch/trev_300-wood.pdf )
And take a look at: http://www.ebu.ch/trev_home.html for more.
Best regards,
-- Jeroen.
|-----------------------------+---------------------------------------|=
| From: Jeroen H. Stessen | E-mail: Jeroen.Stessen@xxxxxxxxxxx |=
|-----------------------------+---------------------------------------|=
| Building: SFJ-5.22 Eindhoven| Philips Digital Systems Laboratories |=
|-----------------------------+---------------------------------------|=
| Phone: ++31.40.27.32739 | Visiting & mail address: Glaslaan 2 |=
|-----------------------------+---------------------------------------|=
| Fax: ++31.40.27.32572 | NL 5616 LW Eindhoven, the Netherlands |=
|-----------------------------+---------------------------------------|=
| Pager: ++31.6.6513.3818 | Visit us: http://www.pdsl.philips.com/|=
|-----------------------------+---------------------------------------|=
=
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts:
