[Albert Manfredi[_2_]]

On Feb 21, 12:48*pm, Doc wrote:

Actually, it also has composite jacks. I assume it's oriented toward
those like myself who have analog tv's and no cable. Mine has
composite jacks but some older sets don't. *I would imagine there are
few people with a 16:9 High-def tv running off rabbit ears. I would
assume they mostly have cable or whatever. Don't most of the newer
16:9's have tuners built in?

Yes, all new TVs have to have a built-in digital receiver, in the US
that is, if they have any built-in receiver at all. Meaning that plain
monitors are still permitted, but not TVs with only an NTSC (i.e.
analog) tuner. Just as was the case years ago, though, plain old
monitors don't seem to sell well. The receiver part just isn't that
much more of an expense.

Something that's not clear to me, some of the stations specify "HD"
but can you get a high-def image over the air or do you have to have
cable, satellite etc? I realize it's not applicable to an analog tv in
any case.

Absolutely. Just about all prime time shows on OTA stations are
transmitting glorious HD these days. Have been for years, actually.
And PBS stations seem to transmit HD 24/7. Lots of sports are HD too.

However, I'm struck by how much better the image is even on this cheap
tv than I ever saw with even the strongest analog station in the area.

True, if you are downconverting the digital signal to feed a standard
CRT TV, you'll get the equivalent of a really clear, ghost-free, good
color signal. Not much different from a pristine NTSC station. It's
just that this will be the rule for all your local and even not-so-
local stations, as opposed to just the one or two "best" analog
stations in your market.

And the box will also give you the multicast channels, so you get more
choice.

Bert

[Richard C.]

"Doc" wrote in message
...
:

I would imagine there are
few people with a 16:9 High-def tv running off rabbit ears. I would
assume they mostly have cable or whatever. Don't most of the newer
16:9's have tuners built in?

==================================
For several years, the sets did not have tuners built in.
Only in the past year did smaller sets have tuners.

OTA is the best way to receive DTV/HDTV - and it is FREE!

I dumped cable and sat long ago.

[Richard C.]

"Doc" wrote in message
...

However, I still don't like how HD TV looks when viewed on actual HD
TV's and computer monitors, all kinds of artifacts that I find I don't
see on an analog set - at the top of the list is the "swarm of bees"
effect whenever there's motion.

============================
I have NONE of that on my HDTV!
You must be referring to LCD sets, which have that problem.

My Pioneer 710 64" RPTV still has a great picture.
================================

Everyone says "you haven't seen a good
one set up correctly" etc. I have yet to see one, either in someone's
house or set up in special display areas to showcase them in the
stores, whether the source is Blu-Ray, HD disc, or off whatever feed
they have that I haven't had a problem with. Is it possible there
isn't ONE competent setup person in ANY of the stores around here?

==========================
100% possible!
===========================

I'll find a big analog TV so I can get a decent-sized letterbox image
until it dies and I can't find another one and have no other choice.

[dmaster]

On Feb 21, 5:24*pm, Doc wrote:
On Feb 21, 3:39*pm, dmaster wrote:

Something that's not clear to me, some of the stations specify "HD"
but can you get a high-def image over the air or do you have to have
cable, satellite etc? I realize it's not applicable to an analog tv in
any case.

Did I miss something? *Are you asking if HD broadcasts are available
over-the-air? *

Nope, you've grasped what I was asking. It wasn't clear to me if HD is
available over the air or if you had to pay for it. Your answer
clarified this.

However, I still don't like how HD TV looks when viewed on actual HD
TV's and computer monitors, all kinds of artifacts that I find I don't
see on an analog set - at the top of the list is the "swarm of bees"
effect whenever there's motion. Everyone says "you haven't seen a good
one set up correctly" etc. I have yet to see one, either in someone's
house or set up in special display areas to showcase them in the
stores, whether the source is Blu-Ray, HD disc, or off whatever feed
they have that I haven't had a problem with. Is it possible there
isn't ONE competent setup person in ANY of the stores around here?

I wouldn't count on finding any well setup HDTVs in any Big Box
store. Nor would I count on finding a well distributed signal. I'm
not saying it doesn't happen, I just wouldn't count on it. Even then,
the low priced LCD HDTVs, despite their array of wonderful specs, are
the most likely to process signals poorly, especially fast motion
sequences. Couple that with large screen sizes and the likelihood
that you are examining the pictures from much closer than you normally
would watch, and it doesn't surprise me that you'd find artifacts in
most every store you visited.


I'll find a big analog TV so I can get a decent-sized letterbox image
until it dies and I can't find another one and have no other choice.

Sure, that's your choice, so go ahead and enjoy. Personally, I have
no problems with artifacts when watching OTA HDTV or SD-DVDs on my
EDTV Plasma (480p) or my 720p bedroom LCD. Digital SD can show
artifacts on either set, particularly the 720p LCD. I attribute that
to 1. poor processing in the LCD 2. low bit rate encoding of the
original analog SD source into digital SD and 3. the magnification of
the low resolution picture to fit the HD LCD. Since I also feed the
same signals through an ATSC equipped DVD-R to a 32 inch analog CRT
TV, I can do direct comparisons. While the low resolution of the CRT
hides a multitude of sins, I would always choose to watch the other
sets first. Although we refer to analog CRTs as "480i", there is such
an immense impovement going from the CRT to the 480p Plasma, that I
have to believe the CRT's effective resolution is even lower. Having
gotten used to the higher resolution of the Plasma and the LCD, I've
really begun to notice the really low screen resolution of the CRT;
the individual lines are so apparent from close to the set.

Now consider letterboxed wide screen material. You are probably
getting no more than 320i on the CRT. And that assumes the CRT can
really resolve 480i, which in my case looks pretty doubtful. The
lowly ED Plasma gives 480p for 16:9 material, and something like 400p
for 2.35:1 material. No contest what-so-ever. The Plasma totally
smokes the CRT every time. Again, since I was used to poor quality
analog OTA, I used to think ordinary DVDs looked pretty good on the
32" CRT. Once I got the Plasma, I realized just how much of the
detail from an *ordinary* DVD was lost by displaying it on the CRT.
Or consider showing photos on the CRT. It was hardly worth the
effort. The ED Plasma is worlds better, and a true HD set is better
yet.

The only "advantage" I can truly give my CRT is the ability to hide
the poor quality of some source material. But that occasional slight
benefit is completely outweighed by the higher resolution of the
Plasma or LCD in the vast majority of the cases. Don't get me wrong,
OTA digital will make your CRT look better than it ever has, but if
you view the same signal on a decent HDTV, you will never go back.

Dan (Woj...)

[~consul[_2_]]

and thus Albert Manfredi inscribed ...
True, if you are downconverting the digital signal to feed a standard
CRT TV, you'll get the equivalent of a really clear, ghost-free, good
color signal. Not much different from a pristine NTSC station. It's
just that this will be the rule for all your local and even not-so-
local stations, as opposed to just the one or two "best" analog
stations in your market.

That is actually a benefit I hadn't thought about. Right now, there are some analogue channels that are fuzzy, and I still have my VCRs hooked up to tape some of them, and it will make it a lot clearer to record those signals. That was a huge reason why I went to OTA HDTV and recording to my desktop, because what I taped on the VCR was fuzzy at times. Now it won't.
--
"... respect, all good works are not done by only good folk. For here, at the end of all things, we shall do what needs to be done."
--till next time, consul -x- poetry.dolphins-cove.com

[Sal M. Onella]

"Doc" wrote in message
...

snip

However, I still don't like how HD TV looks when viewed on actual HD
TV's and computer monitors, all kinds of artifacts that I find I don't
see on an analog set - at the top of the list is the "swarm of bees"
effect whenever there's motion.

===============================================

Could this be from sitting too close? I have heard (but not tested) the
rule that 3 times the screen diagonal is the optimum viewing distance for
HDTV. Mightn't the artifacts disappear at that distance?

No doubt the artifacts are real and essentially unavoidable. Worst I ever
saw was some tape made at a big swimming pool. Those random ripples do not
compress!

[William Sommerwerck]

"Sal M. Onella" wrote
in message news
However, I still don't like how HD TV looks when viewed on actual
HD TVs and computer monitors, all kinds of artifacts that I find
I don't see on an analog set -- at the top of the list is the "swarm
of bees" effect whenever there's motion.

I've seen HD many times (don't yet own it), but have never seen such an
effect. (I have seen "mosquito noise" on DVDs.) There's obvious smearing on
some LCD sets during fast motion, but that doesn't seem to be what you're
referring to.

[jwvm]

On Feb 21, 7:42 am, "Arny Krueger" wrote:

snip

Interestingly enough this rule breaks down for HDTV QAM channels on cable.
My local Comcast system has 333 digital channels, but as promised, only
about a dozen of them are watchable without a magic decoder ring. At this
point the magic decoder ring for PCs requires total re-engineering of the
whole PC from the BIOS on up.

A number of Comcast's 333 digital channels have sound, but put up a nifty
blue graphic of a FM radio, even though the sound is clearly from one of the
cable channels. Comcast's technical staff are about as much on the ball as
ever, I guess. ;-)

Of course some cable companies manage to hose up the QAM channels so
even for those you need a magic box. The signal is there per the FCC
requirement but it fades in and out just like OTA. :-(

[jwvm]

On Feb 21, 10:19 am, Don Stauffer in Minnesota
wrote:
On Feb 20, 11:25 am, Doc wrote:

I just picked up an Insignia NS-DXA1 digital converter box at Best
Buy. The picture comes in clear but on some channels, I've noticed
what I'd call an aliasing artifact in the sound - overly shrill and
harsh on the highs. It seems to be more prevalent on certain
channels.

Is this an issue with this particular model/brand or a common problem?
Anything that can be done about it?

Thanks

I have similar problems with my new HDTV. One of my local channels
has that problem, but only for live broadcasts, mainly of news
programs. All other channels are fine. I have not yet pinged on the
station, but I am assuming that they are having some problems with
their audio setup. Their analog station is fine, it is only on their
digital channel that I notice this.

Also, I notice this on the national news broadcast on that channel.
So I assume the problem is at the transmitter, not at their local news
studio.

It may be a compression artifact. I have noticed a similar effect with
especially severe distortion on sibilants when a track has been
compressed and decompressed multiple times. If you hear the problem
with remote broadcasts (live at the scene of a crash!) this could be
the cause.

[Albert Manfredi[_2_]]

On Feb 21, 3:44*pm, wrote:

S-video has exactly the same bandwidth as NTSC. The only difference
from composite is the modulated subcarrier is not added onto the Y
signal and instead is routed out separately.

I might have said that S-video has the same luminance bandwidth as
COMPOSITE video, but not the same as OTA NTSC.

Take the case of a DVD player connected to an analog TV via composite
or S-video cable.

The reason DVDs look better than standard NTSC OTA shows in this case
is that the DVD player can actually squeeze through at least 5.5 MHz
of video bandwidth, over composite or S-video. Compared with only 4.2
MHz of luminance bandwidth via the NTSC tuner of the TV. This is
because there is no audio carrier to stay away from, in the composite
or S-video interfaces, as there is in the analog 6 MHz OTA channel. So
no need to roll off luminance as fast. Theoretically, since NTSC video
is scanned at 483 * 30 visible horizontal lines/sec, or 14.5 MHz, a
baseband video interface should be able to approach half that figure.

I think you'll find that if you use a composite or S-video cable, the
HDTV image on the older set should look as good as a DVD player's
signal, when played over either of these baseband interfaces. Some SD
programs, on the other hand, will be very grainy, similar to analog
TV. Especially those that the broadcaster upconverted from an analog
feed.

I won't get into a debate of composite vs S-video. I think there are
many variables there, but S-video is probably at least marginally
better, in some cases. Reason being, you don't have to depend on the
color difference signal being perfectly 90 degrees out of phase with
the luminance.

If instead you're using the RF cable ouput from box to TV, then the
image should be as good as the very best OTA NTSC stations. Which is
still a step up, IMO.

Bert

[Albert Manfredi[_2_]]

On Feb 26, 5:56*pm, Albert Manfredi wrote:
Theoretically, since NTSC video
is scanned at 483 * 30 visible horizontal lines/sec, or *14.5 MHz, a
baseband video interface should be able to approach half that figure.

Ooops. forgot part of what's involved here.

DVDs have 720 pixels in the horizontal. Which means that a line of
video luminance consists of max 360 cycles (alternating light and dark
pixels).

So, including the blanking interval, that makes the bandwidth:

525 lines/frame * 360 cycles/line * 30 frames/sec = 5.67 MHz of video
luminace bandwidth maximum. Too much for the NTSC RF to handle, given
the audio carrier at 4.5 MHz, but achievable with baseband interfaces.

Bert

[Bob Howes]

"jwvm" wrote in message
...

It may be a compression artifact. I have noticed a similar effect with
especially severe distortion on sibilants when a track has been
compressed and decompressed multiple times. If you hear the problem
with remote broadcasts (live at the scene of a crash!) this could be
the cause.

It almost certainly IS a compression artifact. If you're ever lucky enough
to see raw, uncompressed HDTV pictures, they're simply stunning even up
close to the display. The problems that have been described come from
trying to squeeze too many channels into too little bandwidth.

Bobbsy

[Alan[_2_]]

In article Albert Manfredi writes:

Take the case of a DVD player connected to an analog TV via composite
or S-video cable.

The reason DVDs look better than standard NTSC OTA shows in this case
is that the DVD player can actually squeeze through at least 5.5 MHz
of video bandwidth, over composite or S-video. Compared with only 4.2
MHz of luminance bandwidth via the NTSC tuner of the TV. This is
because there is no audio carrier to stay away from, in the composite
or S-video interfaces, as there is in the analog 6 MHz OTA channel. So
no need to roll off luminance as fast.

True, also the signal/noise is generally much better.

Theoretically, since NTSC video
is scanned at 483 * 30 visible horizontal lines/sec, or 14.5 MHz, a
baseband video interface should be able to approach half that figure.

Huh? This is dimensionally incorrect, and numerically suspect.

483 lines/frame * 30 frames/second = 14490 lines/second

Nowhere does this come out to units of Hertz. Similarly, it doesn't come
anywhere near 14.5 * 10^6 for the mega part.


I think you'll find that if you use a composite or S-video cable, the
HDTV image on the older set should look as good as a DVD player's
signal, when played over either of these baseband interfaces. Some SD
programs, on the other hand, will be very grainy, similar to analog
TV. Especially those that the broadcaster upconverted from an analog
feed.

I won't get into a debate of composite vs S-video. I think there are
many variables there, but S-video is probably at least marginally
better, in some cases. Reason being, you don't have to depend on the
color difference signal being perfectly 90 degrees out of phase with
the luminance.

This detail is not possible. The color signal is two parts, and
they are at 90 degrees phase differenc with each other. They cannot
be at 90 degrees difference with the luminance, even if the luminance
had a phase reference, which it does not.

Alan

[Arny Krueger]

"Albert Manfredi" wrote in message

On Feb 21, 3:44 pm, wrote:

S-video has exactly the same bandwidth as NTSC. The only
difference from composite is the modulated subcarrier is
not added onto the Y signal and instead is routed out
separately.

I might have said that S-video has the same luminance
bandwidth as COMPOSITE video, but not the same as OTA
NTSC.

I'm glad you corrected yourself. I had written a correction but it got lost.

The bandwidth of the chrominance component of OTA NTSC is far less than that
of the luminance - about half.

The bandwidth of the luminance component of OTA NTSC is about 4 MHz.

The bandwidth of either component of a S-Video connection can be 4 MHz.

The bandwidth limitations of broadcast NTSC as compared to the limitations
of in-studio NTSC are clearly visible when you can view both the studio
video and the video that is sent over the air side-by-side with a good live
source.

Take the case of a DVD player connected to an analog TV
via composite or S-video cable.

Good example.

The reason DVDs look better than standard NTSC OTA shows
in this case is that the DVD player can actually squeeze
through at least 5.5 MHz of video bandwidth, over
composite or S-video. Compared with only 4.2 MHz of
luminance bandwidth via the NTSC tuner of the TV.

The video on a DVD is not frequency division multiplexed in analog domain,
like NTSC video is. It is time division multiplexed in the digital domain.

This is
because there is no audio carrier to stay away from, in
the composite or S-video interfaces,

There is no audio in either composite or S-video interfaces, either in the
home or in the studio. The problem with a studio or home compsite signal is
that the chroma has to be separated out in the analog domain. Since this
process uses frequency division multiplexing, there are naturally bandwidth
limitations. For example, in a composite connection, the chroma signal
frequency components are interleaved with the luninance signal frequency
components in the same band.

as there is in the analog 6 MHz OTA channel. So no need to roll off
luminance as fast. Theoretically, since NTSC video is
scanned at 483 * 30 visible horizontal lines/sec, or
14.5 MHz, a baseband video interface should be able to
approach half that figure.

This bandwidth is pretty much inherent in the S-video connection, but has to
be threaded through a maze of frequency divisions and demodulators for the
composite signal.

I think you'll find that if you use a composite or
S-video cable, the HDTV image on the older set should
look as good as a DVD player's signal, when played over
either of these baseband interfaces.

That is at least a possibility. HDTV signals have greater bandwidth than
DVD signals, and by quite a bit. It takes about 2 hours of full
DVD-bandwidth signal to create a ca. 4 GB dataset, while it takes only about
30-40 minutes of a full-channel HDTV signal to create a ca. 4 GB dataset.

Some SD programs, on
the other hand, will be very grainy, similar to analog
TV. Especially those that the broadcaster upconverted
from an analog feed.

Agreed. This can be observed with present-day OTA DTV signals. The big thing
about MPEG-coded video is that there are no hard numbers - bandwidth and
visual quality can be compromised every which way including loose.

I won't get into a debate of composite vs S-video. I
think there are many variables there, but S-video is
probably at least marginally better, in some cases.

One of the problem with S-video is that we observing it as it comes from a
number of sources. Originally, it was intended to come from S-Video tapes,
which were not all that wonderful. In many cases it comes from an OTA
source. The S-video output of a DVD player could be a thing to see, just
like a NTSC signal from a DVD player could be quite a thing to see. However,
a NTSC signal from a DVD player is going to be decoded in a TV set, and I
know of no TV sets that have a separate chroma decoder for the back-panel
S-video input. It's easier and more economical to just put S-Video and
composite video jacks there.

[jwvm]

On Feb 27, 3:33 am, (Alan) wrote:
In article Albert Manfredi writes:

Take the case of a DVD player connected to an analog TV via composite
or S-video cable.

The reason DVDs look better than standard NTSC OTA shows in this case
is that the DVD player can actually squeeze through at least 5.5 MHz
of video bandwidth, over composite or S-video. Compared with only 4.2
MHz of luminance bandwidth via the NTSC tuner of the TV. This is
because there is no audio carrier to stay away from, in the composite
or S-video interfaces, as there is in the analog 6 MHz OTA channel. So
no need to roll off luminance as fast.

True, also the signal/noise is generally much better.

Theoretically, since NTSC video
is scanned at 483 * 30 visible horizontal lines/sec, or 14.5 MHz, a
baseband video interface should be able to approach half that figure.

Huh? This is dimensionally incorrect, and numerically suspect.

483 lines/frame * 30 frames/second = 14490 lines/second

Nowhere does this come out to units of Hertz. Similarly, it doesn't come
anywhere near 14.5 * 10^6 for the mega part.

I think you'll find that if you use a composite or S-video cable, the
HDTV image on the older set should look as good as a DVD player's
signal, when played over either of these baseband interfaces. Some SD
programs, on the other hand, will be very grainy, similar to analog
TV. Especially those that the broadcaster upconverted from an analog
feed.

I won't get into a debate of composite vs S-video. I think there are
many variables there, but S-video is probably at least marginally
better, in some cases. Reason being, you don't have to depend on the
color difference signal being perfectly 90 degrees out of phase with
the luminance.

This detail is not possible. The color signal is two parts, and
they are at 90 degrees phase differenc with each other. They cannot
be at 90 degrees difference with the luminance, even if the luminance
had a phase reference, which it does not.

Alan

Very true. The luminance component (Y) contains the fine detail. The
chrominance components (I and Q) modulate the color carrier at 3.58
MHz and are in quadrature. The presence of the color carrier and color
signals in the luminance signal further degrades the bandwidth of the
luminance component resulting in an additional loss of spatial
resolution. S-video has separate channels for luminance and
chrominance eliminating the loss of spatial resolution. The wider
bandwidth from a DVD player, of course, also improves spatial
resolution.

[jwvm]

On Feb 27, 12:32 am, "Bob Howes"
wrote:
"jwvm" wrote in message

...



It may be a compression artifact. I have noticed a similar effect with
especially severe distortion on sibilants when a track has been
compressed and decompressed multiple times. If you hear the problem
with remote broadcasts (live at the scene of a crash!) this could be
the cause.

It almost certainly IS a compression artifact. If you're ever lucky enough
to see raw, uncompressed HDTV pictures, they're simply stunning even up
close to the display. The problems that have been described come from
trying to squeeze too many channels into too little bandwidth.

Bobbsy

The sad thing here is that the bandwidth required for audio is much
less than video. Compressing the audio excessively is not going to
provide significant additional bandwidth for the video signal.

[William Sommerwerck]

"jwvm" wrote in message
...

The luminance component (Y) contains the fine detail. The
chrominance components (I and Q) modulate the color carrier
at 3.58 MHz and are in quadrature. The presence of the color
carrier and color signals in the luminance signal further degrades
the bandwidth of the luminance component resulting in an additional
loss of spatial resolution.

Not at all true! You need to read up on your color-TV theory.

The subcarrier is at a frequency (an odd multiple of half the horizontal
scanning rate) that causes its sidebands (which are not continuous, but at
multiples of the horizontal scanning frequency) to fall _between_ the
sidebands of the luminance signal.

There is no loss of luminance information whatever -- unless the
receiverlops off the upper part of the composite signal to reduce visible
interference.

Most modern TVs use multi-frame comb filtering, and do a very good job of
extracting the fine detail.

[William Sommerwerck]

"Arny Krueger" wrote in message
. ..

One of the problem with S-video is that we observing it as it comes
from a number of sources. Originally, it was intended to come from
S-Video tapes, which were not all that wonderful.

I don't know what you mean by "S-Video tapes". VCRs -- at least, non-studio
machines -- traditionally recorded Y and C separately, long before the S
output became available (at the time JVC introduced Super VHS). It was a
trivial matter to add an S-output jack to any of these machines -- including
Beta.

Laserdisks recorded the composite video signal. Those with S outputs
required an internal comb filter to separate the C. When I got an IDTV, I
used the composite player output and let the TV do the separating.

[Arny Krueger]

"William Sommerwerck" wrote in
message
"jwvm" wrote in message
...

The luminance component (Y) contains the fine detail. The
chrominance components (I and Q) modulate the color
carrier
at 3.58 MHz and are in quadrature. The presence of the
color carrier and color signals in the luminance signal
further degrades the bandwidth of the luminance
component resulting in an additional loss of spatial
resolution.

Not at all true! You need to read up on your color-TV
theory.

The subcarrier is at a frequency (an odd multiple of half
the horizontal scanning rate) that causes its sidebands
(which are not continuous, but at multiples of the
horizontal scanning frequency) to fall _between_ the
sidebands of the luminance signal.

There is no loss of luminance information whatever --
unless the receiverlops off the upper part of the
composite signal to reduce visible interference.

Most modern TVs use multi-frame comb filtering, and do a
very good job of extracting the fine detail.

However, there remains the problem of fine detail in the chrominance
information in NTSC composite signals. There are two chroma signals, one
with 1.5 MHz bandwidth, and one with 0.5 MHz bandwidth. So you can do what
you want to in terms of extending the bandwidth of the luninance signal, but
the color is locked into these numbers. They are not bad choices for a 4 MHz
luminance signal, maybe a bit conservative. They suffer as you extend the
bandwidth of the luminance beyond 4 MHz.

If you send the signal as S-video, then AFAIK a lot more bandwidth can be
packed into the luminance information.

[Arny Krueger]

"William Sommerwerck" wrote in
message
"Arny Krueger" wrote in message
. ..

One of the problem with S-video is that we observing it
as it comes from a number of sources. Originally, it was
intended to come from S-Video tapes, which were not all
that wonderful.

I don't know what you mean by "S-Video tapes". VCRs -- at
least, non-studio machines -- traditionally recorded Y
and C separately, long before the S output became
available (at the time JVC introduced Super VHS). It was
a trivial matter to add an S-output jack to any of these
machines -- including Beta.

I meant S-VHS. I owned a S-VHS VCR and literally ran it 'till the heads were
gone which was about 4 years. I got it on a close-out for a really cheap
price, so I pretty much got the extra quality I paid for. ;-)

Laserdisks recorded the composite video signal. Those
with S outputs required an internal comb filter to
separate the C. When I got an IDTV, I used the composite
player output and let the TV do the separating.

That gets you into that post a few days back from "Trotsky". If you are
using a format that requires that the video undergo NTSC Y/C separation,
the quality difference available from using the S-video cable can be
negligible. It can even favor not using the S-Video connection if the TV Y/C
separator is hot and the one in the video player is not.

That can all change when you start working with digital media. If you start
out with a DVD-V signal, there is potentially a about double the video
information available to the player than OTA NTSC can handle. The situation
is even more extreme for HDTV, where there is potentially about three times
the video information as DVD-V.

A DTV adaptor can potentially have about 6 times as much information
available to it as the video from the tuner in the analog TV it is attached
to. You really want a component video connection, but S-video done right
could be a good kicker.

[William Sommerwerck]

"Arny Krueger" wrote in message
...

However, there remains the problem of fine detail in the
chrominance information in NTSC composite signals.

The chrominance signals don't need fine detail -- but not for the reason you
might think. See below.


There are two chroma signals, one with 1.5 MHz bandwidth, and one
with 0.5 MHz bandwidth. So you can do what you want to in terms of
extending the bandwidth of the luninance signal, but the color is locked
into these numbers. They are not bad choices for a 4 MHz luminance
signal, maybe a bit conservative. They suffer as you extend the bandwidth
of the luminance beyond 4 MHz.

Not necessarily. Again, see below.


I don't have time to discuss this in detail, but have you ever noticed that
virtually every system of storing and transmitting color images uses
color-difference signals?

These are created by subtracting the luminance information (Y) from the raw
color signals (R, G, B). This removes the fine luminance detail from them,
reducing the required bandwidth.

More to the point -- as the color-difference signals represent
_saturation_ -- not hue -- and most things, natural or manufactured, are
decorated with colors of constant saturation, color-difference signals don't
have to be wideband to _fully_ represent the color in the image. This is why
an NTSC receiver can "get away" with displaying only 0.5MHz of
color(-saturation) information.

The only time the "reduced" bandwidth is noticeable is in lettering -- such
as the CNN logo -- where 0.5MHz isn't enough to render the vertical bars.

I am one of the few people in the world who consciously understands all
this. ahem

[William Sommerwerck]

"Arny Krueger" wrote in message
...

Laserdisks recorded the composite video signal. Those
with S outputs required an internal comb filter to
separate the C. When I got an IDTV, I used the composite
player output and let the TV do the separating.

That gets you into that post a few days back from "Trotsky".
If you are using a format that requires that the video undergo
NTSC Y/C separation, the quality difference available from using
the S-video cable can be negligible. It can even favor not using the
S-Video connection if the TV Y/C separator is hot and the one in
the video player is not.

Exactly my point.


That can all change when you start working with digital media. If you
start
out with a DVD-V signal, there is potentially a about double the video
information available to the player than OTA NTSC can handle. The
situation
is even more extreme for HDTV, where there is potentially about three
times
the video information as DVD-V.

A DTV adaptor can potentially have about 6 times as much information
available to it as the video from the tuner in the analog TV it is
attached
to. You really want a component video connection, but S-video done right
could be a good kicker.

It would certainly improve the luminance part. However, the receiver will
have its own bandpass filter for the chrominance, and if it's limited to
0.5MHz... you won't see much of an improvement.

I've often wondered why VHS has such terrible chrominance bandwidth. It's SO
bad that you often see objects -- especially in home videos -- with
uncolored borders. Horrible.

[Arny Krueger]

"William Sommerwerck" wrote in
message
"Arny Krueger" wrote in message
...

However, there remains the problem of fine detail in the
chrominance information in NTSC composite signals.

The chrominance signals don't need fine detail -- but not
for the reason you might think. See below.

There are two chroma signals, one with 1.5 MHz
bandwidth, and one

with 0.5 MHz bandwidth. So you can do what you want to
in terms of extending the bandwidth of the luninance
signal, but the color is locked into these numbers. They
are not bad choices for a 4 MHz luminance signal, maybe
a bit conservative. They suffer as you extend the
bandwidth of the luminance beyond 4 MHz.

Not necessarily. Again, see below.

Role reversal, eh? ;-)

I don't have time to discuss this in detail, but have you
ever noticed that virtually every system of storing and
transmitting color images uses color-difference signals?

No.

I don't see that he

http://www.ntsc-tv.com/ntsc-index-06.htm

I can see colors being subtracted, but there seems to more than simple
subtraction of luminance going on. Furthermore, each luminance component is
low-passed.

These are created by subtracting the luminance
information (Y) from the raw color signals (R, G, B).
This removes the fine luminance detail from them,
reducing the required bandwidth.

I follow the psychovisual explanation in the reference, above:

http://www.ntsc-tv.com/ntsc-index-04.htm

http://www.ntsc-tv.com/ntsc-index-06.htm

More to the point -- as the color-difference signals
represent _saturation_ -- not hue -- and most things,
natural or manufactured, are decorated with colors of
constant saturation, color-difference signals don't have
to be wideband to _fully_ represent the color in the
image.

You still got edges to worry about, particularly when you go from saturated
color to white.

This is why an NTSC receiver can "get away" with
displaying only 0.5MHz of color(-saturation) information.

Well, 1.5 MHz and 0.5 MHz depending on color.

The only time the "reduced" bandwidth is noticeable is in
lettering -- such as the CNN logo -- where 0.5MHz isn't
enough to render the vertical bars.

Lettering ofgten looks a lot better coming from a DVD, even via a S-video
cable.

I am one of the few people in the world who consciously
understands all this. ahem

ummm... ;-)

[jwvm]

On Feb 27, 9:50 am, "William Sommerwerck"
wrote:
"jwvm" wrote in message

...

The luminance component (Y) contains the fine detail. The
chrominance components (I and Q) modulate the color carrier
at 3.58 MHz and are in quadrature. The presence of the color
carrier and color signals in the luminance signal further degrades
the bandwidth of the luminance component resulting in an additional
loss of spatial resolution.

Not at all true! You need to read up on your color-TV theory.

The subcarrier is at a frequency (an odd multiple of half the horizontal
scanning rate) that causes its sidebands (which are not continuous, but at
multiples of the horizontal scanning frequency) to fall _between_ the
sidebands of the luminance signal.

There is no loss of luminance information whatever -- unless the
receiverlops off the upper part of the composite signal to reduce visible
interference.

I assume that you mean bandwidth here and that sounds like a loss of
spatial resolution.


Most modern TVs use multi-frame comb filtering, and do a very good job of
extracting the fine detail.

Indeed they do. However, there is still some loss of fine detail at
least vertically and artifacts are introduced. See, for example:

http://members.aol.com/ajaynejr/vidcomb.htm#CombHorLoss

If comb filtering were as effective as you claim, then there would be
no need for S-video. However, it is very easy to see the quality
difference between composite video and S-video.

[Arny Krueger]

"William Sommerwerck" wrote in
message
"Arny Krueger" wrote in message
...

Laserdisks recorded the composite video signal. Those
with S outputs required an internal comb filter to
separate the C. When I got an IDTV, I used the composite
player output and let the TV do the separating.

That gets you into that post a few days back from
"Trotsky". If you are using a format that requires that
the video undergo
NTSC Y/C separation, the quality difference available
from using
the S-video cable can be negligible. It can even favor
not using the S-Video connection if the TV Y/C separator
is hot and the one in
the video player is not.

Exactly my point.


That can all change when you start working with digital
media. If you start out with a DVD-V signal, there is
potentially a about double the video information
available to the player than OTA NTSC can handle. The
situation is even more extreme for HDTV, where there is
potentially about three times the video information as
DVD-V.

A DTV adaptor can potentially have about 6 times as much
information available to it as the video from the tuner
in the analog TV it is attached to. You really want a
component video connection, but S-video done right could
be a good kicker.

It would certainly improve the luminance part. However,
the receiver will have its own bandpass filter for the
chrominance, and if it's limited to
0.5MHz... you won't see much of an improvement.

But, the chroma bandwidth over a S-video connection is not necessarily as
limited because it does not need to undergo frquency division-multiplesming
based demodulation.

I've often wondered why VHS has such terrible chrominance
bandwidth. It's SO bad that you often see objects --
especially in home videos -- with uncolored borders.
Horrible.

The VHS and S-VHS chroma bandwidth is 0.4/0.4 MHz, which is even worse
than NTSC's 0.5/1.5 MHz.

[jwvm]

On Feb 27, 10:45 am, "William Sommerwerck"
wrote:

snip

More to the point -- as the color-difference signals represent
_saturation_ -- not hue -- and most things, natural or manufactured, are
decorated with colors of constant saturation, color-difference signals don't
have to be wideband to _fully_ represent the color in the image. This is why
an NTSC receiver can "get away" with displaying only 0.5MHz of
color(-saturation) information.

The only time the "reduced" bandwidth is noticeable is in lettering -- such
as the CNN logo -- where 0.5MHz isn't enough to render the vertical bars.

There is actually a more basic reason for reduced resolution of color
information. The human visual system has greatest acuity in the green
portion of the spectrum and much less in the red and blue. Letting the
green portion correspond to the luminance signal provides good spatial
resolution and considerably less bandwidth need be allocated for the
red and green portions of the spectrum. BTW, this trick is also
commonly used for single-chip color cameras. There are often twice as
many green pixels than red or blue.

And yes, colored text can be exceedingly hard to read and looks
terrible.

I am one of the few people in the world who consciously understands all
this. ahem

Congratulations! We should notify the Nobel committee. :-)

[William Sommerwerck]

"Arny Krueger" wrote in message
. ..

It would certainly improve the luminance part. However,
the receiver will have its own bandpass filter for the
chrominance, and if it's limited to 0.5MHz...
you won't see much of an improvement.

But, the chroma bandwidth over a S-video connection is not
necessarily as limited because it does not need to undergo
frquency division-multiplesming based demodulation.

No, not necessarily as limited. But... To increase the chroma bandwidth in
the monitor would reduce the chroma signal's group delay, screwing up the
image.


I've often wondered why VHS has such terrible chrominance
bandwidth. It's SO bad that you often see objects --
especially in home videos -- with uncolored borders.
Horrible.

The VHS and S-VHS chroma bandwidth is 0.4/0.4 MHz, which is
even worse than NTSC's 0.5/1.5 MHz.

Until recently, few NTSC sets demodulated the full 1.5MHz. My classic NAD
MR-20a, which has an excellent image, has only 0.5MHz bandwidth.

[William Sommerwerck]

"jwvm" wrote in message
...
On Feb 27, 10:45 am, "William Sommerwerck"
wrote:

snip

More to the point -- as the color-difference signals represent
_saturation_ -- not hue -- and most things, natural or manufactured, are
decorated with colors of constant saturation, color-difference signals
don't
have to be wideband to _fully_ represent the color in the image. This is
why
an NTSC receiver can "get away" with displaying only 0.5MHz of
color(-saturation) information.

The only time the "reduced" bandwidth is noticeable is in lettering --
such
as the CNN logo -- where 0.5MHz isn't enough to render the vertical bars.

There is actually a more basic reason for reduced resolution of color
information. The human visual system has greatest acuity in the green
portion of the spectrum and much less in the red and blue. Letting the
green portion correspond to the luminance signal provides good spatial
resolution and considerably less bandwidth need be allocated for the
red and green portions of the spectrum. BTW, this trick is also
commonly used for single-chip color cameras. There are often twice as
many green pixels than red or blue.

This is true, but it's not the full -- or best -- explanation as to why
color information doesn't require full bandwidth.

[William Sommerwerck]

"jwvm" wrote in message
...

If comb filtering were as effective as you claim, then there would
be no need for S-video. However, it is very easy to see the quality
difference between composite video and S-video.

S-Video was a cheap and dirty "solution" to the problem of the lack of
expensive multi-dimensional comb filtering in TV receivers.

[jwvm]

On Feb 27, 11:29 am, "William Sommerwerck"
wrote:
"jwvm" wrote in message

...







On Feb 27, 10:45 am, "William Sommerwerck"
wrote:
snip
More to the point -- as the color-difference signals represent
_saturation_ -- not hue -- and most things, natural or manufactured, are
decorated with colors of constant saturation, color-difference signals
don't
have to be wideband to _fully_ represent the color in the image. This is
why
an NTSC receiver can "get away" with displaying only 0.5MHz of
color(-saturation) information.
The only time the "reduced" bandwidth is noticeable is in lettering --
such
as the CNN logo -- where 0.5MHz isn't enough to render the vertical bars.
There is actually a more basic reason for reduced resolution of color
information. The human visual system has greatest acuity in the green
portion of the spectrum and much less in the red and blue. Letting the
green portion correspond to the luminance signal provides good spatial
resolution and considerably less bandwidth need be allocated for the
red and green portions of the spectrum. BTW, this trick is also
commonly used for single-chip color cameras. There are often twice as
many green pixels than red or blue.

This is true, but it's not the full -- or best -- explanation as to why
color information doesn't require full bandwidth.

Huh!!?? The FCC mandated that the new color television signal had to
be compatible with existing monochrome sets and the monochrome
broadcasting standard. This meant that the color information needed to
fit within the existing channel bandwidth. The only way this could be
done and still maintain most of the existing image quality was to take
advantage of human perception limitations and limit the chromaticity
bandwidth severely. What other part is there to this story?

[jwvm]

On Feb 27, 11:31 am, "William Sommerwerck"
wrote:
"jwvm" wrote in message

...

If comb filtering were as effective as you claim, then there would
be no need for S-video. However, it is very easy to see the quality
difference between composite video and S-video.

S-Video was a cheap and dirty "solution" to the problem of the lack of
expensive multi-dimensional comb filtering in TV receivers.

Are you claiming that S-video isn't any better than a good comb
filter? It actually works quite well though not as good as component
color given the need for quadrature demodulation with S-video.

[William Sommerwerck]

"jwvm" wrote in message
...
On Feb 27, 11:31 am, "William Sommerwerck"
wrote:
"jwvm" wrote in message
...

If comb filtering were as effective as you claim, then there would
be no need for S-video. However, it is very easy to see the quality
difference between composite video and S-video.

S-Video was a cheap and dirty "solution" to the problem of the lack
of expensive multi-dimensional comb filtering in TV receivers.

Are you claiming that S-video isn't any better than a good comb
filter? It actually works quite well though not as good as component
color given the need for quadrature demodulation with S-video.

It is, in principle, IF the original signal was created AND STORED as
separate Y & C components. The Y & C laid down on an S-VHS recorder, unless
they come from a camera, were not originally separate.

[Arny Krueger]

"William Sommerwerck" wrote in
message
"jwvm" wrote in message
...
On Feb 27, 11:31 am, "William Sommerwerck"
wrote:
"jwvm" wrote in message
...

If comb filtering were as effective as you claim, then
there would be no need for S-video. However, it is
very easy to see the quality difference between
composite video and S-video.

S-Video was a cheap and dirty "solution" to the problem
of the lack of expensive multi-dimensional comb
filtering in TV receivers.

Are you claiming that S-video isn't any better than a
good comb filter? It actually works quite well though
not as good as component color given the need for
quadrature demodulation with S-video.

It is, in principle, IF the original signal was created
AND STORED as separate Y & C components.

In the case of the S-video output of a DVD player or HDTV receiver, the
S-Video signal can have far more information in it than the NTSC output of
the same player.

[William Sommerwerck]

"Arny Krueger" wrote in message
...

In the case of the S-video output of a DVD player or HDTV receiver, the
S-Video signal can have far more information in it than the NTSC output of
the same player.

Yes, and component outputs have even more.

[Alan[_2_]]

In article "Arny Krueger" writes:

In the case of the S-video output of a DVD player or HDTV receiver, the
S-Video signal can have far more information in it than the NTSC output of
the same player.

Since the s-video output and the composite output are both NTSC, it
is impossible for either the s-video output or the composite output
to have *more* output than the NTSC output -- they *ARE* NTSC outputs.

Alan

[Sal M. Onella]

"William Sommerwerck" wrote in message
...



More to the point -- as the color-difference signals represent
_saturation_ -- not hue -- and most things, natural or manufactured, are
decorated with colors of constant saturation, color-difference signals
don't
have to be wideband to _fully_ represent the color in the image. This is
why
an NTSC receiver can "get away" with displaying only 0.5MHz of
color(-saturation) information.

The human eye perceives color with the less-numerous "cones," compared to
the "rods" which perceive differences in brightness. Thus, small areas are
not perceived in color, hence the reduced need for chroma bandwidth.
Wideband NTSC chroma would be a waste.

When the color difference signals are matrixed, the resultant I & Q signals
are transmitted as a double-sideband color signal whose two quadrature
amplitudes and polarities represent a vector that instantaneously describes
both hue and saturation. The receiver reconstitutes the difference signals.
It's genius.

[G-squared]

On Feb 27, 7:06*am, "Arny Krueger" wrote:
"William Sommerwerck" wrote in
messagenews:V7qdnUmJH6sv5VjanZ2dnUVZ_v6rnZ2d@comca st.com



"jwvm" wrote in message
news:
...

The luminance component (Y) contains the fine detail. The
chrominance components (I and Q) modulate the color
carrier
at 3.58 MHz and are in quadrature. The presence of the
color carrier and color signals in the luminance signal
further degrades the bandwidth of the luminance
component resulting in an additional loss of spatial
resolution.

Not at all true! You need to read up on your color-TV
theory.

The subcarrier is at a frequency (an odd multiple of half
the horizontal scanning rate) that causes its sidebands
(which are not continuous, but at multiples of the
horizontal scanning frequency) to fall _between_ the
sidebands of the luminance signal.

There is no loss of luminance information whatever --
unless the receiverlops off the upper part of the
composite signal to reduce visible interference.

Most modern TVs use multi-frame comb filtering, and do a
very good job of extracting the fine detail.

However, there remains the problem of fine detail in the
chrominance
information in NTSC composite signals. There are two chroma
signals, one
with 1.5 MHz bandwidth, and one with 0.5 MHz *bandwidth. So you can
do what
you want to in terms of extending the bandwidth of the luninance
signal, but
the color is locked into these numbers. They are not bad choices
for a 4 MHz
luminance signal, maybe a bit conservative. They suffer as you
extend the
bandwidth of the luminance beyond 4 MHz.

If you send the signal as S-video, then AFAIK a lot more bandwidth
can be
packed into the luminance information.

The last YIQ wncoder I saw was in an RCA TK45 color camera from around
'75. All I've seen since then are equiband Y,R-Y, B-Y encoders where
the modulation axes are on the 0 and 90 degree axes rather than
rotated 33 degrees. That bandwidth filtering hasn'r been done that way
in a long time. S- video _could_ have more bandwidth if a custom
encoder was built but in fact they are simply an encoder where the
modulated subcarrier and luminance (Y) are not summed together and are
instead sent out separately on their own cables. Why would anyone
bother to make a 'super' S-video encoder since the destination is a
limited response (sompared to a broadcast vtr) VHS deck ?.

OK, you guys talk about "500 lines" resolution on a Betamax or an S-
VHS but those numbers are pure BS. Camera response is defined when the
signal reaches 5% at some high frequency -- 26 dB down !! It was done
that way because cameras suffer from a LOT of ills like optical and
electrical focus and there had to be SOME meaningful way to compare
them. When U-Matic 3/4" hetrodyne VTRs and its poor cousin the
Betamax showed up with a similar frequency response as cameras, THAT
measurement technique was adopted for consumer tape decks. Analog
broadcast VTRs are pretty dang close to 'flat' response so that multi-
generation editing was possible. Digital decks don't even measure the
response because it's never wrong. Anything that messes up the
response will cause other problems way worse so it will get fixed.

PAL and NTSC machines are essentially the same with some changes but
by and large the same. PAL required 5.5 MHz bandwidth so guess what,
the NTSC ended up with more than it needed which is one of the reasons
the component digital DVD is better than analog composite.

BTW, Accom built a digital decoder ( D-221 ) about 20 years ago that
could do a 525 line (full frame) comb filter. This was not used for
moving pictures but for grabbing a still frame, it had 0 cross color
artefacts and was virtually identical to component.

GG

[Arny Krueger]

"Alan" wrote in message

In article "Arny Krueger" writes:

In the case of the S-video output of a DVD player or
HDTV receiver, the S-Video signal can have far more
information in it than the NTSC output of the same
player.

Since the s-video output and the composite output are
both NTSC,

In a home or studio video component, neither s-video nor composite video
must be NTSC, unless broadcast compatibility is required.

[William Sommerwerck]

Since the s-video output and the composite output are both NTSC,
it is impossible for either the s-video output or the composite output
to have *more* output than the NTSC output -- they *ARE* NTSC
outputs.

This might be true in practice, but "it ain't necessarily so".

[William Sommerwerck]

"Sal M. Onella" wrote in message
...
"William Sommerwerck" wrote in message
...

More to the point -- as the color-difference signals represent
_saturation_ -- not hue -- and most things, natural or manufactured,
are decorated with colors of constant saturation, color-difference
signals don't have to be wideband to _fully_ represent the color in
the image. This is why an NTSC receiver can "get away" with
displaying only 0.5MHz of color(-saturation) information.

The human eye perceives color with the less-numerous "cones",
compared to the "rods" which perceive differences in brightness.
Thus, small areas are not perceived in color, hence the reduced
need for chroma bandwidth.

Not so.


Wideband NTSC chroma would be a waste.

We already have wideband NTSC chroma -- the Q signal of 1.5MHz.


When the color difference signals are matrixed, the resultant I & Q
signals are transmitted as a double-sideband color signal whose two
quadrature amplitudes and polarities represent a vector that
instantaneously
describes both hue and saturation. The receiver reconstitutes the
difference
signals. It's genius.

Yes, it's brilliant. (It's one of the great 20th century inventions.) But --
and I will keep repeating this ad nauseum -- the reason color TV systems (of
all sorts) can "get away" with reduced chroma bandwidth has little to do
with the eye's (relatively) limited color resolution and a great deal to do
with what I said -- most objects are colored with constant saturation, which
greatly reduces the required bandwidth for the chrominance signals -- or
more precisely, lets them convey more useful information.