Harsh, "aliased" sound with digital TV converter box.
Doc
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
David Grant
"Doc" <docsa...@yahoo.com> wrote in message
news:20ab4a9d-04d6-4fe6...@s19g2000prg.googlegroups.com...
Need to know more about the rest of the system... what are you interfacing
with? Are you using the digital output, and if so, into what?
David Grant
"David Grant" <REMOVE_M...@msn.com> wrote in message
news:LoCdnY5iXaeD_iHa...@giganews.com...
Sorry, just looked up the box... no digital output. Nevermind then.
Doc
> Need to know more about the rest of the system... what are you interfacing
> with? Are you using the digital output, and if so, into what?
Just to clarify, I'm using a Walmart cheapie antenna going through an
amplifier, just going to an analog TV. Actually, I've got it going
through a VCR but I wouldn't think that would be an issue.
As far as I can tell, it's an "it either works or it doesn't"
proposition. It seems once the signal falls below a certain threshold,
you don't get a picture. If it bounces around that threshold, the
picture gets corrupted with pixelation artifacts.
GregS
If it sounds as good as my cell phone, you could stomp on it.
Let us know what happens. I would try ANOTHER brand unit
and hear what happens. It seems like there are cheaper models ??
greg
GregS
I just love digital picture artifacts.
greg
Arny Krueger
> I just picked up an Insignia NS-DXA1 digital converter
> box at Best Buy.
AFAIK Insignia is a BB house brand.
> 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.
Interesting. I watch HDTV using a HDTV receiver card in my PC. Sound seems
to be OK.
> Is this an issue with this particular model/brand or a
> common problem?
I noticed some mixed reviews of the device at the BB web site.
IMO, it is low-featured and under-specified.
> Anything that can be done about it?
Take it back and get a good one?
Ron Capik
> < ...snip... >
>
> I just love digital picture artifacts.
>
> greg
Green Xenon (aka:Radium) is that you?
Later...
Ron Capik
--
2344...@qq.com
2344...@qq.com
Doc
>
> I noticed some mixed reviews of the device at the BB web site.
>
> IMO, it is low-featured and under-specified.
I'll have to check out the BB site. I got it because I happened to be
in the store. By "under-specified" what are you referring to?
As far as features, the user interface seems pretty well thought out
and intuitive. I don't know what other features I'd expect it to have.
One of the very useful features is a beep tone for adjusting the
antenna so you can tell if the signal is getting stronger or weaker
without having to see the screen, though there is a "thermometer"
display as well.
The only problem I really have is the sound issue, which unfortunately
is a pretty major issue. I'll be taking it back if I don't divine
something that can be done about it.
Arny Krueger
> On Feb 20, 1:15 pm, "Arny Krueger" <ar...@hotpop.com>
> wrote:
>
>>
>> I noticed some mixed reviews of the device at the BB web
>> site.
>>
>> IMO, it is low-featured and under-specified.
>
>
> I'll have to check out the BB site. I got it because I
> happened to be in the store. By "under-specified" what
> are you referring to?
The only specs for it that I can find are like the dimensions of its case,
and other equally obvious stuff.
For me, the lack of anything but a simple coax connector for video is a
stopper. Lots of people have TV sets with S-video and component video
inputs, that would give a better picture.
trotsky
Proof?
Don Stauffer in Minnesota
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.
Doug Smith W9WI
> For me, the lack of anything but a simple coax connector for video is a
> stopper. Lots of people have TV sets with S-video and component video
> inputs, that would give a better picture.
I don't think boxes with component outputs are eligible for the voucher
program.
Arny Krueger
news:pan.2008.02.21....@invalid.nospam
Ya can't buy my videophile card for $40, Mr. Big government. ;-)
Arny Krueger
news:lXevj.43609$yE1.7395@attbi_s21
> Proof?
Are you saying that you can't tell the difference in picture quality betwen
a good TV running off of a plain old baseband coax output, and CV, where the
source material is first rate HDTV?
G-squared
> >> For me, the lack of anything but a simple coax connector
> >> for video is a stopper. Lots of people have TV sets with
> >> S-video and component video inputs, that would give a
> >> better picture.
> > Proof?
>
> Are you saying that you can't tell the difference in picture
quality betwen
> a good TV running off of a plain old baseband coax output, and CV,
where the
> source material is first rate HDTV?
Depending on the quality of the Y/C separator in the set, the
difference could be very subtle. As far as the source being HD, you're
converting it back to NTSC. If you had a really good feed before, how
could you tell a difference? The best source I ever saw on the old 35"
set was the 2003 Rose Parade on analog OTA KTLA down converted from
the HD feed.
GG
pj
Yeah, I've been there with that "aliasing
artifact." Advice: Don't necessarily throw
rocks at the Insignia Box.
Rule of thumb for me: Two channels derived from
OTA seems to work well. Two-channels derived
from whatever is on cable might be harsh (but
not always).
I've had this sound problem with both our
'digital SD' cable box (SA 2100) and with
translation of 5.1 sound into two-channel sound
by our HD box (SA 4240). (Haven't sprung for a
receiver yet as we are in 'technology overload'
around here.)
There are probably better ways of connecting the
4240 to our HDTV but that's way down my list of
things-to-do.
(As an aside, here's Rule #2 for Home
Technology: Do not buy identical Harmony
remotes for two separate systems, then fail to
label them and write down the login's and
passwords for the online profiles.)
'Harsh' was definitely noticeable during the
ProBowl game. We ran sets in three rooms.
One was feeding a R-L signal from the Cox 4240
into our projection HDTV--harsh. When I
switched the HDTV to OTA and let the HDTV
convert the broadcast sound into it's own
"surround-sound" the sound was fine (this HDTV
would probably be classified as a 3.1 with some
screwball, Bose-like, phase shifting to 'fill
the room.')
In another room, Cox's 2100 box was converting
cable's digital (SD) signal into 480i and
two-channel 'stereo' for an older CRT set. That
set also sounded "harsh." (I tend to blame Cox
cable for something it was doing with the Fox
signal.)
In a third room, we had an older EDTV plasma
running from a Samsung 451 tuner, OTA (bunny
ears) with R-L audio (being decoded from
whatever the local Fox station was putting out
on its DTV signal) -- that one sounded clean.
Things will probably be better a year from now.
Meanwhile, keep a cooler of cold beer near the
TV and use it as needed to take the edge off the
sound. {;-)
--
pj
Doc
> For me, the lack of anything but a simple coax connector for video is a
> stopper. Lots of people have TV sets with S-video and component video
> inputs, that would give a better picture.
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?
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.
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.
Arny Krueger
news:e2481639-6a47-4f78...@s37g2000prg.googlegroups.com
> On Feb 21, 7:58 am, "Arny Krueger" <ar...@hotpop.com>
> wrote: > >> For me, the lack of anything but a simple
> coax connector > >> for video is a stopper. Lots of
> people have TV sets with > >> S-video and component
> video inputs, that would give a > >> better picture.
> > > Proof?
>
> > Are you saying that you can't tell the difference in
> picture quality betwen
> > a good TV running off of a plain old baseband coax
> output, and CV, where the
> > source material is first rate HDTV?
> Depending on the quality of the Y/C separator in the set,
> the difference could be very subtle.
CV (YUV) from sources from other than a NTSC TV signal need not pass through
a Y/C separator.
The NTSC Y/C separator inside the TV set can't possibly have as much
bandwidth of a good quality CV decoder driven by a first rate HDTV signal.
Even just S-Video can have more bandwidth than a NTSC signal can handle.
> As far as the source being HD, you're converting it back to NTSC.
CV need not be downconverted to NTSC for analog TVs that can properly handle
CV. I don't know why one would do such a thing unless there was some NTSC
stumbling block in the signal path.
> If you had a really good feed before, how could you tell a difference?
The really good feed can come from a higher bandwidth source than mere NTSC.
> The best source I ever saw on the old 35" set was the
> 2003 Rose Parade on analog OTA KTLA down converted from
> the HD feed.
The old 35" set probably didn't have YUV inputs.
Richard C.
That box has NO digital output!
Richard C.
On Feb 20, 1:15 pm, "Arny Krueger" <ar...@hotpop.com> wrote:
>
> I noticed some mixed reviews of the device at the BB web site.
>
> IMO, it is low-featured and under-specified.
I'll have to check out the BB site. I got it because I happened to be
in the store. By "under-specified" what are you referring to?
============================
See my review on the site.
The unit is one of the higher priced and has NO digital audio out
and it has NO S-Video out!
===============================
trotsky
Not if they're ABX'd, no.
Arny Krueger
We've done video ABX tests and had results that were pretty consistent with
the video state of the art. I take it that you have had some problems?
David Grant
> That box has NO digital output!
>
Hence my follow-up post over 24hours ago.
Jer
Who said anything about a voucher program?
--
jer
email reply - I am not a 'ten'
dmaster
> On Feb 21, 4:57 am, "Arny Krueger" <ar...@hotpop.com> wrote:
>
> > For me, the lack of anything but a simple coax connector for video is a
> > stopper. Lots of people have TV sets with S-video and component video
> > inputs, that would give a better picture.
>
> 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?
Most do have ATSC/QAM tuners. But don't dismiss rabbit ears. }:)
Sure, the majority of people use some cable or satellite service, but
I'm one of the minority who only watches OTA. Since the switch to
digital broadcasts, all the broadcasts in my area are easier to
receive and *far* better quality.
>
> 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? Of course they are. All the networks and most of the
major independents broadcast quite a bit of HD these days. All it
takes is the appropriate antenna (for most people near the broadcast
towers, that could be rabbit ears - Don't forget, unless you live with
a particular reception problem, like a mountain in the way, or live in
an extreme fringe area, (over 40 miles away) digital broadcast are
usually much easier to receive than the analog broadcasts. ) and an
HDTV with an ATSC tuner (which is pretty much all of them for the last
2 years). Of course an external ATSC tuner can be used if needed.
>
> 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.
Certainly. }:) I use a DVD recorder with an ATSC tuner to feed my
remaining analog TV set. Even downconverted for a standard tube TV,
the improvement in quality is obvious except for the most perfect of
analog broadcasts.
Dan (Woj...)
stra...@yahoo.com
> "G-squared" <stratu...@yahoo.com> wrote in message
>
> news:e2481639-6a47-4f78-bbea-
abb04d...@s37g2000prg.googlegroups.com
>
> > On Feb 21, 7:58 am, "Arny Krueger" <ar...@hotpop.com>
> > wrote: > >> For me, the lack of anything but a simple
> > coax connector > >> for video is a stopper. Lots of
> > people have TV sets with > >> S-video and component
> > video inputs, that would give a > >> better picture.
> > > > Proof?
>
> > > Are you saying that you can't tell the difference in
> > picture quality betwen
> > > a good TV running off of a plain old baseband coax
> > output, and CV, where the
> > > source material is first rate HDTV?
> > Depending on the quality of the Y/C separator in the set,
> > the difference could be very subtle.
>
> CV (YUV) from sources from other than a NTSC TV signal need not
pass through
> a Y/C separator.
True but the subject is the subsidized converter box which is NOT
intended to be the best performance
> The NTSC Y/C separator inside the TV set can't possibly have as
much
> bandwidth of a good quality CV decoder driven by a first rate HDTV
signal.
> Even just S-Video can have more bandwidth than a NTSC signal can
handle.
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.
> > As far as the source being HD, you're converting it back to
NTSC.
>
> CV need not be downconverted to NTSC for analog TVs that can
properly handle
> CV. I don't know why one would do such a thing unless there was
some NTSC
> stumbling block in the signal path.
It does if it's for the subsidized converter box. If you're buying a
better grade box you'd be correct.
> > If you had a really good feed before, how could you tell a
difference?
>
> The really good feed can come from a higher bandwidth source than
mere NTSC.
Yeah but those folks have been watching NTSC on an analog set. So how
could they tell the difference?
>
> > The best source I ever saw on the old 35" set was the
> > 2003 Rose Parade on analog OTA KTLA down converted from
> > the HD feed.
>
> The old 35" set probably didn't have YUV inputs.
True, but it doesn't alter the fact that it was the BEST OTA analog I
ever saw on that set.
GG
Doc
> > 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'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.
William Sommerwerck
Hello? The subsidy is a fixed discount. The government doesn't care which
converter you buy, as long as it meets certain minimum standards.
Albert Manfredi
> 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.
:
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.
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!
===========================
dmaster
> On Feb 21, 3:39 pm, dmaster <dan....@sbcglobal.net> 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
> 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
< 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
in message news:DB8wj.52854$ov5....@newsfe15.phx...
> 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
wrote:
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
> 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
> 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" <jw...@umich.edu> wrote in message
news:39589e80-5e54-42db...@f47g2000hsd.googlegroups.com...
>
> 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
>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
news:c455d5f1-1d31-4cab...@v3g2000hsc.googlegroups.com
> On Feb 21, 3:44 pm, stratu...@yahoo.com 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
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
wrote:
> "jwvm" <j...@umich.edu> wrote in message
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
> 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
news:rvadnTvNF5kA81ja...@comcast.com...
> 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
message news:V7qdnUmJH6sv5Vja...@comcast.com
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
> "Arny Krueger" <ar...@hotpop.com> wrote in message
> news:rvadnTvNF5kA81ja...@comcast.com...
>
>> 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
> 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
>> 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.
jwvm
wrote:
> "jwvm" <j...@umich.edu> wrote in message
>
> news:067d997b-499d-4d56...@60g2000hsy.googlegroups.com...
>
> > 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
> "Arny Krueger" <ar...@hotpop.com> wrote in message
> news:er2dndE_B-Bw4Vja...@comcast.com...
>
>> 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 here:
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... ;-)
Arny Krueger
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
<grizzledgee...@comcast.net> 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
>> 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
> <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
> 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
<grizzledgee...@comcast.net> wrote:
> "jwvm" <j...@umich.edu> wrote in message
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
<grizzledgee...@comcast.net> wrote:
> "jwvm" <j...@umich.edu> wrote in message
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
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
> "jwvm" <jw...@umich.edu> wrote in message
> news:c8d71ccc-b31c-4e9c...@41g2000hsc.googlegroups.com...
>> On Feb 27, 11:31 am, "William Sommerwerck"
>> <grizzledgee...@comcast.net> wrote:
>>> "jwvm" <j...@umich.edu> wrote in message
>>> news:6db62321-d53f-49ba...@u69g2000hse.googlegroups.com...
>
>>>> 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
> 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
>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
>
> 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
> "William Sommerwerck" <grizzledgee...@comcast.net> wrote in
> messagenews:V7qdnUmJH6sv5Vja...@comcast.com
>
>
>
> > "jwvm" <j...@umich.edu> wrote in message
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
news:fq5l8a$q50$1...@news.Stanford.EDU
> In article "Arny Krueger" <ar...@hotpop.com> writes:
>
>> In the case of the S-video output of a DVD player or
>> HDTV receiver, the S-Video signal can have far more
>> information in it than the NTSC output of the same
>> player.
> Since the s-video output and the composite output are
> both NTSC,
In a home or studio video component, neither s-video nor composite video
must be NTSC, unless broadcast compatibility is required.
William Sommerwerck
> it is impossible for either the s-video output or the composite output
> to have *more* output than the NTSC output -- they *ARE* NTSC
> outputs.
This might be true in practice, but "it ain't necessarily so".
William Sommerwerck
news:BOtxj.10845$kp6....@newsfe07.phx...
> "William Sommerwerck" <grizzle...@comcast.net> wrote in message
> news:C-idnUOgZ9rhGFja...@comcast.com...
>> 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.
William Sommerwerck
> 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
I can see generating and recording R-Y and B-Y signals, as it makes it
easier to convert the signal to PAL and SECAM. But you can't transmit
_equal-bandwidth_ NTSC signals, as that would push the bandwidth beyond
5MHz. At some point NTSC color has to be converted to IQ.
> OK, you guys talk about "500 lines" resolution on a Betamax
> or an S-VHS but those numbers are pure BS.
I don't remember any of us talking aobut such things.
My memory is that Betamax originally had 240 lines, which equates to about
3MHz, the minimum bandwidth needed for an "acceptably" sharp picture on a
20" screen. 500 lines of resolution on American TV would require more than
6MHz of luminance bandwidth, an impossibility for Beta or VHS.
Scott Dorsey
How would they not be NTSC?
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Arny Krueger
> 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
If we extrapolate this discussion to audio, then we have Willaim
Sommerwerck, MP3 advocate! ;-)
> has little to do
> with the eye's (relatively) limited color resolution and
The eye's limited color and intensity resolution is at least an effect that
is at least perceivable.
We have to distinguish between sensory limits like those due to rods and
cones in the eye, and perceptual limits due to the brain's data crunching
bottlenecks and limited training.
> a great deal to do with what I said -- most objects are
> colored with constant saturation,
This is like saying that all objects are each well-modeled as being painted
all over with the same paint, with such variations in coloration that exist
being related to things like the location of light sources.
> which greatly reduces
> the required bandwidth for the chrominance signals -- or
> more precisely, lets them convey more useful information.
The "same paint" model works well for a lot of artificial objects, some
natural objects, and many more natural objects if viewed from a distance.
But it doesn't work for everything.
Many objects, both artificial and natural, don't follow the "same paint"
rule.
I contemplate the evolution of special effects in movies. In their day, I
found the special effects in the better early-1950s science fiction movies
to be compelling. Today, my brain is trained by experience to see through
many of them, and I perceive them as being hopelessly toy-like. I can do a
binary search of sorts and contemplate the special effects in the hottest
movies of the early 80s. While the better movies of those days are not so
toy-like, they are also not nearly as compelling as the latest-greatest.
Arny Krueger
news:fq6fj4$6a7$1...@panix2.panix.com
> William Sommerwerck <grizzle...@comcast.net> wrote:
>>> Since the s-video output and the composite output are
>>> both NTSC,
Only if the source is NTSC. Today we have many common video sources that
exceed NTSC limits in many ways.
>>> 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".
> How would they not be NTSC?
Only broadcast video *must* be NTSC, right?
Albert Manfredi
wrote:
> 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.
If it's true that colors are applied with constant saturation, then
that would explain why many consider NTSC to provide cartoon-like
images. Perhaps what you're really saying is that the eyes are not as
sensitive to color intensity variations as they are to luminance
variations, in which case it's back to "limited color resolution."
Bert
William Sommerwerck
news:d6fcd81b-38c4-41c6...@f47g2000hsd.googlegroups.com...
> 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.
If it's true that colors are applied with constant saturation, then
that would explain why many consider NTSC to provide cartoon-like
images. Perhaps what you're really saying is that the eyes are not as
sensitive to color intensity variations as they are to luminance
variations, in which case it's back to "limited color resolution."
I'm saying nothing of the sort.
Those who consider NTSC to provide "cartoon-like images" know nothing about
color television. PAL and SECAM use color-difference signals as well.
When you subtract the luminance signal from the color-primary signals, you
get a signal that represents ONLY the saturation of the color -- nothing
else. As most colors are of constant saturation, there is less "information"
in the color-difference signal -- redundant information that already appears
in the luminance -- thus requiring less bandwidth.
If you don't understand this, think of a cube illuminated from the side, and
what the color-primary image and color-difference image would look like.
Then you will understand.
William Sommerwerck
> "William Sommerwerck" <grizzle...@comcast.net> wrote in
> message news:zoqdnZ3Q2_7RJFva...@comcast.com
>> 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
> If we extrapolate this discussion to audio, then we have Willaim
> Sommerwerck, MP3 advocate! ;-)
God, no. I hate compressed audio. (Dolby Digital, at least.)
>> has little to do
>> with the eye's (relatively) limited color resolution and
> The eye's limited color and intensity resolution is at least
> an effect that is at least perceivable.
In the NTSC system, this difference shows up in the bandwidth of the color
signals. The researchers determined that (for a 480-line, 30-frame system,
on a 21" screen, presumably) you could see full red/green/blue-primaries
color only up to 0.5MHz, while only colors that could be synthesized from
red-orange and blue-green primaries were visible from 0.5MHz to 1.5MHz. *
Above 1.5MHz, the eye saw only B&W. So the 0.5MHz I signal is yellow/purple,
and the 1.5MHz Q signal is red-orange/blue-green.
But this has _nothing whatever_ to do with what I'm talking about.
One of the nice things about NTSC (and PAL for that matter -- they're
basically the same system) is that the use of color-difference signals
(rather than color-primary signals) removes any redundancy. The color
signals contain ZERO information about brightness. Which is good, because
the brightness information is already conveyed by the Y signal. It doesn't
need to be contained in the color signals.
* Note that these are roughly the early two-strip Technicolor primaries. The
overall gamut is not very broad, but "Secret of the Wax Museum" is
surprisingly good.
> This is like saying that all objects are each well-modeled as being
> painted all over with the same paint, with such variations in coloration
> that exist being related to things like the location of light sources.
Not coloration (hue), but lightness (value). Otherwise, that is absolutely
correct.
> The "same paint" model works well for a lot of artificial objects, some
> natural objects, and many more natural objects if viewed from a distance.
> Many objects, both artificial and natural, don't follow the "same paint"
> rule.
But it's true for most objects, natural or artificial. If you don't believe
this, try to find any colored object -- natural or artificial -- that is
_not_ "constant saturation". You're most likely to find it in flowers and
fabric patterns.
I don't want to argue this too much, because most people don't have a good
understanding of color analysis and systhesis.
Arny Krueger
> "Arny Krueger" <ar...@hotpop.com> wrote in message
> news:i9ydnfwAHqhiUlva...@comcast.com...
>> "William Sommerwerck" <grizzle...@comcast.net> wrote
>> in
>> message news:zoqdnZ3Q2_7RJFva...@comcast.com
>
>>> 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
>
>> If we extrapolate this discussion to audio, then we have
>> Willaim Sommerwerck, MP3 advocate! ;-)
> God, no. I hate compressed audio. (Dolby Digital, at
> least.)
(1) Dolby Digital is really old-old tech, predating MP3 by lots.
(2) Usual grousing about what people say and what they actually hear with
their eyes closed.
>>> has little to do
>>> with the eye's (relatively) limited color resolution and
>
>> The eye's limited color and intensity resolution is at
>> least
>> an effect that is at least perceivable.
> In the NTSC system, this difference shows up in the
> bandwidth of the color signals. The researchers
> determined that (for a 480-line, 30-frame system, on a
> 21" screen, presumably) you could see full
> red/green/blue-primaries color only up to 0.5MHz,
Unfortunately, by the end of the NTSC era, 32 and 36 inch sets were
mainstream, even average. 42 inch sets were common. NTSC looked like $#@!!
on large screens - barely tolerable on 32 inch sets.
> while
> only colors that could be synthesized from red-orange and
> blue-green primaries were visible from 0.5MHz to 1.5MHz.
> * Above 1.5MHz, the eye saw only B&W. So the 0.5MHz I
> signal is yellow/purple, and the 1.5MHz Q signal is
> red-orange/blue-green.
This would all be sensory-based, no doubt tested with what amounted to be
synthetic, worst-case test patterns.
> But this has _nothing whatever_ to do with what I'm
> talking about.
Right, you're talking about perception.
> One of the nice things about NTSC (and PAL for that
> matter -- they're basically the same system) is that the
> use of color-difference signals (rather than
> color-primary signals) removes any redundancy. The color
> signals contain ZERO information about brightness. Which
> is good, because the brightness information is already
> conveyed by the Y signal. It doesn't need to be contained
> in the color signals.
OK.
> * Note that these are roughly the early two-strip
> Technicolor primaries. The overall gamut is not very
> broad, but "Secret of the Wax Museum" is surprisingly
> good.
OK.
>> This is like saying that all objects are each
>> well-modeled as being painted all over with the same
>> paint, with such variations in coloration that exist
>> being related to things like the location of light
>> sources.
> Not coloration (hue), but lightness (value). Otherwise,
> that is absolutely correct.
Ah yes, I need to reload some long-unused parts of my vocabulary.
>> The "same paint" model works well for a lot of
>> artificial objects, some natural objects, and many more
>> natural objects if viewed from a distance.
>> Many objects, both artificial and natural, don't follow
>> the "same paint" rule.
> But it's true for most objects, natural or artificial. If
> you don't believe this, try to find any colored object --
> natural or artificial -- that is _not_ "constant
> saturation".
By this you mean constant saturation of a given color hue, no?
> You're most likely to find it in flowers and
> fabric patterns.
And certain trees and rocks. Water with certain lighting and/or degrees of
activity. Artificial objects with exposed frameworks. Artificial objects
designed to be highly visible. Much text.
> I don't want to argue this too much, because most people
> don't have a good understanding of color analysis and
> systhesis.
One point is that the DVD was one of the larger beginnings of the end for
NTSC TV.
Scott Dorsey
Well, in terms of the fact that the FCC will only come after you if your
broadcast waveform doesn't match the NTSC specs, yes. But in fact, just
about everything in use today meets the NTSC specs, other than VHS machines
which need a time base corrector to meet timing specifications and which
are going away very fast.
William Sommerwerck
>> In the NTSC system, this difference shows up in the
>> bandwidth of the color signals. The researchers
>> determined that (for a 480-line, 30-frame system,
>> on a 21" screen, presumably) you could see full
>> red/green/blue-primaries color only up to 0.5MHz,
> Unfortunately, by the end of the NTSC era, 32 and 36 inch sets
> were mainstream, even average. 42 inch sets were common.
> NTSC looked like $#@!! on large screens -- barely tolerable on 32"
> sets.
I remember the early 25" Sony consoles. They had really weak color, though I
don't know why.
However, I own a 32" Toshiba IDTV and and Sony 36" IDTV. They display
spectacularly good NTSC images. Both digitally goose the luminance, and (as
far as I know) both have full-bandwidth chroma demodulation.
By the way, the original Advent projector had full-bandwidth color.
>> But this has _nothing whatever_ to do with what I'm
>> talking about.
> Right, you're talking about perception.
No, I'm talking objective fact. Color-difference signals require less
bandwidth than color-primary signals.
>>> Many objects, both artificial and natural, don't follow
>>> the "same paint" rule.
>> But it's true for most objects, natural or artificial. If
>> you don't believe this, try to find any colored object --
>> natural or artificial -- that is _not_ "constant
>> saturation".
> By this you mean constant saturation of a given color hue, no?
Yes. It would be meaningless to talk about different hues.
> And certain trees and rocks. Water with certain lighting and/or degrees
> of activity. Artificial objects with exposed frameworks. Artificial
objects
> designed to be highly visible. Much text.
Text? Are you referring to illuminated manuscripts? <grin>
> One point is that the DVD was one of the larger beginnings
> of the end for NTSC TV.
I don't want to be too quick to defend NTSC, but it can be exceptionally
good. It's not that NTSC is of lower quality than DVD, but rather that DVD
is better.
trotsky
> "William Sommerwerck" <grizzle...@comcast.net> wrote in
> message news:ydKdnRAcDvkFf1va...@comcast.com
>> "Arny Krueger" <ar...@hotpop.com> wrote in message
>> news:i9ydnfwAHqhiUlva...@comcast.com...
>>> "William Sommerwerck" <grizzle...@comcast.net> wrote
>>> in
>>> message news:zoqdnZ3Q2_7RJFva...@comcast.com
>>>> 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
>>> If we extrapolate this discussion to audio, then we have
>>> Willaim Sommerwerck, MP3 advocate! ;-)
>
>> God, no. I hate compressed audio. (Dolby Digital, at
>> least.)
>
> (1) Dolby Digital is really old-old tech, predating MP3 by lots.
What difference does it make when it was created?
pj
> William Sommerwerck <grizzle...@comcast.net> wrote:
>>> 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".
>
> How would they not be NTSC?
> --scott
A case for S-Video in preference to Composite:
Let's consider the case of cable delivery of a
480i broadcast.
The signal originates from the station as a
digital feed from the station to the cable
company. (In San Diego, Cox maintains fiber
feeds from each 'must carry' station.) At this
point, the signal it subject to the limitations
of the NTSC spec and Cox is receiving something
better than the OTA signal.
The cable company produces two distinct products:
1) A conventional NTSC analog signal that it
delivers to the customer (via format conversions
as it travels through the cable infrastructure).
This RF signal is delivered directly to the
customer's TV receiver or, is demodulated in the
STB and presented to the customer as either an
R.F. signal, in NTSC format on Channel 3/4; or,
a composite video signal -- essentially the
baseband version of the NTSC signal; or, an
S-Video output of luminance and color. All
three of these outputs are limited in quality by
the limitations inherent in NTSC.
2) A digital signal applied, along with one or
more other signals, to an RF channel compatible
with the STB. This signal will have been
sufficiently compressed to fit in the allocated
bandwidth.
This signal is detected and made available to
the customer by perhaps four outputs; RF (NTSC),
Composite (NTSC-baseband), Component and
S--Video. The RF and Composite outputs are
subject to the limitations inherent in the NTSC
spec. The S-Video and Component outputs may be
slightly superior since NTSC wasn't imposed
between the originating station and the
customer's STB.
I've also seen this work in reverse where the
cable company heavily compressed the digital
feeds for less popular media (to fit three or
four signals into a single RF slot). The analog
signals (raw NTSC-RF) were superior to the
output from the STB.
--
pj
Arny Krueger
news:vACxj.1219$TT4.358@attbi_s22
Perceptual coding was and is a work in progress. Progress was pretty rapid
at the time that DD was introduced and the decade following it.
Dolby AC-3 AKA Dolby Digital was introduced in 1991. It is a proprietary
standard, and has not changed a lot over the years.
MP3 has remained a work in progress since 1989. The rate at which MP3 coders
were improved slowed down quite a bit after ca. 1998, but improvement may
still be possible.
trotsky
You're not making sense. Did Dolby do their homework and do sufficient
blind tests to "prove" that their codec was transparent to people?
Maybe you're a different Arny Krueger and have come to realize that
these blind tests are ineffective.
Scott Dorsey
>Scott Dorsey wrote:
>> William Sommerwerck <grizzle...@comcast.net> wrote:
>>>> 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".
>>
>> How would they not be NTSC?
>
Oh, there are many strong cases for S-Video over composite. But both are
NTSC. The S-Video is also NTSC, it's just not RS-170.
William Sommerwerck
> You're not making sense. Did Dolby do their homework and do
> sufficient blind tests to "prove" that their codec was transparent
> to people? Maybe you're a different Arny Krueger and have come
> to realize that these blind tests are ineffective.
It doesn't matter. Dolby Digital is so bad that you can hear its problems
without comparing it with anything else.
Before Arny objects... I was accustomed to listening to CD-format stereo
from my LaserDisks. I was continually surprised and pleased with the great
transparency, cleanliness, and "ease" of the sound.
The first time I decoded a Dolby Digital signal ("The Incredibles") I could
hear the difference -- flat, grainy, dry, blah sound.
The audibility of lossy codecs varies with the quality of the playback
system. Over my computer speakers (Monsoon planar magnetics), KUOW sounds
fine. Not only is it clean and transparent, but I've never heard anything
that I interpreted as an artifact. (This is the Microsoft codec.)
Arny Krueger
Please clarify, because the questions that follow are not requests for
clarification.
> Did Dolby do their homework and
> do sufficient blind tests to "prove" that their codec was
> transparent to people?
AFAIK, Dolby never claimed that DD was perfectly transparent. The MPEG group
coder tests in the late 1990s showed that Dolby Digital was not sonically
transparent and generally inferior to other, more modern codecs.
> Maybe you're a different Arny Krueger
Nope. Just older and wiser. ;-)
> and have come to realize that these blind tests are ineffective.
How so? The fact that AC-3 was a substandard codec based on the MPEG
Group's blind tests was pretty well publicized by the MPEG and the AES. This
was no doubt a bit of an embarrassment to Dolby. Dolby has been doing their
own blind tests for decades.
Dolby subsequently came out with a new multimodal system for coding and
decoding audio known as Dolby TrueHD. In some modes, TrueHD is definitely
sonically transparent.
trotsky
> "trotsky" <gms...@email.com> wrote in message
> news:MbDxj.54004$yE1.27091@attbi_s21...
>
>> You're not making sense. Did Dolby do their homework and do
>> sufficient blind tests to "prove" that their codec was transparent
>> to people? Maybe you're a different Arny Krueger and have come
>> to realize that these blind tests are ineffective.
>
> It doesn't matter. Dolby Digital is so bad that you can hear its problems
> without comparing it with anything else.
>
> Before Arny objects... I was accustomed to listening to CD-format stereo
> from my LaserDisks. I was continually surprised and pleased with the great
> transparency, cleanliness, and "ease" of the sound.
>
> The first time I decoded a Dolby Digital signal ("The Incredibles") I could
> hear the difference -- flat, grainy, dry, blah sound.
Agreed on all counts.
> The audibility of lossy codecs varies with the quality of the playback
> system. Over my computer speakers (Monsoon planar magnetics), KUOW sounds
> fine. Not only is it clean and transparent, but I've never heard anything
> that I interpreted as an artifact. (This is the Microsoft codec.)
What are you trying to say? Are you saying the lossiness of DD would be
audible over your computer speakers or not?
William Sommerwerck
>> system. Over my computer speakers (Monsoon planar magnetics),
>> KUOW sounds fine. Not only is it clean and transparent, but I've never
>> heard anything that I interpreted as an artifact. (This is the Microsoft
>> codec.)
> What are you trying to say? Are you saying the lossiness of DD would
> be audible over your computer speakers or not?
No, I'm saying that the Monsoons, good as they are, aren't Apogees.
Anim8rFSK
"William Sommerwerck" <grizzle...@comcast.net> wrote:
> The first time I decoded a Dolby Digital signal ("The Incredibles") I could
> hear the difference -- flat, grainy, dry, blah sound.
Ooo, nice description.
--
Star Trek 09:
No Shat, No Show.
http://www.disneysub.com/board/noshat.jpg
Alan
>No, I'm talking objective fact. Color-difference signals require less
>bandwidth than color-primary signals.
I rather doubt that. Even small changes in small areas would produce the
same bandwidth. However, the magnitude of the signal may be less.
The critical thing is that *because of the limitations of human vision*,
one can get away with reducing the bandwidth of the color difference signals.
The original color difference signal may well have full bandwidth --
since the color difference signals vary with hue variation, even when
the saturation remains the same. (They also vary with saturation change
even with hue remaining the same.)
Alan
William Sommerwerck
> In article "William Sommerwerck" <grizzle...@comcast.net> writes:
>> No, I'm talking objective fact. Color-difference signals require less
>> bandwidth than color-primary signals.
> I rather doubt that. Even small changes in small areas would produce the
> same bandwidth. However, the magnitude of the signal may be less.
Correct. But color-difference signals DON'T HAVE THOSE SMALL CHANGES.
> The critical thing is that *because of the limitations of human vision*,
> one can get away with reducing the bandwidth of the color difference
signals.
Not so. Think about it. Saturation (which is what the amplitude of the
color-difference signal represents) "never" (well, hardly ever) changes as
rapidly as luminance.
> The original color difference signal may well have full bandwidth --
> since the color difference signals vary with hue variation, even when
> the saturation remains the same. (They also vary with saturation change
> even with hue remaining the same.)
See above.
This is the typical knee-jerk reaction to something someone hasn't bothered
to think through.
trotsky
And your working definition for "perfectly transparent" is what, exactly?
The MPEG group
> coder tests in the late 1990s showed that Dolby Digital was not sonically
> transparent and generally inferior to other, more modern codecs.
At all bit rates? And again, what is the definition or "sonically
transparent"--when people with Radio Shack stereos can tell the difference?
>> Maybe you're a different Arny Krueger
>
> Nope. Just older and wiser. ;-)
Yeah, I'll buy the older part. If you're saying blind test results have
to be taken with a grain of salt then I'll buy the wiser part.
>> and have come to realize that these blind tests are ineffective.
>
> How so? The fact that AC-3 was a substandard codec based on the MPEG
> Group's blind tests was pretty well publicized by the MPEG and the AES.
Perhaps you can show us a cite for these results, then.
This
> was no doubt a bit of an embarrassment to Dolby. Dolby has been doing their
> own blind tests for decades.
>
> Dolby subsequently came out with a new multimodal system for coding and
> decoding audio known as Dolby TrueHD. In some modes, TrueHD is definitely
> sonically transparent.
Again, you are using a term without defining its meaning.
Arny Krueger
> Arny Krueger wrote:
>> "trotsky" <gms...@email.com> wrote in message
>> news:MbDxj.54004$yE1.27091@attbi_s21
>>> Did Dolby do their homework and
>>> do sufficient blind tests to "prove" that their codec
>>> was transparent to people?
>> AFAIK, Dolby never claimed that DD was perfectly
>> transparent.
> And your working definition for "perfectly transparent"
> is what, exactly?
Passes a bypass test with under any relevant test condition without audible
alternation.
> The MPEG group
>> coder tests in the late 1990s showed that Dolby Digital
>> was not sonically transparent and generally inferior to
>> other, more modern codecs.
> At all bit rates?
As typically used.
> And again, what is the definition or
> "sonically transparent"--when people with Radio Shack
> stereos can tell the difference?
Relevant tests are used with selected, trained listeners. Listener
sensitivity is essential. Please see ITU recommendation BS 1116, which is
availble through the web.
>>> Maybe you're a different Arny Krueger
>>
>> Nope. Just older and wiser. ;-)
>
>
> Yeah, I'll buy the older part. If you're saying blind
> test results have to be taken with a grain of salt then
> I'll buy the wiser part.
Let's put it this way - every test result must be considered in its context.
Sighted tests for signal quality are generally so invalid that they need not
be taken seriously at all. Blind test results are at least worth
considering.
>>> and have come to realize that these blind tests are
>>> ineffective.
>> How so? The fact that AC-3 was a substandard codec
>> based on the MPEG Group's blind tests was pretty well
>> publicized by the MPEG and the AES.
> Perhaps you can show us a cite for these results, then.
Check the AES web site. They were published in the JAES some years back.
> This
>> was no doubt a bit of an embarrassment to Dolby. Dolby
>> has been doing their own blind tests for decades.
>> Dolby subsequently came out with a new multimodal system
>> for coding and decoding audio known as Dolby TrueHD. In
>> some modes, TrueHD is definitely sonically transparent.
> Again, you are using a term without defining its meaning.
Which term? I've used tons of them. I would expect that the readers of the
newsgroups we are posting to know what most common audio terms mean.
Arny Krueger
> "trotsky" <gms...@email.com> wrote in message
> news:p%Txj.55219$yE1.36314@attbi_s21
>> Arny Krueger wrote:
>>> "trotsky" <gms...@email.com> wrote in message
>>> news:MbDxj.54004$yE1.27091@attbi_s21
>
>>>> Did Dolby do their homework and
>>>> do sufficient blind tests to "prove" that their codec
>>>> was transparent to people?
>
>>> AFAIK, Dolby never claimed that DD was perfectly
>>> transparent.
>
>> And your working definition for "perfectly transparent"
>> is what, exactly?
>
> Passes a bypass test with under any relevant test
> condition without audible alternation.
Correction:
Passes a bypass test with under any relevant test
condition without audible alteration.
jwvm
wrote:
<snip>
> Not so. Think about it. Saturation (which is what the amplitude of the
> color-difference signal represents) "never" (well, hardly ever) changes as
> rapidly as luminance.
You need to be careful here. While saturation is a function of color
differences, it needs to be normalized by the intensity. Simple color
differences are functions of both saturation and luminosity. Consider,
for example, calculating saturation in the HSI coordinate system as
illustrated in this link:
http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/OWENS/LECT14/lecture12.html#tth_sEc3
Similar examples for other coordinate systems can be found here:
William Sommerwerck
>> color-difference signal represents) "never" (well, hardly ever) changes
as
>> rapidly as luminance.
> You need to be careful here. While saturation is a function of color
> differences, it needs to be normalized by the intensity.
That is PRECISELY the point. Subtracting Y from R, G, or B provides the
normalization and produces a saturation -- color-difference -- signal from
which brightness information has been removed. This is what we want.
More than 50 years ago, Electronics magazine published pictures of the NTSC
color signals, based on real scenes. The colors are completely "flat" --
they are of constant saturation, with no variation in brightness.
It's important to understand that NTSC and PAL are non-redundant systems.
None of the three signals contains information present in another.
trotsky
> "trotsky" <gms...@email.com> wrote in message
> news:p%Txj.55219$yE1.36314@attbi_s21
>> Arny Krueger wrote:
>>> "trotsky" <gms...@email.com> wrote in message
>>> news:MbDxj.54004$yE1.27091@attbi_s21
>
>>>> Did Dolby do their homework and
>>>> do sufficient blind tests to "prove" that their codec
>>>> was transparent to people?
>
>>> AFAIK, Dolby never claimed that DD was perfectly
>>> transparent.
>
>> And your working definition for "perfectly transparent"
>> is what, exactly?
>
> Passes a bypass test with under any relevant test condition without audible
> alternation.
A) That sentence makes no sense grammatically, and B) you have provided
no definition for "passes a bypass test". You keep making the same
mistakes over and over.
>> The MPEG group
>>> coder tests in the late 1990s showed that Dolby Digital
>>> was not sonically transparent and generally inferior to
>>> other, more modern codecs.
>
>> At all bit rates?
>
> As typically used.
>
>> And again, what is the definition or
>> "sonically transparent"--when people with Radio Shack
>> stereos can tell the difference?
>
> Relevant tests are used with selected, trained listeners. Listener
> sensitivity is essential. Please see ITU recommendation BS 1116, which is
> availble through the web.
This is silly. I'll ask again: can you define what you're talking
about? What's a passing grade for "sonically transparent"--100%? 90%?
If 80% of the trained listeners can't tell the difference, is it then
"sonically transparent"?
Then there's Bill Somerwerck's point: how good is the resolution of the
equipment they're using? If the speakers are mediocre, that will skew
the results.
>>>> Maybe you're a different Arny Krueger
>>> Nope. Just older and wiser. ;-)
>>
>> Yeah, I'll buy the older part. If you're saying blind
>> test results have to be taken with a grain of salt then
>> I'll buy the wiser part.
>
> Let's put it this way - every test result must be considered in its context.
> Sighted tests for signal quality are generally so invalid that they need not
> be taken seriously at all. Blind test results are at least worth
> considering.
Sure, if you believe in Jesus I guess you can believe in the vagaries of
blind testing.
>>>> and have come to realize that these blind tests are
>>>> ineffective.
>
>>> How so? The fact that AC-3 was a substandard codec
>>> based on the MPEG Group's blind tests was pretty well
>>> publicized by the MPEG and the AES.
>
>> Perhaps you can show us a cite for these results, then.
>
> Check the AES web site. They were published in the JAES some years back.
Nice try, Arny. AES charges $5 for members and $20 for non-members for
each paper on the topic. Again, please provide a cite for what you're
talking about.
>> This
>>> was no doubt a bit of an embarrassment to Dolby. Dolby
>>> has been doing their own blind tests for decades.
>
>>> Dolby subsequently came out with a new multimodal system
>>> for coding and decoding audio known as Dolby TrueHD. In
>>> some modes, TrueHD is definitely sonically transparent.
>
>> Again, you are using a term without defining its meaning.
>
> Which term? I've used tons of them. I would expect that the readers of the
> newsgroups we are posting to know what most common audio terms mean.
You have no working definition of "sonically transparent".
Intellectually, you are about as credible as a crack addict. Are you a
crack addict?
trotsky
You're still not speaking English.
Richard Crowley
>>> Not so. Think about it. Saturation (which is what the amplitude of the
>>> color-difference signal represents) "never" (well, hardly ever) changes
> as
>>> rapidly as luminance.
>
>> You need to be careful here. While saturation is a function of color
>> differences, it needs to be normalized by the intensity.
>
> That is PRECISELY the point. Subtracting Y from
> R, G, or B provides the normalization and produces a
> saturation -- color-difference -- signal from which
> brightness information has been removed. This is what
> we want.
The original reason for even matrixing R-G-B into Y-*-*
was to preserve monochrome compatibility for those who
have B&W receivers. And then to produce a color signal
that could be relatively easily compressed and encoded
onto a subcarrier and then decoded at the receiver.
This is not necessarily "what we want" for those who prefer
their video ucompressed and uncompromised. It was a
kludge workaround to fit the 15-pound color signal into the
5-pound monochrome sack (channel bandwidth).
> More than 50 years ago, Electronics magazine published
> pictures of the NTSC color signals, based on real scenes.
> The colors are completely "flat" -- they are of constant
> saturation, with no variation in brightness.
Then you were looking at the color difference signals. (Pb,
Pr, etc.) after the luminance (Y) had been removed. We
will have to disagree whether to call those signals "color".
They are maybe "color difference" at best. If you saw the
original Red Green Blue signals, they have plenty of contrast.
> It's important to understand that NTSC and PAL are
> non-redundant systems. None of the three signals contains
> information present in another.
It is equally important to remember that in NTSC and PAL,
the color-difference part of the signal is artifically frequency-
limited to save bandwidth during transmission/storage. Any
time you execute this kind of lossy compression, you irretrievably
lose information. Whether you are talking about audio or video.