Why does one Line require a pair of Lines (black & white) ??

Discussion in 'Amateur Video Production' started by ken, Jul 20, 2003.

1. kenGuest

I see that in television, you have 480 visible lines. This is
frequently given as 240 lines of resolution, stating that for every
line, you must have one black line and one white line. Huh?

I have seen this stated frequently - everywhere it seems.

If my image is all blue, or all white, then the line pairs will be
the corresponding color - of course you would not use one black line
and one white line.

ken, Jul 20, 2003

2. Gareth ChurchGuest

As stated in another thread, I believe the number of 240 comes because of
the quality loss that occurs when the original 485 visible lines are RF
modulated (converted to waves that you can pick up with an antennae).
I'm not sure, so welcome being corrected here, but think that the black and
white lines being referred to are the measurement used to determine the
horizontal resolution of the analogue video signal (as opposed to the
vertical resolution, which both the 485 and 240 numbers are indicating).

The horizontal resolution of analogue video is hard to determine. One common
way is to take the number of alternating black and white vertical lines that
can be displayed before things become unacceptably blurry (when does that
happen? It's a very subjective thing, so the number doesn't mean all that
much).

Gareth

Gareth Church, Jul 20, 2003

3. kenGuest

Well that's an odd one. My guess is that someone was trying to say
that you need two lines from the two fields to represent a single line
of one frame. Then they, for some reason, used an odd example of one
white line and one black line - which just doesn't quite make sense.
Then multiple websites on Television copied that info.

I'll bet that came from Professor Kuhn's article:

http://www.ee.washington.edu/conselec/CE/kuhn/ntsc/95x4.htm

That article includes info that appear on many other websites - even
wrong info. For example, he states that ""the vertical retrace
(flyback) uses lines 248-263 for Field 1, and lines 511-525 for Field
2." Several other websites quote the same, wrong numbers.

This cannot be correct - There are actually 3 fallacies here:

The flyback lines he lists total 31 lines - huh ? He lists lines
248-263 which is a total of 16 lines for Field 1, and lines 511-525 is
a total of 15 lines for Field 2. Each field have different numbers of
flyback lines ?? No, they should be the same.

Also the flyback sequence of lines must end with a half line at the
end so that the next field will trace lines "between" the last field.
Each field has a total of 262.5 lines for a total of 525 lines.
Therefore the number of flyback lines per field MUST end with ".5" -
but his are Integers.

Also, there are missing lines in that line budget. He states that
there are 31 vertical retrace lines. Subtract that from 525 lines and
you are left with 494 visible lines. All sources I have seen say there
are anywhere from 480 to 486 visible lines.

ken, Jul 20, 2003
4. RGBakerGuest

I see that in television, you have 480 visible lines. This is
First, you mix scan lines or vertical resolution with measured or horizontal
resolution -- the 480 value is the fixed vertical resolution of NTSC (the
number of scan lines per frame of picture data) -- the 240 line value is the
horizontal resolution, the only variable resolution in a video signal, and
is measured by counting the number of vertical lines, like fence pickets,
that can be resolved across a width of the screen no wider than the height
of the screen -- 'within a square'. The value for horizontal resolution
runs from a low of 240 (or worse) for VHS up to a high of 540 for the
somewhere in between, though at source the broadcast cameras may be
delivering even better than 540 lines.

You must have two contrasting lines or you can't see them -- if your line is
white, and your background is white ... it is safe to say you wouldn't know
what your are resolving. No point in using colour, as most systems resolve
less colour than b&w (the colour resolution for the consumer digital formats
is 1/4 of the b&w resolution by design) -- hence a white background with
black lines laid on it. If you were measuring resolution in photography,
you would count the number of black lines that were resolved and call this
your resolution -- seems sensible. But if you were doing the same in video,
you would count the number of black lines resolved ... add in the number of
white 'gaps' resolved and call that the resolution. Why? Most likely
because describing resolution as 120 lines for VHS would sound so poor as to
frighten the VP of sales. It doesn't matter much, as long as everyone
agrees to do it the same way, as the resolution values are useful only in
comparison. In practice there are plenty of factors that determine if an
image is pleasing, or accurate ... and the ability to define b&w vertical
lines is not really one of them.
If you could count blue lines on a blue background -- and you can't, can
you? -- the result would be awful as the chroma sampling has quartered the
image before it gets to tape in digital recordings, and analog recordings do
something similar by reducing the percentage of recording data to one third
of the space allocated to luma data.

GB

RGBaker, Jul 20, 2003
5. kenGuest

I am sorry to differ, but the resolution of VHS has typically been
stated at 320x240. Since the aspect ratio of the screen is 4:3 then
it would make sense that the 240 is "vertical resolution", or
horizontal line pairs - not horizontal resolution.

Also, I don't see why you need black lines to act as the "background".
For the example of an all-white image - wouldn't all white lines
suffice to show that ?? If you took your example, anbd had an all
black background with "white gaps" - the picket fence scenario - then
your eyes would see gray, since they can't detect each separate line.

In the past, I have used a magnifying glass, and looked at an
all-white portion of a video (of the arctic for example). The RGB
dots were all be brightly lit, meaning that every "line" is white.
There should be no black lines. It there was a picket fence then
every other "line" would consist of RGB dot triads that are shut off.
Displaying black for half of the lines, would only lessen the white -
i.e. gray.

Alternatively you can create an image in Photoshop with alternating
black and white lines, and it will be gray. I made one ar 320x240 and
placed it on my website: www.infocellar.com/temp1.jpg - it does
indeed look gray.

I am still open to any reasoning that explains this, and am open to
being called completely wrong - especially since I have an unfinished
Television web page that has several mysteries I am trying to solve.

ken, Jul 20, 2003
6. Mike KujbidaGuest

Ken;
With all due respect, you're getting in way over your head here and making
statements without the technical knowledge (eg. - the resolution of VHS has
typically been stated at 320x240) to back them up.
Here are a number of web sites (from fairly straight forward to very
technical) for you to go through in the hopes that at least some of it will

Mike

http://www.bealecorner.com/trv900/respat/

http://www.evdemo.com/page.asp?id=15

http://videoexpert.home.att.net/artic1/201res.htm

http://www.videotek.com/notes.html

http://www.uwasa.fi/~f76998/video/

http://toolbox.sgi.com/TasteOfDT/documents/video/lurker/

http://www.snellwilcox.com/reference.html

Mike Kujbida, Jul 20, 2003
7. David McCallGuest

You aren't paying attention. The answers you have been given
are mostly right on target. That 320 x 240 measurement has
nothing to do with reality, and everything to deal with marketing
new formats. I think, although I wouldn't take it to the bank, that
VHS has the same number of active video scan-lines as
broadcast video, which is technically 486 in NTSC. The 480
number didn't appear until digital came along. Full screen video
on a computer was 640 X 480 for quite a while. It looked a little
soft compared to analog video, so when profesional digital video
came along the upped the horizontal resolution to 720, and put
back in the 6 lines the computer people had chucked to yield
a full 720 X 486. Most broadcast level equipment uses this value
today, but there was an interim format made to compete with
1" video tape, called D2 that had a higher horizontal resolution,
but I think they dropped those 6 lines there too (752 x 480).
When digital formats were designed for consumer video
(DV and DVD), they decided on 720 x 480. When broadcasters
saw how good the DV format looked, they said "make us some
of those too", so along came the slightly more robust DVcam,
and DVCpro. Now you could buy a camera for the same sort of
price as profesional S-VHS or Hi-8 cameras (~\$7,000 to \$15,000).
Then they realized that most people just don't care enough about
video quality, and that the high end consumer DV cameras were
actually good enough for many uses. They used to spend \$80,000
on Betacam SP cameras, and now they realized they could
buy 10 of those \$4,000 cameras and still have \$40,000 left over to
show as profit. They could just hand these things to reporters and
let them loose. If they broke it, it wasn't such a big deal.

Whoa, was that a tangent or what? Sorry. At any rate, all of those
resolution numbers are finite. 480 (or so horizontal lines, then
each line is digitally devided into 720 pixels. This is the type of
resolution you keep wanting to talk about, but these numbers
are all fixed, and not aty all subjective.

Long before digital they were measuring resolution by counting
lines printed on a standards based chart. You already got a good
explanation of this. Re-read it. It is correct information. This
number, while not precise is a measure of the amount of detail
a system can reproduce. It is not directly about pixels, or scan lines.
It includes thae capabilities of the lens, the pickup, and the encoder,
and even the recording quality. The point being, that you could put a
truly crappy lens on a camera that had 720 x 480 resolution, and
wind up with a system that could only resolve 150 lines. I know, it's
very confusing because there are scan lines, and lines of resolution.
They don't relate to each other directly.
You are still confused. Those aren't the lines we are talking about when
we talk about lines of resolution. You are looking for scan lines. if every
other line was black on a white field, the interlace flicker would be awful.
See, now you are on a tangent. those are dots of phosphor that
glow when an electron beam hits them. They aren't lit all of the
time on a tube TV (LCDs work a little differently) they are only lit
when the beam hits them. Your eye/brain puts the imagfe together.
Those dots only relate to the resolution of the tube itself. Those
colored dots are not a part of the video, and have little to do with
OK, but it is a question of the resolving power of your eyes, that
makes it look grey. The lines are still there and can be resolved
if you look close enough. It is that point where the lines from a
resolutuion chart become grey that define the lines of resolution.

Try this. Draw a set of 10 vertical lines, on a white field, in Photoshop.
Make the lines and spaces about 1/4" each. The lines should look
pretty clear. Now go to transform/distort. squeeze the bottom of your
set of lines so that it becomes a point. So, now you have a triangle
that has 10 vertical lines that converge to a down to a point. At some
point you will see that the lines are no longer distinguishable, no matter
how close you look at the picture. That is the limit of your lines of
resolution, for this chart. Of course this was just a little demonstration,
and your "resolution chart is not calibrated, but it should give you an idea.
Then reread the other post and see if you understand it now.
I hope somebody will be able to help you see the difference between
scan lines and lines of resolution. Good luck in your search for the truth.

David

David McCall, Jul 20, 2003
8. RGBakerGuest

Please leave it 'unfinished' 'til you've mastered the subject.

http://www.digitaltelevision.com/
has some good primers, and

is worth a close look.

http://videoexpert.home.att.net/
has a collection of pages, some of which are worth reading -- the Utz page

Frankly, you are way off base right now, and though you describe yourself as
open to any reasoning I can't imagine how you can visualize white lines on a
white background ... nor the believe that 320x240 represents anything other
than a minimum acceptable digital capture of a VHS source -- it surely has
nothing to do with a television format.

Cheers,
GB

RGBaker, Jul 20, 2003
9. kenGuest

Oh I will - it will take a while before I get the info and
double/triple-check it all. I do appreciate your explanation of the
B/W picket fence. I think if someone made a web page showing how some
example of how it works when an all-white or all-blue, or even an
all-black image would help. Then one could take that info and also
understand how these black and white lines can duplicate flowing
images of color.

ken, Jul 21, 2003
10. RGBakerGuest

http://www.evdemo.com/page.asp?id=15
No, he says HORIZONTAL resolution.
No. In fact he is so clear in this (though I'd stick to reading the first
As it should. Horizontal resolution varies in analog devices, and though
720 samples has become a nom in digital devices there is no reason why this
has to be the value ... even some digital devices use variations on the 720
standard, leaving black bars on the sides of the image.

GB

RGBaker, Jul 21, 2003
11. RGBakerGuest

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

There are examples of standard resolution charts here. I don't follow your
question about flowing images of colour ... except to say that luma
resolution charts don't measure chroma resolution, chroma is typically only
tested for purity.

GB

RGBaker, Jul 21, 2003
12. MattieGuest

Then the sites are mistaken... I believe you're confusing vertical
resolution (which is fixed in NTSC video) with horizontal resolution
(which can vary). Bear in mind that vertical resolution is determined
by horizontal scan lines and horizontal resolution is measured by
vertical lines. This might be where the confusion lies. I think what
you might have seen stated in other articles is something to the
effect of, "VHS can produce 240 vertical lines of resolution." This is
stating a measure of horizontal resolution, not vertical resolution.
And, while VHS, Broadcast and DVD video all have the same vertical
resolution, they vary in horizontal resolution. One more site that
might help is:

http://www.hometheaterhifi.com/volume_6_3/essay-video-resolution-july-99.html

Mattie

Mattie, Jul 21, 2003
13. Mike KujbidaGuest

To rephrase, vertical resolution is determined by the number of horizontal
lines from the top of the screen to the bottom which, in NTSC, is 525 lines.
I'm ignoring the lines lost to the verticla blanking portion of the signal
which, as Ken has discovered, can vary somewhat.
In the analog days, this was fixed. In the digital world, active video is
anywhere from 480 to 486 lines.
Horizontal resolution resolution can only be determined by using test charts
and measuring equipment similar to what Peter Utz describes in his article
at http://www.evdemo.com/page.asp?id=15

And that would be because the author is (unintentionally) misleading people.
For example, http://www.evdemo.com/page.asp?id=15 states "If you ever need
to make a quick mental calculation, simply take the number of pixels, divide
by 1000, lower the number just a tad, and call it lines of horizontal
resolution. Thus a chip with 300,000 pixels would give a horizontal
resolution of about 290 lines". The key words here are "horizontal
resolution".

http://www.hometheaterhifi.com/volume_6_3/essay-video-resolution-july-99.html

Good article. The first graphic (IMO) does a nice job of explaining vertical
resolution.

Mike

Mike Kujbida, Jul 21, 2003
14. -Guest

But he also says vertical resolution further down the article. It's
the kell factor. Although there 483 (or 486 or whatever) horizontal
scan lines, their ability to capture detail is reduced by approx 0.7

See http://members.aol.com/ajaynejr/kell.htm

-, Jul 21, 2003
15. MattieGuest

Mike... This last article is terrible. Utz is confusing his terms
throughout....

"Although vertical lines of resolution always stay the same because of
the way the pictures are made, horizontal lines of resolution vary
with the quality of the equipment."
AND
"Horizontal lines of resolution, measure how sharp a picture is from
left to right." - Peter Utz

The above sentences are simply incorrect. I wouldn't recommend using

Mattie

Mattie, Jul 22, 2003
16. MattieGuest

several factual errors in both versions (although Utz does correct
himself on the laughable math in the first version, "The real
vertical resolution available from 483 scan lines is .7 x 483 = 240
lines of resolution.")

No wonder Ken is confused.

An article that is certainly more accurate and easier for a novice to
understand is:

http://www.hometheaterhifi.com/volume_6_3/essay-video-resolution-july-99.html

Mattie

Mattie, Jul 22, 2003
17. RGBakerGuest

RGBaker, Jul 22, 2003
18. Mike KujbidaGuest

Mattie, I have no problem with Utz's article as, IMO, it is correct.
Somewhat oversimplified at time, but accurate nonetheless.

In the first statement you quote, the quality of the equipment has a
definite impact on horizontal resolution. There is no way a single chip
3-chip camera.

As far as the second statement is concerned, you have me puzzled. What
about it is incorrect? Think horizon when you're talking horizontal
resolution. The horizon runs side to side (or left to right, if you prefer)
and that's exactly what he's saying.

Mike

Mike Kujbida, Jul 22, 2003
19. Mike KujbidaGuest

I agree that Utz's choice of words could've been better. I remember
studying this as part of basic TV electronics in college (way too many)
years ago. It was just as frustrating for most of us back then as it is for
Ken today.

My other problem with Utz is when he first tries to explain the Kell factor.
He states "The real vertical resolution available from 483 scan lines is .7
x 483 = 240 lines of resolution." Huh? I do the math and it comes out to
338 lines. He does to on to correct his math later in the article but these
are the kinds of problems I'm concerned about.

My concern is that Ken says he's going to eventually post this on his web
site. I'd hate for the information to be incorrect as so many other web
sites are.

Mike

Mike Kujbida, Jul 22, 2003
20. RGBakerGuest

My other problem with Utz is when he first tries to explain the Kell
factor.
The worst thing Utz says IMHO is this:

"It would seem that 483 scan lines would give you a vertical resolution of
483 vertical lines of resolution (483 rungs on the ladder). This is not the
case. If one scan line saw one rung, the next scan line would need to see
the space between the rungs, and the following line would see the next rung
in order for the rungs on the ladder not to merge together. Put another way,
if each scan line saw a rung, then the ladder would look like it was made of
solid rungs with no spaces. Thus, an image that goes "rung-space-rung-space"
is defined as 4 lines of vertical resolution and it took four scan lines to
do it. Thus, 483 scan lines can show 120 actual rungs on the ladder, but
still the vertical resolution is called 483 lines."

It starts off as perhaps the best simple explanation of the concept of why
it takes a line pair to resolve a line ... and then blows it completely by
concluding that the maximum number of 'rungs' countable in a 483 line
example is 120! No -- it is 240 ... which is rather the point of the rest
of his explanation. But in fairness to Utz I've seen versions of this
article posted where that oddity was corrected -- maybe his .7 error is
corrected someplace too.

It starts with distinguishing vertical from horizontal, a trick a remarkable
number of posters seem to miss. And it doesn't hurt to see that black text
on a black background wouldn't have much legibility ...

Cheers,
GB

RGBaker, Jul 22, 2003