Searching \ for '[OT] Monitor Question' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/index.htm?key=monitor+question
Search entire site for: 'Monitor Question'.

Exact match. Not showing close matches.
PICList Thread
'[OT] Monitor Question'
1999\04\07@235916 by Sean Breheny

face picon face
Hi all,

       Recently, a friend of mine asked me exactly how a PC monitor matches up
logical pixels to physical pixels in the raster on the screen. Of course, I
know that the electron beams from the R,G,and B guns are scanned
horizontally and vertically. However, it would seem to me that the monitor
would have to ensure that the start of the signals on each scan line
actually began on a real pixel,otherwise various odd patterns (akin to
aliasing) would occur if the data was out of sync by even a fraction of a
physical pixel. Viewing a monitor using a magnifier,we were also able to
see that as we turned the "horizontal position" knob, the horizontal
position actually seemed to only take on certain discrete positions,such
that pixels in the image always lined up exactly with physical pixels in
the raster.
       While I understand how horizontal and vertical sync signals can ensure
that each scan line and frame begin at the right time in the signal stream
coming from the video card, I don't understand how the monitor electronics
ensures that it lines up with physical pixels. This would seem to be a
great engineering problem,because a slight drift or jitter in the
positioning could have grave effects on the image.

Can someone please explain this to me?

Thanks,

Sean

|
| Sean Breheny
| Amateur Radio Callsign: KA3YXM
| Electrical Engineering Student
\--------------=----------------
Save lives, please look at http://www.all.org
Personal page: http://www.people.cornell.edu/pages/shb7
spam_OUTshb7TakeThisOuTspamcornell.edu ICQ #: 3329174

1999\04\08@004303 by Wagner Lipnharski

picon face
There is a metallic mask in front of the screen (before the screen), so
the red electron cannon only "sees" the red phosphor dots, the green
only see green dots and the same to the blue.  It is arranged in some
manner that just one screen is enough to the three cannons. The cannons
are assembled in a triangular fashion, so the electron beans from one
cannon crosses the same hole in the mask, in different angle, hitting
its right color right after the mask.

It goes crazy when you magnetize this screen, or if the screen moves...
then the mess is complete.

There is nothing to do with timing.

Wagner.


Sean Breheny wrote:
{Quote hidden}

1999\04\08@005718 by Sean Breheny

face picon face
Hi Wagner,

Thanks for responding.

Yes, I understand about the color determination,but that isn't a complete
answer to my question. What I mean is this: imagine that you have a grid of
holes (i.e., the holes in the metallic mask) and you have a electron gun
which is meant to scan over the screen and shoot through the holes, thereby
causing the phosphor to iluminate at those points. Now, imagine that the
gun was delayed in its sweep slightly so that the signal meant to govern
the brightness of pixel #1 was actually sent while the gun was aimed half
way between pixel #1 and #2. Now, instead of having the signal level meant
for #1 affect only #1, you affect #1 less than normal,but also affect #2.
The result would be (it seems to me),that you would get a dimmer image, and
maybe even some type of repeating light and dark bands across the screen
(like what happens when you take two window screens and don't line them up
exactly), because the beam would be "masked" by the metal screen at
different points than it was supposed to me.

I must be thinking of this wrong, but this is the way it seems to me.

Sean

At 12:36 AM 4/8/99 -0400, you wrote:
{Quote hidden}

| Sean Breheny
| Amateur Radio Callsign: KA3YXM
| Electrical Engineering Student
\--------------=----------------
Save lives, please look at http://www.all.org
Personal page: http://www.people.cornell.edu/pages/shb7
shb7spamKILLspamcornell.edu ICQ #: 3329174

1999\04\08@005726 by Wagner Lipnharski

picon face
You can do this experiment:

Tape a blank paper page on the table, take two books, aprox 1 inch tick,
and put one at the left, another at the right side of the paper,
covering aprox 1 inch of the paper margins. Now tape another blank page
over the books. Both pages would be 1 inch apart.

Make a pencil hole in the middle of the top paper. Without moving your
head, paint a red spot in the bottom paper where you can see with your
right eye through the hole, paint blue where you can see with your left
eye.  Make another hole, repeat the spots... with some luck your left
eye will never see a red spot.

Doesn't matter how fast your open or close your eyes...

The monitor has 3 eyes (cannons) and millions of holes in the mask.

Wagner.

1999\04\08@011320 by Sean Breheny

face picon face
Hi again Wagner,

Thanks again for responding.

My question really has nothing to do with color,it would apply even to a
monochrome monitor.

My point is this: as the beam scans across the matrix of physical pixels
(holes in the mask),its intensity is varying due to the analog signal
coming from the video card. At a certain time,the signal is meant to
correspond to a logical pixel in video memory (call it pixel #1) and then
one pixel clock later,it corresponds to the next pixel in video memory
(pixel #2). Where these pixels occur on the matrix is irrelevant,BUT,we
expect that the electron beam will be aimed STRAIGHT THRU one of the holes
when the signal is present for pixel #1 and STRAIGHT THRU the next hole (or
some INTEGER number of holes later) when the signal is present for pixel #2.

If the beam is delayed by enough time so that it is aimed at the mask
inbetween pixels when the signal is present for some pixel (instead of
STRAIGHT THRU a hole), then the intensity of that pixel will be reduced
(partially masked by the mask).

Thinking about it some more, I guess that this problem could be minimized
by making the space between pixels very small, so that it would be
impossible for the beam to spend a majority of its time "inbetween" pixels.

Sean

At 12:53 AM 4/8/99 -0400, you wrote:
{Quote hidden}

| Sean Breheny
| Amateur Radio Callsign: KA3YXM
| Electrical Engineering Student
\--------------=----------------
Save lives, please look at http://www.all.org
Personal page: http://www.people.cornell.edu/pages/shb7
.....shb7KILLspamspam.....cornell.edu ICQ #: 3329174

1999\04\08@020723 by Eric Oliver

flavicon
face
Wagner,

I think I understand what Sean is getting at.

Let's say the mask hole ( pixel #1 ) at the top left of the monitor _image_
should be R255 G175 B55 and the pixel directly to the right should be R157
G65 B123. Now the cannons ( as you called them ) must shoot through the
mask hole for pixel #1 at the correct intensity (?) and then change
intensities for pixel #2.  Now all modern monitors have he ability to move
the image left or right and make it wider or thinner.  It would seem that
if the cannons were off just a slight amount they would not hit pixel #1
directly thus changing the color of pixel #1. If the positioning were off
enough the beams meant for pixel #1 could have an effect on pixel #2.


I too am interested in this as it seems a marvel of engineering to make
something that needs such exact positioning so _cheap_.

Eric

{Original Message removed}

1999\04\08@021229 by Sean Breheny

face picon face
Yes, Eric, that is exactly what I meant. I am just trying to (temporarily)
keep color out of the question because the same effect should (AFAIK) occur
in monochrome monitors and it is easier to think about.

Sean

At 12:53 AM 4/8/99 -0500, you wrote:
{Quote hidden}

>{Original Message removed}

1999\04\08@023059 by Dave VanHorn

flavicon
face
>My question really has nothing to do with color,it would apply even
to a
>monochrome monitor.


Mono monitors don't have pixels or shadowmasks.

1999\04\08@023538 by Dave VanHorn

flavicon
face
>Let's say the mask hole ( pixel #1 ) at the top left of the monitor
_image_
>should be R255 G175 B55 and the pixel directly to the right should be
R157
>G65 B123. Now the cannons ( as you called them ) must shoot through
the
>mask hole for pixel #1 at the correct intensity (?) and then change
>intensities for pixel #2.  Now all modern monitors have he ability to
move
>the image left or right and make it wider or thinner.  It would seem
that
>if the cannons were off just a slight amount they would not hit pixel
#1
>directly thus changing the color of pixel #1. If the positioning were
off
>enough the beams meant for pixel #1 could have an effect on pixel #2.


This happens, it's an adjustment called Purity.
But, once this is done, a red gun can only hit a red dot, and so on.

1999\04\08@024849 by Brian Duran

flavicon
face
hi guys

Here is my attemped to explain your problem, and some of what I say might
be repeating items already noted, but it 2:30 in the morning.

       Yor have a horizontal sync, which is extreamly stable.  It is
controlled by thermisters(temparture variabled resistors), which allow the
circuit to only have, lets say a 15% drift.

       There is a mask, which through each hole a trio (three pixels,
red, green, blue) are let through.  In recent years these masked are
actually only wire grids, so that the holes are relatively large compared
to the metal grid cross sections(since anything that it this cross section
only heat up the grid).

       The hope here is that the grid sections are wide enough that it
would cover the 15% drift, yet still extreamly small.  Thus you would
always get your the pulse(three guns shooting at a trio) hitting the right
trio,a dn the only thing that is left to worry is if the whole beam is
going to hit, but you know that atleast 85% of it will.  So this is only a
brightness question, you just increase the intesity to make it brighter.

       After the monitor has been assemebled, it will pass through a
bunch of test.  Only is purity, which helps make sure that you red beams
are only hitting red pixels, and so on.  The other is brightness/whiteness
balance, which I don't completely understand since a computer does it, but
you read the center and four corners for brightness and whiteness, to make
sure that the beams in the center are hitting the same intensity as the
one towards the corner, and this varies because of exactly what you were
mentioning.

In conclusion, YES it is a problem, and there is not nifty trick that is
used to fix it(I think), it is just that the circuit is really accurate.
You will also get tubes, which can not be adjusted, because when you get
one beam to hit the pixel, the other is hitting too much of the grid, and
then is a manufacturing problem in the tube construction.

I hope that this helps, and is atleast a little clear.

brian

1999\04\08@031133 by Clyde Smith-Stubbs

flavicon
face
On Wed, Apr 07, 1999 at 11:57:50PM -0400, Sean Breheny wrote:
>         Recently, a friend of mine asked me exactly how a PC monitor matches u
p
> logical pixels to physical pixels in the raster on the screen. Of course, I
...
> physical pixel. Viewing a monitor using a magnifier,we were also able to
> see that as we turned the "horizontal position" knob, the horizontal
> position actually seemed to only take on certain discrete positions,such
> that pixels in the image always lined up exactly with physical pixels in

I dunno what you thought you were seeing there, but there is no attempt
to line up pixels with dots. Apart from anything else, the video card has
no idea how big the dots are, or even if in fact they are dots and not
stripes (as used in Trinitron and some other picture tubes).

Depending on the size of the pixels, the bandwidth of the video circuitry,
the scan rates used and the size of the dots, a single pixel may correspond
to one dot, part of a dot, or more than one dot. If you look at the screen
of a hi-res (i.e. wide video bandwidth and sharp focus) monitor it is possible
to see that a single dot may not be fully illuminated - one side of the dot
can be darker than the other. If the dots are actually stripes, it is
normal for one scan line to illuminate only part of a stripe (vertically).

As long as the dots are too small for the eye to distinguish, the whole
screen just looks like a smooth field that the pixels are painted on. Having
said that, it is quite possible to get interference (Moire) patterns with
the right combination of video resolution and monitor. This is particularly
apparent with e.g. 1152x864 on a 17" monitor, and a dithered pattern (like
an X-windows background).

Incidentally, this whole discussion is irrelevant to monochrome monitors, which
have no shadowmask and therefore no dots. It's also irrelevant to LCD displays,
where pixels must match dots (hence severely restricting the choices of resoluti
on).

--
Clyde Smith-Stubbs               |            HI-TECH Software
Email: EraseMEclydespam_OUTspamTakeThisOuThtsoft.com          |          Phone            Fax
WWW:   http://www.htsoft.com/    | USA: (408) 490 2885  (408) 490 2885
PGP:   finger clydespamspam_OUThtsoft.com   | AUS: +61 7 3355 8333 +61 7 3355 8334
---------------------------------------------------------------------------
HI-TECH C: compiling the real world.

1999\04\08@085553 by Mike Keitz

picon face
On Thu, 8 Apr 1999 01:11:58 -0400 Sean Breheny <@spam@shb7KILLspamspamCORNELL.EDU> writes:

>Thinking about it some more, I guess that this problem could be
>minimized
>by making the space between pixels very small, so that it would be
>impossible for the beam to spend a majority of its time "inbetween"
>pixels.

There is no attempt made to synchronize in the electronics.  A clear
high-resolution picture is based on the phosphor dots (what you call
"physical pixels") on the screen being very small.  Better monitors have
smaller dots.  The electron beam, even at it's best focus, is large
enough to hit more than one dot (of the same color) at a time.  Also
since faster amplifiers cost more, the amplifiers in a cheap monitor with
large dots on the screen will be slow, probably slow enough to smear each
pixel from the card over more than one dot.

Monochrome monitors don't have this limitation at all since the phosphor
is a uniform continuous coating.



___________________________________________________________________
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com/getjuno.html
or call Juno at (800) 654-JUNO [654-5866]

1999\04\08@103321 by Wagner Lipnharski

picon face
.... [snip]

Sean, this is why bigger TV screens and monitors "as usual" give you
better image resolution, not because it has more dots per horizontal
line of image, but because it has more physical screen color dots per
image color dot.

For example, a 17 inches .26 dot pitch monitor has aprox 13 inches of
horizontal image, what means 330 mm times 0.26mm/dot = 1269 horizontal
dots.  A 14 inches .28 monitor has only 996 horizontal dots.  Which one
do you think would show better a 1024 x 768 image?

There is something nice about it, our eye & brain recompose the image
that "we are supposing to see", and we actually "see" it, even if it is
somehow failing at the screen. This is also helped by the fact that the
monitor is not perfect, and sometimes the horizontal instability (that
we can't see easily), end up covering the flaws on the image.  It means,
that pixel not covered in the previous image field, probably can be
shows in the next, because the image is not so steady.

Also, the mask creates a shadow for the red cannon right over the blue
and green phosphor dots, but the phosphor does not offer so much
definition, once it is hit by the electron bean of only 0.10mm diameter,
it would generate a 0.20 or more round color light dot. It means the
spot grows by itself. The same effect of diffusion can be seen by the
sunlight through a small hole, it is never perfect, it is dimmed around,
even with its parallel rays.

Now, a common TV image has a horizontal rate of 15725 lines per second,
with a bandwith of aprox 4MHz, it gives you only steady horizontal
pixels.  So, a TV screen doesn't need to have a dot pitch so small,
since the image quality doesn't require it.  In a way, you can think
that your computer monitor has an image much better than your family
room big TV.

1999\04\08@103929 by Sean Breheny

face picon face
Thanks to all who responded.

I think I finally understand what is going on,due to your explanations.

Yeah, it was dumb of me to think that a monochrome monitor had separate
phosphor dots,after all,an oscilloscope is essentially a type of "mono
monitor" and I always knew that they don't have dots or a mask.

Essentially, my mistake was to assume that moire patterns or other
problems would result under most resolutions, due to the guns hitting
partial pixels, combinations of pixels at once, etc. One of the posters
said that such problems DO occur,but only at higher resolutions. This
makes sense,and it is probably because the pixel (phosphor dot) dimensions
were smaller than I thought they were, and have even less space between them.

Thanks again,

Sean

1999\04\08@130215 by Richard Martin

picon face
Electrons in a color tv tube arrive at the face at a 'point' (set
by the main x-y deflection) AND (for sake of argument )are
grossly at an angle 'normal' to the tube face (the whole magic of
short tubes and flat faces are 'solutions' to accomplishing this.)
IN ADDITION 'at the face' they also have a local <short> angular
deflection which encodes the color. Think of it as 'red' electrons
arrive at the face ALL aimed to the upper left. An analogy
is 'rain' which may arrive in my town (a pixel) all slanted
to the east (local deflection) while in your nearby town, the
rain (if raining) slants to the west. Depending on how we
carry an unbrella we are wet or dry. In the TV case the pixel
(bright spot) is much larger than the mask hole, due to
'splatter'  at the phosphors in addition to focus so the tube
face is more evenly illuminated than the granular mask would
suggest.

Wagner's example (the books} doesn't work since it fails
to distinguish the 'gross' (head position :: x-y deflection ::
whose town) from the 'local' (angle AT mask :: rain slant)
geometries. In the limit I can see any spot with any eye
if I move my head and look 'sideways', the example works
only if I fix my gaze at a particular ANGLE and 'scan'
by TRANSLATING my head. I.e. gross angle always = 'normal'
with local angle due to eye separation.

Hope this helps, it's not really obvious. The technology of making
masks and phosphor patterns is fairly exotic and proprietary
and its' language 'Trinitron' etc. creeps into marketing buzz.

R.M.M.

R.M.M.

1999\04\08@132503 by Richard Martin

picon face
I hope my 'explanation' made sense. The whole point is that
the 'gross' issues (controlled by timing, synchs, overall
deflection circuits) can't practically be controlled well enough
to make a (mask) color TV work. By making color entirely (well,
mostly) dependent on what appears to be a local (at the
mask/phosphor plane) off-axis deflection , which conveniently
is controlled by another 'gross' parameter (R,G B gun positions)
the whole thing becomes manufacturable/settable.
Purity controls have to do with making small adjustments
to the (e.g local 'red' deflection angle).
If you undersatnd all this you will see why the transition from
tubes that had the face as roughly the surface of a sphere centered
at the 'guns', to modern short/flat tubes took several decades.

Monochrome tubes in general don't have masks.

R.M.M.

1999\04\08@133601 by Wagner Lipnharski

picon face
By the same way that if you move the color cannon it makes a mess on the
screen, and this is why I said "without moving your head...", don't
worry about ball eye movement when moving the eye sight from the first
to the second hole, since it would be a repeated action when painting
the spot or when just looking at it later... :) Did you already tried
it?

There is an very nice experiment that you make several different
drawings using a small 10x10 holes mask, and you can see a different
image in each eye, removing the mask you almost can't identify it, since
they are painted with the same color pen.  It was in an explanation
about stereoscopy using perfurated masks instead of polarized light or
separated images.


Richard Martin wrote:
> Wagner's example (the books} doesn't work since it fails
> to distinguish the 'gross' (head position :: x-y deflection ::
> whose town) from the 'local' (angle AT mask :: rain slant)
> geometries. In the limit I can see any spot with any eye
> if I move my head and look 'sideways', the example works
> only if I fix my gaze at a particular ANGLE and 'scan'
> by TRANSLATING my head. I.e. gross angle always = 'normal'
> with local angle due to eye separation.

1999\04\08@134535 by Nigel Goodwin

flavicon
picon face
In message <KILLspam3.0.3.32.19990408011158.00e33030KILLspamspampostoffice2.mail.cornell.ed> u>, Sean Breheny <RemoveMEshb7TakeThisOuTspamCORNELL.EDU> writes
>Hi again Wagner,
>
>Thanks again for responding.
>
>My question really has nothing to do with color,it would apply even to a
>monochrome monitor.

It wouldn't apply to monochrome, only colour tubes have shadow masks.

{Quote hidden}

The method used is very simple, make sure that the shadowmask holes, and
phosphor dots on the screen, greatly outnumber the require resolution,
then none of your problems occur. This is why monitor tubes are more
expensive than TV tubes, they have much greater resolution.

Anyway, it all works, why wory about it :-).
--

Nigel.

       /--------------------------------------------------------------\
       | Nigel Goodwin   | Internet : spamBeGonenigelgspamBeGonespamlpilsley.demon.co.uk     |
       | Lower Pilsley   | Web Page : http://www.lpilsley.demon.co.uk |
       | Chesterfield    | Official site for Shin Ki Ju Jitsu         |
       | England         |                                            |
       \--------------------------------------------------------------/

1999\04\08@183201 by paulb

flavicon
face
Sean Breheny wrote:

> Essentially, my mistake was to assume that moire patterns or other
> problems would result under most resolutions, due to the guns hitting
> partial pixels, combinations of pixels at once, etc.

 I think that's the point.  The *critical* requirement for a shadow-
mask is that the luminosity is the same whatever the position of the
electron beam.  How this has been achieved over the development of CRT
displays represents a technology in itself, closely linked with the goal
of shadow-mask efficiency.

 The trick is to have the shadow-grid *not* intercept the electron beam
to any significant extent, but by its charge, to split and focus the
beam into slivers to match the dots.  If this is achieved, then a half-
way position of the beam will be split into adjacent pixels, still
totalling almost the same intensity as it it neatly struck either one.

 Also, MoirŽ patterns are by-and-large, more discernible in chroma
information.

>  One of the posters said that such problems DO occur, but only at
> higher resolutions.  This makes sense, and it is probably because the
> pixel (phosphor dot) dimensions were smaller than I thought they were,
> and have even less space between them.

 I think that the only critical adjustment is likely to be the dot
focus.  When this exceeds the pixel pitch by a certain value, the loss
from the shadow-mask becomes independent of the precise position.
--
 Cheers,
       Paul B.

More... (looser matching)
- Last day of these posts
- In 1999 , 2000 only
- Today
- New search...