LCD scanning frequency

  • Thread starter Thread starter Guest
  • Start date Start date
| On 5 Oct 2007 19:36:58 GMT, (e-mail address removed) wrote:
|
|
|>| Where is the $200 savings? Practically free integrated
|>| video can already do this, and it is not as though a really
|>| old system otherwise has the processing muscle to handle HD
|>| decompression.
|>
|>I should have said save as much as $200 on computer hardware AND software.
|
| So what costs $200?

Some of that cost is the cost of buying certain commercial software.
Some of that cost is the consequences of managing they license keys of
that commercial software ... something businesses have to do that a
home user never worries about.

Some of that cost is the cost of buying greater hardware capacity than
is otherwise needed (more RAM, faster CPU) because that commercial
software is bloated and inefficient.

So maybe the LCD monitor business is only able to get their little
stickers that say they are certified for that commercial software by
ensuring that the operating ranges are limited to only the upper end
of performance.


|>HD decompression is not involved in an office desktop environment.
|>You're still thinking in terms of a home media center.
|
| What's the point then? If the office isn't doing any HD,
| it'll be fine at 60Hz with integrated video or low end video
| card (unless a special need besides running at 60Hz
| refresh).

Again, you have missed the whole argument. By using a lower fps rate,
such as 24 Hz (I picked that number for use in office purpose because
it is reasonable to expect LCDs for HD TV viewing to also be able to do
that rate because there is such a thing as 720p24 and 1080p24 formats),
the video card can be extended to have higher resolution geometry. HD
TV viewing is _not_ the _only_ reason to have a geometry of 1920x1200
or more. It also increases work productivity.


| However, if the office desktop needed to play back HD
| content on a 60Hz LCD, it can do that as well.

They can do that regardless. It's a non-issue.


|>Actually, 60 Hz degrades 24 fps video.
|
| Far less than that it was 24FPS in the first place instead
| of higher framerate.

Upconverted 60 Hz is not the ideal way to display 24 Hz.
Downconverted to 24 Hz is not the ideal way to display 60 Hz.

So the answer is, make sure all LCD displays can handle everything from
24 Hz (actually 23.976 Hz) to 60 Hz. If they can handle even higher, as
high as 120 Hz, that's even better. They seem to be able to do most of
that now, if not all (certainly as high as 85 Hz, since when I first put
an LCD monitor on my computer, it was running at the 85 Hz rate I had it
set to for the CRT I had just pulled out, and the LCD handled it well,
even though it was at a limited geometry due to needing to run at a high
frame rate to avoid flicker on a CRT).


|>Recording 24 fps in a 24 fps format,
|>or transmitting it as such over ATSC, renders motion more correctly.
|
| More correctly than what? We're talking about a PC, we
| don't have these limitations.

It's there so you can understand that there is such a thing as 24 fps video
that is part of the standards. LCDs that expect to be able to display TV
video for both SD and HD really should be able to display all TV video formats.
Do you know of _any_ LCD display that can? Perhaps it may well be only found
in units marketed as TV sets (with or without tuner). But to the extent that
computer monitors cannot go down to 23.976 Hz vertical frequency, they are not
able to display the full scope of all TV video formats.

|>CRTs, which must upconvert 24 fps, doing so to 48 or 72 Hz instead of 60 Hz
|>would avoid the problem.
|>
|>http://en.wikipedia.org/wiki/Telecine#Telecine_judder
|
|
| The problem is the 24FPS rate, not 2X that rate or higher.
| If you can't perceive the 24FPS lag, it is unreasonable to
| claim a shorter duration lag is then important.

24 fps displayed AT 24 fps in a static display system (like LCD) will NOT have
any judder appearance at all. Judder is NOT the slow frame rate. Yes, there
is an effect that 24 fps can cause to certain video content, but that is not
what the term "judder" refers to. Judder is the other effect caused by doing
any non-integral frame rate conversion, where "3:2 pulldown" that is done with
24 fps to 60 fps is an example of. By displaying one frame 2 times and the
next frame 3 times, it distorts the time smoothness of the motion. It can
have the effect of making things look like they slow down then speed up then
slow down then speed up at rate of 12 times per second. Had the 24 fps been
converted instead to 48 fps or 72 fps, then while you would still see the
steps of motion at 24 times a second, it would be uniform, and not have any
judder.

So my point there is that displaying at 60 fps some content shot at 24 fps is
a _reduction_ of quality relative to displaying it at a whole multiple frame
rate ... or just displaying it as it is which LCD is capable of doing.


|>|>What should be done is to have a new video protocol that transmits blocks
|>|>of video that identify the portion of the screen to update, and all the
|>|>pixel values that go there. Then the video controller card can send the
|>|>image updates however it likes. It could still update everything in a
|>|>legacy raster scan. But the advantage will be realized by being able to
|>|>update only the portions that actually change, when they change. So when
|>|>you open a new window on your dekstop, the act of it opening would send a
|>|>block identifying the reference point of the window, its size, and its
|>|>contents. The video card would detect this based on what was written to
|>|>video memory as it happens.
|>|
|>| Why change what already works?
|>
|>It's a choice between progress or stagnantion.
|
| yes, going to extra effort to reduce LCD monitor refresh
| rates would be stagnation. Progress is moving higher than
| 24FPS.

Stagnantion is sticking to "why change what works". Progress is making things
work better.

Waste is scanning something that doesn't move (much) way more often that it
actually needs.

Progress is being able to handle all video formats (23.976 to 60 Hz).

Opportunity is being able to take advantage of the fact that an LCD monitor
that can do all video formats allows slowing the scan rate down to make more
use of other resources, and saving money as a result.

Do you think that making computer software more bloated is progress?


|>| I encourage you to hook a monitor capable of the resolution
|>| up to any lowest-end modern video card. Try it. Plenty of
|>| people do watch HD this way.
|>
|>Watching HD TV programs is not my interest with regard to the computer.
|>I speak of 24 fps movies because that is an issue that should _also_ be
|>addressed.
|>
|>Let me know of a particular lowest-end modern video card that works on all
|>software, if you know of one. I've not yet found one. The last one I have
|>seen is the Matrox G550 (not low-end, but most certainly works universally).
|
|
| What are you talking about "software"?
| A video card doesn't need to support software, besides
| having an OS driver. You could get lower CPU utilization
| if the video card driver plus some software titles support
| hadware overlays, and MPEG2, 4/2, or 4/10 acceleration, but
| that does not keep it from "working on all software" unless
| you have a very strange definition of working?
|
| What is it you claim makes a G550, "work"?

If I needed features like MPEG conversion, rendering acceleration, etc, then
I'd choose a video card that works that way. I do not need those things.
What I need is a video card that just displays all the computer desktop
content I have, and can do so at the resolutions and modes I need. That
also requires an "open interface" video card so I can use the software that
supports my other needs. Those video cards you are probably referring to
do not have any kind of "open interface" (though at least one such company
is now making some efforts that may result in that in a year or two).

Still, we have LCD displays that cannot display _all_ HD TV formats. Any
that could, in the sizes suitable for desktop use (a 42" HD TV certainly
is too big for _my_ desktop), would work for me, because they would be able
to handle frame rates as low as the 23.976 Hz you can get from some of the
HD formats.

========================================================================
Small matrix of digital video formats recognized by ATSC:

width in pixels -----> 1920 1280 704 704 640
height in lines -----> 1080 720 480 480 480
aspect ratio --------> 16:9 16:9 16:9 4:3 4:3
pixel shape ratio ---> 1:1 1:1 40:33 10:11 1:1

frame field line | | | | |
rate rate scheme | | | | |
------ ------ ------
23.976 23.976 progr YES YES YES YES YES
24.000 24.000 progr YES YES YES YES YES
29.970 29.970 progr YES YES YES YES YES
30.000 30.000 progr YES YES YES YES YES
59.940 59.940 progr no YES YES YES YES
60.000 60.000 progr no YES YES YES YES
29.970 59.940 inter YES no YES YES YES
30.000 60.000 inter YES no YES YES YES
========================================================================

========================================================================
Large matrix of digital video formats recognized by ATSC:

width in pixels -----> 1920 1280 1024 768 704 704 640
height in lines -----> 1080 720 576 576 480 480 480
aspect ratio --------> 16:9 16:9 16:9 4:3 16:9 4:3 4:3
pixel shape ratio ---> 1:1 1:1 1:1 1:1 40:33 10:11 1:1

frame field line | | | | | | |
rate rate scheme | | | | | | |
------ ------ ------
23.976 23.976 progr YES YES no no YES YES YES
24.000 24.000 progr YES YES no no YES YES YES
25.000 25.000 progr no no no no no no no
29.970 29.970 progr YES YES no no YES YES YES
30.000 30.000 progr YES YES no no YES YES YES
47.952 47.952 progr no no no no no no no
48.000 48.000 progr no no no no no no no
50.000 50.000 progr no no no no no no no
59.940 59.940 progr no YES no no YES YES YES
60.000 60.000 progr no YES no no YES YES YES

23.976 47.952 inter no no no no no no no
24.000 48.000 inter no no no no no no no
25.000 50.000 inter no no no no no no no
29.970 59.940 inter YES no no no YES YES YES
30.000 60.000 inter YES no no no YES YES YES

576 line geometry and 25/50 Hz scan rates are included for comparison.
They are not recognized by ATSC. Maybe they should have been since a
lot of content will be available in such formats.
========================================================================
 
| On Fri, 05 Oct 2007 10:22:07 -0400, Wolf Kirchmeir
|
|
|>I think what you are proposing is a smart display, one that performs
|>many of the operations now done in software (ie, CPU) and video hardware
|>(ie, GPU). I think it's a good concept. It would really, really speed up
|>games. Would make humungous screens possible. 16,0000x90,000? Wowser!
|>
|>[...]
|
|
| Processing by CPU and video card is much higher bandwidth
| than outputting the finished frames, for example video cards
| often have roughly 5 to over 30GB/s.

Agreed.

Two options include:

1. Transmit the finished pixels of the processing to the display. If the
processing was done on a subset of the image, then only a small block
needs to be sent. It would know how much that involves.

2. Shift the video architecture model to have the GPU processing done in
the display itself. A lot of things would have to change before this
approach would be practical.

Note that the method described does not rule out always transmitting blocks
that represent the entire image frame, at whatever rate the content is
changing, or at some arbitrary rate even if the content does not change.
But it does give the opportunity to update portions of the display in real
time, even in special cases of reduced bandwidth between source and display,
such as remote KVM access in large data centers, or over the internet to a
colocated server (often done now by things like VNC which the idea is based
on). FYI, this is already done with some KVM switches over IP (internet
technology) to a client application program. Many of these use some other
protocols. But VNC represents an open protocol anyone can use.

It would be easy for LCD technology to do this. It would just require an
extra layer of software or hardware to understand these data blocks and to
update the displayed image in place at the right place. For CRT, though,
it would require adding a layer of frame buffering with RAM having the speed
or dual port capability so it can be updated in place at it is scanned out
to the CRT beam modulation. I don't know enough about other technologies
like Plasma, DLP, or OLED, to tell you how they will be impacted.
 
| kony wrote:
|> On Fri, 05 Oct 2007 10:22:07 -0400, Wolf Kirchmeir
|>
|>
|>> I think what you are proposing is a smart display, one that performs
|>> many of the operations now done in software (ie, CPU) and video hardware
|>> (ie, GPU). I think it's a good concept. It would really, really speed up
|>> games. Would make humungous screens possible. 16,0000x90,000? Wowser!
|>>
|>> [...]
|>
|>
|> Processing by CPU and video card is much higher bandwidth
|> than outputting the finished frames, for example video cards
|> often have roughly 5 to over 30GB/s.
|
|
| I take it you mean that putting a GPU in the monitor doesn't promise
| enough of an advantage to make it worthwhile. I agree that's so for
| current HD standards and monitors. But suppose a 120 x 67.5 inch monitor
| with a .25mm pixel pitch. That would be 12192 x 6858 resolution at 16:9.
| Call it Ultra-HD. It would just fit at one end of my living room....
|
| And I could fit a Cinemascope proportioned screen (18242 x 6858) on the
| long wall of the basement rec room....

We may get to that some day. When we do, it may well make sense to not
just have one GPU in the display itself, but perhaps several just to avoid
having to fan out 125,103,636 pixels just from one, or because the way of
the future is not faster CPUs/GPUs, but multiple cores, multiple units,
and more parallel architectures.

BTW, Where did you get those numbers? We might want to choose numbers that
are multiples of 16 so we have whole 16x16 blocks to work the compression
in without losing a few pixels on the egde. Could you live with 18432 x 6912
(2.667:1 or 8:3) instead?
 
| On 5 Oct 2007 19:36:58 GMT, (e-mail address removed) wrote:
|
|
|>| Where is the $200 savings? Practically free integrated
|>| video can already do this, and it is not as though a really
|>| old system otherwise has the processing muscle to handle HD
|>| decompression.
|>
|>I should have said save as much as $200 on computer hardware AND software.
|
| So what costs $200?

Some of that cost is the cost of buying certain commercial software.

That has nothing to do with anything. If you need certain
commercial software it is a cost not required by a monitor
refresh rate.

Some of that cost is the consequences of managing they license keys of
that commercial software ... something businesses have to do that a
home user never worries about.

Some of that cost is the cost of buying greater hardware capacity than
is otherwise needed (more RAM, faster CPU) because that commercial
software is bloated and inefficient.

You're not making any point relative to playing HD video.
Monitor refresh rate importance was the claim, and these are
not factors, the monitor is only the output device.

I feel this thread would have no productive outcome and so I
will leave it, leaving you more time to develop this
solution you're sure is needed, then we can let the industry
decide. If you're right, it'll sell itself.
 
kony said:
I think what you are proposing is a smart display, one that performs
many of the operations now done in software (ie, CPU) and video hardware
(ie, GPU). I think it's a good concept. It would really, really speed up
games. Would make humungous screens possible. 16,0000x90,000? Wowser!

[...]


Processing by CPU and video card is much higher bandwidth
than outputting the finished frames, for example video cards
often have roughly 5 to over 30GB/s.


I take it you mean that putting a GPU in the monitor doesn't promise
enough of an advantage to make it worthwhile.

I mean that given available cost-effective technology, there
is no video output port than accomplish it (bandwidth) at
even a fraction of your 24FPS target result, without even
more limitations in what could be output than are already
present.
 
| On 6 Oct 2007 15:01:17 GMT, (e-mail address removed) wrote:
|
|>| On 5 Oct 2007 19:36:58 GMT, (e-mail address removed) wrote:
|>|
|>|
|>|>| Where is the $200 savings? Practically free integrated
|>|>| video can already do this, and it is not as though a really
|>|>| old system otherwise has the processing muscle to handle HD
|>|>| decompression.
|>|>
|>|>I should have said save as much as $200 on computer hardware AND software.
|>|
|>| So what costs $200?
|>
|>Some of that cost is the cost of buying certain commercial software.
|
| That has nothing to do with anything. If you need certain
| commercial software it is a cost not required by a monitor
| refresh rate.

Using a video card with less choice in supported software results in
a greater software cost. Using a video card with a wider choice in
supported results in a lower software cost. It could be as much as
$200 and maybe more.


|>Some of that cost is the consequences of managing they license keys of
|>that commercial software ... something businesses have to do that a
|>home user never worries about.
|>
|>Some of that cost is the cost of buying greater hardware capacity than
|>is otherwise needed (more RAM, faster CPU) because that commercial
|>software is bloated and inefficient.
|
| You're not making any point relative to playing HD video.
| Monitor refresh rate importance was the claim, and these are
| not factors, the monitor is only the output device.
|
| I feel this thread would have no productive outcome and so I
| will leave it, leaving you more time to develop this
| solution you're sure is needed, then we can let the industry
| decide. If you're right, it'll sell itself.

The industry does not always know what is going on, especially when it is
filled with people that are closed minded and cannot see this. If you
cannot see it, how can they?
 
| On 6 Oct 2007 15:01:17 GMT, (e-mail address removed) wrote:
|
|>| On 5 Oct 2007 19:36:58 GMT, (e-mail address removed) wrote:
|>|
|>|
|>|>| Where is the $200 savings? Practically free integrated
|>|>| video can already do this, and it is not as though a really
|>|>| old system otherwise has the processing muscle to handle HD
|>|>| decompression.
|>|>
|>|>I should have said save as much as $200 on computer hardware AND software.
|>|
|>| So what costs $200?
|>
|>Some of that cost is the cost of buying certain commercial software.
|
| That has nothing to do with anything. If you need certain
| commercial software it is a cost not required by a monitor
| refresh rate.

Using a video card with less choice in supported software results in
a greater software cost.

A video card is an output device, there is rarely (if ever)
a factor of "supported software". All you need is a driver
for the same platform upon which the software runs.

Using a video card with a wider choice in
supported results in a lower software cost. It could be as much as
$200 and maybe more.

This is arbitrary nonsense. Quit making vague comments and
describe any specific use common enough the entire industry
should degrade monitors to 24Hz in order to suit that end.
 
| On 7 Oct 2007 00:18:51 GMT, (e-mail address removed) wrote:
|
|>| On 6 Oct 2007 15:01:17 GMT, (e-mail address removed) wrote:
|>|
|>|>| On 5 Oct 2007 19:36:58 GMT, (e-mail address removed) wrote:
|>|>|
|>|>|
|>|>|>| Where is the $200 savings? Practically free integrated
|>|>|>| video can already do this, and it is not as though a really
|>|>|>| old system otherwise has the processing muscle to handle HD
|>|>|>| decompression.
|>|>|>
|>|>|>I should have said save as much as $200 on computer hardware AND software.
|>|>|
|>|>| So what costs $200?
|>|>
|>|>Some of that cost is the cost of buying certain commercial software.
|>|
|>| That has nothing to do with anything. If you need certain
|>| commercial software it is a cost not required by a monitor
|>| refresh rate.
|>
|>Using a video card with less choice in supported software results in
|>a greater software cost.
|
| A video card is an output device, there is rarely (if ever)
| a factor of "supported software". All you need is a driver
| for the same platform upon which the software runs.

Pick one of your modern video cards. Now show me drivers that take full
advantage of the video (or at least full advantage of its range of geometry
settings ... at least out to 1920x1200), for all the software system choices
(or at least as many for which the Matrox G550 works with at 24 Hz).

Windows (XP, Vista)
Mac with x86 (OX/X)
Linux (Debian, Fedora, Slackware, SuSE, Ubuntu)
BSD (FreeBSD, NetBSD, OpenBSD)
any others


|>Using a video card with a wider choice in
|>supported results in a lower software cost. It could be as much as
|>$200 and maybe more.
|
| This is arbitrary nonsense. Quit making vague comments and
| describe any specific use common enough the entire industry
| should degrade monitors to 24Hz in order to suit that end.

It is not a "degrade" to be able to support video at 24 fps. Support for
low frame rates can be accomplished without any reduction in support for
higehr frame rates. It does not result in the reduction of quality. With
support for frame rates down to 23.976 Hz/fps such a monitor will then be
able to directly display two kinds of video sources: 24 fps motion picture
films converted directly to 24 fps video, and extended geometry capability
of lower cost office computer systems.

It seems you continue to think that by having the capability to display a
24 fps video source, it will somehow no longer be able to display a video
source at 60 Hz or at some higher frame rate that some computer may be
using. This is entirely untrue and seems to be one of the many factors
that is clouding your ability to understand all of this.

If you think extending the range of video support will disable ot degrade
support for other video formats, then please explain how that can possibly
happen. I think it would not.

And if your meaning was really a poorly worded reference to a claim that
24 fps is just a a degraded format, then explain why it is that most movies
are still shot at 24 fps. I've already explained that in an office desktop
environment where screens will typically have slow changing document content
and financial spreadsheets, where people need to read what is actually on
the screen (so it needs to be still rather than moving around at any frame
rate faster than 1 fps), 24 fps has not effect of degrading at all.
 
Back
Top