pc <-> display bandwidth

"alexandre" <alexc RemoveThis @iprolink.ch> wrote in message
news:441b02fa.0306290639.6af41f89@posting.google.com...
 > Hello,
 >
 > just out of curiosity, i was wondering about the needed bandwidth
 > between pc and display. If you have a resolution of 1600*1200 and
 > color depth of 32 bits with a game runing at 48 FPS... then the
 > bandwidth is more than 3Gb/s right ?

The PC->display bandwidth is not depending on FPS, just refresh rate, which
is about 85 Hz.

As VGA is analogue the "bandwidth" is (I'm guessing here) something like
1600x1200x85 Hz =~ 163 MHz.

I don't know how DVI works, but let's say it transfers 8 bits of colour
information. As 32 bit is really 24 bit with alpha channel (which is removed
before displayed), you get 1600x1200x60x24 bits/sec (the lower refresh
because TFT screens only need 60 Hz) =~ 2.8 Gb/s

 > Is the signal between a dispay adapter and the display digital ?

Only when running DVI or EGA/CGA (heh heh).

 > Will there be wireless glasses displaying such signal before 100 years
 > ?

I guess it's possible to make them right now, but to outrageous costs,
especially if you want a sleek design still displaying 1600x1200.

I guess 10 years +- 10 years =)

/M<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

"alexandre" <alexc.TakeThisOut@iprolink.ch> wrote in message
news:441b02fa.0306290639.6af41f89@posting.google.com...
 > Hello,
 >
 > just out of curiosity, i was wondering about the needed bandwidth
 > between pc and display. If you have a resolution of 1600*1200 and
 > color depth of 32 bits with a game runing at 48 FPS... then the
 > bandwidth is more than 3Gb/s right ?

Technically, the figure you're looking for is not "bandwidth"
(which is just the frequency range that is required) but "channel
capacity" or "data rate", which is something slightly different.

The raw data rate required, at a minimum, for any given format
and refresh rate may be found by multiplying the number of
pixels per frame times the number of frames per second times
the number of bits per pixel. "32 bit" color depth generally
becomes 24 bits at the display interface (8 bits per color, which
is sufficient for the majority of applications). So we get:

1600 x 1200 x 48 x 24 = 2.211 Gbit/sec.

This is the raw data rate in terms of the actual information being
produced, IF the image is truly being generated at a steady
48 FPS. As another response pointed out, the FPS rate may not
be the same as the display refresh rate. It IS, however, the important
rate in terms of the actual amount of "new" information that would
have to be delivered per second. Some new display interface
proposals take advantage of this by requiring frame storage in
the display, such that the display interface only has to run at the
rate at which new frames are being produced, regardless of the display's
refresh rate. There are other techniques, such as "conditional updating"
(i.e., only transmit those portions of the image which have actually
changed) and of course digital compression methods which in the
future will further reduce the data rates needed on the display
interface.

 >
 > Is the signal between a dispay adapter and the display digital ?
 > Will there be wireless glasses displaying such signal before 100 years
 > ?

Possibly, although they'll have to wait until these new techniques
are available. Remember, digital HDTV currently provides a
1920 x 1080 image at 30 frames per second while requiring only
about 20 Mbit/sec (and transmitted over a 6 MHz channel) - which
gives you some idea of where we are right now with compression
technology. And that's definitely a "wireless" interface...

 >
 > ... any web ressources on this subject would also be appreciated.
 >

I don't know of any resources that cover all of this on the web, but
if you don't mind the self-promotion, it IS covered (hopefully fairly
well!) in my book:

"Digital Interfaces: Fundamentals & Standards" from John Wiley &
Sons.

Amazon has it, I know. The only problem is that it IS a bit
pricey - sorry, they don't let the writers set the price!


Bob M.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

Wow, i wasn't expecting so much answers... not to mention the quality.
Even the writer of a book on the subject!

Thanks to all of you, i appreciate.

Bob, your book isn't that pricey I'll probably buy it.

Thx again and regards,
Alexandre


"alexandre" <alexc DeleteThis @iprolink.ch> a écrit dans le message de
news:441b02fa.0306290639.6af41f89@posting.google.com...
 > Hello,
 >
 > just out of curiosity, i was wondering about the needed bandwidth
 > between pc and display. If you have a resolution of 1600*1200 and
 > color depth of 32 bits with a game runing at 48 FPS... then the
 > bandwidth is more than 3Gb/s right ?
 >
 > Is the signal between a dispay adapter and the display digital ?
 > Will there be wireless glasses displaying such signal before 100 years
 > ?
 >
 > ... any web ressources on this subject would also be appreciated.
 >
 > Thanks
 > Alexandre<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

"Alexandre" <ac007 RemoveThis @bluewin.ch> wrote in message
news:3f007f25$1_4@news.bluewin.ch...
 > Wow, i wasn't expecting so much answers... not to mention the quality.
 > Even the writer of a book on the subject!
 >
 > Thanks to all of you, i appreciate.
 >
 > Bob, your book isn't that pricey I'll probably buy it.

Thanks!

Bob M.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

Arg... this damn curiosity :
So if you can transmitt 1920*1080*30(Hz)*24(assuming HDTV uses this color
depth) =~1.5Gbit/sec over 6MHz radio channel, then i would assume you can
transmitt 1 byte at each "cycle"(1/250 ratio), right ?

If that's so, that would be uncompressed ?... And why would the signal over
each cycle be limited to 1 byte ?...

And, you say HDTV only requires 20Mbit/sec ? how come i find 1.5Gbit/sec ?
And if it's realy 20Mbit/sec, why does it require a full 6MHz... or is it
the other way around ? : like the digital signal is compressed to 20Mbit/sec
first, then transmitted over the radio signal with 2 bits each cycle ?

k, k, i realy have to buy your book i guess

But if you feel like answering this one it would be cool

Regards,
Alexandre

 > Possibly, although they'll have to wait until these new techniques
 > are available. Remember, digital HDTV currently provides a
 > 1920 x 1080 image at 30 frames per second while requiring only
 > about 20 Mbit/sec (and transmitted over a 6 MHz channel) - which
 > gives you some idea of where we are right now with compression
 > technology. And that's definitely a "wireless" interface...
 > Bob M.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

"Alexandre" <ac007 RemoveThis @bluewin.ch> wrote in message
news:3f00b984$1_5@news.bluewin.ch...
 > Arg... this damn curiosity :
 > So if you can transmitt 1920*1080*30(Hz)*24(assuming HDTV uses this color
 > depth) =~1.5Gbit/sec over 6MHz radio channel, then i would assume you can
 > transmitt 1 byte at each "cycle"(1/250 ratio), right ?

Nope - squeezing 1920 x 1080 at 30 FPS (60 Hz interlaced, actually)
into a standard 6 MHz TV channel is done through some pretty serious
compression; under the ATSC TV standard, it's a variant of MPEG-2
compression that achieves a final transmitted bit rate, from that original
image data stream, of a tad over 19 MBit/sec. In other words, if you
look at just what's going over the air, there's LESS than one bit of
information per pixel of the original image.

How this is achieved is a bit more complex than I can get into in a
forum such as this one - but obviously, it works with pretty good results.
(Although there ARE examples of types of programming that will play
merry hell with this sort of compression system.)


 > And, you say HDTV only requires 20Mbit/sec ? how come i find 1.5Gbit/sec ?
 > And if it's realy 20Mbit/sec, why does it require a full 6MHz... or is it
 > the other way around ?

Ah, that's where the distinction between "data rate" and "bandwidth"
becomes very important, and why I tend to avoid the misuse of the latter
term to mean the former. "Digital" does NOT imply that you're necessarily
using a simple one-bit-at-a-time ("one bit per symbol") encoding
scheme for transmission, and in fact such systems are actually among the
least efficient in terms of using the available channel bandwidth for
maximum data capacity. (For instance, a modem does NOT do this -
note that it gets up to 56 kbit/sec - that's a BIT rate, note - through a
channel (the phone line) that only provides a bandwidth of about 3 kHz
or so (which limits the BAUD, or "symbol" rate).) You make better use of
the channel by transmitting multiple bits per transmitted symbol. U.S.
broadcast digital television normally employs a system called "8-VSB",
in which three bits at a time are encoded as eight distinct amplitude levels
in a vestigial- sideband (VSB) transmission. Digital cable takes advantage
of the lower expected noise level in THAT channel, and sends more bits
per symbol using a combination of phase and amplitude modulation
called "QAM".


Bob M.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

On Mon, 30 Jun 2003 11:34:59 -0600, "Bob Myers"
<nospamplease.RemoveThis@addressinvalid.com> wrote:

 >I don't know of any resources that cover all of this on the web, but
 >if you don't mind the self-promotion, it IS covered (hopefully fairly
 >well!) in my book:
 >
 >"Digital Interfaces: Fundamentals & Standards" from John Wiley &
 >Sons.

Ahh, so THAT'S why you never did the Web page... <!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

"chrisv" <chrisv DeleteThis @nospam.invalid> wrote in message
news:3mv2gvsmbjk5pk6vpmncstj7e34pmr0hev@4ax.com...

  > >"Digital Interfaces: Fundamentals & Standards" from John Wiley &
  > >Sons.
 >
 > Ahh, so THAT'S why you never did the Web page...

Yeah, and I just noticed I mis-typed the title of MY OWN
#!^$&#**$!! BOOK!!!! Isn't THAT the height of stupidity?
The book is actually entitled "DISPLAY Interfaces", not
"Digital Interfaces". Guess I was just thinking a bit too much
about the particular subject at hand. It covers a lot more than
just the "digital" display interfaces...

Bob M.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

"Bob Myers" <nospamplease.TakeThisOut@addressinvalid.com> wrote in message
news:3f01c935$1_3@hpb10302.boi.hp.com...
 >
 > "chrisv" <chrisv.TakeThisOut@nospam.invalid> wrote in message
 > news:3mv2gvsmbjk5pk6vpmncstj7e34pmr0hev@4ax.com...
 >
   > > >"Digital Interfaces: Fundamentals & Standards" from John Wiley &
   > > >Sons.
  > >
  > > Ahh, so THAT'S why you never did the Web page...
 >
 > Yeah, and I just noticed I mis-typed the title of MY OWN
 > #!^$&#**$!! BOOK!!!! Isn't THAT the height of stupidity?
 > The book is actually entitled "DISPLAY Interfaces", not
 > "Digital Interfaces". Guess I was just thinking a bit too much
 > about the particular subject at hand. It covers a lot more than
 > just the "digital" display interfaces...

ROFL!

(=

/M<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

Thx a bunch for your time and science Bob.
And erm... for the title i didnt dare to mention it

Regards,
Alexandre

"Bob Myers" <nospamplease.RemoveThis@addressinvalid.com> a écrit dans le message de
news:3f01c88a_1@hpb10302.boi.hp.com...
 >
 > "Alexandre" <ac007.RemoveThis@bluewin.ch> wrote in message
 > news:3f00b984$1_5@news.bluewin.ch...
  > > Arg... this damn curiosity :
  > > So if you can transmitt 1920*1080*30(Hz)*24(assuming HDTV uses this
color
  > > depth) =~1.5Gbit/sec over 6MHz radio channel, then i would assume you
can
  > > transmitt 1 byte at each "cycle"(1/250 ratio), right ?
 >
 > Nope - squeezing 1920 x 1080 at 30 FPS (60 Hz interlaced, actually)
 > into a standard 6 MHz TV channel is done through some pretty serious
 > compression; under the ATSC TV standard, it's a variant of MPEG-2
 > compression that achieves a final transmitted bit rate, from that original
 > image data stream, of a tad over 19 MBit/sec. In other words, if you
 > look at just what's going over the air, there's LESS than one bit of
 > information per pixel of the original image.
 >
 > How this is achieved is a bit more complex than I can get into in a
 > forum such as this one - but obviously, it works with pretty good results.
 > (Although there ARE examples of types of programming that will play
 > merry hell with this sort of compression system.)
 >
 >
  > > And, you say HDTV only requires 20Mbit/sec ? how come i find 1.5Gbit/sec
?
  > > And if it's realy 20Mbit/sec, why does it require a full 6MHz... or is
it
  > > the other way around ?
 >
 > Ah, that's where the distinction between "data rate" and "bandwidth"
 > becomes very important, and why I tend to avoid the misuse of the latter
 > term to mean the former. "Digital" does NOT imply that you're necessarily
 > using a simple one-bit-at-a-time ("one bit per symbol") encoding
 > scheme for transmission, and in fact such systems are actually among the
 > least efficient in terms of using the available channel bandwidth for
 > maximum data capacity. (For instance, a modem does NOT do this -
 > note that it gets up to 56 kbit/sec - that's a BIT rate, note - through a
 > channel (the phone line) that only provides a bandwidth of about 3 kHz
 > or so (which limits the BAUD, or "symbol" rate).) You make better use of
 > the channel by transmitting multiple bits per transmitted symbol. U.S.
 > broadcast digital television normally employs a system called "8-VSB",
 > in which three bits at a time are encoded as eight distinct amplitude
levels
 > in a vestigial- sideband (VSB) transmission. Digital cable takes
advantage
 > of the lower expected noise level in THAT channel, and sends more bits
 > per symbol using a combination of phase and amplitude modulation
 > called "QAM".
 >
 >
 > Bob M.
 >
 ><!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth 

How does RAMDAC speed factor into all this?
 >> Stay informed about: pc <-&gt; display bandwidth 

"Shiranui Gen-An" <shiranuigenan.TakeThisOut@supamu-ga-kirai-da.bellsouth.net> wrote in
message news:jVhNa.499$rE3.368@fe02.atl2.webusenet.com...
 > How does RAMDAC speed factor into all this?

It would depend on just what is meant when someone
is talking about "RAMDAC speed factor". The number
most often quoted for a RAMDAC is its maximum clock
rate, which is simply how fast data can be clock into the
thing for conversion. While not completely uninteresting,
that number may be irrelevant if the output rise/fall times
and other factors aren't up to the task of providing
acceptable video at that rate.

Bob M.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: pc <-&gt; display bandwidth