Flat TV?
jug...@holly.nospam.colostate.edu
so, how much it will cost? I have been unable to find any information on
it besides the sadly unhelpful website.
--
DANsapienza
GUIL...@webtv.net
THE PRICE IS $ 14,999.99 AND PHILLIPS HAS TOLD ME THAT THEY
COME TO YOUR HOUSE AND INSTALL
IT. I THINK IT'S WAY OVER-PRICED
AND THE SCREEN IS NOT THAT BIG.
MSCHULZE
Jerry Roush
...jerry <><
Jeroen Stessen
MSCH...@prodigy.net wrote:
> Is that Philips a HDTV format?
http://www.flat-tv.com/index2.html
http://www.sv.philips.com/vision/flattv.html
The Philips 42PW9982 is built around the Fujitsu Plasma Display
Panel which has 852x480p resolution. The USA version has (afaik)
YPbPr inputs to directly display 480 lines progressive.
All versions have VGA inputs to display 480 lines progressive.
If you define HDTV as 720 lines progressive or 1080 lines
interlaced then this is obviously not an HDTV display.
We've made it so that it can display 960 lines from a 1080i
signal as a 480p image, so it is at least HDTV compatible.
In my biased opinion even 480p can look very very good.
Most integrated receiver-decoders will output 480i or 480p.
I would seriously consider that option if it were my money.
Matrix displays are not auto-scaling like CRTs, never will be.
Do not expect the arrival of 42" HDTV plasma panels because
they would be very inefficient. PALC is a better candidate for
flat HDTV displays of such small size (42" is small for HDTV).
Regards,
Jeroen.
William Sweeney
including the brand new higher contrast one) and as lovely as they are, they
don't hold a candle to a REAL HDTV....
480P (via RGB) in widescreen makes NTSC look bad for sure (and while a DVD
is nice, watching a S-VHS movie is downright painful) but the pixels are the
panel are large and isolated enough to destroy any illusion of a photograph
like you get with a 720P or a 1080i. Also, even on the enhanced contrast
unit, the contrast is still not as good as a direct view CRT. Additionally,
the panel is easy to burn and the pulse modulated grayscale system for the
pixels sometimes has a VERY annoying flicker on a fixed image.
The 42" panel is a set of training wheels for the real thing. According to
some sources, a 60" 1080 panel is in the works for '99... this panel is
supposed to have a IEEE1394 input instead of the RS170 RGB to allow digital
to digital connections to a set top or computer.
I know I put $$$$$ in my budget!!!!
Randy Sweeney
Alan Roberts
: Matrix displays are not auto-scaling like CRTs, never will be.
Not really rrue, but auto-scaling bumps up the price a little.
: Do not expect the arrival of 42" HDTV plasma panels because
: they would be very inefficient. PALC is a better candidate for
: flat HDTV displays of such small size (42" is small for HDTV).
The Pioneer 50" panel is 1280*768, easy for HDTV at 720p, and should scale
well for 1080i or p.
--
******* Alan Roberts ******* BBC Research & Development Department *******
* My views, not necessarily Auntie's, but they might be, you never know. *
**************************************************************************
Jeroen Stessen
Alan Roberts wrote:
> Jeroen Stessen wrote:
> : Matrix displays are not auto-scaling like CRTs, never will be.
>
> Not really true, but auto-scaling bumps up the price a little.
That is not what I meant with "auto-scaling".
What you mean is that a scaling processor is added (Genesis,
Pixelworks, Limesco, TriMedia) to a matrix display. That is
not quite the same as a CRT that can be varied to display
various resolutions without scaling.
Scaling (spatial sample-rate conversion) generates some
artefacts, of which some loss of sharpness is IMO the least
objectionable one. Most methods generate horrible aliasing.
Whoever is interested can still request a small .pdf file of our
IPA'97 paper about scaling, particularly about good down-sampling
with a transposed polyphase filter. I can e-mail it (~ 100 kB).
> : Do not expect the arrival of 42" HDTV plasma panels because
> : they would be very inefficient. PALC is a better candidate for
> : flat HDTV displays of such small size (42" is small for HDTV).
>
> The Pioneer 50" panel is 1280*768, easy for HDTV at 720p, and
> should scale well for 1080i or p.
Indeed, 50" is larger than 42". The efficiency of plasma cells
goes down with their size, so high-resolution displays will
necessarily be big. Not everybody wants big HDTV, even if it
makes little sense to view HDTV on a smaller high-res screen.
Of course one can scale anything to 720p, or 768p.
Which brings me to the issue of overscan.
If you say that a 768-lines display should be easy for 720p,
do you mean that all 720 lines are shown, plus 48 black lines ?
On our Flat-TV we show the same picture as is shown on a CRT,
this implies the same ~6% overscan. Of NTSC we show only 49 us
x 450 lines, up-scaled to 640x480 or 852x480. Of PAL we show
49 us x 540 lines (not 576), down-scaled to 640x480 or 852x480.
Obviously, the 480 lines display benefits neither standard and
is only good for VGA. Traditionally VGA is never overscanned.
I could explain why overscan is/was a good thing, but I would
be preaching to the choir. I'll sum up the reasons anyway:
- CRT deflection tolerances, including EHT variations
- VCR timing instabilities (jitter, head take-over)
- digital signal processing (filters) run-in and run-out
- motion compensation, the vectors could point off-screen.
Never ever show the entire signal, TV needs overscan !
My point is this: it does not make much sense to build
(discrete) matrix displays with the same resolution as any of
the ATSC standards if you want to (must !) display the same
part of the picture that we are used to. The display resolution
should be ~6% less.
If they continue to build 480, 720 or 1080 lines displays then
we need scaling, even if the resolution "fits". What a pity.
Regards,
Jeroen.
--
----------------------------------------------------------
Jeroen Stessen, Philips Research Laboratories Eindhoven
mailto:ste...@natlab.research.philips.com (Research Labs)
----------------------------------------------------------
Randell Jesup
>Whoever is interested can still request a small .pdf file of our
>IPA'97 paper about scaling, particularly about good down-sampling
>with a transposed polyphase filter. I can e-mail it (~ 100 kB).
Is it available on the web?
>> The Pioneer 50" panel is 1280*768, easy for HDTV at 720p, and
>> should scale well for 1080i or p.
...
>Which brings me to the issue of overscan.
>If you say that a 768-lines display should be easy for 720p,
>do you mean that all 720 lines are shown, plus 48 black lines ?
>
>On our Flat-TV we show the same picture as is shown on a CRT,
>this implies the same ~6% overscan. Of NTSC we show only 49 us
>x 450 lines, up-scaled to 640x480 or 852x480. Of PAL we show
>49 us x 540 lines (not 576), down-scaled to 640x480 or 852x480.
>Obviously, the 480 lines display benefits neither standard and
>is only good for VGA. Traditionally VGA is never overscanned.
>
>I could explain why overscan is/was a good thing, but I would
>be preaching to the choir. I'll sum up the reasons anyway:
>- CRT deflection tolerances, including EHT variations
This is a valid point in that there are analog CRT HDTV's (direct
view and CRT-based projection units). Why 6%?
>- VCR timing instabilities (jitter, head take-over)
This is irrelevant in that there are no (consumer) analog HD VCRs,
nor are there likely to be. Overscan can be provided in 480i (legacy NTSC)
modes of course.
>- digital signal processing (filters) run-in and run-out
This is new to me. What's the issue?
>- motion compensation, the vectors could point off-screen.
What's the issue with motion compensation? Missing blocks viewable
on the edge due to MPEG motion vectors? Doesn't the MPEG2 spec state how
to deal with this? Does the MPEG2 spec state that the outermost x% should
be assumed to not be visible? What about the HDTV/ATSC specs/etc? How
much can be affected by this and the previous issue?
>Never ever show the entire signal, TV needs overscan !
Analog scanning TV needs (some) overscan IF it's important to go
bevel to bevel (which, slightly unfortunately, it is, at least for analog
sets). However, why must HD modes have the "classic" NTSC 6%? Aren't
tolerances tighter nowadays than on old, cheap NTSC sets, so we can use a
smaller %? So HDTV needs a "HD-Safe" area of slightly less than the full
width/height of the screen - this is nothing new. However, the area
outside the "safe" section must have content, since the reason for having
the overscan is that the area _might_ be seen on some sets.
For example, I have my ProScan (i.e. Thomson/RCA/etc) TV set for
well under 6% overscan. (I think the lowest I could get it with an equal
overscan/correct geometry was 3 or 4%.) Sure, that area is outside the
"safe" NTSC area, but there is content there.
>If they continue to build 480, 720 or 1080 lines displays then
>we need scaling, even if the resolution "fits". What a pity.
Well, you could just not use the "extra" pixels - or allow the user
to adjust the "overscan" level by turning on and off those areas (what a
strange idea!)
I also assume that Phillips has enough clout to ask for displays in
the size you want. I just don't want you to stop me from watching the
entire signal if I have the hardware to do so. Given that there must be
content in the overscan area, whatever that might be, (which I'm assuming),
why not have a fixed-pixel TV show all the overscan area? Are the reasons
you mentioned above a good enough reason not to? If so, how much of the
edges do those reasons apply to?
--
Randell Jesup, Worldgate Communications, ex-Scala, ex-Amiga OS team ('88-94)
rje...@wgate.com
CDA II has been passed and signed, sigh. The lawsuit has been filed. Please
support the organizations fighting it - ACLU, EFF, CDT, etc.
Ed Ellers
">- VCR timing instabilities (jitter, head take-over)
"This is irrelevant in that there are no (consumer) analog HD VCRs, nor are
there likely to be."
There certainly are consumer analog HD VCRs -- in Japan that is, JVC's W-VHS
models (which are also sold in North America as quasi-industrial products).
I don't know what they do about head switching, though.
"Overscan can be provided in 480i (legacy NTSC) modes of course."
True, but I agree with you that 6% overscan may be excessive. The consumer
VCR service manuals I've seen specify head switching to take place three
lines before the start of vertical blanking, which if you extend that to the
top and bottom of the screen makes 2.5%. You could be conservative and
extend that to 4% to allow for misaligned VCRs.
Randell Jesup
>Randell Jesup <rje...@wgate.com> wrote:
>"This is irrelevant in that there are no (consumer) analog HD VCRs, nor are
>there likely to be."
>
>There certainly are consumer analog HD VCRs -- in Japan that is, JVC's W-VHS
>models (which are also sold in North America as quasi-industrial products).
>I don't know what they do about head switching, though.
Sorry, I should have specified US. Certainly they're not a good
reason for giving up 6% of the screen.
>"Overscan can be provided in 480i (legacy NTSC) modes of course."
>
>True, but I agree with you that 6% overscan may be excessive. The consumer
>VCR service manuals I've seen specify head switching to take place three
>lines before the start of vertical blanking, which if you extend that to the
>top and bottom of the screen makes 2.5%. You could be conservative and
>extend that to 4% to allow for misaligned VCRs.
True - I run 3-4% today. But for (US, digital) HDTV in HD modes
(i.e. better than 480i) being displayed on a fixed-pixel display, this is
irrelevant. (The W-VHS industrial VCR's notwithstanding.)
Alan Roberts
: Alan Roberts wrote:
:> Jeroen Stessen wrote:
:> : Matrix displays are not auto-scaling like CRTs, never will be.
:> Not really true, but auto-scaling bumps up the price a little.
Hi Jeroen,
: That is not what I meant with "auto-scaling".
: What you mean is that a scaling processor is added (Genesis,
: Pixelworks, Limesco, TriMedia) to a matrix display. That is
: not quite the same as a CRT that can be varied to display
: various resolutions without scaling.
Ok, I understand, but even crts don't really auto scale unless the spot
size is adjusted for optimum at each scan rate. I've never seen any multi-
standard crt display that did that, so at some rates we see gaps between
lines.
:> The Pioneer 50" panel is 1280*768, easy for HDTV at 720p, and
:> should scale well for 1080i or p.
: Indeed, 50" is larger than 42". The efficiency of plasma cells
: goes down with their size, so high-resolution displays will
: necessarily be big. Not everybody wants big HDTV, even if it
: makes little sense to view HDTV on a smaller high-res screen.
: Of course one can scale anything to 720p, or 768p.
: Which brings me to the issue of overscan.
: If you say that a 768-lines display should be easy for 720p,
: do you mean that all 720 lines are shown, plus 48 black lines ?
Overscanning is an option on the Fujitsu as it arrives from Delphi. I'm
suggesting that a 768-line panel can display 720 lines, pixel mapped,
or can rescale 720 to 768 or overscan as the viewer wishes. One of the
nicest features of the Fujitsu panel (in the form we get from Delphi) is
that it doesn't need overscan, we can see all the picture if we want to.
: I could explain why overscan is/was a good thing, but I would
: be preaching to the choir. I'll sum up the reasons anyway:
: - CRT deflection tolerances, including EHT variations
: - VCR timing instabilities (jitter, head take-over)
: - digital signal processing (filters) run-in and run-out
: - motion compensation, the vectors could point off-screen.
: Never ever show the entire signal, TV needs overscan !
Almost agreed. Your first reason is why overscan _was_ essential, and
the other three have taken advantage of it. I'd love to throw it out
now that the fundamental need is going.
Anyway, the difference between analogue and digital blanking is
supposed to be enough to cover most of the horizontal problems. These
days it should be possible to get things right rather than live with
a neccessity dictated by old unstable eht supplies.
: My point is this: it does not make much sense to build
: (discrete) matrix displays with the same resolution as any of
: the ATSC standards if you want to (must !) display the same
: part of the picture that we are used to. The display resolution
: should be ~6% less.
Fundamentally yes, but I don't want to get into a deep discussion over the
magnitude of overscan :-)
: If they continue to build 480, 720 or 1080 lines displays then
: we need scaling, even if the resolution "fits". What a pity.
I think we need scaling anyway:
1 so that the viewer can decide whether overscan is good or not
2 to future-proof the display
Jeroen Stessen
Randell Jesup <rje...@wgate.com> wrote:
> Jeroen Stessen <ste...@natlab.research.philips.com> writes:
> >Whoever is interested can still request a small .pdf file of our
> >IPA'97 paper about scaling, particularly about good down-sampling
> >with a transposed polyphase filter. I can e-mail it (~ 100 kB).
>
> Is it available on the web?
No.
> >I could explain why overscan is/was a good thing, but I would
> >be preaching to the choir. I'll sum up the reasons anyway:
> >- CRT deflection tolerances, including EHT variations
>
> This is a valid point in that there are analog CRT HDTV's
> (direct view and CRT-based projection units). Why 6%?
6% is the typical figure in Europe. It is more difficult to get a
figure for the USA, but it can be as large as 9-11 % !
Part of the difference is due to a different definition. When we
say 6% we mean that the active video is 106% of the visible video.
When they say 11% in the USA they probably mean something else that
translates to our 5-6%. I forgot what it was but it had to do
something with the overscan test picture I saw on "Video Essentials".
> >- VCR timing instabilities (jitter, head take-over)
>
> This is irrelevant in that there are no (consumer) analog HD VCRs,
> nor are there likely to be. Overscan can be provided in 480i (legacy NTSC)
> modes of course.
That will be a valid point, for ATSC. Unless it is recorded as NTSC.
> >- digital signal processing (filters) run-in and run-out
>
> This is new to me. What's the issue?
I was referring mainly to analog signals being sampled and then
processed digitally, but it applies in the other cases too. When you
apply a filter to a signal then the duration of the output signal is
longer by the amount of the duration of the impulse response of the
filter. Specifically, some part of the blanking is transferred to the
active part of the signal. This is e.g. important for multi-tap
polyphase filters, the first and last pixels of an output line will
contain part of the blanked input pixels and are therefore not
totally valid pixels. The overscan will normally mask this.
> >- motion compensation, the vectors could point off-screen.
>
> What's the issue with motion compensation? Missing blocks viewable
> on the edge due to MPEG motion vectors? Doesn't the MPEG2 spec state how
> to deal with this? Does the MPEG2 spec state that the outermost x% should
> be assumed to not be visible? What about the HDTV/ATSC specs/etc? How
> much can be affected by this and the previous issue?
This was not about MPEG, it was about motion-compensated field-rate
up-conversion, like when you do a decent 24p -> 60i interpolation.
Near the edges of the picture some errors will be made and this can
be covered by the blanket of overscan. I would say that the extra
(overscanned) pixels are not meant to be seen but to help improve
the quality of the pixels that must be seen.
> >Never ever show the entire signal, TV needs overscan !
>
> Analog scanning TV needs (some) overscan IF it's important to go
> bevel to bevel (which, slightly unfortunately, it is, at least for analog
> sets). However, why must HD modes have the "classic" NTSC 6%? Aren't
> tolerances tighter nowadays than on old, cheap NTSC sets, so we can use a
> smaller %? So HDTV needs a "HD-Safe" area of slightly less than the full
> width/height of the screen - this is nothing new. However, the area
> outside the "safe" section must have content, since the reason for having
> the overscan is that the area _might_ be seen on some sets.
Agree, 6% is maybe too much. But it depends on the TV chassis !
- many cheap NTSC sets have raster-correction-free picture tubes,
there is no east-west modulation and thus it is often not possible
to adjust the picture width, this adds a large tolerance !
- many sets do not have adequate EHT stabilization or compensation,
EHT variations affect the picture size
- many traditional displays do not respond well to VCR-induced phase
jumps in the horizontal sync, the line-PLL(s) need about 25 lines
to settle, some of those lines are blanking and some are overscan.
> For example, I have my ProScan (i.e. Thomson/RCA/etc) TV set for
> well under 6% overscan. (I think the lowest I could get it with an equal
> overscan/correct geometry was 3 or 4%.) Sure, that area is outside the
> "safe" NTSC area, but there is content there.
There is content, but it is guaranteed by the director to be
unimportant content. It could even be nuisance content.
> >If they continue to build 480, 720 or 1080 lines displays then
> >we need scaling, even if the resolution "fits". What a pity.
>
> Well, you could just not use the "extra" pixels - or allow the user
> to adjust the "overscan" level by turning on and off those areas (what a
> strange idea!)
Scaling is not a principal problem but it causes some loss of quality.
Putting black borders around the picture to mask the overscan would
be even sillier. Still, there is this fundamental difference between
CRTs and matrix displays and I wonder how this difference is going to
show. But I should not wonder, I should help solve these issues...
> I also assume that Philips has enough clout to ask for displays in
> the size you want. I just don't want you to stop me from watching the
> entire signal if I have the hardware to do so. Given that there must be
> content in the overscan area, whatever that might be, (which I'm assuming),
> why not have a fixed-pixel TV show all the overscan area? Are the reasons
> you mentioned above a good enough reason not to? If so, how much of the
> edges do those reasons apply to?
Your guess is as good as mine. But I think that showing the entire
signal
(e.g. 1920x1080) is NOT automatically the best answer. We'll see.
Bye,
Jeroen.
Randell Jesup
>> This is irrelevant in that there are no (consumer) analog HD VCRs,
>> nor are there likely to be. Overscan can be provided in 480i (legacy NTSC)
>> modes of course.
>
>That will be a valid point, for ATSC. Unless it is recorded as NTSC.
When showing NTSC material it can default to overscan (since it has
to scale it anyways this is no big deal).
>> >- digital signal processing (filters) run-in and run-out
>>
>> This is new to me. What's the issue?
>
>I was referring mainly to analog signals being sampled and then
>processed digitally, but it applies in the other cases too. When you
>apply a filter to a signal then the duration of the output signal is
>longer by the amount of the duration of the impulse response of the
>filter. Specifically, some part of the blanking is transferred to the
>active part of the signal. This is e.g. important for multi-tap
>polyphase filters, the first and last pixels of an output line will
>contain part of the blanked input pixels and are therefore not
>totally valid pixels. The overscan will normally mask this.
One would assume that the cameras (the analog HD signals) have
enough pixels to compensate for any expected processing, and the resultant
stream of digital video is all valid content. I don't see this as a good
reason for overscanning fixed-pixel HD displays (or scanning ones, for that
matter).
>> >- motion compensation, the vectors could point off-screen.
>>
>> What's the issue with motion compensation? Missing blocks viewable
>> on the edge due to MPEG motion vectors? Doesn't the MPEG2 spec state how
>> to deal with this? Does the MPEG2 spec state that the outermost x% should
>> be assumed to not be visible? What about the HDTV/ATSC specs/etc? How
>> much can be affected by this and the previous issue?
>
>This was not about MPEG, it was about motion-compensated field-rate
>up-conversion, like when you do a decent 24p -> 60i interpolation.
>Near the edges of the picture some errors will be made and this can
>be covered by the blanket of overscan. I would say that the extra
>(overscanned) pixels are not meant to be seen but to help improve
>the quality of the pixels that must be seen.
This is a reason for doing something about the edges. You are
assuming, however, that the original had pixels that "were not meant to be
seen" to improve video quality. Personally, I want to be able to see all
of an image, especially one that's taken from a movie, not have (say) 6%
cropped off each dimension. Doubly so with a fixed-pixel display. Now,
that might not be possible with a real 1080@60i display - but real 60i
displays will probably scanning displays (CRT's involved somewhere), not
fixed-pixel displays. A 24p signal on a fixed-pixel display normally
should not need a 24p->60i conversion phase.
Again, analog (scanning) displays argue for a safe area, not
necessarily for overscan in all display devices.
>> smaller %? So HDTV needs a "HD-Safe" area of slightly less than the full
>> width/height of the screen - this is nothing new. However, the area
>> outside the "safe" section must have content, since the reason for having
>> the overscan is that the area _might_ be seen on some sets.
>
>Agree, 6% is maybe too much. But it depends on the TV chassis !
>- many cheap NTSC sets have raster-correction-free picture tubes,
> there is no east-west modulation and thus it is often not possible
> to adjust the picture width, this adds a large tolerance !
Not the issue we're discussing - the fact that NTSC signals need
overscan is not in question.
>- many sets do not have adequate EHT stabilization or compensation,
> EHT variations affect the picture size
Again (if I understand this correctly) a reason for HD scanning
displays to overscan slightly (how much?), but not a reason for fixed-pixel
displays to do so.
>- many traditional displays do not respond well to VCR-induced phase
> jumps in the horizontal sync, the line-PLL(s) need about 25 lines
> to settle, some of those lines are blanking and some are overscan.
Only an issue for NTSC signals.
>> For example, I have my ProScan (i.e. Thomson/RCA/etc) TV set for
>> well under 6% overscan. (I think the lowest I could get it with an equal
>> overscan/correct geometry was 3 or 4%.) Sure, that area is outside the
>> "safe" NTSC area, but there is content there.
>
>There is content, but it is guaranteed by the director to be
>unimportant content. It could even be nuisance content.
That's basically the definition of the "safe" area - if it's
important, put it in the safe area. In fact, given 4:3 converters and
sets, the horizontal safe area is a small fraction of the total width
(annoying TV set designers and directors to no end, I'm sure).
>> >If they continue to build 480, 720 or 1080 lines displays then
>> >we need scaling, even if the resolution "fits". What a pity.
>>
>> Well, you could just not use the "extra" pixels - or allow the user
>> to adjust the "overscan" level by turning on and off those areas (what a
>> strange idea!)
>
>Scaling is not a principal problem but it causes some loss of quality.
>Putting black borders around the picture to mask the overscan would
>be even sillier.
From the issues above I really don't see a need for fixed-pixel
displays to overscan, unless the reason is that "so it looks exactly the
same as it does on a scanning display". I.e. the picture looks "bigger"
because the talking heads in the center are larger. (I'm talking about a
pair of fixed-pixel displays, same physical size, one (say) 1920x1080, the
other (say) 1811x1018 (roughly 6% smaller)). This may be a valid argument
(though I disagree with it), in that if those pixels are outside the safe
area, and if (in reality) scanning displays won't show them, then the
same-size fixed-pixel panel can be slightly brighter and slightly cheaper
(maybe) if there are slightly fewer but larger pixels.
>Still, there is this fundamental difference between CRTs and matrix
>displays and I wonder how this difference is going to show.
There are certainly fundamental differences. One of the several
quite appealing things for me about fixed-pixel displays is that you can
get rid of overscan. (As I said, I have my 27" ProScan TV adjusted for
minimum overscan today.)
> But I should not wonder, I should help solve these issues...
:-)
Thanks for the thoughtful and informative response.
Alan Roberts
: Jeroen Stessen <ste...@natlab.research.philips.com> writes:
:>- digital signal processing (filters) run-in and run-out
: This is new to me. What's the issue?
:>- motion compensation, the vectors could point off-screen.
: What's the issue with motion compensation? Missing blocks viewable
Any spatial filtering process uses data from an area of the signal to
contribute to each pixel. So, at the edge of the image, where part of
the contributing area is blanking level (black), the filters will give
incorrect results. The same is true for motion compensation in that there
can be no valid vector available for an object moving newly into the
image. So there's always a small border round the image that has incorrect
processing.
Alan Roberts
: Sorry, I should have specified US. Certainly they're not a good
: reason for giving up 6% of the screen.
It isn't 6% of the screen, it's 6% of width and height taken from each
edge, i.e. the visible _area_ is 0.88*0.88 of the entire screen, or only
77.44%
Alan Roberts
: Randell Jesup <rje...@wgate.com> wrote:
: : Sorry, I should have specified US. Certainly they're not a good
: : reason for giving up 6% of the screen.
: It isn't 6% of the screen, it's 6% of width and height taken from each
: edge, i.e. the visible _area_ is 0.88*0.88 of the entire screen, or only
: 77.44%
Following Jeroen's explanation of the 6% figure, by my calculations the
visible area is 1/1.12 width by 1/1.12 height, or 79.72% os the total
area.
Jeroen Stessen.
Alan Roberts <al...@sung1.rd.bbc.co.uk> wrote:
> It isn't 6% of the screen, it's 6% of width and height taken from each
> edge, i.e. the visible _area_ is 0.88*0.88 of the entire screen, or only
> 77.44%
I understand what you are leading us to, but I certainly did not
intend those numbers as a reference for PAL in Europe !
I have come up with some figures:
The reference is the standard "test-circle" pattern from you-know-who.
Each display shall be adjusted so that the "castellations" around the
picture are just not visible. That defines "visible video".
I've just checked a synthetic test circle on my computer system and
the visible video (ITU-R.601, 13.5 Ms/s x 64.0 us, 576 lines) is
658 pixels wide and 548 lines high. Related to an active video of
(13.5 Ms/s x 52.0 us) 702 pixels x 576 lines this means that the
screen should show 93.7% (658/702) horizontally and 95.1% (548/576)
vertically. In area that means that 89.2% of the video signal fills
100% of the screen.
In our lab we used to say that the overscan is 702/658 = 106.7% or
576/548 = 105.1%. This is what WE mean by "5-6% overscan".
I now seem to recall that where the overscan test picture on the
DVD "Video Essentials" defines "5% overscan", it means that the
5% and 95% markers on the picture touch the borders of the screen.
It means that 90% horizontally and vertically is shown, or only
81% of the video signal fills 100% of the screen.
In our local definition, the overscan would be 100/90 = 111.1% .
Sadly, this seems to be current practice for NTSC in the USA,
at least for the usual non-calibrated sets.
When Randall Jesup says that he sets his display to 3-4% overscan,
let's say 3%, then it might be that he means that he sees the
3% and 97% markers of a test picture. In our Philips definition
that means that he has 100/94 = 106.4% overscan, a good number !
In that case, by that definition, our PAL sets (showing 658 x 548)
have 3.1% horizontal "overscan" and 2.4% vertical "overscan".
Are these the numbers that people can live with ?
Because 3% "overscan" implies that a 1920x1080 signal should be
shown 1:1 on a 1805x1015 matrix display, and that was my point.
Then the argument has turned into reasons why we don't currently
need this amount of overscan anymore, and some of those reasons
are indeed valid. But that does not answer the question of how
much Mr. Director in the TV studio really *wants* us to see.
Please don't say "all of the signal" because I don't think so.
Regards,
Jeroen.
Jan Panteltje
Philips pushes PAL plus in a bad (IMNSHO) way.
Who the fuck cares about overscan in the Netherlands, when you have UNDERSCAN
from the PAL plus (in vertical) at least 30 % of the time (at the moment, and
getting worse).
Now they even send cinemascope without the bars, that is vertically streched.
Other times they have the bars, then is is distorded on 16:9.
You have totally raped PAL.
What the fuck is Philips talking about artistic issues, while the viewers
have to live with cut of heads etc.
You fucked up a perfectly good system at the cost of everybody (the viewers
that is).
Only a small amount of people have a 16:9 reciever, and to be precise, only the
rich.
The next time Philips tries to blackmail the government by threatening to move
production to the Asian countries (cheap labor), I would say : please go.
Jan Mourer
Randell Jesup
>In our lab we used to say that the overscan is 702/658 = 106.7% or
>576/548 = 105.1%. This is what WE mean by "5-6% overscan".
>
>I now seem to recall that where the overscan test picture on the
>DVD "Video Essentials" defines "5% overscan", it means that the
>5% and 95% markers on the picture touch the borders of the screen.
>It means that 90% horizontally and vertically is shown, or only
>81% of the video signal fills 100% of the screen.
Right.
>In our local definition, the overscan would be 100/90 = 111.1% .
>Sadly, this seems to be current practice for NTSC in the USA,
>at least for the usual non-calibrated sets.
Yes - in fact, many are set for more overscan than that.
>When Randall Jesup says that he sets his display to 3-4% overscan,
>let's say 3%, then it might be that he means that he sees the
>3% and 97% markers of a test picture. In our Philips definition
>that means that he has 100/94 = 106.4% overscan, a good number !
Right - 3-4% was the lowest I could adjust it to and keep the
geometry correct. This was on a ProScan 27123 if I remember correctly
(27").
>Are these the numbers that people can live with ?
>Because 3% "overscan" implies that a 1920x1080 signal should be
>shown 1:1 on a 1805x1015 matrix display, and that was my point.
I'd prefer to see 1920x1080 (or as close to it as possible) on a
matrix display. That's one of the reasons I want a matrix display. If
I could adjust my direct-view CRT for 0% overscan, I'd do it.
>Then the argument has turned into reasons why we don't currently
>need this amount of overscan anymore, and some of those reasons
>are indeed valid. But that does not answer the question of how
>much Mr. Director in the TV studio really *wants* us to see.
>Please don't say "all of the signal" because I don't think so.
Well, in the past there's been a common assumption about how much
one could see, with directors taking the conservative approach (keep things
inside the minimum safe area, while making sure that anything in the
overscan area is reasonable, so if people do see it there won't be problems).
(They may care less about the last few percent.) A few technologies have
always been able to ignore overscan if they wanted to - FPTV's for example
(front-projection). Most were probably set/calibrated to overscan,
however, so it wouldn't look that different than a regular TV.
With HDTV, to a certain extent there will need to be rethinking of
this. If some of the old restrictions can be relaxed (and they can), then
a larger safe area may make sense. Also, flat-panel displays showing
digital signals bring the opportunity to really show the entire signal
safely. The displays people are watching the shows on will affect how the
directors format shots. Since most HD displays will be some form of
projection for some time, some of the old requirements for overscan/safe
areas probably will continue. (Actually for HDTV, the safe area
horizontally is MUCH smaller than the full width, because of display on 4x3
screens, unless the director plans to add Pan & Scan information to the
signal after the fact (or live?) There may be two "safe" reticles on the
viewfinder/whatever for HD; one for 16x9 and one for 4x3.)
So, if directors are going to use a safe area (of whatever size),
one has to determine what to do with them for fixed-pixel displays - more
smaller pixels and show the full signal (or close), or less larger pixels
and show only the safe area? I argue for more smaller pixels for a few
reasons: first, it's hard to know which overscan value directors will
actually end up using. Second, when the image is not live, but instead
comes from a movie, one would assume that the image will go outside the
"safe" area, so if the overscan areas are (partially or fully) visible
there will be content there. I want to see that content! I don't want to
lose 10-20% of the movie image! (After all, that's a big part of why one
would buy an HD set - movies, not seeing razor stubble on a newscaster.
Another would be sports, where all the action is likely to be in the safe
area, though there may be more in the overscan area.)
Of course, in either case you can have an option to rescale the
image to the other - rescale 1920x1080 to say physical 1700x950 to let you
see the full signal on the less-pixels display, or rescale 1700x950 of the
full signal to 1920x1080 to show just the safe area on a full-width panel.
If any of you do work for a fixed-pixel HD manufacturer, please include the
option to scale the full image to be visible if for some reason you decide
not to have a 1920x1080 (or 1280x720) panel. And keep the number of
physical pixels as close to the full range as you can justify.
IMHO.
Jim Easterbrook
>Jeroen, this I really have to get of my mind:
This might have been better said in Dutch, that well known encryption
system. (-:
To be fair, Philips did not invent PALplus on their own. The system is
an amalgam of ideas from several manufacturers and broadcasters. None of
my ideas was included, but I'm not bitter. Oh no.
I think Jan's problems are the fault of the broadcasters. 16:9 letterbox
is very unpopular with many viewers, which is why PALplus has not taken
off in the UK. Digital broadcasting offers a slightly more satisfactory
way of coping with the mixture of aspect ratios we are forced to
endure.
--
Jim Easterbrook
BBC Research & Development <http://www.bbc.co.uk/rd/>
*** All opinions are mine and might not be shared by the BBC ***
David Crossman
wrote:
> 16:9 letterbox is very unpopular with many viewers, which is why
> PALplus has not taken off in the UK.
With respect, this view is confused.
What has been unpopular is the display of widescreen material
on 4:3 screens. The untechnical audience feel that black borders
top and bottom are giving them *less* picture rather than more.
The picture definition benefits of PALplus were always going to
be lost to the man in the High Street. Couple only Channel Four
seeming to support the system in the UK and a higher price for a
PALplus-equipped set and everyone waits for digital distribution.
> Digital broadcasting offers a slightly more satisfactory
> way of coping with the mixture of aspect ratios we are forced to
> endure.
4:3 displayed 'pillarbox' on a 16:9 screen is wasting picture area
just as 16:9 'letterbox' does on a 4:3 A bad combination is
4:3 archive material 'pillarboxed' against black into an otherwise
16:9 originated programme. Worse would be zooming-in the archive
to fill the 16:9 frame. There are no easy answers Jim.
--
David Crossman DGGB, LRPS
-------------------------
dareks at dircon.co.uk
http://www.dareks.dircon.co.uk/
albert.e...@boeing.com
Jim Easterbrook <jim.eas...@rd.bbc.co.uk> wrote:
>
> To be fair, Philips did not invent PALplus on their own. The system is
> an amalgam of ideas from several manufacturers and broadcasters. None of
> my ideas was included, but I'm not bitter. Oh no.
>
> is very unpopular with many viewers, which is why PALplus has not taken
> way of coping with the mixture of aspect ratios we are forced to
> endure.
I know that Jan doesn't like 16:9, but he does make a valid point. I sure
_hope_ that set-top box manufacturers keep in mind that people with 4:3 sets
DO NOT NECESSARILY WANT TO PUT UP WITH LETTERBOXING!!!
In the digital world, it would be possible to permit the viewer to view a
16:9 program in full-frame 4:3, the heck with lost info on the edges. Set-top
boxes which decode MPEG-2 should permit this, even if dumb
(technically-speaking) PAL receivers can't today.
By the way, I saw an HDTV being demoed just the other day. But they were
showing a standard NTSC broadcast distorted to fill the wide screen. I asked
the guy if he couldn't just put some stripes left and right, and he did at the
push of a button. Aaah. Undistorted image.
Bert
manf...@arl.bna.boeing.com
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
albert.e...@boeing.com
David Crossman <dar...@NOJUNKdircon.co.uk> wrote:
> 4:3 displayed 'pillarbox' on a 16:9 screen is wasting picture area
> just as 16:9 'letterbox' does on a 4:3 A bad combination is
> 4:3 archive material 'pillarboxed' against black into an otherwise
> 16:9 originated programme. Worse would be zooming-in the archive
> to fill the 16:9 frame. There are no easy answers Jim.
That's why viewers should be given the option. And set-top boxes used with 4:3
displays ought to give us those options.
But I do want to make this point: letterboxing in a 4:3 screen is evil,
because it throws away what little resolution the screen has. Whereas
pillarboxing (interesting term) in 16:9 displays is virtuous, because it
gives the viewer the whole image without distortion and without throwing away
any resolution.
Somehow, with PAL or NTSC, I think that retaining what little resolution the
displays provide is the only reasonable course. Edges are usually not all
that important anyway. But, as I said, I think the best of all is to give the
viewer the (difficult) choice.
Jim Easterbrook
>> PALplus has not taken off in the UK.
>
It is a simple truth about our audience, which gave the UK broadcasters
cold feet about going into PALplus.
>What has been unpopular is the display of widescreen material
>on 4:3 screens. The untechnical audience feel that black borders
>top and bottom are giving them *less* picture rather than more.
And they are right, when a programme has been made in 16:9 instead of
the 4:3 that they feel they've paid for. Even when the material is a
widescreen film, the majority of viewers prefer to have their screen
filled. You may not have much respect for the "untechnical audience" but
they pay our wages. Early adopters of squat screen TVs are not a
significant audience.
>There are no easy answers Jim.
I've been here long enough to have discovered that.
Charles Poynton
wrote:
> letterboxing in a 4:3 screen is evil, because
> it throws away what little resolution the screen has ...
If you're viewing test pictures, then I agree. But of you're viewing
_programs_, I disagree. Because letterboxing is liable to throw away the
_plot_, and that is much worse than throwing away "resolution."
Example: At the climax of the movie, the two central proponents of the
story appear at the extremes of the 16:9 picture width. The 4:3
letterboxed viewer sees an empty wall. This is not what the director
intended. Some directors, Woody Allen for one, refuse to give permission
to have their films letterboxed, to avoid this dilemma. I won't enter into
a discussion of what production techniques are necessary to make a show
dual-purpose (16:9 and 4:3), except to say that the techniques range from
difficult to impossible, and many directors won't compromise their primary
aspect ratio.
Second example: Watch a sitcom, and correlate the laughs with action at
the edges of the picture. Jokes designed for 16:9 aren't funny when
letterboxed.
A definitive history of aspect ratios is available in Mark Schubin's
article "Searching for the Perfect Aspect Ratio," in the August 1996 SMPTE
Journal.
--
Charles Poynton
<mailto:poy...@poynton.com> [Mac Eudora/MIME/BinHex/uu]
<http://www.inforamp.net/~poynton/>
albert.e...@boeing.com
poy...@poynton.com (Charles Poynton) wrote:
> In article <749te0$59c$1...@nnrp1.dejanews.com>, albert.e...@boeing.com
> wrote:
>
> > letterboxing in a 4:3 screen is evil, because
> > it throws away what little resolution the screen has ...
>
> If you're viewing test pictures, then I agree. But of you're viewing
> _programs_, I disagree. Because letterboxing is liable to throw away the
> _plot_, and that is much worse than throwing away "resolution."
Charles,
My wife always accuses me of watching as if the program were test patterns
rather than getting involved in the more etherial (artistic) aspects of the
program. Maybe that explains my bias, then.
> Example: At the climax of the movie, the two central proponents of the
> story appear at the extremes of the 16:9 picture width. The 4:3
> letterboxed viewer sees an empty wall. This is not what the director
> intended. Some directors, Woody Allen for one, refuse to give permission
> to have their films letterboxed, to avoid this dilemma.
I avoid the dilemma by avoiding Woody Allen movies. That's even easier.
> I won't enter into
> a discussion of what production techniques are necessary to make a show
> dual-purpose (16:9 and 4:3), except to say that the techniques range from
> difficult to impossible, and many directors won't compromise their primary
> aspect ratio.
However, most movies are distributed in both formats, and in any event I did
say that the choice should be up to the viewer. And it can be, because MPEG-2
even carries the required information to make sure the action is centered on
the display aspect ratio in use, in case people don't like letterboxing. Or
you can use letterboxing for those special Woody Allen movies, if you prefer.
> Second example: Watch a sitcom, and correlate the laughs with action at
> the edges of the picture. Jokes designed for 16:9 aren't funny when
> letterboxed.
Like Woody Allen, I'd just as soon steer clear of insipid sitcoms too. I know
that isn't a proper response, but again, it should be up to the viewer. I get
indignant when the broadcaster forces me to accept 270 scan lines in a
483-line screen (horizontal scan lines).
Ultimately, even a 16:9 display would occasionally have to be letterboxed. If
you want to watch blockbusters on HDTV, they tend to be shot at 2.35:1 aspect
ratio, so some amount of letterboxing would be required for those who want to
view the whole image.
The point is, it will no longer be technically mandatory to force viewers to
accept letterboxing. So since some viewers have small screens to begin with,
this enforced loss of resolution should not longer be tolerated.
Jeroen Stessen.
> If you're viewing test pictures, then I agree. But of you're viewing
> _programs_, I disagree. Because letterboxing is liable to throw away the
> _plot_, and that is much worse than throwing away "resolution."
> Example: At the climax of the movie, the two central proponents of the
> story appear at the extremes of the 16:9 picture width. The 4:3
> letterboxed viewer sees an empty wall. This is not what the director
> intended. Some directors, Woody Allen for one, refuse to give permission
> to have their films letterboxed, to avoid this dilemma.
> Second example: Watch a sitcom, and correlate the laughs with action at
> the edges of the picture. Jokes designed for 16:9 aren't funny when
> letterboxed.
Hello Charles,
Three times you have used the word "letterbox" where I think you
should have used the word(s) "pan-and-scan" or "sidepanel".
Letterbox = a widescreen picture on a 4:3 display with black bars
at the top and bottom. You see the entire plot, but with reduced
vertical resolution.
Sidepanel = a widescreen picture on a 4:3 display with the left
and/or right sides of the picture clipped off. You don't see part
of the plot, but you have full vertical resolution and increased
horizontal resolution.
The discussion has moved from overscan to 16:9 - 4:3 compatibility.
16:9 signals can be transmitted in several ways:
- anamorphic (looks distorted on a 4:3 screen)
- letterbox (loss of vertical resolution on all screens)
- Pal-plus (vertical resolution is restored by 16:9 receivers)
The 16:9 signals can be viewed in several modes on 4:3 screens.
1.
Letterbox (including PAL-plus) is obviously a default mode and
nothing much will be done to show it any better on 4:3 screens.
2.
Showing anamorphic is not a real option because everything will be
tall and narrow. There are some variants for anamorphic signals:
2a.
Vertical compression. All our high-end 4:3 receivers in Europe can
be put into a widescreen mode by reducing the vertical amplitude.
This creates a letterbox picture from an anamorphic signal.
Vertical sample-rate is not affected (if you do the compression by
reducing the vertical scan height of the CRT) but the picture is
of course smaller and so is the vertical resolution (that is also
determined by the spot size). There are some other minor problems.
2b.
Sidepanel allows the full horizontal and vertical resolution that
the display is capable of, but you lose 1/4 of the signal width.
Sidepanels with static clipping is unwise, as you have indicated.
By means of automatic pan-and-scan the 4:3 viewer can see the most
important part of the 16:9 signal, to be decided by the director.
This has always been a feature of the D2MAC standard, although
with its short lifespan it has never really been used. I would
assume this is also a feature of the recent Digital TV standards.
2c.
There is a new option, that we call "norapama" or "amaronap",
an anagram of "panorama" mode. Panorama mode is the method we use
(conceived by JVC but patented by RCA) for showing 4:3 pictures on
a 16:9 screen by stretching the left and right sides. We've shown
you panorama mode on our Flat-TV and you hated it. It is possible
to do the reverse for showing 16:9 pictures on a 4:3 screen, by
compressing the left and right sides and/or expanding the top and
bottom sides. I think that Sony was first to bring this on the
market. I guess you would hate that feature even more.
Nevertheless, there IS a market for solutions to the compatibility-
problem between 16:9 and 4:3. Not only because it may look better
to fill up a screen but also because some display technologies
don't like it if you put black bars on their side.
Most displays with heavily loaded phosphors will have serious
ageing problems. My 16:9 RPTV at home refuses to display black side
panels in 4:3 mode. Instead it puts some video information there,
which it hides by closing 2 black doors behind the screen. This
ensures equal phosphor ageing, at the cost of a 33% higher energy
consumption. For the Flat-TV, panorama mode is the better choice.
Regards,
Jeroen.
Ed Ellers
"Most displays with heavily loaded phosphors will have serious ageing
problems. My 16:9 RPTV at home refuses to display black side panels in 4:3
mode. Instead it puts some video information there, which it hides by
closing 2 black doors behind the screen. This ensures equal phosphor ageing,
at the cost of a 33% higher energy consumption. For the Flat-TV, panorama
mode is the better choice."
JVC showed a prototype direct-view 16:9 set (for NTSC at least) a few years
ago that had motorized masks at the left and right sides that could move in
to cover the unused areas of the tube for 4:3 viewing, which sounds about
the same as what you describe. (Panasonic's answer, on their new HD-capable
16:9 RPTV in North America, is to display gray side panels. UGH!)
Jeroen Stessen
Jan Panteltje <j...@panteltje.demon.nl> wrote:
> Jeroen, this I really have to get of my mind:
> Philips pushes PAL plus in a bad (IMNSHO) way.
I don't even know for a fact that Philips IS pushing PAL-plus !
> Who the fuck cares about overscan in the Netherlands, when you
> have UNDERSCAN from the PAL plus (in vertical) at least 30 % of
> the time (at the moment, and getting worse).
You're supposed to buy a 16:9 PAL-plus TV to get rid of the underscan.
There is NO friendlier way of introducing 16:9 software than PAL-plus !
PAL-plus is one big compromise to satisfy the owners of 4:3 displays
at least partially. I was there when the alternatives were considered.
If it had been Philips' choice then we would have had D2MAC instead
of PAL-plus. Then the 4:3 viewer would have seen widescreen programs
as pan&scan (or alternatively as letterbox with black borders). Alas,
the broadcasters wanted PAL-plus to satisfy their existing audience,
compatibility with existing 4:3 receivers had to cost *nothing*.
> Now they even send cinemascope without the bars, that is vertically
> streched. Other times they have the bars, then is is distorded on
> 16:9. You have totally raped PAL.
Please complain with the broadcasters, not with Philips.
PAL-plus was not our idea, we just helped to improve certain aspects.
The basic idea is from the University of Dortmund. PAL-plus, like
PAL, came from Germany. So go blame the Germans, if you must.
> What the fuck is Philips talking about artistic issues, while the
> viewers have to live with cut off heads etc.
> You fucked up a perfectly good system at the cost of everybody
> (the viewers that is).
> Only a small amount of people have a 16:9 receiver, and to be
> precise, only the rich.
Bullshit, and you know it. I have 2 16:9 receivers and the extra
cost is not very much anymore. Even you can afford one. Within 10
years maybe there will be *only* 16:9 receivers anymore, who knows.
True, only the most expensive ones have built-in PAL-plus decoders,
the rest have only "widescreen-plus" (vertical zoom without helper).
But on a small screen, nobody is going to notice the difference.
(I consider anything smaller than 32" a small display.)
And, for Philips, PAL-plus is always combined with 100 Hz Natural
Motion, which is an expensive feature. The separate PAL-plus
set-top decoders from Nokia were never a success.
> The next time Philips tries to blackmail the government by
> threatening to move production to the Asian countries (cheap
> labor), I would say : please go.
The sad fact is that the *customers* do not want to pay for European
labor. Given the choice they will always buy products from low-wages
countries. Philips did not write the laws of economics.
Philips is one of the very very few companies who produce in the USA
or Europe at all. Most other companies have vanished or moved
already. Blame the customers, the unions, the governments, but not
Philips. These are hard times and we didn't write the rules.
All I wanted to know was whether there is an official convention
about the amount of overscan that is assumed for (digital) television
systems. Now look what you have made out of this thread !
Bye,
Jeroen (at home).
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Jeroen Stessen, Eindhoven, Nederland
mailto:Jeroen....@iae.nl
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Waiting for Windows 98 OSR 2, one more year
albert.e...@boeing.com
Jeroen....@iae.nl wrote:
[ ... ]
> The sad fact is that the *customers* do not want to pay for European
> labor. Given the choice they will always buy products from low-wages
> countries. Philips did not write the laws of economics.
> Philips is one of the very very few companies who produce in the USA
> or Europe at all. Most other companies have vanished or moved
> already. Blame the customers, the unions, the governments, but not
> Philips. These are hard times and we didn't write the rules.
As an Original European myself, I completely, totally, utterly agree with
Jeroen. It is truly gratifying to hear someone speak sense on this subject.
By the way, the same can be said on this side of the pond. People always moan
and groan about "exporting jobs," but doggone it don't they go looking for the
lowest possible price every chance they get.
Funny how consumers seem blind to the realities of what their preferences
mean.
Jan Panteltje
They need to sell, it makes no difference to them WHATSOEVER what it is,
if it is for the good of anybody, improves anything, as long as it SELLS.
>You're supposed to buy a 16:9 PAL-plus TV to get rid of the underscan.
Exactly, and that IS the intention, to get into your purse.
>There is NO friendlier way of introducing 16:9 software than PAL-plus !
Do it on digital channels only, leave the existing PAL channels alone.
You could fade these out (if people wanted 16:9) over time.
Time in this case should be say 10 years, so sets can be written of.
>PAL-plus is one big compromise to satisfy the owners of 4:3 displays
>at least partially. I was there when the alternatives were considered.
>If it had been Philips' choice then we would have had D2MAC instead
Thank GOD no D2MAC
>Please complain with the broadcasters, not with Philips.
Philips dictates the broadcasters.
>
>PAL-plus was not our idea, we just helped to improve certain aspects.
>The basic idea is from the University of Dortmund. PAL-plus, like
>PAL, came from Germany. So go blame the Germans, if you must.
Like I said it does not matter where it came from or how bad it is,
as long as it sells.
No need to complain to the Germans, they develop a lot of good things,
it is the one who uses the wrong thing, stuffs EVERY shop to the brim
with the junk.
>
>Bullshit, and you know it. I have 2 16:9 receivers and the extra
>cost is not very much anymore. Even you can afford one. Within 10
For you, maybe, but then maybe they pay you to much?
>years maybe there will be *only* 16:9 receivers anymore, who knows.
No idea.
>And, for Philips, PAL-plus is always combined with 100 Hz Natural
The first 100Hz sets that came out were terrible ('moving wallpaper)'
I never had need for a 100 Hz set, same reasoning.
maybe the newer ones are better, but who needs them?
>
>The sad fact is that the *customers* do not want to pay for European
>labor. Given the choice they will always buy products from low-wages
>countries. Philips did not write the laws of economics.
Right, I just bought a microwave (really cheap) from that part of the world,
but not because of the price only, mainly because it came out best in
the consumer test, saves me having to buy 2 or living with problems.
>Philips is one of the very very few companies who produce in the USA
>or Europe at all. Most other companies have vanished or moved
>already. Blame the customers, the unions, the governments, but not
>Philips. These are hard times and we didn't write the rules.
Blame the government?
Let me tell you something.
In 79 I worked for Philips military in Huizen (writing tech docs from lab
reports).
They had a government grand to develop a digital telephone system (via
sattelite).
They could not get it right.
So they also build an analog system.
When the government officials came to see what Philips did with the money,
they let them listen to the analog system, telling them it was the digital
one.
I stood there, and was told to keep my mouth shut.
I did, for almost 20 years.
After all I signed a secrecy act.
At this moment my government is more important to me then Philips.
I hope someone reads this, and Philips has to pay back the grand with 20
years interest.
You see, that is how they work.
That is how they got the money.
Cheat and deceit.
>
>All I wanted to know was whether there is an official convention
>about the amount of overscan that is assumed for (digital) television
>systems. Now look what you have made out of this thread !
Overscan on digital TV? come on!
It would be nice if the pixels in the camera would correspond to those
in the display.
But of cause many practical problems arise.
But overscan? why? the borders are exactly defined, unlike in a CRT.
So the broadcaster knows this.
And should take that into account.
mmm, other opinions may be better then mine on this one.
>Waiting for Windows 98 OSR 2, one more year
>
I use Linux, who needs Microsoft, I do not have to wait.
Ed Ellers
"There is NO friendlier way of introducing 16:9 software than PAL-plus!"
Well, maybe. The David Sarnoff Research Center had a project in the late
1980s to develop a compatible 16:9 EDTV system which would display a
full-screen 4:3 picture on existing sets. It was called Advanced Compatible
Television, and Thomson Consumer Electronics and NBC were participating in
the project, essentially reuniting three former divisions of RCA. (I later
found out that elements of the ACTV-1 system came out of an RCA project in
the early 1970s to develop a compatible 3-D TV system that would display one
image from the pair on existing receivers.)
"Philips is one of the very very few companies who produce in the USA or
Europe at all. Most other companies have vanished or moved already. Blame
the customers, the unions, the governments, but not Philips. These are hard
times and we didn't write the rules."
FWIW, Philips Consumer Electronics sold off its Tennessee television factory
a while back and I don't know where their sets are made now. Thomson also
quit making TVs in the U.S.earlier this year, in their case moving all
production to Mexico.
Charles Pope Jr.
FWIW, Sony's XBR200 36" 'Wega' is made in California.
CPJ.
Alan Roberts
:>Please complain with the broadcasters, not with Philips.
: Philips dictates the broadcasters.
Now that's _really_ silly. Philips don't dictate to us or to any other
broadcaster. They are suppliers of hardware, and if we don't like their
kit, we buy from others.
: Overscan on digital TV? come on!
: It would be nice if the pixels in the camera would correspond to those
: in the display.
But I agree with you here.
Marc Mueller
>
> In article <367049E8...@iae.nl>,
> Jeroen....@iae.nl wrote:
>
> [ ... ]
>
> > labor. Given the choice they will always buy products from low-wages
> > countries. Philips did not write the laws of economics.
> > or Europe at all. Most other companies have vanished or moved
> > already. Blame the customers, the unions, the governments, but not
> > Philips. These are hard times and we didn't write the rules.
>
also
makes CRT's for both those plants in both those countries.
Also, Sharp, Sanyo and Sony have television plants in the U.S.
Also, the Toshiba plant has been exporting sets to Japan, Taiwan,
Mexico, Chile and Brazil.