[opendtv] Re: Microsoft Exec: 1080p HDTV Is Meaningless

  • From: "peter wilson" <peter.wilson@xxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Mon, 21 Aug 2006 16:28:37 +0100

Hi All,

This list is sometimes very entertaining in the way it brings up past issues
as if they are in the future and prejudges the future as if it will never
happen.

Non Square pixels were never a real problem, TV 2D and 3D DVE's worked fine
in the 1980's. Now even the computer guy's who didn't understand the math
can use square or rectangular pixels at will. The great majority of Flat
displays are using odd numbers of rectangular pixels and yes you do get a
significant number of incompetent processors but most of the main brands are
now ok.

16-9 was a done deal for TV some years ago, ok some vendors with obsolete
glass still try it on with 5-4 or 4-3.

14-9 was only ever a safe active area reasonably compatible with both 4-3
and 16-9.

Computer guys still confuse frame rate with picture refresh rate.

Mobile phone trials are underway in London already and I have seen a really
neat Samsung clamshell phone where the lid swivels to make a 16-9 screen
rather than 9-16.Picture quality is remarkable fed from the BBC and ITV.

All the noise about 720P only for Europe was based on bad science, 1080P50
was never tested for coding efficiency.

Hans's question which was only answered by one or two on the list is a trick
question due to the coding efficiency with high temporal rate progressive
images. 
There are now many 1080P native displays on the market and there are also
many 720P displays on the market. You pay your money and make your own
choice.

Apparently most CBS Episodic's made in Hollywood today are shot with the
Sony HDC 1500 1080P60 camera's and SR Decks which Sony have been quietly
shipping in the hundreds all over the world including Europe. This gives the
benefit of better 720P and 1080I as it is easier to build a proficient
digital down-converter than an optical filter. 
When asked why they are working in reverse, that is not hyping vapour ware
but shipping product with little noise it's because they don't yet have the
whole product line ready to ship.

The BBC's objective is to switch to Terrestrial 1080P50 broadcasts by 2010
in time for the Olympics in 2012.
In fact there is an interesting bun fight brewing about who gets the
Analogue spectrum at switch off. 
Should there be HD freeview or should HD be the province of the Satellite
and cable operators. 

For any of you visiting IBC in Amsterdam September 7-12, please look in at
the EBU village where I am sure Hans Hoffman will be showing remarkably
efficient 1080P50 coding. 

The BBC will also be showing several versions of their Dirac and new Dirac
Pro open technology Codecs operating at 1080I25 and 1080P50.Using Wavelets
with motion compensation efficiencies equal or greater than h264 / AVC have
been acheived.

EBU Village, RAI Hall 10 stand number 10.410.

More information can be found on Dirac at http://dirac.sourceforge.net/

If you wish to have in depth discussions with the BBC engineers please mail
me on Peter.wilson@xxxxxxxxxx

Bye for now Peter www.hddc.co.uk



-----Original Message-----
From: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx] On
Behalf Of Craig Birkmaier
Sent: 17 August 2006 15:07
To: opendtv@xxxxxxxxxxxxx
Subject: [opendtv] Re: Microsoft Exec: 1080p HDTV Is Meaningless

Not sure I ever thought I would be saying this, but the exec at 
Microsoft and Kilroy have it right.

1080P is clearly better than 1080i as a camera acquisition format, 
however, there are practical limits on the frame rates that can be 
supported and encoded for emission. The reality is that 1080P on the 
acquisition side of the video chain is becoming synonymous with 
1080@24P. This is what is being used by Hollywood for digital post 
production, and in some cases for acquisition. 1080@60P is for all 
practical purposes a display format, and as Kilroy noted, it can just 
as easily be 1080@50P or 1080@72P.

1080@60P as an acquisition format is ludicrous today. It may take 
another decade to develop cameras that can do a decent job with 
1080@xxxxxx

To explain, the problem with cameras is that you need to oversample 
by a factor of 1.5 to 2, to produce images that can take full 
advantage of the available spectra in a 1920 x 1080 format.  We are 
not there yet even for 1080@24P. Doing this at 1080@60P is even 
further out in the future. Then there is the other reality that the 
sensitivity of the camera is limited by the sample area on the 
sensors and the frame rate. Building an oversampling camera to 
produce high quality 1280 x 720 is far more practical; on an 
oversampling 1080P display the quality will be outstanding.

Kilroy touched on the issue of getting 1080@60P through the HDMI 
connections to commercial displays. This is also true for the entire 
video production chain. Virtually all of the HD production gear in 
use today is limited to the clock rates used for 720@60P/1080@30i. 
There are now over 40 remote trucks on the road that can easily 
switch between these formats; video production chains that could 
support 1080@60P do not exist today, and there is little reason to 
believe that this equipment will be available in the near future, or 
that the industry will replace the existing HD gear that is just 
coming on line.

Another interesting side story here is that we are talking about 
gaming platforms. Clearly, it is desirable to render the graphics for 
a game using progressive formats. I don't know what the render specs 
are for the current generation of game platforms, or what is planned 
for the PS3, but I am very confident that the graphics ARE NOT being 
rendered at 1920 x 1080@60P.

Bottom line, I believe that most of the hype about 1080@60P is being 
generated by those who are in the 1080@30i camp as an acquisition 
format. The long term belief (hope) that 1080@60P will become the one 
format to rule them all, helps to prop up 1080@30i for today.

Microsoft had it right in the '90s, suggesting that all emission 
formats be progressive. The issues related to entropy coding alone 
justify this, not to mention the elimination of interlace artifacts 
in acquisition gear.

The emphasis should be on building oversampled 720P cameras, and 
giving the emission channel enough headroom to deliver the full 
potential quality of this format. The oversampled 1080P displays can 
deal with the small number of displays sold that will be large enough 
to need 2 million screen samples.

Meanwhile, interlaced SD is still the primary economic engine behind 
the television industry. If we had listened to Microsoft, Apple et 
all in the '90s, we would at least have a digital TV system that 
delivers high quality images to ALL modern progressive displays.

Regards
Craig




At 6:16 PM -0700 8/16/06, Kilroy Hughes wrote:
>[TB] "This doesn't work well with (confused) telecined material, causes
>artifacts, and likely will involve some additional filtering by the
>display to hide those."
>
>Published HD DVD-V discs have had their P24 cadence carefully preserved.
>That has been SOP for DVD-Vs by the major Studios and authoring
>facilities for several years.  Video editing gear has become cadence
>aware to preserve "edited on tape" TV.  Production chains handle native
>P24 at most stages now, such as D5 tape machines, video editing, etc.,
>so P24 can be encoded Progressive_Sequence=1, or =0 with repeat field
>flags and consistent frames/field pairs.
>
>HD DVD-V players should render graphics at P24 as well as the video, and
>only apply field scanning and 3:2 pulldown to the composited output
>(only necessary because display manufacturers don't support HDMI 1080P24
>mode).  A display won't be confused by P24 output with 3:2 pulldown.
>
>The interlace problems you mention used to be a big problem for DVD-V
>for the first few years when 480P output players and progressive
>displays were rare.  Now there are many tens of millions of progressive
>display systems that have resulted in mostly good production behavior.
>
>The industry still has no clue how to make HDMI and 1080P work.
>Displays are still forced to guess what kind of video content is coming
>over the wire and what signal processing is being done by the source
>device and the display device.
>
>Kilroy Hughes
>-----Original Message-----
>From: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx]
>On Behalf Of Tom Barry
>Sent: Wednesday, August 16, 2006 16:04
>To: opendtv@xxxxxxxxxxxxx
>Subject: [opendtv] Re: Microsoft Exec: 1080p HDTV Is Meaningless
>
>
>
>Kilroy Hughes wrote:
>  > The 1080i30 output of the Samsung BD player will be "inverse
>telecined"
>  > by the 1080 displays out there, 12 fields will be thrown away, and
>the
>  > other 48 will be losslessly combined into 24 frames, which can be
>  > blinked by the display at 60, 72, 120Hz, etc. for motion and contrast
>  > improvement.  No scaling/resampling, filtering, "deinterlacing", etc.
>  > needed.
>  >
>
>That is sort of a best case scenario.  If there are bad edits, moving
>video menus, or video (or different cadence) PIP then it is very hard to
>
>just throw away fields.  And even harder to automatically determine that
>
>you should be doing that.
>
>But if the deinterlacer is uncertain it may fall into an attempt at
>motion compensated or adaptive deinterlacing, assuming the worst.  This
>doesn't work well with (confused) telecined material, causes artifacts,
>and likely will involve some additional filtering by the display to hide
>
>those.
>
>So while deinterlacing is absolutely necessary in our new world of all
>progressive displays I still won't trust it much.
>
>- Tom
>
>
>
>>  The writer didn't understand, so you drew the wrong conclusion.
>>
>> 
>>
>>  The fact is that the HD displays worth considering are progressive
>>  display technology (DLP, PDP, LCD, LCOS, etc.).  If they have
>1080lines
>>  of resolution, they always DISPLAY 1080P, even if you input an
>>  interlaced signal, even if the video was sampled interlaced.  The only
>
>>  question is how much damage is inflicted by source and display
>>  processing to create that progressive image.
>>
>> 
>>
>>  The 1080i30 output of the Samsung BD player will be "inverse
>telecined"
>>  by the 1080 displays out there, 12 fields will be thrown away, and the
>
>>  other 48 will be losslessly combined into 24 frames, which can be
>>  blinked by the display at 60, 72, 120Hz, etc. for motion and contrast
>  > improvement.  No scaling/resampling, filtering, "deinterlacing", etc.
>>  needed.
>>
>> 
>>
>>  The 1080i30 HDMI connection produces better progressive video than the
>
>>  1080P60 output.
>>
>> 
>>
>>  The 1080P60 output applies 3:2 pulldown to generate 60 fields in the
>>  player, then deinterlaces with a Genesis chip.  Ugliness happens when
>>  resampling, motion compensating, and filtering 60 540 line fields, to
>>  synthesize 60 1080 line frames.  60 frames of judder and distortion
>are
>>  then passed over the video interface (HDMI) and processed by the
>>  display's image processor (random behavior different for different
>>  displays) and displayed at 60 frames per second (with 32ms/48ms
>judder).
>>
>> 
>>
>>  As many reviewers have noted, the i30 signal connection from this
>player
>>  produces cleaner, higher resolution progressive display with the same
>>  disc, player, and display than the 60P signal connection.
>>
>> 
>>
>>  Microsoft is a big fan of encoding and displaying progressive images,
>>  but the hype about P60 vs. i30 signal interconnect is a red herring. 
>>
>> 
>>
>>  Kilroy Hughes
>>
>>
>------------------------------------------------------------------------
>  >
>>  *From:* opendtv-bounce@xxxxxxxxxxxxx
>>  [mailto:opendtv-bounce@xxxxxxxxxxxxx] *On Behalf Of *Mark Aitken
>>  *Sent:* Wednesday, August 16, 2006 08:32
>>  *To:* OpenDTV
>>  *Subject:* [opendtv] Microsoft Exec: 1080p HDTV Is Meaningless
>>
>> 
>>
>>  I find it "interesting" that MicroSoft NOW finds the P vs. I argument
>>  "meaningless"! Perhaps marketing hype only works in one direction???
>>
>>  http://www.tvpredictions.com/1080p081506.htm
>>
>>  News & Commentary
>>  Microsoft Exec: 1080p HDTV Is Meaningless
>>  The company's XBox strategist attacks Sony's decision to include the
>new
>>  format in the PlayStation 3.
>>  By Phillip Swann
>>
>>  Washington, D.C. (August 15, 2006) -- A key Microsoft strategist says
>>  the industry is 'hyping' 1080p, the new format that purportedly offers
>a
>>  sharper High-Definition TV picture.
>>
>>  Andre Vrignaud, Microsoft's chief strategist for the XBox game
>console,
>>  says the current 1080i format ("i" stands for interlaced; "p" for
>>  progressive.) provides a picture just as good.
>>
>>  Sony, Microsoft's chief gaming rival, is launching a new game console
>>  (Play Station 3) in November that will support 1080p for HDTV movies
>and
>>  games while Microsoft's XBox 360 will not. The latter console displays
>
>>  games in 1080i and will include a 1080i HD-DVD adapter later in the
>year.
>>
>>  "What's interesting is that a lot of folks don't realize how
>meaningless
>>  1080p actually is in this generation," Vrignaud writes at his blog,
>>  Ozymandias.com. "Most modern HD displays (Plasmas, LCD, DLP, etc.)
>>  display content progressively, //even if they first received an
>>  interlaced signal (so) //when you're watching a 1080 signal on a
>modern
>>  HD display, you're //almost always// watching a 1080p signal."
>>
>>  Vrignaud, Microsoft's director of technical strategy for XBox Live,
>>  added that gamers, and HDTV owners, should not be "sucked into all the
>
>>  1080p hype."
>>
>>  Swanni Sez:
>>  The battle over picture formats is just starting. Sony (and TV
>>  manufacturers who are launching 1080p sets) will say the new format
>>  offers a better picture. But Microsoft, and perhaps some network
>>  programmers who would like to keep filming in 1080i, will say you
>can't
>>  tell the difference.
>>
>>  It's too early to say which side will win, But the short term loser is
>
>>  the American consumer who is already confused enough by high-def.
>>
>>  Click TVPredictions.com <http://www.tvpredictions.com> to see the rest
>
>>  of today's Swanni Sez.
>>
>>  (c) TVPredictions.com
>>
>
>
>
>----------------------------------------------------------------------
>You can UNSUBSCRIBE from the OpenDTV list in two ways:
>
>- Using the UNSUBSCRIBE command in your user configuration settings at
>FreeLists.org
>
>- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
>unsubscribe in the subject line.
>
>
>
>----------------------------------------------------------------------
>You can UNSUBSCRIBE from the OpenDTV list in two ways:
>
>- Using the UNSUBSCRIBE command in your user configuration settings 
>at FreeLists.org
>
>- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the 
>word unsubscribe in the subject line.

 
 
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.


 
 
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at 
FreeLists.org 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: