[opendtv] Re: Pan-scan-zoom
- From: "Peter Wilson" <peter.wilson@xxxxxxxxxx>
- To: <opendtv@xxxxxxxxxxxxx>
- Date: Sat, 17 Oct 2009 23:22:03 +0100
-----Original Message-----
From: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx] On
Behalf Of Craig Birkmaier
Sent: 09 October 2009 13:42
To: opendtv@xxxxxxxxxxxxx
Subject: [opendtv] Re: Pan-scan-zoom
At 4:42 PM +0100 10/8/09, Peter Wilson wrote:
>My memory might be getting rusty but I believe Hollywood forced the
Japanese
>to move from 5-3 to 16-9.If they wanted 2-1 I don't remember them being
very
>vocal. This was very expensive as all the CRT's needed to be reworked.
Not exactly. The TV industry pushed 16:9 because of the mathematical
relationship to 4:3, and the belief that it was close enough to the
1.85:1 AR that Hollywood was using for a large percentage of their
theatrical releases. Japan dropped 5:3 in favor of 16:9 because most
of the world wanted HDTV to have a wider aspect ratio.
>>I think the input from Hollywood was stronger than you think. Also the
change in aspect ratio is so small that it doesn't make any real difference
except that the Japanese have to redesign most of theie equipment. I think
Phillips and Thompson had their part to play too.
Unfortunately, reliance upon CRT displays played a major role in
these decisions. Going wider than 16:9 was very difficult technically
and the weight of the CRTs alone was a huge problem. And the
preservation of interlace in HDTV was directly related to scanning
CRTs...
>>Agreed
Can you say 1080i?
>>?
>
>The main supporter of 2-1 today is Vittorio Storaro with the Univisium
>system which also runs at 25 Fps.
2:1 was proposed by Hollywood producers in the early 90's. They held
to this position until the FCC approved the ATSC standard in 1995. If
needed I can look up the name of the Hollywood group that advanced
this position.
>>That might be the case but Vittorio is still supporting this format.
>
>I spent a lot of time in Washington in the late 90's at the ATTC and I
>believe the only people who supported 4-3 for HD were the computer
>industries who wanted progressive scan, a laudable wish but most of them
did
>not know the difference between frame rate and refresh rate and the
>processing power capability was limited to de-interlacing 525.
The ATTC was disbanded in the mid '90s when the ATSC standard was approved.
>>I stayed in Alexandria many times. One of the interesting aspects of this
was demonstrating to various influential people an Asic designed by Snell
and Wilcox and fabricated by IBM to carry out De-interlacing and scaling.
This Asic was mainly used for I-P conversion and resolution conversion. Its
range was limited by the Fab tehnologies of the time but it found use in the
Snell and Wilcox interpolator Gold and Supervisor Products. It is still used
today in many LED wall Display's fifteen years later.
The computer industry DID NOT propose ANY aspect ratio for HDTV. They
specifically proposed that issues such as scanning formats, frame
rates and aspect ratios be left to the marketplace at the
applications layer. They DID argue that ALL sources use progressive
scanning and orthogonal samples (square pixels).
>> I can see the point about square pixels and progressive but I was
surprised that the computer guy's were paranoid about the math of
rectangular pixels. The other point with progressive was that very few of
the staff from the big Computer players understood the difference between
frame rate and refresh rate.
The computer industry was already moving away from the CRT to
lithographed display technologies by the early '90s. They knew that
the CRTs were going away and that panel displays could be made in any
resolution or aspect ratio ( this is why the marketplace offers 16:10
panels for computer applications.
>>I think mid 90's was a bit early, maybe in the lab.
As an industry they had already moved away from building applications
for a specific resolution or aspect ratio. Computer displays became
blank canvases that allowed windows of any size and multiple windows
to be displayed simultaneously.
>>I am not sure this is a benefit for quality programming, for general use I
use these qualities myself.
Most of the technology in modern display systems is based on the
graphics engines developed for the computer industry. Media
architectures such as QuickTime have all the hooks necessary to
display ANY source on any computer display, including transforming
the colorimetry of the source to the characteristics of the display.
>>I think its rather that the computer industry finally realised that upping
the processor clock rate was a hiding to nothing and finally conceded that
image processing hardware co-processors were the future. Until PCi express
these processors could only output to the display.
>There was
>already a significant population of 16-9 TVs in Europe at this time so 4-3
>for a new service was already obsolete. As to the-Blu ray issue I think its
>Cest la vie. I remember visiting I think the Warner DVD plant at LAX at the
>beginning of DVD's and they had every brand and model on the planet as
>nearly every designer had interpreted the menu stuff differently.
The CE and video equipment industries really struggled with the
transition resolution independent. hardwired black boxes to software
driven products.
>>I think the cost was the main issue.
They caused the MPEG-2 standard to become "frozen"
because they did not understand how to deal with issues such as
reserved extensions. And the video production business was lost to
new companies who understood that software driven applications are
inherently resolution independent.
>> This was a very painful process with many of these new players
going to the wall before reaching market acceptance.
The TV types believed that they could fend off the approaching
computer hoards with HDTV - that it would take million dollar HD edit
bays to deal with the increased processing requirements for HD. IN
the end only a handful of these million dollar behemoths were even
built.
>>I think rather more linear suites were built than a handful though now are
mostly replaced with the afore mentioned Avid or Apple.
Avid and Apple just scaled up the applications to deal with HD
- you don;t even need any special hardware today other than an HD-SDI
card.
>>This was a long and difficult path lasting nearly 20 years not one or
two as the marketing guy's kept saying.
>
>As I understand it the 702 / 720 debate is about Digital Filtering.
Analogue
>line length is approximately 702 pixels long but you have to allow for
>Nyquist and ringing when you convert to and from Digital so the system
>designers put in some slack which would be buried by overscan. If you have
>active picture to 720 you may well find you are in trouble. You can
>digitally generate full amplitude Black to white transitions on adjacent
>pixels but it does not relate to reality.
Yes, the 704/720 debate was directly related to the ITU-R BT-601
standard. This standard required controlled rise times for the
transition from picture to blanking to prevent ringing.
Blanking is now a rather obscure concept in an all digital world, and
720 active samples do not cause problems for most display systems.
That is, we now deal with rasters of samples, not formats with all
kinds of supporting crap that was needed for scanning display systems.
>> The samples are still scanned and incorrect filtering will allow alias.
The legacy of Shannon and Nyquist is not dead
Regards
Craig
ANd did I mention the unnecessary imposition of Rec 709 colorimetry.
We had huge discussions about this in 1992 in the SMPTE Task Force on
Digital Image Architecture.
>>This is an interesting topic in itself as with the disappearance of CRT's
there are not many 709 monitors about. Interestingly I believe that the high
end print industry still attaches pantone metadata to each colour in a page
to avoid the problems of anarchic colour space.
Fortunately, this is all history now. What is left of the video
equipment industry no longer has the clout to impose nonsensical
standards to prevent interoperability with the rest of the world.
>>This is a very strange approach as the point of standards is to facilitate
interoperability not prevent it. It might help if the computer world had
read and understand the Video standards before entering the market. Look at
the MAC and PC colour space issues for example.
----------------------------------------------------------------------
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:
