poynton@vector.Sun.COM (Charles Poynton) (03/11/89)
This posting is followed by two separately-posted documents, TN21 and TN28
(both of whose Subject: lines begin 'HDTV--').
This is a plea for contributions which could convince the HDTV standards
community to adopt square pixels for 1125-line HDTV. The current proposal
calls for pixels that are 4% narrower than they are high. This fact is of
little significance for traditional television but I believe it would
severely limit the appeal of the standard to commercial, industrial,
scientific, and medical applications, and in particular to computer
graphics.
Attached to this note is a letter which outlines the state of the 1125-line
standards discussions, and presents names and addresses of committee
chairmen and such. There is precious little participation in this effort
from the computer community, and the television industry recognizes no
particular advantage to square pixels. This document recommends a course
of action for those of you who would like to contribute.
The second posting (TN21) is my current view of HDTV in general. The
Europeans have scuttled the adoption of a single international standard
and have proposed their own 1250-line standard, and the U.S. television
networks are muddying the waters with their 1050-line proposal (for which
no hardware exists). Consumer and especially terrestrial broadcast HDTV
is a long way off in America. However, the parameters of an 1125-line
system for Production HDTV were adopted in the U.S. as ANSI/SMPTE 240M a
few weeks ago, commercial 1125-line hardware is abundant (albeit
expensive), and industrial and commercial application of this system is
imminent. Colour pre-press equipment exploiting 1125-line HDTV display
technology is already in the field.
Discussions are currently underway to establish the standard "digital
representation" of the 1125-line system. The parameters currently
favoured are based on a picture with 1035 lines of 1920 pixels, and a
picture aspect ratio of 16:9, giving pixels which are 4% narrower than
they are high. The third posting (TN28) is a detailed description of the
pixel aspect ratio issue, and the choices available for the adoption of
square pixels. A 4% reduction of the proposed sampling frequency, or any
of a number of other parameter changes, would make the pixels square.
Warning: this one is tough slogging for anyone not intimate with
television standards.
Technical contributions would be most effective; for example, issues of
fonts, image processing, anti-aliased rendering, filter and transform
algorithms, raster file exchange, input devices (scanners, CCDs), output
devices (hard copy, plasma, LCD), window system issues, etc. I'm
especially looking for technical contributions from some of the large
organizations who have recently announced various HDTV initiatives, for
example, IBM, H-P, DEC, Apple, AT&T, MIT, and Tektronix. Organizations
such as these have much to gain from the convergence of HDTV and computer
graphics, but have been surprisingly absent from the standards
discussions.
This is a rather liberal use of net bandwidth, but if you wade through
this set of documents you will be very well informed regarding HDTV.
If you make a short contribution describing why square pixels are
desirable, we stand a chance of adopting an HDTV standard which promises
to unify the heretofore disparate fields of television and computer
graphics.
Thanks.
Charles A. Poynton <poynton@sun.COM>
Sun Microsystems, Inc.
-----
Charles A. Poynton <poynton@sun.COM>
Sun Microsystems, Inc.
2550 Garcia Avenue, MS 8-04
Mountain View, CA 94043
415-336-7846
89.03.10
The basic analog parameters of the 1125-line HDTV system have been
accepted internationally (as Annex II of CCIR Report 801-2), and in the U.S.
as the recently-adopted ANSI/SMPTE 240M. Standardization of the digital
representation of this system is proceeding rapidly. Although it seems that
European obstructionism has prevented the 1125-line system from being
adopted as a single worldwide standard for HDTV, it is clear that all users
of 1125-line HDTV will conform to a common standard.
There is a once-in-a-lifetime opportunity to base this standard on a
sampling structure which has exactly equal horizontal and vertical sample
spacing, which I believe will greatly benefit both the television industry
and other industries (in particular, computer graphics), by facilitating the
design of hardware and software, and the exchange of images.
The parameters contained in CCIR Report 801-1 and ANSI/SMPTE 240M do
not specify a sampling frequency. The Japanese manufacturers' Broadcast
Television Association (BTA) have recommended a sampling frequency
(74.25 MHz) which produces a sample aspect ratio of 120:115, only 4% off-
square. There are great benefits in having exactly square pixels, and an
alternate proposed sample rate which is just 4% lower (70.875 MHz)
achieves precisely square pixels.
The Society of Motion Picture and Television Engineers (SMPTE) is
accredited by the American National Standards Institute (ANSI) to develop
national standards for studio television, and also to recommend to the U.S.
State Department the position to be taken by the U.S. in international
standards development. The U.S. Advanced Television Systems Committee,
ATSC, also contributes to this process.
The SMPTE Working Group on High Definition Electronic Production
(WGHDEP), has formed the Ad-Hoc Group on Digital Representation of 1125-
line HDTV (AHG-DR1125). There has been some discussion within the
WGHDEP and the AHG-DR1125 regarding the square-pixel issue, but the
various SMPTE groups have representation that is essentially only from the
broadcast industry, which does not view square pixels as being important
for their particular application. In the absence of any contribution from
other industries, the current proposals will almost certainly be adopted.
The feeling at the various SMPTE standards committees is that if industries
other than television do not contribute to the standardization process, then
the work will be done without them, and they can take or leave the result.
The international standardization for HDTV is proceeding through Study
Group 11 of the CCIR (International Radio Consultative Committee). The
chairman of the U.S. National Study Group to the CCIR is Bernie L. Dickens.
The chairman of Canadian National Study Group 11 is Ken P. Davies. These
groups are fundamental to the international standards process, and the
individuals have influence within SMPTE, but the international process lags
the national work and this particular issue is likely to be a fait accompli
before the CCIR study groups have had a chance to consider it.
If you would like to make your views known, then I would suggest a short
letter to Hugo Gaggioni, the chairman of the AHG-DR1125. A copy should be
sent to Dick Stumpf, the chairman of the Working Group on High Definition
Electronic Production (WGHDEP), and copies could optionally be sent to the
other individuals listed below. The next meeting of the WGHDEP is on
April 5 in Los Angeles, and the AHG-DR1125 meets on April 29 in Las Vegas,
immediately prior to the National Association of Broadcasting (NAB)
convention. It is likely that the square pixel issue will be finally decided
at one of these meetings.
I am a member of both committees, and I will convey to the committees any
e-mail replies which I receive regarding this issue. A large volume of
replies from the net might make the committee sit up and take notice, but
e-mail contributions will carry less weight than a written submission to
the committee.
Please review these documents, distribute them to anyone that you think
may have comments, suggestions, or influence, and let your views be known!
Charles A. Poynton <poynton@sun.COM>
-----
HDTV STANDARDIZATION -- NAMES AND ADDRESSES
SMPTE COMMITTEES
Richard J. Stumpf, Chairman SMPTE WHGDEP
Vice-President R&D, Universal Studios
100 Universal Plaza
Universal City, CA 91608
818-777-3198
Hugo Gaggioni, Chairman SMPTE AHG-DR1125,
Sony Advanced Systems
1600 Queen Anne Road
Teaneck, NJ 07666
201-833-5715, fax 201-833-9321
Keith Field, Chairman SMPTE AHG-HDSS
C.B.C. Engineering Headquarters
7925 Cote St. Luc Road
Montreal, Quebec H4W 1R5 Canada
514-485-5570
SMPTE HEADQUARTERS
Stanley N. Baron, SMPTE Engineering Vice-President
NBC
30 Rockefeller Plaza (1600W)
New York, NY 10112
212-664-7557
Barry C. Detwiler, Television Engineer
SMPTE
595 West Hartsdale Avenue
White Plains, NY 10607
914-761-1100, fax 914-761-3115
Sherwin H. Becker, Director of Engineering
SMPTE, 595 West Hartsdale Avenue
White Plains, NY 10607
914-761-1100, fax 914-761-3115
CCIR STUDY GROUPS
Bernard L. Dickens, Chairman USNSG-11
C.B.S
555 West 57th Street
New York, NY 10019
212-975-2003
Ken P. Davies, Chairman CNSG-11
C.B.C. Engineering Headquarters
7925 Cote St. Luc Road
Montreal Quebec H4W 1R5 Canada
514-485-5474
ATSC
Dr. Robert Hopkins, Executive Director
A.T.S.C
1771 N Street N.W.
Washington, DC 20036
202-429-5345
-----