[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: HDTV: CBS and 16mm

Regarding Chris Bacon's witty and devastating comments about the CBS High
Definition Film Tests:  

Chris could have witnessed these results firsthand had he attended the local
New York SMPTE chapter's meeting of June 19 at David Niles' "Show Me New
York" HDTV theater on East 58th Street, a few blocks east of Du Art.

Niles' theater claims to be the first commercial HDTV theater, albeit
specialized, and houses a short, splashy big-screen presentation that extols
the splendors of Gotham, mostly from safe aerial distances.  All visuals in
Nile's feature derive from 1125-line interlaced cameras; no film.  The HDTV
projectors are Barcos, their CRTs converged and corner-registered to within
an inch of their lives.  As meticulous a job of set-up as I've seen anywhere.

In other words, the perfect venue in which to visually scrutinize CBS's
claims.  With an extra added attraction, however.  Some evident prankster at
SMPTE programmed a double bill with a hard-to-follow opening act.  Preceding
the CBS presentation was a demonstration of transfers of 16mm, Super-16, and
35mm Eastman Color Negative to 1125/60 courtesy of John Dowdell of The Tape

Naturally, John showcased the capabilities of his Spirit DataCine.  It was an
unprecedented, if inadvertent, conjoining of the best film negative and best
commercial HD telecine (for the moment) and best HDTV projector.  Think of
cascaded MTF curves that match ideal system performance thresholds.  Think of
no weak link in the proverbial chain.

For this observer, it was a revelatory experience.  Stunning is not too
strong a word.

Clips of transfers from Super-16 and 35mm Vision neg's and PrimeTime 640T,
even an old contact-printed 16mm dupe negative (!), were shared with the
capacity audience.  On Niles' copious screen they were steady, sharp,
detailed, textured--for all the world, visually equivalent to projected 35mm
film prints.  No aliasing, no artifacts.  A negative's full tonal range, yet
somehow newly mediated, not simply film-to-tape, with satisfying deep yet
detailed blacks, creamy midtones, and gentle highlights that typify the
ineffable "film look" (or, rather, film print look) with, as a bonus, the
pristine cleanliness associated with film dailies.

These old eyeballs are dire skeptics, but they know a cinematic experience
when they see one.  Never before in an HDTV projection room had they bought
the results on the screen as film-like.  What I learned that night was:  the
material cause of the film look is not systemic and compound as I had assumed
for two decades.  Film printing and projection are nonessential.  It boils
down to the capture medium alone.  Hats off to Eastman Kodak and all of our
superb film camera and lens manufacturers.

Then it was CBS's turn.  Carefully calibrated tests comparing 16mm, Super-16,
35mm, and an HDTV camera shot under controlled conditions on a Hollywood
backlot were presented with the same Barco projector by a good soldier from
CBS (not author of the paper).  Variables like lens apertures and exposures
were impressively accounted for, and picture images as well as test images
with resolution indices like Siemens stars were shown.  Charts tracking
resolution fall-off for each medium and format were overlaid, providing
multiple comparisons of performance.  It was oh-so-methodical, scientific,
and authoritative.  And total baloney.

The 16mm examples, both standard and Super, looked akin to something a
consumer VHS camcorder might produce.  OK, I exaggerate, but not by much.
 Resolution targets reproduced on the big screen were soft and broken up.
 The 16mm images overall looked devoid of higher-order detail and were
clearly low-res.

The gentleman from CBS announced, at the conclusion, that all film production
for the coming CBS television season would be 35mm only.  It's not surprising
that white collar CBS executives who examined these tests arrived at the
inescapable conclusion that the quality of 16mm falls far short of that
needed for their 1080-line DTV broadcasts from the Empire State Building.  Or
that HDTV origination is marginally yet demonstrably superior to 35mm. 

In psychology there's a useful concept called cognitive dissonance.
 Sometimes conscious contradictory thoughts will co-exist peaceably; but I
think I speak on behalf of all others in the audience when I say that this
evening was not one of those times.  Dowdell's superb demonstration flatly
preempted the flawed conclusions of the CBS white paper.  

I spent the late '70s and early '80s directing Du Art Film Lab's optical
printing department.  I was directly involved in all Du Art seminars that
demonstrated side-by-side comparisons of original 35mm with blow-ups from
16mm and Super-16mm.  I did Super-16 blow-ups in the late '70s from original
7247 to both 35mm reversal CRI and neg/pos 5243 that virtually rivaled the
35mm of the day.  Today's T-grain Vision products, enhanced 5244
intermediates, and sophisticated 16mm lenses have only served to raise the
bar.  Here's the reality check:  these days at any moment there's a Super-16
blow-up in theatrical distribution somewhere masquerading as a 35mm feature:
 check out Orion/MGM's current release "Ulee's Gold," by Victor Nunez,
starring Peter Fonda.  (Incidentally, the title sequence background is
Super-16 scanned by Cineon:  an industry first.  Looks astounding, too.)

I also set up and operated Du Art's camera & lens test facility in the early
'80s.  I spent several years analyzing 16mm and 35mm zooms and primes, both
per se and as integral elements of a compound lens/camera imaging system.
 Singular specialists in blow-up that we were at the time, we were in a
unique position to notice and attend to variances in resolution, contrast,
and graininess that obtained from shifting combinations of 16mm and 35mm film
stocks, processing, printers, cameras, and lenses.  Inspecting 16mm via the
blow-up process was like holding a giant magnifying glass to the 35mm screen.
 Nothing escaped our attention.

As Director of New Technology, I was also somewhat involved in Du Art's
telecine operation.  I wrote and gave the paper at the 1984 SMPTE National
Technical Conference entitled "Single-Telecine Transfer of 16mm Negative A &
B Rolls to a 1" Type-C Master," which detailed our development of Super-Sync,
the first computer system to synchronize on a field basis a Rank Cintel with
a Type-C VTR.  (It was a modification and extension of Jack Calaway's AVRS,
and it preceded TLC by several years.  Du Art never marketed it.)
 Coincidentally, at that same SMPTE conference I gave a second paper on Du
Art's camera & lens testing facility.

As Super-16 pioneers, we had to build the optics and cut out the skid plate
for our Super-16 Mark III gate.  We had no choice in 1983; none existed.  (Du
Art did a lot of things first in those days.  We imported the first Rank
Cintel into the U.S. in '76 and first telecine liquid gate from Geyer-Werke
in '82.)

I am taking the trouble to trot out some of my credentials because I think I
have something to contribute by way of explaining how CBS's otherwise probing
technology demonstration could have turned out results that ultimately take
leave of reality, at least as far as this individual with 20 years of
real-world experience with the 16mm format is concerned.

It's obvious that everything about 16mm compared to 35mm is scaled down,
including tolerances:  circles of confusion, depth of focus, field flatness,
flange focal depth, weave and jitter in the gate--be it camera, printer, or
telecine.  Or, to put it another way, all optical and mechanical errors are
magnified, including lapses in cleanliness and handling that produce
abrasions, scratches, and negative dirt.

Most film systems have historically been designed for 35mm with its
larger-scale tolerances, with 16mm as an afterthought.  To wit:  contact
printers, optical printers, continuous processing machines, and film scanners
up to and including Cineon.  16mm was introduced, after all, as an amateur
Kodak format in 1923 and for decades thereafter designated in professional
circles as "substandard."  Not where the big money was, mind you.

Flying spot telecines were something of an exception, since 16mm had played a
role in TV since the '50s heyday of TVR kinescopes and '60s heyday of 16mm TV
news, pre-Betacam.  Europe had a post-war legacy of 16mm dramatic TV
production, as well.

All of this is to say that, in general, more attention has been paid to the
needs of 35mm than 16mm--despite that fact that 16mm systems require greater
optimization to produce optimum results.  (It even requires more skill to
shoot 16mm well.  Focus and exposure are more critical.  And, from personal
experience, try watching a jaded 35mm AC attempt to pull focus using the
tight markings on a comparatively puny 16mm lens barrel.)

Chris Bacon remarked: "Since the film image has to pass through at least two
sets of lenses (one on the film camera and then whatever they use in the
telecine), while the video image only has to pass through one lens on the
video camera, the deck is stacked!"

The man's onto something.

Putting aside questions of which 16mm lenses CBS used, how they were mounted
and what condition they were in, the larger issue raised is:  how does one
evaluate the optics of a telecine?  No optics are perfect.  Optical design is
an exercise in damage control, as far as image fidelity is concerned.  This
issue has been a sleeping dog for years.  Most telecine owners just assume
that, optically speaking, a gate is a gate is a gate.  The requirements of
high-definition 16mm are too critical for such assumptions.  Assuming that
Rank Cintel's 16mm gate optics were always ideal and diffraction-limited and
incapable of further improvement:  if five of them were lined up in a row,
how consistent would they prove to be in terms of construction, assembly,
alignment, and performance?

(I haven't touched upon the mechanical issue of traveling film flatness and
steadiness under dynamic conditions of transfer, and the consistency and
repeatability of such over time.)

I'm not picking on or singling out Cintel.  Questions of consistent
manufacture are intrinsic to the world of optics.  Take the Hubble Space
Telescope, for example.

Try this recipe at home:  a certain someone at a Swedish lab recently took a
16mm negative and made a "direct blow-up" to 35mm 5385 low-contrast print
stock.  He transferred the result via his 35mm Rank gate, then compared the
outcome to a transfer of the original 16mm negative via the 16mm Rank gate.
 The blow-up transfer was noticeably sharper.  He shared this tidbit with his
buddy, a certain New York lab owner, who promptly made his own direct low-con
blow-up and replicated the comparison on his own Rank.  With the same
results.  Draw your own conclusions.

Now consider this:  optics are not Sony's forte.  

Aside from the key esthetic issue Chris Bacon so thoughtfully addresses
("fallacy of the 'more is better' school of thought") and the controversies
surrounding deployment of ATV or DTV to which he alludes, including his
correct stating of the FCC's actual ATV policy and his fingering of DTV's
awful chicken-and-egg quandary ("a competitively priced digital TV receiver!")
--and putting aside for the moment the merits or demerits of Sony's
filmchain-on-testosterone HD telecine approach--perhaps Bacon has done us all
a great service to raise the basic 16mm issue.  It's overdue.

Here's my question.  In electronics, smaller is smarter.  Why is film the

D. W. Leitner

++thanks to Digital Vision USA for support of the TIG in 1997
mailinglist digest available
TIG subscriber count is 842 on Wed Aug 13 13:51:30 PDT 1997
the Telecine Internet Group  <http://www.alegria.com/telecinehome.html>