Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

The Adventures of Robin Hood digital restoration now playing in L.A. theatre

7 views
Skip to first unread message

dino w.

unread,
Sep 29, 2003, 11:11:55 PM9/29/03
to
The "Ultra-Resolution" digital restoration of "The Adventures of Robin
Hood" is now playing at the ArcLight Cinemas in Hollywood. I saw it
there on Sat. and was really amazed by how good it looked (it's being
shown in digital projection). I'm hoping that they'll do the same for
"The Wizard of Oz". Here's a little info on the Ultra-Resolution
process from the Warners Home Video site:

The Ultra-Resolution Process
Leading the collection will be The Adventures of Robin Hood, presented
in Warner Bros. Pictures dazzling new "Ultra-Resolution," which allows
today's viewers to see parts of the images which were never visible
before and sharper detail than in conventional Technicolor release
prints.

WHV first used this process for Singin' in the Rain; and Ned Price,
Warner Bros. Vice President of Mastering, Technical Operations, says
"Ultra-Resolution" has been improved for The Adventures of Robin Hood.
"As good as Singin' in the Rain turned out, Price said, "Robin Hood
takes us one step further."

Sixty-five years later, Warner Bros. Studios is employing a process
which begins with scanning the original Technicolor 3-strip black and
white "records" at extremely high (2k) resolution. The black and white
records are then combined electronically to create the color images,
which are also electronically re-registered, steadied and cleaned
before the final DVDs are produced.


- D

Michel Hafner

unread,
Sep 30, 2003, 6:31:11 AM9/30/03
to
"dino w." wrote:
>
> The "Ultra-Resolution" digital restoration of "The Adventures of Robin
> Hood" is now playing at the ArcLight Cinemas in Hollywood. I saw it
> there on Sat. and was really amazed by how good it looked (it's being
> shown in digital projection). I'm hoping that they'll do the same for
> "The Wizard of Oz". Here's a little info on the Ultra-Resolution
> process from the Warners Home Video site:
>
> The Ultra-Resolution Process
> Leading the collection will be The Adventures of Robin Hood, presented
> in Warner Bros. Pictures dazzling new "Ultra-Resolution," which allows
> today's viewers to see parts of the images which were never visible
> before and sharper detail than in conventional Technicolor release
> prints.
>
> WHV first used this process for Singin' in the Rain; and Ned Price,
> Warner Bros. Vice President of Mastering, Technical Operations, says
> "Ultra-Resolution" has been improved for The Adventures of Robin Hood.
> "As good as Singin' in the Rain turned out, Price said, "Robin Hood
> takes us one step further."
>
> Sixty-five years later, Warner Bros. Studios is employing a process
> which begins with scanning the original Technicolor 3-strip black and
> white "records" at extremely high (2k) resolution.

2K is not extremely high but a bare minimum for 35mm camera nagatives.

Scott Dorsey

unread,
Sep 30, 2003, 2:39:15 PM9/30/03
to
"dino w." wrote:
> Sixty-five years later, Warner Bros. Studios is employing a process
> which begins with scanning the original Technicolor 3-strip black and
> white "records" at extremely high (2k) resolution.

2K is not ultra high resolution. 2K is cheesy, minimally acceptable for
CGI sort of resolution.

See, it's like gasoline in Mexico. You can call 75 octane gas "Ultra"
but it still doesn't run any better in a V-8.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Richard

unread,
Oct 1, 2003, 10:26:32 AM10/1/03
to
On 29 Sep 2003 20:11:55 -0700, mrth...@yahoo.com (dino w.) wrote:

>
>Sixty-five years later, Warner Bros. Studios is employing a process
>which begins with scanning the original Technicolor 3-strip black and
>white "records" at extremely high (2k) resolution.

I am very sorry, but it is my humble opinion that "2K resolution"
is very far from being "extremely high" resolution.

Not even "high" in my opinion.


Scott Norwood

unread,
Oct 1, 2003, 11:11:08 AM10/1/03
to

In article <3f7be3ae...@news.stben.net>,

Agreed, though I thought that the new print of "Singin' in the Rain" looked
stunninly good. I'm looking forward to seeing Robin Hood (complete with
screwey soundtrack) when it plays here in a month or so.

--
Scott Norwood: snor...@nyx.net, snor...@redballoon.net
Cool Home Page: http://www.redballoon.net/
Lame Quote: Penguins? In Snack Canyon?

Lincoln Spector

unread,
Oct 1, 2003, 1:14:16 PM10/1/03
to

"Richard" <*core*len...@nsa.org> wrote in message
news:3f7be3ae...@news.stben.net...
For 3-strip Technicolor, which was never very sharp to begin with, it's
plenty.

Lincoln


Scott Dorsey

unread,
Oct 1, 2003, 1:25:06 PM10/1/03
to

The 3-strip originals were plenty sharp.... the imbibition prints were
not so sharp. Now, the question is whether we want improved resolution
in this case or not.

Peter H.

unread,
Oct 1, 2003, 1:41:06 PM10/1/03
to

>>
The 3-strip originals were plenty sharp ... the imbibition prints were not so
sharp.
>>

Nor were the early TV transfers, some of which were made by contact printing,
with two strips being emulsion-to-emulsion, and one strip being
base-to-emulsion, thereby making an out-of-register (and out-of-focus) IP.

Three-strip composited negs were optically printed to the matrices in order to
account for the slight differences in each camera.

Correction for those differences is effected in the optical printer.

Lincoln Spector

unread,
Oct 1, 2003, 6:24:49 PM10/1/03
to
> The 3-strip originals were plenty sharp.... the imbibition prints were
> not so sharp.
I'm not so sure the prints were the problem. Dye transfer prints from
Eastmancolor were plenty sharp.

From what I've read, the negatives weren't that sharp because of the beam
splitter the light had to pass through. But the real culprit was imperfect
registration between the three negatives. There was simply no way the images
on the three negatives would line up perfectly. The alignment was good
enough to avoid fringing, but not good enough to not fuzz the focus a bit.

> Now, the question is whether we want improved resolution
> in this case or not.

The Singin' in the Rain restoration got some criticism for having too
improved a resolution, showing facial wrinkles on the glamorous movie stars
that no one would have wanted visible.

Lincoln


David Mullen

unread,
Oct 1, 2003, 7:32:31 PM10/1/03
to
At Cinecon, I saw the projected comparisons between the earlier reprint of
"Singin' in the Rain" versus the new digitally restored version and I was
sort of undecided.

The earlier version, though it DID look somewhat soft in comparison, with
less rich blacks, also had richer colors (which seems like a contradiction
since rich blacks help keep the colors rich).

The digitally restored version was sharper with better blacks, but harsher
also with a lot more contrast yet slight less rich colors (or at least, the
colors were darker in general). So the earlier version seemed to look
"prettier" to my eyes, color-wise, although slightly blurrier compared to
the digitally restored version, which had a "dupey" look although not any
grainier.

I don't think the problem though is the use of the digital restoration so
much as the aesthetic choices made to make the restored version have more
contrast. It just looked too harsh. I would have preferred to keep the
colors of the earlier photo-chemical restoration but the sharpness of the
digitally restored version.

David Mullen


Kram Sacul

unread,
Oct 2, 2003, 12:49:19 AM10/2/03
to
"Lincoln Spector" <Notr...@myemailaddress.com> wrote in message news:<RuIeb.464$Ix.68...@newssvr21.news.prodigy.com>...

> The Singin' in the Rain restoration got some criticism for having too
> improved a resolution, showing facial wrinkles on the glamorous movie stars
> that no one would have wanted visible.

I thought the goal was to get as much image information out of the
filml elements as possible. Now it's too detailed? Yeah, right.

gbravy

unread,
Oct 2, 2003, 9:49:02 AM10/2/03
to
>
> I thought the goal was to get as much image information out of the
> filml elements as possible. Now it's too detailed? Yeah, right.

I have no 'professional qualifications' to speak about this
(disclaimer) but yet and still...

Some of the recent comments in this group bring up the issue of what,
exactly, is 'restoration'of a film...is it to retrieve all of the data
available in the original or is the goal to recreate as nearly as
possible the experience of seeing the film in the best possible
environment when the film was originally shown in theaters?

My belief is that it is the second option - to recreate or restore the
film to what it would have looked and sounded like when shown in the
best possible venue when the film was first released. But this itself
gets into the sticky issue of what indeed did the film look and sound
like when first presented - memories of this get pretty hazy not to
mention how many people are even around that remember what films
looked and sounded like in the 30's and 40's. I myself remember the
original 'Cinemascope' sound as being richer and having more depth
then todays presentations - but is that correct or just a shading of
memory? In some instances (see the recent Vertigo discussion in this
group) the original may be damaged or have elements missing so that it
is impossible to recreate exactly the intentions of the original. In
other instances, especially with films from the late 30's and 40's, I
have read that the some of the sound tracks are very rich and pure,
and that it was the limitations of exhibition equipment of the era
rather then the tracks themselves that prevented the best possible
audio presentation in theaters.

I also think one needs to be very careful when being evaluating a
restoration - without fairly detailed knowledge of why certain
decisions were made it is difficult to label something as 'right' or
'wrong'. In the example given - in Singing in the Rain (which I
havn't seen) I would tend to agree that sharpening an image so much
that it shows something that would not have been seen in theaters is
going a bit too far. (And..do we really KNOW it wouldn't have been
seen in theaters.) But again, w/o knowing exactly why that decision
was made it is a bit hard to be too critical.

Joseph Goodman

unread,
Oct 2, 2003, 3:19:34 PM10/2/03
to
I wonder if there has been any thought to using such digital re-registering
process on titles where there aren't any b&w negatives or seperations, just
old dye transfer prints.Would it even be possible?


Lincoln Spector

unread,
Oct 2, 2003, 7:56:38 PM10/2/03
to

"Joseph Goodman" <grreas...@yahoo.com> wrote in message
news:blhtk8$ssr$1...@bob.news.rcn.net...
The first stage, of course, would be to make B&W seps from the dye transfer
prints. I'd imagine there would be the old contrast problem. Those prints
were very contrasty, which made for great-looking projection but wasn't very
good a source for new copies.

Lincoln


Joseph Goodman

unread,
Oct 2, 2003, 10:08:45 PM10/2/03
to

> The first stage, of course, would be to make B&W seps from the dye
transfer
> prints.


Why do that when you could just seperate the color scan in *insert image
manipulation program here* to CMY or RGB?

>I'd imagine there would be the old contrast problem. Those prints
> were very contrasty,


In the situations I'm imagining, like two-color Technicolor features or
Cinecolor features, the old prints are all that's left, so there wouldn't be
much choice. And if scanners can reproduce reversal films just fine, why not
projection-contrast motion picture prints?


Steve Kraus

unread,
Oct 2, 2003, 11:51:14 PM10/2/03
to
Kram Sacul wrote:
> I thought the goal was to get as much image information out of the
> filml elements as possible. Now it's too detailed? Yeah, right.

Yes, that's possible. Personally I like as much detail and rez as can be
had but filmmakers often have other ideas to artistically get the image
they want which is why diffusion and other filters may be employed. The
decisions made on-set and in post to control what is seen on the theatre
screen are going to be influenced by what they know the photographic
process being used is going to do.

If some new system came along that doubled the resolution of the resulting
prints it's a safe bet that depending on the situation and the look they
want, many filmmakers would change the filtration to compensate. Therefore
it's also safe to say that if we are now able to get substantially better
sharpness off 3-strip originals than they were able to get via the original
IB prints, if the filmmakers were around to see the results while on many
shots they might be thrilled with the sharp detail they might also cringe
at some shots being too sharp--like closeups of the stars revealing
wrinkles--and, if they were in charge, ask to soften them up a bit, in the
same sense that if the prints had been this sharp back in the day they
would have used filters to soften it a bit.

Richard May

unread,
Oct 3, 2003, 12:08:29 AM10/3/03
to
"Joseph Goodman" <grreas...@yahoo.com> wrote in message news:<blhtk8$ssr$1...@bob.news.rcn.net>...
> I wonder if there has been any thought to using such digital re-registering
> process on titles where there aren't any b&w negatives or seperations, just
> old dye transfer prints.Would it even be possible?

It would be necessary to make separations from the print, then scan
these and re-register the images. It is doubtful that it would make
any improvement over a photochemical restoration, except to fix
misregistration. Any other problems, such as print damage, color
mis-timing, contrast of the print, etc. would just be carried on thru
the new elements.

DM

David Mullen

unread,
Oct 3, 2003, 12:41:28 AM10/3/03
to
> In the situations I'm imagining, like two-color Technicolor features or
> Cinecolor features, the old prints are all that's left, so there wouldn't
be
> much choice. And if scanners can reproduce reversal films just fine, why
not
> projection-contrast motion picture prints?

Because the end goal of the digital restoration is to output a new negative
for making prints, so you want something that has as close as possible the
information on the original camera negative, not a print.

David Mullen


Michel Hafner

unread,
Oct 3, 2003, 9:04:22 AM10/3/03
to

This will become less an issue as prints become obsolete and digital
projection is the standard. Digital restoration should be seen
direct digital anyway as additional film steps just degrade picture
quality again. For archival purposes a positive element can be used
that can be scanned back in later should the digital master ever get
lost.

Michel Hafner

unread,
Oct 3, 2003, 9:07:33 AM10/3/03
to

I see no need for b&w separations. With a good enough scan the
color separations can be done easily from the color master in the digital
domain. And the rest of the problems can be dealt with digitally anyway.

David Mullen

unread,
Oct 3, 2003, 10:35:39 AM10/3/03
to
> This will become less an issue as prints become obsolete and digital
> projection is the standard. Digital restoration should be seen
> direct digital anyway as additional film steps just degrade picture
> quality again. For archival purposes a positive element can be used
> that can be scanned back in later should the digital master ever get
> lost.

(1) It will be years before digital projection becomes commonplace; and (2)
I'm not sure you can technically call it a "restoration" if it never ends up
back in something like its original state, i.e. on film; (3) long-term
digital-only storage of movies is unproven & risky and should be backed-up
with an archival film master; (4) digital restoration work is still best
done using elements with more information on them than a projection-contrast
prints, which is basically a LAST resort, particularly for color work.

So basically, the answer is that a dye-transfer print is still a poor
element to base a restoration on but if nothing else exists, you don't have
much choice...

David Mullen


Lincoln Spector

unread,
Oct 3, 2003, 1:46:54 PM10/3/03
to
> So basically, the answer is that a dye-transfer print is still a poor
> element to base a restoration on but if nothing else exists, you don't
have
> much choice...
So the question becomes...if all you have is a dye transfer (or other early
color) print, what's your best method of restoring it? And maybe Joseph has
a point: Scan the image, separate the colors digitally, than re-align these
digital seps.

And you should also, of course, make b&w prints from the digital seps for
archival purposes.

Lincoln


David Mullen

unread,
Oct 3, 2003, 4:41:20 PM10/3/03
to
I'm sure that if all you have to work with is a misregistered dye transfer
print, digital restoration would be, by far, the most flexible way of
restoring the image and creating a new negative. You could record back to
film from the digital master both a color negative and b&w separations
(negative or positive) if you wanted to.

David Mullen


Peter H.

unread,
Oct 4, 2003, 1:44:39 AM10/4/03
to

>>
By 1955, Technicolor had refined the printing to the point that it was far
better than anything you're seeing in a theatre today. Sharpness was NOT an
issue from 1955 to the day they shut down the IB equipment.
>>

From 1955 on, register was not an issue.

Had those 1955 improvements been available for Fox's 1953 introduction of
CinemaScope, DeLuxe probably would have gone with IB printing.

As it was, the 1955 improvements coincided with VistaVision's big push, and
with its need for very good registration when projected onto "the largest
screen in the world".

Also, "flat wide screen", including Vista, had taken over from Academy, and
good register was needed for all products, not just Vista, as every aperture,
except for true 4:3 Academy (declining or already dead) and 2.55/2.35 'Scope
was being magnified far greater than previously.

Michel Hafner

unread,
Oct 4, 2003, 11:42:56 AM10/4/03
to
David Mullen wrote:
>
> > This will become less an issue as prints become obsolete and digital
> > projection is the standard. Digital restoration should be seen
> > direct digital anyway as additional film steps just degrade picture
> > quality again. For archival purposes a positive element can be used
> > that can be scanned back in later should the digital master ever get
> > lost.
>
> (1) It will be years before digital projection becomes commonplace; and (2)
> I'm not sure you can technically call it a "restoration" if it never ends up
> back in something like its original state, i.e. on film; (3) long-term
> digital-only storage of movies is unproven & risky and should be backed-up
> with an archival film master; (4) digital restoration work is still best
> done using elements with more information on them than a projection-contrast
> prints, which is basically a LAST resort, particularly for color work.
>

Well, any negative made from a digital master has by definition not more
information than the digital master. The digital master should have
sufficient bit depth, of course, to allow for further adaptation to various
masters derived from it (including an anachronistic negative),
but any required prints can be made with better quality directly from the
digital master (of course mass production is cheaper via a negative, but
that's the only reason to make one long term). Analogue backups should
be color separations, put in a salt mine and not be touched again. :-)
And the film elements from which the digital master was made should always
be kept as well, of course.

Peter H.

unread,
Oct 4, 2003, 1:51:27 PM10/4/03
to

>>
Peter, shame on you. You said "4:3" when referring to film. You're confined to
quarters until 16:9 is standardized in the U.S.
>>

Agreed, Marty ... and I'll take that Buck Auricon model with me.

Early Film

unread,
Oct 4, 2003, 5:21:37 PM10/4/03
to
Dick May replied to:

>"Joseph Goodman"

>
>> I wonder if there has been any thought to using such digital re-registering
>> process on titles where there aren't any b&w negatives or seperations, just
>> old dye transfer prints.Would it even be possible?
>
>It would be necessary to make separations from the print, then scan
>these and re-register the images. It is doubtful that it would make
>any improvement over a photochemical restoration, except to fix
>misregistration. Any other problems, such as print damage, color
>mis-timing, contrast of the print, etc. would just be carried on thru
>the new elements.

When one starts preservation from a print instead of from either a negative or
finegrain, photochemical restoration always compresses the shadow and highlight
details. This is an inherent problem in the beast that labs have fought for
years.

Digital can correct the grey scale compression in those problem areas if a
print must be the source.

Earl.

David Mullen

unread,
Oct 4, 2003, 6:40:37 PM10/4/03
to
Laser recording a feature-length digital master out to film costs in the
range of $70,000 so we are not at the point yet where it makes financial
sense to record directly to print stock. You can record a negative on Estar
base for the same cost and make 35mm prints for a couple of thousand each.

David Mullen


David Mullen

unread,
Oct 4, 2003, 6:49:07 PM10/4/03
to
All I can say Michael is that it is a well-known fact that doing digital
scanning from a low-contrast, fine-grain film source like a negative,
interpositive, or internegative yields superior results and allows maximum
flexibility in making digital corrections than using a high-contrast source
like a projection print, especially something as contrasty as a dye-transfer
print.

Talk to anyone doing telecine work or special effects work or restoration
work -- anything that involves converting film into a digital form -- in the
film industry. Ask them which is more ideal, using a negative or a
projection-contrast print. It's a no-brainer. Now if ALL that exists is a
print, then there really isn't any options anyway but to use it.

Perhaps it would be less necessary if all you had to do is scan directly to
print stock, but like I said elsewhere, that is not cost-effective for
features. Ultimately it would make more sense to create a new negative and
then make as many prints as needed off of that.

David Mullen


Lincoln Spector

unread,
Oct 5, 2003, 1:28:42 PM10/5/03
to

"Martin Hart" <martinB...@widescreenBLOCKmuseum.com> wrote in message
news:wJpfb.36945$_N1....@twister.austin.rr.com...
> In article <IXDeb.8321$3k2....@newssvr27.news.prodigy.com>,
> Notr...@myemailaddress.com says...
> This is untrue. The PRINTS were not exceptionally sharp, but the camera
> negatives were quite sharp and had a tonal range that was beyond the
> ability of the printing process to duplicate. Doing 4k scans of
> Technicolor three-strip elements will yield a much better image than 2k.
Yes, but if 2k produces a better image than the printing process the film
was shot for, it is definitely good enough. I can't fault a restoration for
not spending huge amounts of money to make a movie look much, much better
than it was ever expected to look when it was made.

Lincoln


Lincoln Spector

unread,
Oct 5, 2003, 1:33:43 PM10/5/03
to
> Talk to anyone doing telecine work or special effects work or restoration
> work -- anything that involves converting film into a digital form -- in
the
> film industry. Ask them which is more ideal, using a negative or a
> projection-contrast print. It's a no-brainer. Now if ALL that exists is
a
> print, then there really isn't any options anyway but to use it.
Ah, but if all you have is a high-contrast release print, how do you use it?
What's the best way to restore an image from a release print if you don't
have something better?

Lincoln


Lincoln Spector

unread,
Oct 5, 2003, 5:38:16 PM10/5/03
to
> I hear what you're saying and I agree in principle, but if the goal is to
> capture as much information as was on the original negative then 2k isn't
> going to cut it. If preservation is part of the goal of the project then
> you have to try to obtain the content of every last grain of the original
> negatives.
Since it's impossible to make an exact copy of a film image--digitally or
analog--that goal can never be achieved. All the more reason to do
everything you can to preserve the oneg itself.

Making a 4k scan was probably economically unfeasible. If your goals are to
create restoration prints that look as good or better than the original
release, AND to preserve the film to as great an extent as is possible, it
probably makes more sense to do a 2K transfer for the restoration and also
make a really excellent IP optically on 35mm B&W film.

One question I find myself wondering, btw, is why 4k scans are still
prohibitively expensive. 2k has been the standard for several years now, and
Moore's Law would suggest that a 4K scan today should be no more expensive
to create or work with than a 2k scan was two years ago. I can think of two
possible explanations: 1) There are price issues involved here that aren't
effected by Moore's Law, and 2) 2K is now so much an accepted standard that
no one wants to pay for 4K, even if it's cheaper than 2K was when it became
a standard.

Lincoln


Michel Hafner

unread,
Oct 5, 2003, 9:03:26 PM10/5/03
to
David Mullen wrote:
>
> All I can say Michael is that it is a well-known fact that doing digital
> scanning from a low-contrast, fine-grain film source like a negative,
> interpositive, or internegative yields superior results and allows maximum
> flexibility in making digital corrections than using a high-contrast source
> like a projection print, especially something as contrasty as a dye-transfer
> print.
>
I don't doubt that. My original point was only that the negative as the final
result of a digital restoration is long term anachronistic since the archival
element is better the digital master and color separations made from it (plus
the elements the restoration was made from) and the distribution and projection will be digital.

Michel Hafner

unread,
Oct 5, 2003, 9:06:46 PM10/5/03
to
Martin Hart wrote:
>
> In article <Q9lfb.320$Qy2...@newsread4.news.pas.earthlink.net>,
> dav...@earthlink.net says...
> I agree, but only if the resolution is at least 4k, AND the people doing
> the work know what the hell they're doing. We've sure seen lots of
> examples of the opposite.
>
> The day will come, sooner than we can imagine, that a 4k scan will be
> considered low rez.

4K will never be low resolution for 35mm negatives up to now. But there will
be different quality levels of 4K as we have now for 2K. High quality 4K
requires scanning at > 4K and downsampling to 4K, of course. This step
is not yet implemented, I think, but better quality 2K via 4K is an option
right now.

David Mullen

unread,
Oct 5, 2003, 9:29:08 PM10/5/03
to
4K scanning is becoming more common -- the digital intermediates at EFILM,
for example, begin with a 4K scan (on an Imagica XE) that is down-rezzed to
2K for doing the actual color-correction work and then uprezzed
(temporarily) back to 4K for the final transfer out to film.

Companies using the Spirit Datacine for the scanning are limited to 2K
though.

The idea is that you get better quality by scanning at 4K and downsampling
rather than scanning at 2K.

The main problem with doing all the work at 4K is the amount of data being
stored during that period. So that problem will also go away as data
storage get cheaper and data processing gets cheaper. A number of telecines
are coming out with 4K capabilities to, so 4K is definitely going to be more
and more feasible in the next few years. I expect most places will soon be
scanning at 4K and recording out at 4K, so the question is when they will be
able to stop downsampling inbetween at 2K.

As for 3-strip Technicolor films though, you're talking about three
elements, not one -- that's a lot to scan at 4K. Perhaps combining three 2K
scans of different b&w elements make up a little for working at 2K... it
might be more like a 3K scan in terms of practical resolution.

Having seen the tests of true 2K digital projection, however, it seems that
2K digital projection for 4K scanned material will give you pretty decent
quality on a big screen. 4K digital projection, if that happens, of
4K-scanned 35mm material might feel like 70mm projection... since we've
never seen 35mm print projection achieve that level of resolution. It would
be cool but it also means that 35mm movies would have to be as carefully
made as a 65mm production in terms of focus, exposure, etc.

David Mullen


William Hooper

unread,
Oct 6, 2003, 1:24:53 AM10/6/03
to
> Agreed, though I thought that the new print of "Singin' in the Rain" looked
> stunninly good.

Lucky Scott! I saw it at a Performing Arts Center which no longer had
film projection equipment. They contracted a company which brought in
two 35mm projectors, put them in the old booth, & ran changeovers.

The color temperature difference between the two lamps was incredible.
Up would come the cues, & the picture would go from brown to green.
Then back to brown at the next changeover. I got conditioned to
become distressed at the cues. And the lenses! They were so soft &
crappy they looked like shooting through molded gelatin.

What an awful show. The previous 'new' rep prints I'd seen in the
last 5 years looked better, because they were on decent equipment!

Those guys. They were *very* cavalier or oblivious about the size of
the loop ahead of the sound reproducer! The first reel had the sound
so far out of sync, folks complained to the mgmt who said that it had
happened at earlier shows! When the gag in the film came up about
difficulties with sound sync, there was NO laughter. It was probably
perceived as a exposition of how sound films have sync problems - even
now.

Restoration made to look worse than prior issue by negligence. Sync
gag made irrelevant by negligence. It would be funny, if you didn't
have a stake in it.

William Hooper

unread,
Oct 6, 2003, 1:28:11 AM10/6/03
to
> Making a 4k scan was probably economically unfeasible. If your goals are to
> create restoration prints that look as good or better than the original
> release, AND to preserve the film to as great an extent as is possible, it
> probably makes more sense to do a 2K transfer for the restoration and also
> make a really excellent IP optically on 35mm B&W film.

You could just *ask* Richard May, who sometimes pops in here. I hope
he's not getting put off by the uninformed quacking.

Michel Hafner

unread,
Oct 7, 2003, 1:36:03 PM10/7/03
to
Martin Hart wrote:
>
> In article <3F80C026...@mhafner.ch>, mha...@mhafner.ch says...

>
> > 4K will never be low resolution for 35mm negatives up to now. But there will
> > be different quality levels of 4K as we have now for 2K. High quality 4K
> > requires scanning at > 4K and downsampling to 4K, of course. This step
> > is not yet implemented, I think, but better quality 2K via 4K is an option
> > right now.
> >
>
> Mike,
> Never say never. Computer technology is changing so fast that we can
> barely see what is going to be possible, much less common, a year ahead.

I was not talking about what we can do (we can probably scan at real 10K
if we need to) but about the fact that 4K is never low res concerning
35mm negatives because there simply is not (much) more spatial detail
on these negatives, even the sharpest, to begin with. Maybe future emulsions
and lenses will change that. And 70mm is another area where 4K is probably
low res long time.

Michel Hafner

unread,
Oct 7, 2003, 4:08:17 PM10/7/03
to
Martin Hart wrote:
>
> In article <3F82F983...@mhafner.ch>, mha...@mhafner.ch says...
> You're underestimating the resolving power of today's film stocks. A
> graph comparing JVC's and Texas Instrument's projection systems with
> prints shows that there's a lot more in film than you imagine. A 4k scan
> doesn't exceed the resolution of the negative, though it may be equal to
> an unsteady print.
>
> I've placed a copy of the graph at the address below:
> http://www.widescreenmuseum.com/film_vs_digital_resolution.jpg

The graph shows what I said. The Kodak number of 12 Mio Pixels is 4K
resolution (~3000*4000). It's coming from Kodak research when they
developed the Cineon system. A true 4K scan gives you basically what's on
the negative. Project it at 4K digitally and you see something you have
not seen before except from good 70mm prints of 70mm originals. It's way beyond
any standard release print and better as a print from the ON as well.
The IMAX version of Matrix uses presumably 4K scans of the negative and
the result concerning sharpness was incredible in non sfx scenes that were
sharp on the negative. The sfx stuff was limited by CGI resolution (2K).

Scott Dorsey

unread,
Oct 7, 2003, 4:40:20 PM10/7/03
to
Michel Hafner <mha...@mhafner.ch> wrote:
>> I've placed a copy of the graph at the address below:
>> http://www.widescreenmuseum.com/film_vs_digital_resolution.jpg
>
>The graph shows what I said. The Kodak number of 12 Mio Pixels is 4K
>resolution (~3000*4000). It's coming from Kodak research when they
>developed the Cineon system. A true 4K scan gives you basically what's on
>the negative. Project it at 4K digitally and you see something you have
>not seen before except from good 70mm prints of 70mm originals. It's way beyond
>any standard release print and better as a print from the ON as well.
>The IMAX version of Matrix uses presumably 4K scans of the negative and
>the result concerning sharpness was incredible in non sfx scenes that were
>sharp on the negative. The sfx stuff was limited by CGI resolution (2K).

Actually, the graph doesn't tell much of the whole story at all, because
with film the actual resolution is not fixed and depends on the contrast
of the original image.

To find the actual resolution of the image you get on the original negative,
you have to take the modulation transfer function of the film, multiply it by
the MTF of the lens, plug in the contrast level, and then approximate by
finding the point where the curve passes a given point. You can have two
images on the same film processed the same way and get different numbers,
and you can have images where you have better resolution in the highlights
than in the shadows.

Calculating resolution with film is MUCH more difficult than doing so with
digital systems, and that's where all of these arguments begin.

As for me, I am more annoyed by compressed tonal ranges than limited resolution.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Michel Hafner

unread,
Oct 7, 2003, 7:02:44 PM10/7/03
to
Scott Dorsey wrote:
>
> Michel Hafner <mha...@mhafner.ch> wrote:
> >> I've placed a copy of the graph at the address below:
> >> http://www.widescreenmuseum.com/film_vs_digital_resolution.jpg
> >
> >The graph shows what I said. The Kodak number of 12 Mio Pixels is 4K
> >resolution (~3000*4000). It's coming from Kodak research when they
> >developed the Cineon system. A true 4K scan gives you basically what's on
> >the negative. Project it at 4K digitally and you see something you have
> >not seen before except from good 70mm prints of 70mm originals. It's way beyond
> >any standard release print and better as a print from the ON as well.
> >The IMAX version of Matrix uses presumably 4K scans of the negative and
> >the result concerning sharpness was incredible in non sfx scenes that were
> >sharp on the negative. The sfx stuff was limited by CGI resolution (2K).
>
> Actually, the graph doesn't tell much of the whole story at all, because
> with film the actual resolution is not fixed and depends on the contrast
> of the original image.

The upper limit is fixed. Below anything is possible.

> To find the actual resolution of the image you get on the original negative,
> you have to take the modulation transfer function of the film, multiply it by
> the MTF of the lens, plug in the contrast level, and then approximate by

Yes, the lens. Another reason the super sharp images are often an illusion.
The lens does not provide the sharpness, or only in the middle, or with
chromatic aberrations etc.
The first pictures with the Foveon chip on the net showed the great color
resolution of the chip, and the errors of the lens used for taking them. Hm.

> finding the point where the curve passes a given point. You can have two
> images on the same film processed the same way and get different numbers,
> and you can have images where you have better resolution in the highlights
> than in the shadows.
>
> Calculating resolution with film is MUCH more difficult than doing so with
> digital systems, and that's where all of these arguments begin.

Digital systems have their own problems. The theoretical limit is trivial
and given by the pixels and bits. But how to reach it with minimal artifacts is
not trivial.

RP Faiola

unread,
Oct 8, 2003, 8:10:04 AM10/8/03
to
>To find the actual resolution of the image you get on the original negative,
>you have to take the modulation transfer function of the film, multiply it by
>
>the MTF of the lens, plug in the contrast level, and then approximate by
>finding the point where the curve passes a given point. You can have two
>images on the same film processed the same way and get different numbers,
>and you can have images where you have better resolution in the highlights
>than in the shadows.
>


"Tell me that again!"
TOWED IN A HOLE, 1933

0 new messages