[Tig] 4K digital intermediates using video technology

Harding, Rick Rick.Harding
Fri Jun 24 01:26:59 BST 2005


>When you say HD I think of tape (D5, HDcam, SR). Typically the data is
>coded in linear 10bit.  That is my biggest >concern.

There seems to a little confusion with respect to Log recording in HDCAM SR,
so I thought I'd jump in to clarify. 

HDCAM SR records 1920 x 1080 linear OR Log, in 10 bit. It does not
pre-filter any input signal, nor convert Log input signals to linear for
record, or otherwise alter the Log input signal in any way. It records what
you give it using intra-frame compression for progressive pictures and
intra-field compression for interlace. 

Hope that answers some questions.

Regards
Rick Harding



-----Original Message-----
From: tig-bounces at tig.colorist.org
[mailto:tig-bounces at tig.colorist.org]On Behalf Of Izhak Hinitz
Sent: Thursday, June 23, 2005 1:25 PM
To: 'tig at tig.colorist.org'
Subject: RE: [Tig] 4K digital intermediates using video technology


Thanks to James Braid of Oktobor for supporting the TIG
--
>OK, I don't like to get in but, this is a great thing to talk about and,
>I have some great people writing and reading so here goes.
>We have been doing D.I work for 3 years now using HD.
>Please give me your thoughts about using HD RGB ( No compression ) as the
>way to offer good quality at low coast to our clients.
I think we all do DI using RGB and I am less bothered by 1920x1080 than I am
by 
the color issues.  It is fairly well accepted in the film community that
10bit log
encoding is sufficient to reproduce the full range of OCN.(I almost agree
but not quite,
a whole different discussion)  When you say HD I think of tape (D5, HDcam,
SR).  Typically the data is coded in linear 10bit.  That is my biggest
concern.  It was well proven several years ago that a digital neg could be
produced from 8bit YUV data that would in turn make a print that was
indistinguishable from a film duped neg > print.  The failure of that
product and system was caused by the fact, the data and the neg it produced
had no range to print up or down.  To explain another way.  The o-neg stock
has a gamma of .6 the print is around 1.8 which means much of the range of
the negative is lost when it is printed.  It was (and still is by many)
believed that this print range is all that is required to do a DI.  I
believe it is necessary to retain the full range (or as much as possible) of
the original neg in the DI files and hopefully the dupe neg.  This is
extremely important in the scanning stage as some scenes will require large
color corrections and if the range has already been limited the outcome will
be an inferior perhaps un-usable shot.  This of course is a waste of time
and money for a movie that was exposed perfectly, has no effects and will
never be seen in the future on a more advanced display medium, be it a
better film stock or a better digital projector.  History shows us that some
films have a very long life and keep coming back.  It just depends on how
much the filmmakers value the future of their products.  The financial
problem is probably related more to the fact that DI comes at the very end
of the process when the money has all been spent.

>We have done this for many years now and are getting to the point of it
>being a " no big deal " as we just treat it as video.
>We do pay special attention to use of color and lookup tables but other
then
>that, what is the big difference in resolution. 
>If the screen was 4X3 then I say we scan and deal with 2k files but, as I
>see
>the screens at our local theatres, they look far closer to 16X9 then the 2k
>dose.
Yep  That's a tough one to argue.  If we say "2k" is good enough, and we
take our full ap scan at 2048x1556 and print it out in 1.85:1, we will see
1828x988 on the screen!  Scope is a different argument and there is all
those scaling artifact discussions...  Oh man this takes a lot of time for
someone that types with 2 fingers...  I shuda' kept my mouth shut...  It is
a subject I am very passionate about!


>and, PS. I love your take on it Bob. 

 
Izhak Hinitz
Eyes Post Group
 


-----Original Message-----
From: Micheletti, Bob (NBC Universal) [mailto:Bob.Micheletti at nbcuni.com]
Sent: Thursday, June 23, 2005 2:59 PM
To: 'tig at tig.colorist.org'
Subject: RE: [Tig] 4K digital intermediates using video technology


Thanks to James Braid of Oktobor for supporting the TIG
--
>What is the general view among this group of what a "Digital Intermediate"
>really is?

I am new to "the group" this will be my first post.  I hope I don't come on
too strong and alienate myself.  I am not new to film.  I have worked in
film effects / post production for 20+ years.  In that time I have been told
several times "film is dead - video is taking over".  I am sure this is true
as is evident in the large number of video post houses entering the "digital
intermediate" market.  My interest is to preserve all of the information
available from the original camera negative.  Hopefully Kodak has not given
up on trying to improve the amount of scene information that can be captured
on the film.  The tools to view digital images (projectors and monitors)
will no doubt improve over the next 100 years as film has over the last
hundred.  The proposition of us going back to the original negative rolls
and rebuilding the picture in the new "best format" is slim to none.  Enough
ranting...here are my answers.


>Does it start and end with film?
Yes.  The term "digital intermediate" was coined by film people to describe
a process of digitally creating a printing negative.  In the good old days
(when the earth was flat) visual effects, titles and opticals went through a
long process of generation loss.  Orig. neg. > Inter-positive > effect/title
composited > dupe neg. > cut in with the rest of the orig. neg. >
inter-positive of the whole picture > printing negs. > print.  If I counted
right that's 5 generations!  The advantages for an effects film were
obvious.  As it became popular for visual effects to be done digitally the
natural progression was to process the whole film through the digital
pipeline.  We all know the advantages of the digital tools now available.
Now that said...does this mean you can't use a digital camera to capture
your images?  Of course you can and many have.  At this point I don't
consider it "digital intermediate" unless it goes to film in the end.  If it
starts as video and is processed as video and finishes as video I think it
should be called "video".


>Does it have to be at "film resolution" - whatever you accept that to be?
Guess you know my answer.  4K is better but for the guy paying for it 2k is
usually adequate.

>Is it only for longform projects?
Film trailers are being done as 3 ? minute DI's.

>Is it allowed to used compressed data anywhere along the path?
I'll stay out of that.

>Does it have to scan the full tonal depth available on the negative?
Yes!  Yes!   Yes!  This may be the only complete version of the film for
future display technologies.  If we throw away tonal depth at this point,
IT'S GONE!

>Is it simply using data files not video?
Not that simple.

>Where would that leave the Sony presentation referred to at the start of
>this thread
That may be the future (film is dead) and I'm ok with that but you asked
what I think "digital intermediate" means to me today.  One thing I have
learned is "the customer is always right".  It is my job to provide the best
possible product in whatever format the filmmakers decide to use.


>Is the term generally used accurately?
"Generally" is a pretty general term ;)

Feel free to comment...I'm not easily offended.

Bob Micheletti
Universal Pictures
participate in the tig wiki pages at http://tig.colorist.org/wiki
_______________________________________________
Tig mailing list
Tig at tig.colorist.org
http://tig.oktobor.com/mailman/listinfo/tig

participate in the tig wiki pages at http://tig.colorist.org/wiki
_______________________________________________
Tig mailing list
Tig at tig.colorist.org
http://tig.oktobor.com/mailman/listinfo/tig
participate in the tig wiki pages at http://tig.colorist.org/wiki
_______________________________________________
Tig mailing list
Tig at tig.colorist.org
http://tig.oktobor.com/mailman/listinfo/tig






More information about the Tig mailing list