Robert Heiber
Chace Productions
“Sound Preservation: Restoration or Remastering?”
Heiber: We
are going to talk about sound. We live
now in this digital age and we have kind of become used to a mindset where if
we have a problem, we can digitally correct it.
Digital will solve all of our problems.
What I want to talk about today a little more conceptionally, so we are
not going to be real technically oriented.
I will talk about some of the hardware that we have available or that we
use, I don’t want to imply that we have any endorsement or that this is the
only way to do it. The things that I am
going to show and talk about come from our experiences—king of the real world
application; and while it is not exactly specific to orphan films, it is really
specific to sound preservation and restoration in general. I will give you some idea as to the
challenges and problems that we face.
I want to really get
into the concept of what is analog and what is digital and examine the digital
mindset that says, “digital, we can take care of it.” We all know that in the audio world digital
is very well embraced, certainly in the post production community, in current
post production, and when we go digital with sound, it is a much easier
component than working with video or images just because of the size of the
storage requirements.
What I want you to understand first is what I mean by our terms
“preservation”, “restoration”. We will
try to come up with a concept of analog and digital for the film preservation
and restoration world. We will look at
some applications of how we use this technology, and then if time permits, we
will try to draw some conclusions. I
have also brought along some good sound examples, and rather than spend a lot
of time talking about things, I am going to try to breeze through some of the
definitions so we can get into the meat of the program.
Preservation for us in the sound community and even in the image community
is strictly what we would call a technical copying process. We take material that is deteriorating. We want to make a technically correct and
accurate copy as good as we can, so we will have access to the material in the
future. Of course, this presents a
number of challenges. The challenges
that we face in preservation have to do with the condition of the elements, all
the wide variety of formats that have been used for sound recording over the
years, and sometimes we have to develop our own technology to be able to now
retrieve this information.
What kind of challenges do we face in preservation? I think there is a children’s rhyme about an
egg, and as you can see, here is a piece of 35 mm full coat film on
acetate. It has had a complete failure
of the base material. It is actually a
unique piece. I have never had a piece
of film that was this fragile and any time you touch it, it just absolutely
shatters like glass. Another piece of
film, does anyone know what we call this in the sound business? We call that a drop out. What we have here is another piece of 35mm
magnetic film and the binder has completely failed, so the emulsion has peeled
away from the film, leaving just a clear acetate base. They are good for leader, but very bad for
sound.
How do we fix this kind of stuff?
Just like you saw on the Library tour this morning, we have film
preparation inspection areas. I would
call this an analog process, because we are going to go through the material
literally frame by frame. This is a real
challenge. Of course, on the very bad
days with vinegar syndrome requires special areas to work on and obviously
special OSHA required gear in terms of avoiding mag dust and such. It is not a lot of fun to have to clean
10,000 feet of film dressed up like that.
We also are required to have lots of additional formats to look at all of
these old sod elements. These are some of the formats that we
currently have. Actually this slide is a little bit older. We have added about four more formats since
then, and even this is definitely not enough.
Literally, we have a hallway that is somewhat of a museum for analog
transports.
As I mentioned, we also have developed our own kind of technology. We have a custom piece of technology that
plays back soundtrack negatives. We
developed that and we also move on to the commercial marketplace, and we have a
special machine that was built to transfer very distressed magnetic film. Believe it or not, I actually have some
examples of material that looked that bad and was transferred, and I can’t
believe the audio quality. There are all
of these different components. These are
all really analog solutions primarily, and I guess the question is “Is
preservation important?”
We were very fortunate last year.
We were able to restore, preserve, and remaster The Wizard of Oz, and I
would like to point out that that picture was completely done from what I call
first generation preservation elements because none of the original elements
still existed. If the people back at MGM in the 1960s hadn’t taken care of
these elements, we wouldn’t have been able to achieve the results that we were
able to get.
Then we go into restoration,
which is where we try to remove the flaws of age and use. These are things that come in from either
improper storage; many multiples running, so we have tears and breaks. We are just looking to repair the film. When you listen to old movies, they weren’t
produced to sound scratchy, poppy and noisy.
They came out of a laboratory and they sounded great, but over the years
these kinds of flaws have crept into either the masters or the protection
copies, and so the restoration process is strictly to remove these problems and
bring us back to the day that the film came out of the laboratory. The challenges that we face in restoration,
basically I summarize all of the issues as you saw in preservation as just this
term “playback”, cause we have got to get it off the original material. Then we look at different processing
technologies that we would be able to apply to it, and that I call this thing
our archival access which is really being able to research the material and
find the most original or the best undamaged work copy in someone else’s
collection that we can get our hands on to begin the restoration process.
Now we get into these other kinds of other
metamorphous ?? concepts. The first one is “What is analog?” The best description I can give you without
getting very technical is where we do some kind of conversion but we are always
in a one to one relationship. In other
words, if we do a transfer from an optical print to a piece of magnetic film,
that is an analog transfer. What would
be our analog processes then in the film preservation restoration world? We look at the laboratory where we have what
we call the photochemicals. We make
prints and we make duplicates of positives to negatives, and we can print new
prints. In our place we do sound
transfer. The types of equipment we use
– we have analog sound processing equipment.
Then, of course, there is a whole other issue of analog recording media. As you can see, it kind of embraces a very
large range of disciplines, and that is what we want to look at in this
somewhat abstract concept of analog.
Then we have
digital. What is digital? We all know about computers. Primarily in sound and then what we call
digital picture restoration, I call it a conversion where we have the original
material then we make a numerical representation. This, of course, would be the sample rate,
where we are taking something analog and turning it into bits of data. Why are we doing that? Really, the only reason we are doing it is at
some point we want to try to manipulate those bits of data to be something
different or to take flaws out and change those bits into unflawed bits. Of course, the bits themselves are not flawed
but they carry the information. So what
would be our digital processes? Of
course, we have digital processing equipment.
This is the new era of the Dr. Killdeer of audio where we can cure
anything.
Then, of course, we have this area of word processing or audio
recognition. There are some interesting
things I have been exposed to in the past few meetings where there are
cataloguing systems now that will actually do word recognition. We have, of course, archiving. Karen has talked about us starting to move
into this whole world of digital archiving.
What is that going to represent?
Then there is this area called the internet. Which, I don’t know what it is. It really gives the ability to research and
access and communicate with people who are on the AMIA list. It is a fabulous way of suddenly being
connected to 1000 people who may be able to solve your problems. To me that is a digital process and a digital
application. To summarize, you can see
that it is very easy to see how we have different disciplines and yet they are
all going to interact in this very beautiful butterfly pattern. I want you to
understand that nothing is exclusive.
Certainly, preservation becomes the foundation for the whole
process. If we don’t do the preservation
work, forget the other three blocks.
They just don’t even exist. So
preservation really builds the foundation which then these other three
disciplines are able to interrelate in.
We will jump in now into our audio examples. The first area is this thought about analog
copying. In analog copying we have our
laboratory processes, and those would be the photochemicals where we are making
dups of negatives, and making positives, and in the sound transfer business
obviously, we can do this analog. The
analog copying is really perhaps one of the big negatives of the analog world,
because in the digital world they would have us believe the digital copy is
really as every bit as good as the original.
In fact, we call them clones, and there are people who are big
supporters of digital work who would say that actually the digital copy is
better than the original. If the original digital master was made well, the
error correction schemes that are built into all digital recording devices have
less work to do, so you actually have more error correction overhead on your
second copy. So, certainly, digital has
a wonderful application in the preservation and restoration area, and I have a
very nice example of how the analog copying process is not necessarily the best
way to go in certain areas.
What I want to talk about is making what I call budget sound
preservation. Now this is actually a
piece of film from It Should Happen to
You with Judy Holiday in 1954, but it is not the sound that we are going to
hear. What we have here was a client who
actually had in the original, it was a nitrate positive that had shrunk, about
2 ½ - 3% shrinkage. We had some Gene
Autry trailers from the early 1930s. So
in the laboratory process he had a nitrate positive print, but he couldn’t
really use it. He didn’t want to project
it. Went into the laboratory, made a
safety dup negative, which would be the first generation. You can see the little soundtrack area in the
squiggly lines here, so in the negative they are right here and they carry
right along. Actually this is a great
example of digital copying. You can tell
our original looks every bit as good as our second generation, because in the computer
you just use copy and paste and you have a new copy. Then, of course, in our second generation we
have again another copy of the sound track.
Of course, here looks every bit as good as the original, but what you
are going to hear is that when we do this kind of budget sound preservation, we
end up really with some severe problems in the audio world. Just because you can do this work and the
track comes along for free, there is no free lunch.
John, play #2. Listen to the hiss
that is around the modulation of the dialog.
This is the transfer now of the second generation print: Music………..cowboys,
ponies……..
That is the transfer off the
second generation. That is a brand new
print, it was easy to transfer, put it up on the machines, let it go. It is easy work, but it sounds lousy. So we modified some sprockets on our equipment
and we went back to the original nitrate positive, and we just transferred
that. And that’s, John, #3………….Music……………cowboy saddle up…………..
There is a little hiss
modulation, but that is kind of normal for early 30s recording work, but it is
certainly a lot less excessive than what we have on the second generation. Now if the client had made his safety copy,
got a good looking image, but the sound is not very good and if he has thrown
away that nitrate, what would the laboratory do? They would say, fix it digitally. So, we took a transfer off the second
generation, gave it to the engineers and asked how much they could knock down
the hiss modulation? We start to get into
a little fine area, so I want you to
really tune your ears here for it, but again, listen around the dialog and the
modulation of the voice and you will still hear some noise come up with the
voice. So this would be #4…………Music……………….
Now, we knocked it down an
appreciable amount. It almost sounds as
good as the transfer of the original track, but we had to go in there and
digital sound clean up is not inexpensive work.
So, why do we want to make our job harder.
John, if you could skip over to #6. I then had our engineers take our transfer of
the original and with less process that was required to give you the quality of
the second generation. If we had this
material to work with and then put it through the same digital process, what
does our end result sound like? If you
could do #6, sir…………………………Music……………….
So you can see the final track
is something you would be very happy with, and certainly wouldn’t have you get
up and turn off the TV set or change channels if you had to sit down and watch
that old Gene Autry movie. So I think it
is very safe to say here that, if we wanted to draw a conclusion in our world,
we would say certainly for our processes, we always want to say that preserve
our most original sources, and then, obviously, in our analog, digital world, I
think we can prove fairly conclusively that we don’t want to look for budget
sound preservation. We want to avoid
photochemical processes for sound preservation when we are doing our work. I think these are safe conclusions. Of course, this is a university and I felt
professorial, so we are going to change our minds.
This is part of this idea of having a flexible and open
mindframe always to looking at your problems and the solutions you are going to
use. If we just said we are not going to
do any laboratory processes when we want to do audio, and now I am saying maybe
we don’t want to make that blanket statement.
So, I have another really interesting example. This one just cropped up for us, and is an
example of when we want to get into a laboratory setting and do some
photochemical work before we do our sound work.
This is actually a piece of a thumbprint. This is the sound track area here. You can see that is one piece of a perf, so there
are 4 perfs to a piece of 35 mm filmframe, we are down within ¼ of a frame, and
the sound track are is only 1/10 “ wide here.
This is about 30 power under the microscope. The problem here is one called
microspotting. These little tiny pin
holes. They are pretty tiny little dots,
and you wouldn’t think a big powerful postproduction company would have a
problem with these little tiny dots. I
must say, they brought us to our knees because when this print plays back over
a light and photo cell, it creates a terrible noise flourish of a sound we call
bacon frying ……John, #7……………..Clip……………………
That was actually The Job Interview. I hope you could here the bacon frying
sound. Well that is what all of these
little micro dots cause. We tried all of
our analog and digital processing, and all we could do literally, was kill the
track to try to reduce that noise. As
part of preservation and restoration, you sometimes have to have special tools
and tricks. We happen to have a piece of
equipment that will play variable density negative tracks. This is a bad density example. The actual modulation goes this way. These straight lines are not problems. It is actually a form of masking that was
used for noise reduction. So we sent the
print to a laboratory to get a dup negative.
Here is our first generation copy, but now we have a special little
player to play it in, and when we played back the track, on the negative these
light dots now turned black so they don’t transmit light through them, and
hence, we don’t get the noise………….John, play #8…………..Clip…………………….
I think you can see there are
still some slight noise issues, but we have reduced our problem
considerably. Now if we decide to go to
work on this picture for restoration or even preservation, we have made a copy
that has taken out these flaws that were likely processing problem. I have been talking with several laboratory
people, I think somewhere along the line something became deposited on the film
that didn’t allow this film to be properly chemically processed, leaving these
voids in the track. So, we really had an
analog solution to solve this, because we printed it photochemically, and then
our little readers, both analog and digital process, where we can set certain
parameters.
Now we move into what I call the digital processes, and
we talked about what they were. There is
processing equipment, word processing audio records and archiving. What we are interested in in the audio world
is what we do with our digital processing equipment and digital equipment. Those break down into two areas. For processing equipment we have our digital
equipment which will take the information and will be able to do our
manipulation with it. Of course, there is also a wide range of analog equipment
which we can also use. The analog
equipment would be things like filter sets and noise gates and such. In the digital processes, we have our work
stations and our work stations basically have two processes there that are
divergent. One is processing
algorithms. These are written by audio
genius’ that can do all sorts of wonderful noise mapping and bring all of the
noise fores down. Of course, a digital
work station has proven in the post production business is the cat’s pajamas
when it comes to editing processes.
I want to focus on today is look at the editorial area
because these areas really bring in some interesting other issues when we start
talking about preservation and restoration.
We are actually able when we get into the digital world to perhaps
overcome obstacles the original filmmakers encountered and maybe even make their
pictures better with sound tracks. Better
than they were actually able to produce it in the early days of
filmmaking. One of the very first things
we encounter often is something we call
splice bloops. In the early days
of filmmaking, everything was recorded optically and when you made an optical
splice and buttered the two pieces of film together. When that went past the light and photocell,
it made a very nasty, big, hard edged pop.
So they blooped the film and it made a much softer, low and thumpy
sound, and what we have here now is #9…..The
Bloop Loop…….
Generally, these are found
between different dialogue cuts when the show is being edited together, and we
do this to remove bloops in one of our digital editing suites. We use some high end equipment from Sonic Solutions, Tom Regal. Here is what these spliced bloops sound like
in programming. Again, this is where
different dialogue edits are made, not always with the picture edit. This will be #10……………………..Clip………………………..Every change of dialogue
is a cue to you by this bloop thump coming through now. The question comes, do we take the bloops out
or leave them in the picture? We are a
service facility, so we do what our client asks us to do, and I have no opinion
on bloops. They are actually very simple
an painless to remove, and they are done in the editorial workstation by
editing and cutting in a little piece of fill between the tracks, or depending
on if it is somewhere too close to music, we will do an audio interpolation and
the bloops disappear. Bloops weren’t a
problem in the early days of motion pictures because the theater speakers
didn't reproduce these low end frequencies.
Here is what the track sounds like with the bloops out……..Clip………..
Certainly, in the commercial
world of distributing motion pictures those are always removed. From a preservation standpoint, they are an
original artifact of the motion picture.
They are an original artifact of the way these pictures were made, and
so there is absolutely a place for bloops in this world.
I talked about surmounting some of the obstacles that the
original filmmakers encountered, and I brought along an example. Here is a very mediocre pot boiler, made in
Canada right after the war. It has a
flaw that was in the original sound track negative. This is the original nitrate sound track
negative. This is the master sound track
that was mixed down and used to print all of the release prints for this epic
and here is as it sounds in its original form…………………Clip……………………………..
You can hear there is a
complete drop out in the track.. We don’t know how that happened. There is no splice, no damage, and no
break. It is a bad job of recording the
obstacle. Do we want to preserve that
artifact if we are going to restore it?
On the preservation side, you would probably keep of the track just the
way it was, flaws and all. When we go to
restore the track, if we have an opportunity to fix it, we are going to. ……………………..Clip…………………It’s not perfect, but we certainly could convey the sense of what was going on. That is better than having a big hole in the
track where someone made a bad track in 1948.
You can see we start to have creative, or technical, input either to
correct things. It is a very fine line
that requires a very important aesthetic to not just trying to correct all of
the problems and begin working without being certain that is what the project
calls for.
By looking at some of these slides, you can understand
analog preservation problems are still going to require analog solutions. Primarily, it is because the original
material is analog and you are always going to have to be able to access that
material and play it back. You can’t
just dump all your analog gear. It is
absolutely more important today than ever before. Then it is clear and digital and analog processes
are completely complimentary and not exclusionary. We always want to have the open mindsets of
problem solving and be able to jump into either world very comfortably, where
ever the best solution is.
Finally, there is no question that as we move into
archives, we are starting to live in the digital age, and there is a lot of
material that is just being created digitally, strictly and solely that
way. These kind of things are going to
create new challenges and new opportunities as we move forward into the next
century.
I would like to thank the people at Chace Productions who
helped put this together ---Jim Young and Tom Regal, and the Chace Transfer
Dept. That is the basics for sound
preservation in the digital age.
Question & Answer :
We don’t want to alter the basic characteristics of the
sound track. We don’t want a 1930s
soundtrack sounding like 1998. It is
very important to us that that occur.
Generally when we do a sound
re-recording project, the first think he makes is a sound master that is
basically level. It has no processing whatsoever, so it is, in
theory, a direct duplicate of the track we supply. Then we decide where to do additional
things. Since most of our tracks are
mono, they could actually go on the same piece of full code, because they are
separate tracks. We could have the
original flat with some type of de-click or depop thing put in. Then we get a new optical sound track negative
that is used for our new film element.
In the Springtime
for Henry track where we have the micro spotting problem we solved. That is a strict preservation job. The client is going to have us correct the
physical splices in the picture from use of the print. What is the real sound characteristic of this
movie because the client doesn’t want that changed. You then very carefully remove the spots and
deliver back basically the same track with corrected micro spotting which was a
manufacturing defect from 1934.
The facility is in a limestone mine, so the mining took
place from 1902 – early 1950’s and all of the whole, in essence, were there,
about 1000 acres. What we do is to go
back and renovate the space so there is no blasting or dynamiting that takes
place.
The number of our division is 202.707.5840 Public
Services.
We have done some work on the experimental documentaries
that are held at the Academy of Motion Pictures, Arts & Sciences and a lot
of 16 mm film production. We listen very
carefully to the original recording and
try to discriminate between what are the problems caused by age, handling and
use, and what was the filmmakers intent or artistic achievement and not to
change it. We encounter this in current
studio motion pictures where source material is meant to be played off an old
scratchy phonograph disk and that is part of the sound track of the movie. We would go to great pains not to fix
it. We go to great pains to educate our
client and QC depts. That we didn’t fix it and it is meant to be there. For films that have no representatives—it is
not so much an issue to us because we are not a repository for material, so
someone is always hiring us to do work.
They are really the one to tell us what extent to improve the track.
If a client wants us to alter it, we will; and we keep an
extensive documentation at our company.
We actually pride ourselves on is we have analog copies of every job we
have ever done. It is an invaluable
research area within our company, because many times we will get material from
our clients and go to the vault and find 8 years ago, we got something
different and a lot better. And we will call
and point this out. We don’t
provide the documentation that would go out with the however the product will
be stored. Our clients are fully
informed of what has happened, and it is their responsibility to inform the end
user. Sometimes we will prepare
production notes that will go on the DVD.
Lund: There have
been talks about site licensing which would enable us to put up copyright
restricted material. It is especially a
problem with sound recordings. We have
been able to put up very little on the web sites. In the case of site licensing, we would
digitize the material and only provide them to certain locations, various
university libraries. People would not
be able to download them. There are no
eminent plans of this, but the idea has been kicked around. There are a lot of collections where they
have had to go through a lot of trouble to clear rights. If there is a letter, they have to track down
the person who wrote the letter or their heirs and get permission. Our legal
counsel is being very careful about this.
She wants us to be exemplary and not get into any trouble. There is a desire to put up more stuff,
especially in the case of the films, obviously we are only showing things made
before 1922, and we have a lot of wonderful things after that that would give
the site more educational or historical value that we can’t put up. There is also a problem with file size
limitation. •
-- Robert Heiber
Robert Heiber is President of
Chace Productions in Burbank, California.
He has been involved in sound preservation since 1990 when he joined
Chace. He is a member of AMIA, SMPTE,
and ITS and has served on the National Film Preservation Board Advisory Task
Force and the Library of Congress panel for the State of American Television
and Video Preservation. He has spoken on
film sound preservation, restoration and remastering at AMIA, ACVL and SMPTE
conferences. Prior to joining Chace
Heiber was Manager of Technical Operations at Warner Hollywood Studios. He is also an award-winning documentary
industrial filmmaker.