00:00:15.210 - 00:00:27.960

BRYONY DIXON: The BFI was set up, the archive of the BFI was set up in 1935. And it had a twin mission, firstly, to collect the best of the art of filmmaking.

00:00:27.960 - 00:00:47.490

And the other mission it had was to collect examples of contemporary life and manners so social history, if you like, the record of life through film. As you can imagine, people like to collect first.

00:00:47.490 - 00:01:01.400

They like to start a story with its origins. So there was a tendency, always, to collect the oldest films first, partly because they were the most at risk so most in need of archiving.

00:01:01.400 - 00:01:16.370

But also because people like to trace the history back to the earliest days. So we began collecting examples of what we're calling Victorian cinema, so everything made between 1895

00:01:16.370 - 00:01:32.960

and 1901, from the very earliest days of the archive's existence. In terms of the survival of material, I suppose, in 1935 when we began collecting,

00:01:32.960 - 00:01:52.610

cinema was 40 or so, 50 years old. Really not that much time had passed. But, already, most of that material was lost. So a bit in the way that digital material now

00:01:52.610 - 00:02:09.200

is a problem because we've lost most of it already. It's the same with very early film. We calculate, broadly speaking, that 80% of all the films that were made in that Victorian era have now been lost.

00:02:09.200 - 00:02:34.400

So we are dealing with the 20% that remains and trying to extrapolate what the rest of the films would have been like from the surviving examples. Once the early films arrive at the archive,

00:02:34.400 - 00:02:48.140

there are processes we have to go through in order to preserve those films. Film is an inherently unstable medium. It starts to decay from the moment that it's manufactured.

00:02:48.140 - 00:03:04.700

It's a rather complicated chemical sandwich. So you have an emulsion, which is coated onto a kind of plastic base, flexible base. And within that emulsion, you have the silver salts,

00:03:04.700 - 00:03:16.110

and what have you, that react to light, which is what gives you your image when exposed to light. Now, all of these elements are fighting each other all of the time.

00:03:16.110 - 00:03:33.260

So it will begin to decompose. Most film of that era is on nitrate cellulose film, which has its own propensity to decompose. It goes sticky.

00:03:33.260 - 00:03:41.760

And then the image will literally dissolve off the base of the film. It is also highly flammable. So it's a dangerous subject.

00:03:41.760 - 00:03:54.060

So we have to handle it in special ways. So the preservation and the restoration of film is a highly specialised job. We have trained people to do this.

00:03:54.060 - 00:04:10.700

And over the years, we have changed the techniques by which we can preserve and make accessible these films. All film has to be shown on a machine. So it's unlike other museum artefacts, sort of books,

00:04:10.700 - 00:04:23.000

or paintings, or physical objects. It has to be run on a machine. So you always have to copy it. So what we do to preserve--

00:04:23.000 - 00:04:37.310

what we did in the originally was to copy the nitrate film, little short Victorian films, onto safety film stock. So that's stock which is made of acetate

00:04:37.310 - 00:05:00.830

or, latterly, polyester, very stable stocks that will survive in vaults for many, many, many years. Recently, more recently, digital technology has come into play. So now most preservation is done both analogue and digitally.

00:05:00.830 - 00:05:16.040

So now we scan film directly. So we take our little piece of Victorian film. And we run it through a scanner. And this gives us a digital file,

00:05:16.040 - 00:05:29.450

which is great, because you lose no quality as you do this transfer process from one medium to another, which is fantastic. And it's very flexible and very easy to manipulate.

00:05:29.450 - 00:05:49.340

However, you then also have to take great care with your metadata, with looking after the digital zeros and ones on whatever format they're being stored on. And so the metadata and the context that goes with the film

00:05:49.340 - 00:05:58.010

is very important. You can't hold it up to the light anymore and see the information. So it's very-- there's a whole different set of skills

00:05:58.010 - 00:06:13.700

comes into play. But what this has meant is that, instead of projecting a Victorian film to a small audience in a cinema, you can now share these worldwide

00:06:13.700 - 00:06:29.690

at the touch of a button. The important thing with Victorian film, in particular, is to marry up the film, the images with the information, which is the purpose of this project,

00:06:29.690 - 00:06:44.210

because Victorian films, of course, don't come with titles. Generally speaking, they are just little individual short strips of film with images on them. And they need some context and explaining.

00:06:44.210 - 00:07:02.480

So once we've copied a film from film onto safety stock, this gives us a preservation master copy from which we can make copies for people to see in projection.

00:07:02.480 - 00:07:18.920

But we don't throw the originals away. So the Victorian films that were duplicated, sometimes, as long ago as the 1950s, we have retained. And in recent years, there have been

00:07:18.920 - 00:07:33.730

huge improvements in how we conserve those original elements. So we've built new refrigerated vaults. And we can now prolong the life of these original elements

00:07:33.730 - 00:07:48.380

by, perhaps, decades or even hundreds of years by reducing the temperature at which they are stored and the relative humidity. So we now have a brand new master film store,

00:07:48.380 - 00:08:09.330

which is where we hold all of these very precious original films at minus 5 degrees and 37 relative humidity, which means that, when we do things like digital processes, we can quite often go back to these original elements

00:08:09.330 - 00:08:28.290

and get the best possible scan or transfer from that, without losing any quality by this intermediate duplication process. So in many ways, we're at a very good time for the--

00:08:28.290 - 00:08:52.270

saving the preservation of film and for making it accessible for audiences quite widely. So the project to digitise all of the Victorian cinema took place over about five years.

00:08:52.270 - 00:09:09.520

It was part of a big funded project that the BFI was doing. And it was a big ambition, was to do all of the very earliest film. It was quite a complex process, in that

00:09:09.520 - 00:09:24.050

Victorian film doesn't come with its own information, necessarily. This is very early days when film was very experimental. It has no titles or context to it.

00:09:24.050 - 00:09:34.900

There's no film industry to supply information about the film. There are no trade papers or anything. So It's quite-- there's a job of work

00:09:34.900 - 00:09:50.990

to do in identifying the films, before you can go ahead and start to duplicate them and bring them to the public. And so, that, we do in two different ways.

00:09:50.990 - 00:10:09.460

There is the information that is held on the artefact itself, the strip of film so what we call forensic data. Then, there is also filmographic data. So this is data that you might find in newspapers.

00:10:09.460 - 00:10:20.920

Or you might find in journals of the time. Or you might find in the memoirs of filmmakers from their very earliest filmmaking days. So we use both of those techniques

00:10:20.920 - 00:10:37.870

in order to try to date the films. Dating is crucial, because we're talking about the development of a medium. And we want to know what happened in what order, all

00:10:37.870 - 00:10:49.570

these different techniques and developments with different filmmakers. So these are-- this is like the archaeology if you like, of the medium in its earliest days.

00:10:49.570 - 00:11:06.880

And so, you have to develop knowledge over time. Now, the BFI has been doing this for very many decades. So we have a body of knowledge that we can use so we can identify, for example, a Robert Paul film.

00:11:06.880 - 00:11:17.680

We know that Robert Paul's camera-- we have the camera. And we know that he produced films with a very straight frame line edge.

00:11:17.680 - 00:11:30.730

Whereas, if you look at a Lumiere film, it's got a slightly rounded edge. So there are those kinds of things. There are no dates along the edge which, in later film,

00:11:30.730 - 00:11:45.610

you have age codes, which gives you a date that the film stock was manufactured. In the early days, you don't have that at all. So you're reliant on trying to identify different film

00:11:45.610 - 00:12:03.410

companies or film makers by other means, so the shape of the frame, any lettering that you might have. Generally speaking, there's no lettering. But you might have that in one or two cases.

00:12:06.190 - 00:12:22.420

And, otherwise, you're reliant much more on things like filmographic data. So whether there's a provenance, very occasionally, if you're really lucky, you get a provenance where, say,

00:12:22.420 - 00:12:37.150

the daughter of a film maker kept her father's films and donated them to the archive, that kind of thing. A great example of this are the films of Queen Victoria's Diamond Jubilee.

00:12:37.150 - 00:12:51.400

There were a great many films of this event. And they're very difficult to differentiate. So there were 20 companies there. There were 40 cameramen taking film of this event in 1897.

00:12:51.400 - 00:13:06.370

How on Earth can we differentiate between Robert Paul's film, or the Lumiere brothers film, or Apple Patton's film, or Arthur Wrench's film? It's quite tricky.

00:13:06.370 - 00:13:21.260

We do have some records, whereby, filmmakers are remembering, in later days, filming this great event. And they might mention where they were located.

00:13:21.260 - 00:13:37.450

So Robert Paul mentions that he was in the square outside St. Paul's Cathedral. So we can, perhaps, deduce that the films taken from that viewpoints are Robert Paul films.

00:13:37.450 - 00:13:52.300

And then we can double check whether they've got these nice straight frame lines. We can deduce the Lumiere brothers films. These are very easy, because so many people

00:13:52.300 - 00:14:06.790

have copies of Lumiere brothers films in other collections. They also had catalogues with images from the films. So those are easy to tell. Other films are not so easy.

00:14:06.790 - 00:14:24.790

One of the best views of Queen Victoria herself in her carriage was, for many, many years unidentified. And we managed to identify it by a comment from the filmmaker Wrench who says he definitely got

00:14:24.790 - 00:14:36.340

a picture of the royal coach. And he was in King William Street. So we go to Google Street Maps, Street View and just look at buildings.

00:14:36.340 - 00:14:50.080

The great thing about the Victorian era is it's left us this legacy of buildings. So it's quite easy to identify films from that point of view. So now we can be pretty sure that that film

00:14:50.080 - 00:15:03.160

was by Wrench, which is great. So that's the kind of work that we have to do, in order to identify dates and filmmakers. Sometimes, we can do this.

00:15:03.160 - 00:15:18.320

Some will remain unidentified. And there are many examples in the BFI Victorian collection that are still unidentified. And we may never know who made them or exactly when.

00:15:18.320 - 00:15:29.080

But we can tell, from the filmmaking technique, that they are definitely from that era. So we can be confident that they belong in the Victorian collection.

00:15:37.660 - 00:15:54.960

So when we digitise the Victorian film, which is this project we've been embarked on, it's-- there are several processes we have to go through. Once we've identified the film, and we have our metadata

00:15:54.960 - 00:16:09.990

and context, we also need to do the technical work in order to bring it to our audience. Now, this is-- can be a painstaking process, because these films are often very fragile.

00:16:09.990 - 00:16:23.460

They're 120 years old now. And they will be brittle. They will have shrunk. So they, quite often, will not go through any modern machine.

00:16:23.460 - 00:16:42.510

So great news, we now have digital equipment where we can scan film, either frame by frame by frame using a kind of, almost like a rostrum camera process. Or we can put it through a scanner which

00:16:42.510 - 00:17:02.370

will deal with shrunken material and where the perforations down the side of the film may have split or broken and are, in any way, too narrow to go through a machine. So digital technology has really come

00:17:02.370 - 00:17:15.780

to our rescue for this project. And we can, generally speaking, scan material that is very damaged and shrunk and get some kind of image from it.

00:17:15.780 - 00:17:34.350

We can also manipulate the grading of the film, how much light we're getting. So we can make some corrections. And we can compensate for misshapen frames

00:17:34.350 - 00:17:52.920

and some damage and, obviously, the speed because silent-era film runs a whole variety of speed. Some film was taken with a motorised camera and will run a very consistent speed.

00:17:52.920 - 00:18:07.500

Some is hand cranked by the cameraman and might vary as it goes along. With digital technology, we can correct that and conform it, so we're looking at something which

00:18:07.500 - 00:18:22.240

is running at the right speed. With some of our films, because not all films of our era were made on standardised 35 millimetre, we have to use different techniques.

00:18:22.240 - 00:18:33.210

So there's one area of film, the films made by the British Mutoscope and Biograph company, which have to be treated differently. Now, these films are much wider.

00:18:33.210 - 00:18:44.460

They're 68 millimetre. So these films are extremely fragile. Film is very thin, because it was originally made to go in a stills camera.

00:18:44.460 - 00:18:58.650

So it's not very robust. So this film, we have to literally copy frame by frame. So you roll the film across a table. You focus on that frame.

00:18:58.650 - 00:19:11.340

You take an image at a very high resolution. They're very, very exquisitely beautiful films, because there is so much camera information in this large format.

00:19:11.340 - 00:19:26.070

So it's worth spending time and effort using these very laborious techniques to reconstitute the film. And they are some of the most beautiful

00:19:26.070 - 00:19:36.410

of the Victorian films. But it's a lot of effort. It's slow. It's painstaking and, of course, expensive.