[From Journal of Educational Television, Vol. 15, No. 3, 1989]

The Principles Used in Selecting, Editing and Transferring Materials for an Archival Videodisc

ALAN MACFARLANE

Department of Social Anthropolog, University of Cambridge

MARTIN GIENKE

Audio-visual Aids Unit, University of Cambridge

Dr Alan Macfarlane, FBA, is a reader in Social Anthropology at the

University of Cambridge, a member of the BBC Domesday Disc editorial

board, and co-director of the 'Cambridge Experimental Videodisc Project.

Martin Gienke is Head of the Audio-visual Aids Unit, University of

Cambridge, and co-director of the 'Cambridge Experimental Videodisc Pro-

ject'.

p.131

ABSTRACT The article explains the methods used in selecting and transferring

materials for the Cambridge Experimental Videodisc Project on the Nagas of Assam.

The selection and editing of black and white photographs, moving film, photographs of

objects, paintings, maps and sound, is described. The ways in which the selected

materials were transferred to Edit Master videotape are explained. This should be of

value to others who are engaged in (or contemplating) making archival videodiscs for

use in teaching, research or public displays.

Introduction

With funding from the Economic and Social Research Council and from the

Nuffield and Leverhulme Foundations, we have recently completed an experimental

videodisc on the Naga peoples of north-east India. This includes 9000 still images,

150 pieces of moving film, and 72 minutes of recorded sound. Each of these items is

indexed and linked to a textual database. Since we were forced to spend a

considerable time selecting and editing and transferring these materials, it may be

useful for others to know what principles we developed during the course of this

exercise.

Selecting the Materials

Photographs

Since there are a finite number of photographs of the Naga and they are of great

132

value, we have tried to include practically all the photographs we have located. Only

a few hundred out of the roughly seven thousand black and white photographs we

discovered have been omitted. These were left out on the following grounds: they

were duplicates of, or very similar to, other images; their quality was poor; they

were outside the delimited geographical area; they were outside our time span; or

they fell on the side of 'private experience' as opposed to 'public experience'. We

did not censor any photographs because their content was embarrassing or shocking

in any way, or might do damage to the reputation of individuals, the British,

anthropology as a discipline, or for any other reasons.

[FIG. 1. Konyak Naga with headhunter's tattoo. Photograph by Christoph von Furer Haimendorf.]

This latitude obviously raises important issues, given that the videodisc system

is intended for use in a variety of educational contexts. Many of the images from the

colonial era do, after all, portray the people concerned in a way which is objectify-

ing, decontextualising, or exoticising. It is undeniably the case that the camera did

133

contribute to the 'normalising gaze', in Foucault's terms, by which a subtle form of

power was exercised over others by classifying them and making them visible. The

Naga videodisc does not attempt to avoid these issues, but rather hopes through the

associated tutorials and courseware to encourage a critical attitude on the part of

users to the historical interaction of anthropology and administration. It must also

be said that the antithesis to decontextualised images is also contained on the disc

itself, in the shape of some of the earliest, and best, examples of a recognisably

'modern' era of empathetic and contextualised anthropological photography.

[FIG. Konyak Nagas preparing arm decorations for the festival. Photograph by Christoph von Furer Haimendorf.]

Moving Film

There were two problems here. First, we found very little film before 1947 (the end

of the period the videodisc was to cover), but since film is so important, we decided

to break our temporal boundaries and include movie film taken on two visits to

Nagaland in 1963 and 1970 by an anthropologist who had worked in the area in

1936-37. When we included these we had some six hours of film. One side of a

videodisc will hold only 36 minutes of moving film (in interactive mode), and since

we needed to allocate at least six minutes equivalent to the stills (that is, 9000 stills

copied at 25 frames a second), we only had about 30 minutes free. This meant that

we had to reduce the film at a ratio of 1: 12. We spent many weeks whittling away

134

material, trying to minimise the loss of valuable archival footage. The principles we

evolved and acted on are set out below.

The first consideration was the content of the moving images. We tried to

include that material which was most intellectually and academically interesting.

This is a subjective matter, our criteria being that material which portrayed events

and processes which were most representative, most revealing, most unusual, and

illuminated the other still images and texts in the most significant way should be

included. The visual images which were unusual included those which were only

preserved in this medium, for instance the eating of dried rats or a shared joke

between anthropologist and Naga. We concentrated on subjects where movement

and action were important, for instance dance, games, postures and gestures, rituals,

and agricultural labour. If long and repeated sequences of film on the same subject

were present, we selected only one or two sequences.

The second set of principles were concerned with form. Taking into account the

interest of the content, we rejected badly filmed sequences, that is to say the few

sequences which were out of focus, badly composed, unsteady, from too great a

distance, and so on. We rejected film that was damaged, the colour fading, or

otherwise unsatisfactory, unless it was particularly interesting. We preferred close-

up shots of detail in most cases to the wider shots of a more static kind, which could

be preserved with. one or two selected stills. We felt close-ups were more effective

on the intimate television screen.

In considering the selection for the videodisc there was one set of factors which

gave us a considerable advantage over the person who is undertaking normal editing

down from raw film. These are connected to the absolute precise controllability with

normal film where one is constrained to save and knit together reasonably long

sequences, five seconds at least (and often three times that length) just to capture

the run up to something important happening, for instance an immobile man before

he starts to run up to do a jump. Action itself, if it is to be appreciated by the viewer

and happens rapidly, may need to be shown at some length or from several different

angles. In other words, a great deal of redundancy has to be built into normal editing

since the viewer will only see the images flashing past once. With videodisc the user

becomes the editor. One effect of this is that one is not constrained in selecting

shots for inclusion on the disc by the usual film conventions, such as keeping the

stage line, matching shots, and so on.

Another important principle is that with videodisc, there is a sense in which film

that has been shrunk or contracted in the editing process can be expanded again at

the viewing stage. It is as if it had been dehydrated. This is because it is possible to

treat each frame in a precise way. One can take just one shot of a view, or non-

moving group of people, and hold it as an establishing shot for a number of seconds;

if necessary, one can easily go back to the start of the sequence and play it through

in slow motion to explain in detail what is happening, and so on. In other words,

instead of merely cutting a long set of moving sequences into shorter pieces and

sticking them in a precise order, one is creating a set of images, some of them still

frames, some of them sequences of still frames taken every few seconds, some of

them moving sequences. The boundaries between still and moving films blur, and

135

one is able to abstract precisely a great deal of the visual information without

apparently losing much. It should be stressed, however, that there is a different 'feel'

as between a frozen frame and a 'still' filmed sequence. However, the general

principle is an important one: when saving moving film on videodisc there are at

least three forms of compression; selection and editing of sequences, extracting

stills, and using a sequence of still frames.

In this way we created out of six hours of film approximately 150 moving

sequences, each lasting between three and about 25 seconds, the mode being about

eight seconds. We also abstracted about one thousand 'stills' taken out of moving

sequences where there was little movement, or randomly every few seconds to

capture a series of events, for instance transactions in the market.

Having made a preliminary selection so that it would fit within 30 minutes, we

tested the selection by the following method. One team member made the selections,

and then a panel of other people watched all the film through, both film experts and

people knowledgeable about the other Naga materials. As it proceeded, they were

told what was being included and what excluded and asked to comment if they felt

that anything important and interesting was being omitted, or anything unnecessary

was being included. Somewhat to our surprise and relief, we discovered that there

was hardly anything that the panel could find to disagree with. Thus, while there is

inevitably some loss of information, it seems tolerably limited. Since videodisc may

well become a very important archival medium, this is of some general importance.

Objects

Here there is again a problem of selection. There are known to be well over 15,000

Naga objects in European museums and private collections, at least 12,000 of them

in Britain. To have located and photographed them all would not only have

absorbed much of the effort of the team, but the final photographs would have used

up over a quarter of the total space for visual images on the videodisc. It would also

have given hundreds of almost exact duplicates. We therefore decided to confine

ourselves to British collections, and within these to select certain of these, ending up

with a little over 1200 photographs of objects.

One general criterion was relative ease of accessibility. That is to say we

confined our work to certain major private and public collections in England,

though we knew of others in other parts of the British Isles and Europe, not to

mention America and, of course, India and Nagaland itself. From the brief

descriptions of other collections, it did appear that we were able to see and select a

fairly representative sample of objects.

Within the ten thousand or so objects which we either examined in themselves or

through catalogue descriptions, we used a number of criteria in our selection

process. We sought a representative selection, in terms of the types and functions of

objects and their origins in different Naga groups. We attempted to use Naga

criteria of significance rather than our own-that is, to ensure the chosen objects did

reveal the key features of Naga social structure and belief (status, head-taking,

kinship organisation, and so on). Where there was duplication we tended to chose

136

objects with superior documentation, unless the condition of the object was very

poor. We sought to photograph as many nineteenth century objects as possible on

the grounds of their rarity. Such objects also were crucial in throwing light on a key

research interest-the colonial encounter and answering the question, what can we

learn from the types of objects collected about how the Nagas were perceived and

classified as the colonial era developed? We also bore in mind the need to

photograph objects which would be interesting from a comparative perspective-for

example, objects indicative of trade amongst Naga groups or between Nagas and

neighbouring peoples; objects bearing strong resemblances to other South East Asian

hill tribes; objects which, when compared with others, revealed continuity or change

over time. We emphasised objects which we knew would tie in well with other

material on the disc, such as objects collected by administrators or ethnographers

whose writings feature prominently in the textual database. We sought to ensure

that mundane, everyday items were as well represented as the most obviously

aesthetic kind. Where there was obvious interest, we sought to photograph from two

angles, but constraints of time and difficulty in placing objects left this task

incomplete.

Painting and Sketches

Here there was relatively little difficulty with selection. Since the absolute number

was not great (in the hundreds) we included all those pictures which we thought

could possibly be of some interest. Many very simple line drawings were included

when they tied up with textual descriptions, and only very occasionally did we leave

out an illustration because it duplicated something else or was so minor and badly

documented that it seemed confusing to include it.

Sound

The 72 minutes of sound data included on the two tracks of the disc are an attempt

to provide very different kinds of data. We sought to combine recordings of a time

breadth to match the photographs. Thus we included early wax cylinder recordings

(1919) and present-day recordings of songs (1987) with their considerable Christian

influence. Examples of several kinds of instruments have been included such as

drums, jew's harps, stringed instruments, as well as singing. Field recordings of

conversation (1970) have also been included.

Maps

We had hoped that we would merely have to photograph the various maps of

Nagaland from the earliest times up to 1947, including the very detailed Survey of

India maps of c. 1910-45. However, when we experimented with the maps by

looking at photographs of the Ordnance Survey maps on a television screen, we

discovered that this was impossible. However much we magnified the maps by

137

photographing them in tiny sections, the mountainous nature of the Nagaland meant

that all we could see were blurry pictures of contour lines with the odd village name,

almost unreadable, dotted among them. This was because of the low definition and

small size of television screens.

We therefore found that we had to trace and redraw the maps, ending up with

165 sketch maps on the videodisc. All contours were left off, but otherwise they

included all rivers, major mountains, borders and the location and names of some

1400 villages and towns which were mentioned by one or more of our photographic

or textual sources. The maps were mainly based on the ' inch Survey of India maps

of the area. Needless to say all maps and mapping were subject to considerable

errors, compounded in this case by the difficult terrain, the shifting character of

many Naga villages, the immense complexity of village names which can vary

radically from author to author, or even within the same author. The maps are

therefore very much to be thought of as sketch maps.

Transferring the Materials

Having decided what to include the next step was to transfer all the different

materials onto film or audio tape as a first step to making a master tape for pressing

the videodisc. Each of the materials described above needed different treatment.

Here, briefly, is how we dealt with each type, transferring them to the one-inch 'C'

format master videotape.

Moving Film

We reckoned that about 6 of 8 available hours of film would be of interest, and

should if possible be included on the disc in full or in 'compressed' form. The

original film was copied to high-quality videotape, in the process one copy receiving

a burnt-in 'time code' which allowed us to watch the tape and decide exactly which

parts we wanted to include on the edit master videotape.

In addition to making time-coded versions for our editing use, the production

centre provided the professional skills of edit suite operators to do the editing work

of putting the desired clips onto the Edit Master videotape. In our case this was

extremely complicated because it involved not only relatively long sequences of

film, but also up to a thousand 'random' edits (taking, say, one frame in every 25, to

give an idea of what was going on in a sequence without including it in full). If very

complicated editing is being done, extra care must be taken to ensure that the field

dominance selected is absolutely consistent throughout, or else the possibility will

occur of 'Jumping' frames appearing on the videodisc.

Sound

This came in various forms, old 78 rpm gramophone records, wax cylinders, reel to

138

reel magnetic tape, modern cassettes and high quality quarter inch magnetic tape.

All had to be transferred onto the quarter inch tape. Most of this was done within

the Audio Visual Unit in Cambridge, though the wax cylinders had to be copied at

the National Sound Archive, London. There are two sound tracks on the videodisc

so it was necessary to be selective, but there were few technical problems. There

was no usable synchronised sound and film, so there was no need to plan the sound

with this constraint in mind.

Having transferred our tape material to quarter inch tape, this was then laid as a

sound-track (or rather, two sound-tracks) to the final master videotape at the BBC

Open University Production Centre.

Rephotographing Original Stills

As is common with many historical photographic projects, the images we identi-

fied in various collections existed in diverse forms-from glass-plate negatives to

modern colour prints. These all had to be transferred to a single medium.

In the case of negatives, we decided that rephotographing as transparencies

would be far too expensive at the tele-cine transfer stage (given that there were

8000 or so individual images). One option considered was to use a 16 mm film

camera, mounted on a copy stand above a flash light source, through which

negatives could be fed. This arrangement was mechanically quite complicated, and

it proved difficult to focus adequately on the image. We decided instead to use a

half-frame camera mounted on a Bowens Illumitran copier. The Illumitran allowed

for some correction of light balance with each negative, in some cases making it

possible not only to copy but also to improve on the quality of the originals. The

half-frame camera altered the direction of the images. Instead of producing images

on the film lying lengthways, they were inverted through 90', to lie across the film;

and they were reduced to about two-thirds of the usual size. In effect, this produced

a short roll of film which had the images lying in exactly the same way as a movie

film. Indeed one could produce short strips, of usually about 60 exposures, of 35

mm movie film. These strips could then be spliced together to create a single reel of

35 mm film with the images of roughly TV-aspect ratio, easily transferable to

videotape. We used reversal film to copy negatives, which produced high grade

positive images when developed. We were able to do our own developing of these

films, substantially reducing the costs.

In the case of positive images (in our case, photographic prints, pencil sketches,

sketch maps), we used the half-frame camera (with ordinary transparency film)

mounted on a regular copy-stand, using either an electric light source or daylight.

The strips could then be spliced together, as described above.

A specific problem which it is impossible to overcome arises from the shape of

the television screen, that is, its 3:4 aspect ratio of height to width. 'Landscape'

negatives are close enough to the aspect ratios of television to need only a little

pruning at the edges. But 'portrait' or upright originals are bound to come out oddly

on the TV screen. Here we exercised our judgement in the rephotography. If only a

part of the portrait is relevant, one can go up close and fill the screen. But in many

cases it was impossible to take off much at the top or bottom, and consequently the

139

image will only occupy the central two-thirds of the screen, with subsequent loss of

detail. We decided to leave roughly equal (white) borders on each side.

As for reframing the landscape photographs, we did a certain amount of this if

there appeared to be parts of the picture which contained most of the visual

information. In general, however, we tried to be careful not to change the image too

much, since the original composition in itself may be important for future analysts.

It was Important to bear in mind, however, that TV will always tend to crop images,

unless they are very carefully taken within the viewfinder's 'safe' area. On occasions

we bracketed two shots together, as 'whole' and 'detail'. Where we were unsure

about copying of very old or otherwise difficult images, we shot two or three times,

at different exposures; but (as explained below) it may prove impractical to edit out

the unwanted versions.

The half-frame method has a number of advantages, but there are three

potential hazards. One is that if a single mistake is made in the splicing, the entire

reel after that point will be 'out' by that amount, and the resultant videotape master

will be useless. The second arises from this need for care with splicing. For non-

specialists, the less splicing, the better. Therefore, it becomes inadvisable to think

about 'splicing out' all your faulty images, which will doubtless figure here and there

in the half-frame film strips; the resultant videotape will, therefore, include your

mistakes, and indeed your codes for where the constituent strips begin and end

(without these, it may be difficult to relate any individual picture to its index record

when it appears on videotape and ultimately on videodisc). Mistakes will of course

be edited out to some extent in the system software (that is, query-based searching

will never take the viewer to a faulty image); but random flicking through the disc

with a keypad will mean that users will see mistakes from time to time.

Third, attention must be given to accurate centring of the images. In many cases

(e.g. landscapes), this may not be important, but where single objects or portraits

are concerned, it is as well to anticipate a possible slight left bias in the tele-cine

process. That is to say, the tele-cine process expects the left side of any film strip to

contain a sound-track. This will of course not be the case with half-frame film. But

potentially this means that the tele-cine process only scans 90% of the film width,

missing out the 10% of 'audio' to the left, which in our case is very much part of the

picture. The effect is that, as is visible in our disc, the true centre of the original

film appears to be shifted by 10% to the left on the videotape (and therefore on the

videodisc). To anticipate and correct this possible bias, it is essential either to

ensure that full-width scanning will be available to you, or else to centre the original

photographs 10% to the right of true centre, when you look through the viewfinder,

and to ensure that nothing of value remains within the left 10% area.

Photographing Original Material: objects

Our project first concentrated on material culture objects from the Naga Hills. This

is the only part of the disc featuring original new material photographed by the

project. Some of the earlier images on the disc relate to the period when we were

developing the technique and remain somewhat unsatisfactory.

140

We were advised to use natural daylight as far as possible, and we did so in

almost all cases. The exception were a few very large objects which could not be

moved outside or properly lit by daylight. We also decided that, since our aim was

mainly to provide a neutral and precise record of the objects, we should try to take

them all from a uniform angle. This was almost always from above, using a copy-

stand set-up. The exceptions were the few objects which because of their shape or

size could not be photographed in that way. We photographed against a variety of

backdrops which we made. It is essential to anticipate cut-off during later stages of

the transferring process, and therefore to frame the objects in the view-finder within

the 'safe area', leaving a generous border.

We tried to take a photograph of the whole object from the top or front, as

appropriate, with extra photographs of other angles or details or close-ups using a

macro lens, if this seemed necessary. We decided to keep out measurement rulers,

but measured each object so that the information could be available in the index

entry about the image.

These colour slides were then transferred by the process described below, with

'tariffing' (colour balancing) to videotape. Depending on the nature of the objects,

the logistical problems in the museum or archives we were working in, the number

of persons working, it is probably roughly the case that with some help from the

curators in searching for and getting hold of the objects, two people working

together could photograph and index between 25 and 50 objects in a full working

day. This was working under pressure, with relatively portable objects, and assum-

ing no hitches. The desired quality of the end product will of course also affect this

calculation. These colour photographs were developed by the University of Cam-

bridge Audio Visual Aids Unit, while the black and white half-frame photographs

were developed by one of our team.

Transferring Black and White Stills to the Master Video

At the BBC Enterprises Centre, we arranged to have our long reel of half-frame film

transferred to a C-format videotape. A trade-off had to be made between target

quality and cost. It is perfectly possible to have each frame checked ('tariffed') for

colour and contrast, and for considerable changes to be effected, improving the

quality of the originals. But this would cost a very great deal of operator's time (at

,(80 per hour). Instead, we opted for 'batch' control, by which a certain amount of

checking for colour and contrast was done for a series of photographs at a time. It is

at this point that the original quality of half-frame photography matters. Batch

monitoring will only be of much use if the half-frame photography is done carefully

enough to produce a fairly uniform standard of image, without wide variations of

brightness and contrast. If there is little tariffing to be done, this process is

remarkably fast: transferring from film to videotape at 25 frames per second

(standard moving film speed) means that 8000 frames will take about 6 minutes. It

also appears to be the case that relatively minor adjustments in exposure in the

telecine process for colour images, produce a larger difference in the final videotape

rendition than is the case with black and white images.

141

Transferring Original Colour Slides

By opting for a slide transparency for our photography of museum objects, we

allowed ourselves the option of slide-by-slide colour monitoring to improve quality,

during the process of transferring to 1 inch C-format tape. This process involved

some very time-consuming mounting of slides (in our case, about 1000) in special

individual mounts and racks, and was also expensive in terms of operator's time. A

slightly less expensive version of this process is available, in which the transparency

film is fed in as continuous strips (i.e. not individually mounted), but this does not

allow individual frame colour control.

Conclusion

Videodiscs are a very new medium. Their use as archival and teaching aids, mixing

different photographic representations, is largely unexplored. Only when a number

have been placed in educational and gallery settings will we know what advantages

they have over previous media, and what their limitations are. We are only at the

start of this work. However, our work on the Naga videodiscs has highlighted a

number of techniques that may be valuable in the creation of future archival

videodiscs.

Acknowledgement

Particular thanks to Julian Jacobs and Sarah Harrison, without whom this article

could not have been written.

Correspondence: Dr A. MacFarlane, Department of Social Anthropology, Free

School Lane, Cambridge CB2 3RF, United Kingdom.