There has been enough films digitally restored in recent years to prove application of digital technology, however most such films have had an immediate, and paying, audience ready to cover the cost of the restoration process - Star Wars, Fantasia, for two obvious examples.

So, this opens the obvious debate as to what the real goal for film restoration is?

Is restoration to be embarked on for future archive reasons, producing a true to original restored master, restoring and preserving material that may otherwise deteriorate further before it sees the light of day?

Or will restoration only be embarked on when there is an immediate and chargeable deliverable requirement that will finance the cost of the work?

Experience of restoration projects that Light Illusion has been involved in suggests the second scenario is the only realistic model available within which restoration has occurred, with companies aiming for DVD release of the restored films, sometimes with limited theatrical distribution too, such as with George Lucas's THX 1138.

For example, The Criterion Collection, which is a USA based company that specialises in DVD releases of classic art-house films, obtains distribution rights for restricted market areas where they can be certain of sales success, and pay for restoration with the single aim of producing a DVD release for immediate sales distribution. Their requirement is not for a long-term archive restoration master, but a limited time-span DVD release. A very sensible business model - and the people involved in Criterion are very nice too, which helps!

Therefore, Criterion is not looking for the ultimate in restoration, by which I mean not restoring the film to a level equal to the original film resolution (4K nominally), but are performing the restoration to a level necessary for their immediate requirements. As a result the restoration level factor is a variable that needs to be understood when designing restoration systems based on the needs of the restored footage.

Additional research of Film Archive environments has also shown that the life expectancy of any archived film (cellulose nitrate, cellulose triacetate - safety film, polyester) can be increased by orders of magnitude, to 100's if not 1000's of years, at ultra-low temperatures and humidity. The temperature required is sub-zero (-2°c) with relative humidity in single digits. With this approach archives are able to continue to preserve and conserve the film held in their care, until such time as better digital alternatives become available. However, such an approach only maintains archived film at its present level of deterioration, and does nothing to answer the question of access to the archived images, or the generation of revenue.

There is also the issue of the resolution necessary for restoration, if performed for long-term archive reasons.

2K has become the defacto standard for Digital Intermediate work performed on new film features, as supported by Light Illusion as the ideal resolution, but there is pressure to use 4K, or even greater, for restoration work if the goal is future-proof archive of restored material.

The reason for this interest in 4K, or greater, resolutions comes from the understanding of celluloid film's relative abilities to hold detail, especially in negative form.

Final print film, as watched in theatres the world over, has an equivalent digital resolution of no more than 1.2K pixels, hence 2K being acceptable for DI on new feature releases. The original camera negative for such features is usually archived for future use, should the 2K deliverable prove to be too low resolution for future applications. However, no one has ever provided a satisfactory answer to the question of vfx shots that don't exist on original camera negative?

When it comes to material already held within archives there is obviously a reluctance to 'drop' resolution to 2K for long-term archive, when the original camera negative holds the equivalent of 4K, or more pixels.

But, if restoring for long-term archival, what of the long-term stability of today's digital storage formats???

A well stored roll of film can be expected to last 10's of years with little degradation within today's archives, and 100's (or even 1000's) of years if the latest suggestions for temp. and RH are followed.

Today's digital storage mediums are not capable of hitting double digit years without the need to carry out rejuvenation copying. Archive's simply cannot afford the cost of such processes, and if part of a digital archive fails the image is lost for good, unlike the progressive failure of film.

What then, is to be the long-term digital archive format and how will it compare to traditional film archive? Today, there is no answer...

I guess we'll just have to continue waiting to see those more obscure archived 'classics'?