For large audiovisual archives, the assessment of the condition of the items in an archive typically begins with an audit of what carrier types are present. The condition of items for each of the carrier type is then determined, often through physical examination of a subset of items of that carrier type.

This information is vital for cost estimation since the cost of digitising clean and playable material differs significantly from that of damaged items [The BBC estimate that it costs 5 times more to transfer an item if it doesn’t playback first time].

Assessment of the overall condition of an archive based on sampling of the contents obviously depends on extracting a representative sample from the collection. Condition of items is typically classified into several levels, for example the five stages of nitrate film decay or the four levels of acidity indicated by A-D strips for vinegar syndrome monitoring. Whilst in reality the classification of condition levels is specific to each carrier type, this report takes a more general approach and introduces a simple five level classification.

If the sample is representative of the archive as a whole, then these percentages can be used as statistical predictors that give the probability of an item taken from the archive being in a particular condition.

The figures derived from a sample are a snapshot of the current state of the archive. The real challenge comes when planning a preservation project in advance, perhaps over ten years, when a projection is needed of what the archive condition will be as a function of time.

The approach we take in the report is to model how items in a collection might move from one condition to another over time. The reasons for items moving between condition categories are varied. For example, the process might correspond to the effects of degradation due to wear and tear. Items move from playable to dirty as a result of initial use.

A statistical degradation model for a specific archive would be built up from sampling over a period of time would be used to populate this model for a specific archive.

For the purposes of the report, we assume that some information is available on the general rates of decay at a broad level and that these figures can be assumed to hold over the lifetime of the planned preservation project. At first sight it appears to be a big assumption that the overall decay rates can be approximated as being constant over time, especially for a long period! However, it is possible that this assumption does hold for large archives.