Processes for creating the digital archive

The process of creating the archive can be broken down into several sub-tasks – audio preparation, the creation of metadata, and indexing this information in meaningful ways.

Audio processing

The Gerhard Revealed project resumed in 2020 with the goal of creating a digital archive of Gerhard’s tape archive. The first task was to assess the condition of the digitisations of the tapes made in 2012. These digitisations were made using the software ProTools, a cutting edge digital audio workstation of the time but in the subsequent decade the version these recordings had been made with no longer operated on current machines.

What’s more, contemporary versions of the software no longer supported file formats that old. As such, the raw audio files were imported into an alternative digital audio workstation called Reaper, a versatile, low cost, open source tool.

The next task step was deciding what the audio should look like. A purist approach to archiving dictated that the archive should contain only the unedited contents of each tape. However, we were conscious that the final format of these audio recordings might not satisfy most listeners who wanted to listen to examples of Gerhard’s work. There were often large periods of silence between sounds, or sections of a tape that were recorded at one speed, only to be followed by a section at a completely different speed.

Similarly, the audio quality was something of a concern – disruptive noise and clicks were present throughout recordings, as well as strong mechanical resonances that Gerhard himself would not have been able to hear on his equipment at the time, but which we can hear very clear now with the advances of speaker technology. All of this presented a possibility to putting people off from engaging with the archive, rather than welcoming them into it.

As a result, we decided to strike a balance. We kept the original digitisations, an as close as possible rendering of the original tape recording as possible, but then also made another, edited version of each recording that smoothed out the listening experience, removing silences, noises, and correcting for changes in speed. These two versions of the audio are clearly marked in the archive as either ‘edited’ or ‘original’, and as such allow different audiences to engage with the archive for different purposes. The digital nature of the audio files allowed for this variation in a way not otherwise possible.

Metadata

Similar to the audio, the metadata collected for the original digitisation was formatted in an application that no longer operated on today’s technology. Thankfully, there was an alternative version of this data, backed up as a PDF file. As such, the team had to painstakingly recreate the archive metadata from this PDF document into a more flexible file format. The metadata was reassembled in an excel spreadsheet, and saved to a .csv file, a file format that over the past ten years has persisted as a robust data format.

Additional metadata was collected from the physical materials where possible, and restructured to fit in accordance with the digital archiving structures offered by Heritage Quay. In assembling information we made sure to attribute where different markings came from, whether made by Roberto or added later by his wife Poldi.

Indexing

Finally, in assembling the digital archive, we set about the task of indexing each tape, so that it was possible to find tapes immediately according to a number of different categories. As well as the linear digital archive, we conceived of three meta-categories: people, works, and kinds.

The first covers all the name that occur in the digital archive, either written on the packaging or acknowledged on the recording itself. We separated these collections of people into distinct subcategories, including composers, orchestras and ensembles, conductors, and performers.

The works index catalogued tapes according to the category of individual creative works, with subcategories collecting works by Gerhard, works by other composers, and works produced for or recorded from the BBC radio.

Finally we indexed the archive according to the kind of work on the tape. We broke each tape down into subcategories of the work type (electronic, acoustic, ensemble, orchestra, and so on), its status as a world, UK or international premiere, and then the type of recording on the tape, whether it was a radio broadcast, a rehearsal, a sketch or assemblage recording, a fragment of a piece, and so on.

In providing an index in this way, visitors to the Roberto Gerhard Digital Archive didn’t need to rely on their own knowledge of Gerhard to find meaningful information. Instead they could look through the available options and find materials that looked interesting to them.

Over to you

What do you think sounds like the most challenging process of constructing the digital archive? What other categories do you think could be used for the indexing process?