The Michael Gerzon Archive and the British Library’s Archival Sound Recordings Project

by Paul Wilson and Adam Tovell

Abstract

This article presents an outline account of an ongoing (as of December 2008) UK project to digitise, preserve and disseminate access to the Michael Gerzon Archive, a collection of original concert recordings (with supporting documentation) held by the British Library Sound Archive (BLSA). Its primary objective is to serve as a guide, or case study, for institutions contemplating the future of comparable collections, or already embarked on similar projects, and to summarise some of the solutions devised in response to problems arising from the unfolding project processes to date.

It begins with a summary of the institutional background: the BLSA, the Joint Information Systems Committee (JISC) Digitisation Programme and the wider objectives of the Archival Sound Recordings Project. Part 2 examines the history of the Michael Gerzon Archive and its originator’s recording methodology; then considers the collection’s potential educational value as both an online audio resource, and as a unique document of late 20th Century recording practice and cultural life. Part 3 is an exposition by the project’s Audio Engineer, Adam Tovell, of the mass digitisation processes deployed in the audio transfer, file creation and validation stages, along with an overview of the software, media and equipment involved. Part 4 summarises other key ASR project processes in relation to the Gerzon Archive component: content selection, cataloguing and metadata creation, rights clearance and the service interface, concluding with a brief look at future plans for the resource, and for the collection as a whole.

1. Institutional and Project Background
2. The Michael Gerzon Archive
3. Audio Transfer, Digitisation and Media Presentation
4. Other Project Processes
Notes | Biography

1. Institutional and Project Background

1.1 The British Library Sound Archive

The BLSA began life in 1955 as the British Institute of Recorded Sound, initially a one-man operation envisaged by its founder, the late Patrick Saul, as a solution to the fact that no British library or museum then accepted responsibility for collecting published sound recordings. The Institute later grew into the National Sound Archive and had by 1983 been absorbed as a department of the British Library. Today the Archive is one of the world’s largest, with millions of individual recordings, published and unpublished, on formats ranging from wax cylinder to solid-state drive, with particular strengths in music (of all kinds), drama and literature, wildlife sounds, radio broadcasting and oral history. The collections have been assembled from an array of sources: record company donations and purchase, radio stations and off-air recording, ethnomusicologists, musicians, zoologists, specialists and enthusiasts, with thousands of field, concert, theatre and speech recordings added by the Library’s own staff.

The Archive is charged with the fulfilment of three fundamental roles: the collection and preservation of the nation’s audio heritage and the provision of study/research access to it. In common with similar institutions around the world, the acquisition of worthy material has historically tended to outstrip its available resources for the essential conservation and cataloguing work, while the provision of access has, for copyright reasons, traditionally been restricted to onsite listening facilities; an arrangement which students and teachers outside the capital (i.e. the vast majority) increasingly find unsatisfactory.

Whilst substantial progress has been made with the preservation of fragile early media, such as cylinders and acetate discs, the Library is now also forced to address the urgent conservation needs of much more modern media, notably defunct formats such as Betamax videotape — previously extensively used by the Library in conjunction with PCM encoding technology to digitally record radio programming, for example — and DAT. While this work has been aided since 2007 by the opening of the Library’s custom-designed Centre for Conservation, with specially equipped audio conservation/digitisation studios, a combination of constantly expanding collections against a backdrop of static or diminishing budgets compels national libraries and archives to seek external support or funding wherever possible.

At the same time, the educational sector which such institutions partially serve is now increasingly voicing demands for access to locally accessible (i.e. online/digitised) primary research materials, including audio, to support teaching and research. Whilst online delivery of digitised audio is now technically relatively straightforward, it continues to be hampered by the thorny matter of intellectual property rights and the fact that libraries, archives and educational institutions rarely own the rights to the material in their possession. By the early 2000s it had become apparent to the higher education community that these issues required national coordination and government backing, and a programme of stakeholder consultation was initiated to plot a collective way forward.

1.2 JISC and Digitisation Strategy in the UK

These developments led to the establishment of the JISC Digitisation Programme, with the primary aim of “unlocking unique, hard-to-access material” within UK collections for the benefit of the UK academic sector. JISC also provides guidance and leadership in the innovative use of information and communication technology, and distributes funding for the creation and delivery of online digitised resources from these previously inaccessible collections of audio, film, photographic and printed media. The programme became a cornerstone of the UK government’s higher education and research policy, which stressed the importance of digital content provision for the next generation of education and research. In addition to this core objective, the programme also aims to foster new institutional partnerships, attract investment to the field, and to establish digitisation standards and guidelines ensuring interoperability and compatibility within the wider education and research communities.

Crucially for the British Library, in addition to serving JISC’s educational access objectives, the project was also designed to serve the wider archival and conservation objectives of the BLSA, which require the parallel creation of archival quality preservation surrogates in addition to the (reduced resolution) online versions. These archival copies, digitised and managed according to the Library’s preferred archival standards, are permanently preserved within its Digital Library.

Needless to say, there is stiff competition for these funding resources at a time when traditional library services are being forcibly reconfigured and repositioned within the wider electronic information sphere. Despite this, the British Library, with an unparalleled body of primary research material to draw upon, secured JISC support for a range of innovative projects, including the digitisation of 17th–19th Century newspapers, UK theses, and the ongoing Archival Sound Recordings (ASR) project of the British Library Sound Archive.

1.3 The Archival Sound Recordings Project

Human resourcing for the project, which followed the constitution of a Project Board, required a combination of specially recruited (temporary) project staff, the Library’s existing technical staff and specialist curators, its Web Services Delivery Unit, and the contracting of various external services/personnel as necessary.

Phase 1 of the project (2005–07) was essentially complete in 2007, with the dedicated ASR website now providing 160 UK colleges and universities with licensed access to an initial batch of almost 4,000 hours (around 12,000 individual recordings) of audio, much of it unique, from the Library’s collections. The selected content packages include themed oral history collections, ethnographic field recordings, African literature/arts radio programming, soundscapes, and a collection documenting the evolution of classical music performance practice. These were later augmented by a further 1,500 wildlife sound, UK dialect and early cylinder recordings, which are being made universally available for unlicensed streamed online access.

Phase 2 (1) which commenced in 2007 and due for completion in March 2009, augments the Phase 1 content with an additional 6,500 hours (31,000 individual recordings) of audio content, including interviews with Holocaust survivors, recorded arts discussions and talks, reproductions of early record catalogues, African popular music of the 1940s–50s, English traditional music and the Michael Gerzon Archive, which we examine in more detail in Part 2.

The many project processes are too complex to detail here but the key stages can be summarised:

• planning and project management;
• content selection and user consultation;
• funding application process;
• audio transfer and file creation;
• cataloguing and metadata creation/augmentation;
• workflow and quality control;
• rights clearance and licensing;
• access management system;
• development of the service interface and content hosting;
• marketing and user feedback;
• maintenance and future expansion of the service.

Details of these aspects of the project can be explored in documents at JISC Digitisation Programme and Archival Sound Recordings webpages.

2. The Michael Gerzon Archive

2.1 Who was Michael Gerzon?

Michael Gerzon (1945–1996) was variously described in his obituaries as “one of the audio industry’s greatest thinkers and writers”, (2) “one of its most prolific polymaths”, (3) and even — put most simply and without embarrassment by Richard Elen — as a “genius”. (4) The respected author of over 120 published papers, many on aspects of audio theory, sound recording/reproduction and technology, he was also a postgraduate student of axiomatic quantum theory, expert in psychoacoustics and auditory perception, leading contributor to the development of ambisonics and co-inventor of the Soundfield microphone (the first capable of directional three-dimensional recording), along with a raft of designs relating to digital audio, signal processing, systems theory, noise shaping and many others. He died in 1996, aged 50, from complications arising from acute asthma, a condition he battled against for most of his life.

Perhaps less well know, until now, was his unflagging near-lifelong interest in recording live music, the legacy of which is one of the subjects of the British Library’s ASR Project.

2.2 History of the Gerzon Collection

Commencing around 1966, Gerzon, a Mathematics undergraduate at Oxford University began making stereo quarter-inch tape recordings at chapels and performance spaces around Oxford and the University campus. These predominantly feature respected local choral groups and early music ensembles such as the Schola Cantorum of Oxford, Oxford Pro Musica and the Oxford Early Music Group. At the same time he began a programme of off-air recording (mostly BBC Radio) of studio or concert performances of works by those modernist twentieth century composers who most interested him: Stockhausen, Webern, Varèse, Bartók, Berg and Stravinsky, along with virtually every new music broadcast by the BBC Symphony Orchestra under the direction of Boulez.

During this period (late 1960s–1976) he also joined the Oxford University Tape Recording Society (OUTRS), which tested their theories concerning stereo sound recording, quadraphonics, microphone technique, surround sound and many others, with practical experiments and concert recordings, (5) many of which survive in the BLSA collection.

Between 1976 and 1981 there was an apparent lull in Michael’s concert recording activities, although his off-air recording continued and indicates a broadening of his musical interests to include free jazz, new wave and avant-pop/rock, and especially free improvisation. A return to occasional recording in 1982 was followed by a surge of activity from 1984 onwards, coinciding with the release of Sony’s acclaimed WM-D6C Professional Walkman, which must have been much easier for Michael — always rather frail on account of his health problems — to carry about.

On an autumn 1984 trip to New York he first used the new PCM technology to record digitally onto Betamax video, documenting in full WKCR-FM’s La Monte Young Festival, which aired dozens of private recordings and little known works for the first time. The following year, realising the potential of the four separate high quality audio channels (digital stereo on the picture track, analogue stereo on the Beta hifi tracks), and undeterred by the difficulty of synchronising the two, Michael began making dual-stereo (or “4-track”) recordings in which the main stereo pair was typically laid down on the digital track, while a pair of spaced PZMs or PA feed (for example) was recorded to the analogue. This allowed the possibility of subsequent remixing, although he seems to have used it sparingly, primarily to correct problems with the stereo image or to enhance intelligibility when poor PA equipment was being used.

In 1988 he also began making DAT recordings using the tiny portable Casio DA-1, usually in conjunction with his preferred near-coincident stereo pair, and the following year also acquired a Yamaha 4-track cassette recorder to record similar configurations to those deployed in the 4-track Betamax recordings, a format he then dropped. Most of these recordings were made at Oxford venues such as the Jericho Tavern, Old Fire Station, and the Holywell Music Room, but he also made frequent trips to London, sometimes recording at two different venues in a single evening before returning by bus to his Oxford flat.

Michael’s concert recordings began to dry up in the early 1990s, although there are occasional recordings through to his year of death, perhaps because he was professionally very busy in this period and researching/writing many of his published papers.

One surprising aspect of the collection now at the British Library is the paucity of surround matrixed recordings, considering Michael’s central and long-term role in the development of surround sound theory, ambisonics and the Soundfield microphone. This was possibly due to the fact that advances in surround recording technology had not then been matched by suitable domestic reproduction equipment, so the creation of hundreds of surround matrixed recordings would have constituted an unacceptable risk. Whilst there are a few matrixed recordings within the collection, which Adam Tovell discusses in Part 3, it seems likely that the results of his ambisonics experiments were retained by (or later given to) associates for further study. The vast majority of his concert recordings were made in stereo, or in the dual-stereo / four-track format he adopted for his Betamax and four-track cassettes.

2.3 Recording Methodology

Gerzon has been unfairly characterised by some as a science and technology geek who obsessed about technical accuracy above all else. They may therefore be surprised to read his definition of the ideal for live amplified music recording as an opportunity to “capture an idealised version of the live sound in which the inherent technical defects and unwanted losses of intelligibility are reduced and rebalanced for domestic playback without losing those virtues of the live “ambience” and “atmosphere” that create the unique “feel” of a live music performance.” The methodologies embodied in the Gerzon Archive represent his solutions to what he perceived as the basic problem of live music recording: “the almost mutual exclusiveness of the two main desiderata : good technical quality and good “live” atmosphere.”

Invaluable to any study of Gerzon’s recording methodology is his paper “Stereo Recording of Live Amplified Music,” (6) in which he stressed the importance of achieving this balance — so often missing in bootleg audience recordings, or faked in multi-tracked record company recordings. Gerzon accomplished this using techniques based in part on his own research into psychoacoustics and directional perception of multisource sounds. His overriding concern was not with technical or theoretical accuracy, but with intelligibility, and the “listener’s ability to hear subtle details of quiet sounds in the presence of loud ones.”

His by no means dogmatic approach recognised the acoustic limitations of most performance spaces, and in particular the ubiquitous problem of multiple sound images created by stage set-ups combining acoustic instruments, electronic instruments with onstage amplifiers, PA speakers and foldback monitors. (7) These and a host of other unpredictable variables threaten the success of the best prepared music recordist on location, yet the Gerzon Archive confirms that Michael had a clever solution for virtually every imaginable scenario.

Despite this, his methods often appear deceptively simple, the most frequently deployed being an adaptation of Blumlein’s coincident-pair technique: a pair of Calrec CM1050C cardioid (not figure-of-eight) microphones at an angle of 120 degrees, with a 5 cm spacing of the microphone capsules. But the collection exhibits many other approaches, including four-track combinations of near-coincident and spaced microphones (such as PZMs), or a stereo pair mixed with the feed from a PA desk. The technical documentation accompanying virtually every recording makes each one a valuable model for the student recording engineer.

Although his primary objective seems to have been the achievement of a satisfactorily intelligible stereo image through appropriate technique and microphone placement, Gerzon also occasionally applied post-processing to flawed recordings, allowing the student engineer to assess his application of such rarely encountered techniques as “stereo shuffling”. (8)

2.4 Cultural and Educational Value

The Gerzon Archive is culturally significant on a number of levels — musically, historically, and technically — and represents a unique opportunity to examine the practical work of one of the 20th Century’s leading audio scientists through one of Europe’s most extensive extant collections of unpublished concert recordings. It broadly documents much of the creative music of two cities over the course of three decades, adopting an unusually objective and inclusive approach: medieval plainchant, choral music of the Renaissance, twelve-tone composition, electronic and electroacoustic music, afro-jazz and free improvisation, psychedelic and minimalist, avant-pop and new wave rock, folk, and many others. The labels were probably of little relevance to Michael, a man of catholic but educated taste, who seems to have been solely concerned with artistry and creativity.

Whereas the collection provides a selective series of snapshots from the left-field pop/rock scenes of the 1980s and 90s, it offers a much more comprehensive account of the development of free improvisation in the same period, particularly the scene centred on Oxford’s Jericho Tavern. It also testifies to the interactions between the local scene and the more interconnected musical milieu of the capital. The opportunity to witness certain individual performers, such as Pete McPhail and Matt Lewis, developing their work on an almost weekly basis is particularly unusual. We get glimpses of success and failure, defined on the musicians own terms: a private recording session abandoned because the performer felt distracted and uninspired; or enrapturing moments of success: electrifying bass and voice duos by Barry Guy and Vanessa Mackness; and the hypnotic alien worlds of all-women ensemble Circle Cycle.

The authentic acoustic honesty of Gerzon’s recordings makes them ideal sources for the study of electroacoustic ensembles, such as AMM, who weave into their performances the ambient sounds that many concert recordings aim to exclude. Likewise the choral music, captured with the natural acoustic reverb of a Gothic chapel, and not conjured from the soundboard of a mixing desk. Then there are opportunities to explore the role of venues and audiences: the warmth and intimacy of the Holywell Music Room, where the musicians felt at home; or the overt hostility elicited by an improvising ensemble at the first (ill-conceived) Oxfordshire Jazz Festival.

Many of the recordings feature one-off associations, short-lived groups, and performers who were never documented in published recordings. But there are also early performances by groups who later became commercially successful: Shake Appeal (who became Swervedriver), My Bloody Valentine, Talulah Gosh.

The collection is also a rich seam for the study of recording technique in relation to both unamplified and amplified acoustic music, mixed acoustic and electronic ensembles — with and without PA systems. The effects of using different types of microphones and recorders can be compared, or of mixing microphone and PA feeds together — in different performance spaces and for different types of music; or the blending of spaced and near-coincident microphones, for example.

3. Audio Transfer, Digitisation and Media Presentation

The Michael Gerzon Archive is the work of one of the most prolific and dedicated recordists in the field of live engineering, consisting of thousands of hours of recordings made over a period of thirty years. As a subject for archival, the transfer of so much audio on a traditional, “one-to-one” basis would require a commitment on an equal and ultimately impractical level. Thankfully, current technology now allows archival-quality audio transfer to take place on a scale previously impossible. Migration of the collection from obsolete carriers to secure, digital files was therefore one of the first — and certainly the largest — to be undertaken by the British Library Sound Archive as part of a programme of mass-digitisation, with audio from up to four discrete sound sources ingested and archived simultaneously.

3.1 Content

While concerns of efficiency and project deadlines were certainly a factor in its selection as a candidate for mass-ingest, the Gerzon collection is notable for both its diversity of, and — conversely — its uniformity within formats. Gerzon’s recorded output was committed to a variety of tape-based media; from analogue 1/4" open-reel recordings and both two and four-track cassettes, to digital recordings on Betamax and Digital Audio Tape. Although diverse in content and form, the quality and structure of his recordings remain largely uniform throughout. As such, while the initial selection and acquisition of hardware for archival proved challenging, the project was aided by an overriding homogeneity seldom seen in personal archives.

The unprecedented level of detail that he, as an engineer, documented alongside each recording is similarly unusual. A trend begun as a member of the OUTRS and upheld throughout his recording career, the meticulous cataloguing of metadata — both technical and otherwise — was as much an element of Gerzon’s engineering as his accurate and considered microphone techniques. With each tape accompanied by detailed descriptions of recording equipment, media and methodology, his documentation leaves little doubt as to his intentions as an engineer, providing valuable support in the decision-making process involved in the collection’s transfer to stable and lasting digital files.

Equally remarkable is the comprehensiveness of each recording in the collection. While the constraints of time, tape lengths and cost could reasonably limit any live recording to the more “vital” components of performance, the majority of Gerzon’s recordings were made with a view to retaining every possible element of the event: from pre-performance rehearsals and sound-checks, to post-performance audience noise and ambience. Moreover, most of his favoured tape stock was evidently selected for its quality, rather than cost or availability, and he largely succeeded in avoiding those 1970s to early 1980s brands which later succumbed to “sticky shed syndrome”. This philosophy of quality from the base elements of the engineering process upwards was extended not only to his recording techniques and meticulous note-taking, but also to storage; all his media being wound and packaged with the utmost care to ensure its long-term integrity. Print-through, a relatively common issue with open-reel archive material, proved to be rare here; possibly due to his storage of tape “oxide-out” on each feed spool (a theory outlined in his 1972 article “Minimising Print-Through,” written for Studio Sound magazine [9]). As such, while an often significant issue in the preparation of media for mass-ingest, pre-transfer care and treatment (tape baking, re-spooling, cassette re-housing, cleaning, etc.) was largely unnecessary.

3.2 Software and Standards

The transfer of the analogue contingent of the Gerzon Archive was one of the first projects at the British Library Sound Archive to use Quadriga as an archival solution. Designed with a view to ensuring archival quality in transfers made as part of a programme of mass-ingest, Quadriga facilitates the input and monitoring of four discrete analogue sources, while producing XML-based quality reports alongside each digitised file; in essence, adding an extra pair of ears to the transfer engineer’s own in the assessment of audio quality — warning of any instances of significant audio “events” (clicks, drop-outs, hum, azimuth errors, etc.) according to a system of pre-defined threshold levels. Further to this (and unlike more conventional multi-track recording systems), Quadriga allows the engineer to save files and change tapes during digitisation without halting concurrent transfers; potentially increasing output to a far more efficient rate, and ensuring accuracy throughout the transfer process.

While the increased throughput resulting from mass-digitisation as an archival solution is of considerable significance, the Gerzon Archive was, above all, treated as any other archival project; with all audio from analogue sources captured as high-resolution 24-bit, 96 kHz MBWF files (enabling file sizes to exceed the 4 GB limit inherent in the standard WAV format, with coding histories and technical metadata stored in the bext chunk of the MBWF header), and full-resolution sample rates and bit depths maintained in the transfer of audio from digital media. Of note here is that the collection has been digitised not solely with a view to preservation, but to facilitate online access for remote users. As such, each of the archival “master” files produced in the transfer process is accompanied by an edited, high-resolution “playback” file; produced through the removal of unnecessary silences and the repair of any correctable flaws in the transfers themselves. Public access is then granted from a lossily-compressed version of this playback file; either in MP3 for downloading or in WMA for streaming from the Archival Sound Recordings website. Given a potential transfer rate of around 24 hours of audio per day, alongside the production of playback and access files, the storage required for such a quantity of data is necessarily vast, and the demands on maintaining file integrity equally significant. Each file was therefore MD5 checksummed at the point of ingest or creation, and backed up simultaneously to an external USB hard drive and networked 24 TB RAID array. Linking each of these files to the British Library’s catalogue is a file-naming system based on 24-string unique file identifiers — comprising of strings derived from the audio object’s curatorial area of origin, accession numbers, channel configuration and file status. Metadata and cataloguing information is then linked to each unique filename, and accessible both internally and by the end user.

3.3 Equipment

Given the regularity and diversity of his recordings, Gerzon’s equipment selection represented a necessary balance between professional specifications, and, ultimately, convenience. Here, while those recordings made under the auspices of the Oxford University campus and as a member of the OUTRS were seemingly bolstered by the local availability of high quality equipment and the luxury of a relatively confined locality, his recordings from the early 1980s onwards were made in a variety of clubs and concert venues in both Oxford and London. The majority of this location work was captured on high-quality portable equipment; such as the Sony WM-D6C Professional Walkman Recorder (a justifiably well-established favourite amongst concert recordists due to its portability, wide frequency response, signal-to-noise ratio and stable tape speed [10]) along with an CM-30 single-point stereo microphone and Dolby B noise reduction — with often remarkable results. Likewise (and in line with his apparent enthusiasm for fidelity and new audio technologies), Gerzon was quick to adopt one of the earliest-available portable DAT recorders, the tiny Casio DA-1, which he used in conjunction with a high quality cardioid stereo pair.

While it was possible in most cases to employ modern, professional replay machines to ensure the best possible signal extraction from the original carriers during archival (in this case, four identically-aligned Studer A807s, with manual azimuth adjustment and RS232 remote control, and an equal number of modified Tascam 122 MKIII cassette decks with manual azimuth adjustment and 25-pin parallel remote control — each feeding four Lake People F444 Analogue to Digital Convertors), the Gerzon collection contains a number of formats which presented a necessity to deviate from “accepted” archival methods.

For example, the four-track cassette: a largely obsolete format in the field of audio recording, an uncommon medium in the British Library collections, and one for which no precedent in terms of archival at the BLSA had been set. While four-track cassette recorders were relatively common as a domestic recording tool from the mid 80s to the turn of the century, their proliferation in a competitive market resulted in a number of diverse and largely incompatible designs; with noise reduction systems, tape speeds and pitch controls varying between (and occasionally within) manufacturers. Sufficiently high quality transfers of the four-track cassette material with compatible equipment were therefore possible only by returning to equipment contemporary with the recordings. Here, Gerzon habitually used two four-track cassette decks; the first a Yamaha MT44, offering “standard” recording and playback (at 1-7/8 ips) with both Dolby B and C noise reduction, and a Yamaha MT100 with “double-speed” playback and recording (at 3-3/4 ips) and DBX Type II noise reduction — with both machines offering adjustable pitch/speed controls. Happily, four-track cassette machines may still be found relatively easily; and both a suitable Yamaha MT44 and MT3X (the latter’s features entirely compatible with the MT100) were eventually acquired second-hand and — following a thorough service — produced excellent results, with four synchronised mono signals taken from the line outputs of each machine, and digitised as two separate stereo stems (in keeping with the typical track configuration of microphone pair and PA feed found on most tapes), and the resulting audio streams time-stamped and saved as discrete, high-resolution digital files, allowing their potential combination as a transparent representation of the original recording.

3.4 Digital Media

Ever an ardent supporter of new technology, Gerzon was keen to exploit the potential of new media; a trait evident in his use of Betamax as a digital recording medium during the late 1980s. While the use of consumer-grade video tapes and machines alongside a compatible PCM encoder was, by that time, a fairly well-established method for capturing digital audio, Gerzon expanded on this through the simultaneous use of newly-available Beta Hifi recording; employing both the linear PCM and Beta Hifi tracks as a carrier for four simultaneous audio streams — in essence transforming a fairly straightforward domestic setup into a versatile multi-track recorder. This he achieved through the use of a Sony SL-HF100UB Betamax recorder — one of only two Beta Hifi-compatible recorders released to the domestic market in the UK — alongside a Sony PCM501ES PCM unit, and it was to comparable hardware that we returned as a means to capture the Betamax contingent of the collection; using two professional, Beta Hifi-capable Sony SLO-1700 machines (with heads re-aligned to correspond with the azimuth used on Gerzon’s original tapes), alongside two compatible Sony PCM701ES units, with the resulting eight audio outputs configured in the same manner as the remainder of the multichannel recording in the collection, and prepared for later reassembly as a direct representation of the original recordings.

Of note, Gerzon was also certainly aware of the limitations of early digital convertors in their ability to maintain an acceptable signal-to-noise ratio at higher frequencies; and he employed EIAJ pre-emphasis as a means of noise reduction throughout his digital recordings. While this was simply dealt with for the Betamax element of the collection through employing the analogue outputs of each PCM unit (essentially stripping the signal of its emphasis curve), the DAT element of the collection proved more difficult to deal with — not least due to its suitability as a subject for mass-ingest.

As an entirely digital format, archival audio extraction from DAT is ideally maintained as a digital process; with transfers via SPDIF or AES avoiding the degradation inherent in analogue signal transfer. Where this approach fails with regard to mass digitisation is the reliance on a single machine as the master clock source of all ingested audio data. While possible with remote-compliant hardware and a dedicated, stable word clock source, the system used in this instance was not equipped for synchronised digital transfer from more than one machine. As such, an alternate means of efficient transfer from DAT was found in VDAT; a relatively niche software solution employing four modified DDS SCSI tape drives, and facilitating the extraction of raw PCM data directly from tape to equal-resolution (in this instance, 16bit, 48kHz) WAV files. This not only facilitates extraction from four tapes at speeds well above real-time, while avoiding potentially inconsistent interpolation and producing frame-accurate error reports alongside each transfer, but also allows the engineer to assess audio quality and re-transfer any problematic areas of recording without incurring any significant loss in productivity. A potential problem here is the fact that maintaining a purely digital signal path in the extraction of pre-emphasised audio retains the presence of pre-emphasis in each extracted WAV file. A solution for this was found in the use of Q10 — a leading software EQ featuring a recognised, accurate de-emphasis filter — developed by Israeli audio company Waves and, fittingly, worked on by Gerzon himself during his later years as a consultant.

3.5 Presentation

In presenting the Gerzon Archive online, the prevailing attitude is that, as an archival project, the transfer and presentation of the collection should remain an objective process. In keeping with this, it is intended that all multi-channel components of the collection will be presented to the end user as separate, individual stems; avoiding any subjective influence through mixing or undue processing. Hopefully, this approach will serve to mark the collection out as a unique research tool — particularly in the field of audio engineering — allowing the end user to download, examine, compare and manipulate each recording alongside its accompanying metadata.

A similar approach is intended for the few surround-compatible recordings present in the collection: twelve RM-compatible four-track cassettes, a single UHJ-compatible two-track cassette, and around twenty BBC H-matrixed quadraphonic offair recordings; the latter currently excluded from the project by copyright issues). For these we intend to make each of the individual stems available for download as standard mono files; allowing the user to freely experiment with software matrixing outside of the archival environment.

4. Other Project Processes

4.1 Making a Case for Selection

In selecting the content packages for the ASR project a number of considerations had to be taken into account, apart from the potential study value of the material. While it was agreed that the project needed to depart from the usual practice of relying solely on out-of-copyright material, which would be unduly restrictive, it was clear that certain content would prove impossible to clear if rights holders were unsympathetic or had other priorities. Indeed this proved to be the case with regard to the substantial broadcast component within the Gerzon Archive, which was consequently excluded [q.v. §4.6]. However, other factors mitigated in its favour, including the assignment by the Gerzon Estate of the mechanical copyright in Michael’s recordings to the British Library, thereby eliminating one of the major rights clearance hurdles. Another logistically helpful factor was perhaps the fact that much of the music was improvised, which meant that the performers would also often be the IPR owners in the featured “compositions”, reducing that additional level of rights negotiation to a more manageable level.

4.2 Cataloguing and Metadata Creation

Cataloguing such a vast collection of unpublished recorded music is a colossal and time-consuming task which could potentially consume a substantial proportion of any library or archive’s documentation resources for a considerable period of time. At the same time, it is clear that adequate access to such collections — which is essentially what the ASR project is about — can never be achieved unless this task is effectively accomplished. Moreover, accurate cataloguing is essential for the creation of high quality metadata for the digitised media since this is, wherever possible, derived from it. Although some cataloguing for the Gerzon Archive had been completed before the project’s inception, a great deal more remained to be done and coordination of this with the metadata creation has proved to be one of the more problematic aspects of the work to date.

However, the Gerzon collection again demonstrates the astonishing foresight of its creator in his determination to document all the essential details of the performances: recording location, date and even time of day, personnel (where known) and instrumentation, as well as the equipment he used to record it. In some cases he even accompanied these details with drawn stage plans plotting performer and microphone positions/angles. This greatly assisted the documentation process and, in view of the special value of the collection as a document of late 20th century recording practice, we have taken the unusual step of including all of this technical information with the cataloguing and metadata.

The project’s metadata requirements are further complicated by the need to produce multiple versions, in a variety of formats, from the already vast quantity of recorded audio, the need to coordinate its creation with the audio transfer/digitisation process, and a requirement to electronically link the new digital media with the BLSA’s existing online catalogue — the latter issue still being addressed at the time of writing.

The project metadata, which of course describes, controls and provides access to the digital media, is being written with a METS schema (Dublin Core with XSD and integrated BLAP-S — a British Library designed application protocol for digitised sound).

4.3 Intellectual Property Rights Clearance

In some respects this remains the most critical aspect of any project aiming for online dissemination of cultural property.

Although the Gerzon collection promised to be more straightforward in certain respects (q.v. §4.1), locating the enormous number of rights owning performers — over two thousand — represented a considerable challenge even before negotiations could commence. Moreover, while all of the recordings were over twelve years old — adding to the difficulty of locating contributors, the oldest was still within the UK’s fifty year copyright period. The uncommercial nature of much of the music also suggested that only a small proportion of the performers would have agency representation or Musicians Union membership. Despite this, it was happily soon discovered that a considerable proportion had some sort of internet profile, in many cases a MySpace page. Others were successfully traced through specialist music websites or those of professional organisations such as Jazz Services. In many cases it was found that a located contributor was both able and willing to provide an up-to-date contact for several others within his/her circle. Although many hundreds of performers were located in this way, inevitably many others have thus far remained elusive, particularly from the older part of the collection.

One encouraging sign to date has been the enthusiastic cooperation of many of the traced performers, who understand the value of the project and wish to assist. This contributed to a substantial revision of the rights clearance process for Phase 2 of the project. Whereas Phase 1 sought clearance only for license holding institutional members of the UK’s Higher and Further Education community, Phase 2 is initially aiming to clear content, wherever possible, for universal online access by users anywhere in the world. Where this is declined the Rights Clearance team will default to a request for password controlled access for licensed UK educational users only. This makes sense when considering that the bulk of the work and costs involved are expended in identifying and locating rights owners, which will be the same whichever approach is taken, and the considerable enhancement of the project’s outreach where universal access is agreed.

A due diligence framework has been implemented and ensures the rigorous documentation of all IPR clearance processes and full compliance with all legislation, as well as ethical considerations in relation to so-called “orphan works”, for which performer(s) remain unidentified.

4.4 Access Management System

Notwithstanding attempts to clear the content for universal access, it is anticipated that many rights owners will only sanction controlled access for licensed educational institutions. In order to effectively secure and protect all such site content during streaming and downloading, user authentication and access is now being centrally controlled by a UK-based federated access management system (11) based on Shibboleth technology.

4.5 Development of the Service Interface

The ASR website was developed by the British Library to ensure compatibility with the Library’s Content Management system, while data storage for the streamable and downloadable media is provided offsite. The XML-driven interface, shaped at an early stage in response to user testing and feedback, was designed to comply with web content and compatibility standards established by W3C and the Open Archives Initiative. (12) Site tools include a “My Project” facility and project team blogs, while forthcoming web 0.2 functionality is expected to include user networking and content tagging. These initiatives have considerable potential in relation to Gerzon Archive content, where the possibility of experimenting with (for example) the re-mixing of the multitrack recordings has exciting educational possibilities with regard to shared resource discovery and learning.

4.6 Future of the collection and conclusion

A key objective of the project is the long-term sustainability of the service, a requirement of the project contract. To this end the British Library is committed to maintaining the online facility for at least an additional ten years beyond its JISC-supported phases (ending March 2009), when it will be entirely dependent on its own resources. The BLSA will continue to expand the resource’s ASR page media content with additional digitised media, including relevant non-audio, where resources and copyright issues are not prohibitive.

A significant proportion of the Gerzon Archive audio was excluded from the ASR project by copyright and IPR issues and it is not currently possible to make this available online. However, all of the excluded Gerzon content, including hundreds of Michael’s off-air radio recordings, will in due course be separately digitised to the same archival standards. All of the Gerzon Archive recordings, and the vast majority of Michael’s published papers, are available for onsite access by Reader’s Pass holders within the British Library’s reading rooms.

We would like to leave the final comment here to a musician who was recorded frequently by Michael Gerzon, guitarist Richard Chapman:

I remember him as an incredibly gentle and enthusiastic character. Seemingly a bit vague… he was always sympathetic and thoughtful. He had a sharp intellect and talked with great knowledge and enthusiasm about music and various gigs that he had recorded. In fact his attendance really made gigs feel important — he would sit and listen and you felt that a real music aficionado was witnessing what you were doing. I also remember that he very much liked events where the vibe was right and where creative chemistry was taking place and was honest when things were not happening. He looked for music which was art. (13)

Further information on the British Library, and British Library Sound Archive  is available from their respective webpages. Much more on Michael Gerzon and the work of the Oxford University Tape Recording Society can be found at Stephen Thornton’s fascinating Michael Gerzon website. For additional information on the Michael Gerzon archive at the British Library and the Archival Sound Recordings project, please contact:

Paul Wilson
The British Library Sound Archive
96 Euston Road
London NW1 2DB
Tel: +44 (0) 20 7412 7446
pwilson@bl.uk

Key Project Personnel

Project Board: Richard Ranft (BLSA Director), Adrian Arthur (Head, BL Web Services Delivery Unit), Alastair Dunning (JISC Digitisation Programme Manager), Kristian Jensen (Head, BL British Collections since June 2008), Chris Clark (Head, BLSA Selection and Documentation), Joanna Newman (Head, BL Strategic Partnerships)

ASR Project Team: Peter Findlay (Project Manager), Michelangelo Staffolani (Project Team Leader), Eva Del Rey (Metadata Editor), Rehanna Kheshgi (Digitisation Assistant/Metadata Editor), Melanie Bourne (Project Support/Metadata Editor), Ellen Hebden (Project Support Officer), Elisa Pettinelli-Barrett (Project Assistant), Ginevra House (Engagement Officer)

BLSA Curatorial Team: Paul Wilson (Curator/Cataloguer)

BLSA Documentation: Antony Gordon (Head, Cataloguing), Glen Thomas (Cataloguer)

BLSA Technical Services Team: Adam Tovell (Transfer and Digitisation Engineer), Andrew Pearson (Equipment Maintenance/Design Engineer), Christine Adams (Production Coordinator)

Notes

1. The Joint Information Systems Committee (JISC), “Archival Sound Recordings 2” (1 January 2007 – 1 January 2009). http://www.jisc.ac.uk/whatwedo/programmes/digitisation/asr2.aspx#downloads [Last accessed 18 December 2008.]
2. Audio Signal, “Michael Gerzon Archive” [Website]. http://www.audiosignal.co.uk/Gerzon archive.html [Last accessed 18 December 2008.]
3. John Borwick, “Michael A Gerzon” (Obituary), Grammophone (August 1996). Reproduced on the Audio Signal website. http://www.audiosignal.co.uk/MAGobit.html [Last accessed 18 December 2008.]
4. Waves Audio [Website]. http://www.ambisonics.net [Last accessed 18 December 2008.]
5. Steven Thorton, http://www.michaelgerzonphotos.org.uk. [Last accessed 18 December 2008.]
6. Michael Gerzon, “Stereo Recording of Live Amplified Music.” Proceedings of Reproduced Sound 3 Institute of Acoustics Conference, Windermere, 5–8 November 1987), pp. 69–87.
7. Michael Gerzon, “The Politics of P.A.” Re Records Quarterly Magazine 1/4 (May 1986), pp. 21–25.
8. Audio Signal, “Stereo Shuffling: New Approach – Old Technique,” Studio Sound (July 1986). Reproduced on the Audio Signal website. http://www.audiosignal.co.uk/Resources/Stereo_shuffling_A4.pdf [Last accessed 18 December 2008.]
9. Michael Gerzon, “Minimising Print-Through,” Studio Sound (September 1972). Reproduced on the Audio Signal website. http://www.audiosignal.co.uk/Resources/Print-through_USL.pdf [Last accessed 18 December 2008.]
10. “Sony WM-D6C Stereo Cassette-corder Operating Instructions”, Sony website. http://www.docs.sony.com/release/WMD6C.pdf [Last accessed 18 December 2008.]
11. The Joint Information Systems Committee (JISC), “UK Federated Access Management.” http://www.jisc.ac.uk/federation.html [Last accessed 18 December 2008.]
12. Open Archives Initiative. http://www.openarchives.org [Last accessed 18 December 2008.]
13. Correspondence with author, May 2008.