Johnathan A Scharf and Professor David A Day, Curator, Music Library Special Collections

Recorded music has existed for more than a century yet, until recently, the media on which this music is preserved degrades with each passing year. Additionally, the very use of the recording itself adds to its degradation and does irreparable damage. It is a major concern for an institution that archives such media to find a method that permits the enjoyment of early recordings without subjecting them to further damage. Moreover, a means of restoring degraded recordings would permit a long-term solution for circulating such work. Through the use of computer technology, these goals can be achieved. It is now possible to electronically convert an analogue waveform produced by anything from a wax phonograph cylinder to an open reel tape into a digital stream of information. This digital stream can be analyzed using computer software to isolate and eradicate noise and other audible imperfections. The audio restoration process not only involves knowledge of the software tools but also a subjective understanding of what “sounds” musical.

The Music Library at BYU has many pieces of equipment to facilitate the playback and capture of music. Foremost among this equipment are the computer-based CEDAR (Computer Enhanced Digital Audio Restoration) and Pro Tools systems. It need not be illustrated how profound an impact the development of the computer has been for this generation. Yet in the areas of audio preservation and restoration, it is only through the advances made in the last few years that have allowed this use of the computer to be practicable. Pro Tools incorporates hardware and software to allow for the capture of the sound wave and its subsequent editing. The CEDAR system is also a union of hardware and software to equalize and “clean-up” this digital information. Like the digitizing of a photographic image, the digitizing of an audio waveform requires high resolution to ensure that nearly all the musical information has been retrieved and to facilitate its manipulation. As such, a separate piece of hardware is used to convert all the subtleties and nuances of the audio signal into a digital stream at up to 192kb/s or nearly five times the resolution of modern CD’s. The other equipment includes turntables, phono preamps, cassette decks, and open reel machines on which the source material is played.

As my project dealt with the education of others on the utility of music preservation and restoration, I chose the 33.3 RPM record as representative media. This piece of media poses several challenges as the audio information is written onto a piece of vinyl as an actual sound wave. Playback involves the use of a stylus to trace this path and the subsequent amplification of the stylus’ oscillations. Since the stylus must make continuous contact with the vinyl to guide its movement, any imperfection in the surface becomes audible.

To communicate the process of music preservation and restoration, a PowerPoint® presentation was created that summarized the steps involved. This presentation can be used to introduce those outside the field to the equipment, time, feasibility, and benefits of doing audio restoration. Additionally, this presentation may be used to introduce to those wishing to work on the CEDAR and Pro Tools equipment, the way in which it is done at BYU’s Music Library. As an overview of the process, the following sections were detailed:

1. Preparation of the media for playback—cleaning of LP and stylus
2. Initial equipment set-up: Pro Tools Hardware/Software
3. Initial digitization and capture of unaltered audio waveform
4. Processing of sound through the CEDAR system: Declick Decrackle  Attenuation  Noise Reduction  Equalization
5. Editing and tracking of clean signal
6. Preparation and writing to CD-R media

The biggest benefit of creating a presentation of the audio preservation and restoration process is to familiarize people with what modern technology can do for older recorded media. Most libraries have a music section that is heavily comprised of CD’s. Sadly, several libraries have totally abandoned the collection of LP’s and even discarded this media altogether. Much of the music on LP has not been transferred to CD nor is this likely in the future because of the small market for old recordings. Thus, if music from bygone generations is to survive, the work of preservation and restoration by digital means must be better understood and communicated.

Although the mechanics of the audio restoration process can be adequately depicted in words and pictures, the results significantly depend upon the sensitivity of the person doing the work. The CEDAR system is capable of modifying many aspects of the digitized audio signal. Some of these modifications can be deleterious. Ideally, an audio engineer should exhibit not only a degree of skill but also some degree of talent and an extended sense of audio acuity. For instance, a person restoring a violin concerto who couldn’t hear above 12,000 Hz may roll off higher frequencies in the noise reduction process thus achieving a very muffled sounding result. Another factor in the restoration process is the introduction of audio artifacts through overprocessing the signal. A click or pop in an LP, for example, replaces the audio information at that spot, its simple removal would leave a hole of silence that is unpleasant to listen to. The CEDAR system uses a complex algorithm for replacing the pop with estimated audio information from the previous good digital sample. However, this process may introduce irritating artifacts that are actually worse than the pop that they replace. Thus, a person with good musical perception as well as a proficiency with the equipment is best suited for doing restoration work.

The previous paragraph exemplifies the frustrations inherent in preparing a presentation of this nature: the inadequacies of language in communicating what sounds good and what sounds bad. Even the manual that came with the CEDAR system was very inadequate in what an artifact sounded like. In the literature, words such as “gurgling,” and “bubbling” are used. Since these artifacts only last a few tenths of a second, the recognition of them is also very difficult but the brain does perceive them. Explaining the difference in surface noise on an LP involves delineating between clicks and crackle which is difficult to put into words. Thus, it becomes frustrating to describe the restoration process in objective terms.

Overall, the experience of creating this presentation not only improved my skills with audio restoration but it also gave the music library a tool to use for training future employees. The experience or audio restoration has awakened in me a profound appreciation for recorded music of past generations and I have begun to acquire works from the early 1900’s over the internet. Interestingly, my degree from BYU will be in Microbiology but my career path may now include something in the audio industry as a result of this research.