I was recently contacted by Dan Gravell, founder and programmer of the server-based music management software, bliss. Bliss received praise from Andrew Everard of What Hi-Fi and their official website calls it a tool “for people who care about their music collection.” Dan posed several questions about my library, and about what differentiates an average personal music collection from a true archive. He suggested that my response might prove useful as a journal entry at Innerspace Labs, so I’m sharing my response for others who might ask the same questions about their own meticulous collections.
So let’s dive right in –
Regarding the difference between run-of-the-mill “playable” music libraries and what one might call an “archive,” there are a few primary factors which could differentiate the two. The first is one of practical function and intent. If a library is for personal use for playback alone it is most likely the former, whereas a consciously organized collection of significant size and scope which is representative of a particular period or culture and which sheds contextual light on that era might serve a greater, almost scholarly purpose as an archive. Uniformity of structure, organization, navigability, and accompanying supplemental metadata enhance a library such as this to greater usefulness than mere playback. And it appears that it is precisely this focus on consistency by which Dan has endeavored to empower users like me with his bliss project. Another important factor is the long-term sustainability of an archive, which I’ll touch upon momentarily.
Next Dan asked whether my source media is exclusively physical. My collection comprises only a few thousand LPs, with a significant focus on the history of electronic sound. This spans the gamut from early notable works of musique concrète to the Moog synthesizer novelty craze, all the way through the international movement of ambient electronic music. I’ve also a predilection for archival box sets, like the Voyager Golden Record 40th Anniversary set with companion hardcover book and the special release from The John Cage Trust, as well as the previously unreleased collection of Brian Eno’s installation music issued earlier this year on vinyl with a new essay by Eno. But the bulk of my library is digital. This is both for practical and financial reasons, as digital libraries are far easier to maintain. (I don’t blog about digital nearly as often, as 450,000 media files are nowhere near as fascinating as a handsome limited edition LP!)
Dan also inquired about my workflow, which is critical to any archive. Early on in the development of my library, (around 2002-3), I began ripping LPs with the following process:
Exclusive analog recordings are captured using a Denon DP-60L rosewood TT with an Ortofon 2M Red cart, powered by a McIntosh amplifier (later replaced with a vintage Yamaha unit), and are saved as lossless FLAC via an entry level Behringer U-Control UCA202 DAC. I previously utilized a Cambridge Audio DacMagic DAC but after it failed I opted for the Behringer and it has been more than sufficient for my needs. Audio is captured using Audacity on my Linux-based DAW and basic leveling and noise reduction are performed but I minimize post-processing to maintain as much of the original audio’s integrity as possible.
Dan specifically inquired as to where the library information was stored (barcodes, etc) and asked about my policies on which metadata are included. This is fairly straightforward, as nearly all of the vinyl recordings I ripped pre-date the use of barcodes or were limited private releases with only a catalog number, which I bracket as a suffix in the release folder path.
Polybagged LPs are stored vertically and organized by primary genre, then by artist, then chronologically by date of issue. Due to the entropic property of vinyl playback, discs are played once as needed to capture the recording and subsequent playback is performed using the digital files. I employed a dozen static local DB applications over the years for my records, but eventually migrated to a Discogs DB which increases accessibility while crate digging in the wild and provides real-time market value assessment for insurance purposes.
But honestly, I almost never need to perform the rip myself, as the filesharing ecosystem has refined itself to the point where even the most exclusive titles are available through these networks in lossless archival FLAC with complete release details. There has never been a better time to be alive as an audio archivist.
Once digitized to FLAC, my assets are organized with uniform file naming conventions with record label and artist parent folders and parenthetical date of issue prefixes for easy navigation. gMusicBrowser is my ideal playback software for accessing large libraries in a Linux environment. Release date and catalog numbers have been sufficient metadata identifiers, as subsequent release details are only a click or a tap away on Discogs. Occasionally I will include a contextual write-up in the release folder where warranted, like in the case of William Basinski’s The Disintegration Loops 9LP + 5CD + DVD set as it related to the events of 9/11.
Next Dan inquired about how my archive is accessed. I employ Sindre Mehus’ Subsonic personal server application on my Linux DAW to make all of my audio and music video film content accessible from my phone, tablet, or any web-enabled device. I use both the official Subsonic app and the independently-developed Ultrasonic fork by Óscar García Amor for remote access of my library, (about eight hours daily). You can see a short video walkthrough of the features of the app that I put together here:
To return to his initial question about what differentiates a playback collection from an archive, my own library incorporates a few key factors which might lend itself to the latter:
– lossless bit-perfect FLAC wherever possible
– index documentation
– a systematic process guide for new acquisitions
– a 76pp manual highlighting special collections and large libraries of the Collection
– disk mirroring in multiple physical locations for preservation and sustainability
– fire protection for further indestructibility
– routine disk operation tests to mitigate risk of data loss
– complete discographic record label chronologies suffixed with catalog numbers
– elementary data visualizations created using Gephi and Prezi web-based tools
– the use of TrueCrypt whole disk encryption to prevent unauthorized access
– and the active use of Subsonic and Ultrasonic for enhanced accessibility
And scale is another noteworthy factor in my circumstances. Just to cite one example, I’ve collected every LP and single issued by the electronic duo Underworld that I’ve been able to get my hands on, and the digital audio branch of my Underworld collection comprises 482 albums, EPs and singles, including 2850 tracks and DJ sessions totaling well over 385 hours of non-stop music, spanning 36 years of Karl Hyde and Rick Smith’s work in all of their many incarnations. This collection is uniformly tagged, organized into a network of categorical root folders, and substructured into chronological subfolders by date of release. And the complete record label collections are a definite differentiator from the majority of casual-listening libraries.
I understand that my archive is small compared to the 12-20 TB libraries of some more seasoned users, but I feel that discretion and selectivity are virtues of my personal collection so that I can focus on only the most exquisite and remarkable recordings of my principle genre foci.
So what about your own collections? Do you employ standardized uniform file naming conventions and organizational standards? Do you supplement your library with relevant documentation to add context to your media? Does your collection offer insight into a particular era or musical culture? And do you take measures to ensure the longevity and sustainability of the work? If so… you might just have an archive.
Supplemental Note:
A good friend was kind enough to offer his thoughts about what sets an archive apart from other collections, and his remark was too good not to share. He said –
I think another major difference between the average personal collection and an archive is retention and adaptation.
A casual listener or collector wouldn’t have the retention of a true archive. The individual may build some playlists or even some advanced structure for locating and listening to music, but there is a very good chance that after some time, that particular music will get buried by the newer, or the most current thing the user is listening to. The casual listener may not want the huge or growing library, so when they feel they have moved on, the music will be removed from their collection. I cannot see someone who is keeping an archive remove anything from their collection. So retaining the entire collection and not removing anything because they are bored with it would be a difference.
I also mentioned adaptation. This is a rather basic idea but would be rather important in the grand scheme of things. Lets say you have a collection of 100 songs, all with 4 points of meta data. You realize as you begin to add more songs to your collection, a 5th point of data is needed. A casual listener may leave those 100 songs in the current state they’re in, with the 4 points of data. The archivist would need to go back, and add that 5th point to all 100 songs, and the new ones. Add another zero to those numbers and that can be a daunting, but necessary task for the archivist.
I really appreciated his input!
The Innerspace Connection 