-
Notifications
You must be signed in to change notification settings - Fork 0
Digital History
Digital files, while fairly standardized now, have a complex history in the department. This document details that history chronologically, and when relevant, by collection, to provide insight when searching for a file, migrating files, or other related tasks.
Callie Holmes (digital archivist from 2016 -- present) standardized many processes and formats in our digitization workflow. If we say that we have a Holmes transfer of an item, that typically means that a transfer was done after her standardization was applied, typically resulting in a preservation, mezzanine, and access-quality copy of the item (only preservation and mezzanine securely retained) that have followed the same general framework that we continue to use to this day.
In the early stages of in-house digitization, most items were digitized using Samma machines that were designed as a sort of all-in-one solution to capture and encode a file from a tape. If we have a Samma transfer of an item, it's a file created from that process. Unfortunately, most of these files are not intact today, and if you pull one from an LTO, playback and re-encoding is usually not possible. It's unclear whether this was due to issues in the creation itself or file decay over time, but almost all exhibit similar symptoms.
Peabody files also have a complex history. Any transfers we've done in-house follow the same general patterns and history as other collections, but there are born-digital items within the Peabody collection. There was a period of several years in the late 2000s into the 2010s when Peabody entries were accepted as digital files on hard drives, flash drives, or other disc storage. Our archival copies of these items, then, are the files that were ingested from these items. They may not follow the same uniform formatting and codec rules that other files do (and that Peabody items we've digitized in-house do), so there is a greater variability in what you'll find.
WSB film in the hands of UGA, like Peabody, predates the creation of the archive, so the history is more knotted. In the 1980s, the journalism school created U-matic dub tapes of all WSB film that we had. Later on, when the archive was outsourcing digitization work to Crawford Digital, these tapes were one of those projects. All of the U-matic tapes were digitized to individual files named FMUGAXXXX.mxf, where the four digit (XXXX) number represents the same reel number that WSB reels have to this day. We keep copies of these files close at hand on a local RAID, so that we can quickly provide clean WSB footage to researchers and licensees when they need it but do not require anything better than SD footage.
Files for wsb-video collections (including most subseries like ac, rndm, olympics, and so on) are some of our oldest digital files. The original videotapes of these programs were digitized by Crawford Digital (which no longer exists in name) and written to TAR-formatted LTO tapes. For a decade or so, this was our source for any of these files, so any requests would be filled by loading the TAR LTO on a Linux machine and writing the files over. Many tapes (roughly 2/3) were pulled in this way either for an order or for proactive migration, and the files were written to the Digital Hub or ARCHive. Occasionally, issues would arise when pulling a specific file, and we would return to the videotape, if available, and complete a new transfer. So, a very small percentage have a Holmes video transfer with resulting files written to LTO, Hub, or ARCHive. In 2025, we outsourced migration of all un-migrated tapes to ProTek, and this vendor wrote all migrated files to new LTO7 tapes (WSB331L7-WSB355L7). This means, that as of 2025, all wsb-video files that we originally received from WSB by way of Crawford transfers can now be located on either an LTO7 tape, a mezzanine share on the Digital Hub, or in ARCHive.
20?? -- 20?? Before the establishment of a digitization unit, all[?] digitization was completed by external vendors, like Crawford Digital. Often, the only versions of those files that exist to this day exist on older generation (such as LTO3) LTO tapes. Migration of those older tapes is difficult and sometimes lower priority, because many times naming conventions have changed, making files difficult to identify, and often the files are substandard compared to what we produce now, or have decayed, no longer allowing playback or re-encoding.
20?? -- 2017 In the early years of the digitization unit, standards were lacking and methods often resulted in files like the Samma files, that in many cases, can no longer be used. Files from this era were written to LTO tape when preserved, but many have decayed and can no longer can be used. If they can be used, we often run into the same file-naming issues. But, some of these files are named correctly are a usable format, and have not decayed. Of these, many have been migrated into our modern systems, ARCHive and the Digital Hub.
- In 2014, we purchased our first film scanner, the MWA Nova. Many (a bit over 1,000) film scans were completed on this scanner. Those scans did not produce a 4K "overscan" quality scan, but most did produce a 2K clean-frame scan that closely matches the quality of mezzanine file that we still produce today and send to licensees in many cases. These scans were written to LTO tapes, but are often in bags, and contain folders of DPX images, a WAV audio file (if sound), and a ProRes MOV file.
2017 -- 2019 Files created in this era are of a standardized format and have little naming issues. However, as ARCHive did not yet exist, preservation-quality files were written to LTO tape. The earlier LTO7 tapes represent this constituency.
2019 -- 2022 Files created in this era are of a standardized format and have little naming issues. In 2019, preservation copies began to be deposited into ARCHive, and mezzanine copies (slightly smaller, slightly lower quality, but still often good enough for productions) were written to LTO, to be more readily accessible.
- In 2020, we purchased the Lasergraphics ScanStation, which we still use today. Uniformly, we produce a 4K "overscan" DPX sequence (which is cooked into an FFV1 Matroska with any present audio) and a 2K ProRes 4444 mezzanine file. These scans also have much less stutter, better lighting, and if returning to the Matroska, higher resolution than our Nova scans. For this reason, on occasion we will still need to complete a new scan of an item that has already been scanned if one of these metrics is vital to the recipient.
2022 -- present day Files created in this era are of a standardized format and have little naming issues. In 2022, the Digital Hub came online, and we transitioned from writing mezzanine copies to LTO tape to writing them to the Hub. Preservation copies continue to be written to ARCHive. Mezzanine copies that predate Hub, or any copies that are not sufficient enough quality to be written to ARCHive, continue to be migrated to Hub.