Bay Area Video Coalition, San Francisco, June 8, 2011
Sarolta Cump and Lauren Sorensen work in the video preservation department of the Bay Area Video Coalition (BAVC)1. BAVC is a non-profit organisation founded in 1976 in San Francisco to allow independent producers and underrepresented communities to use emerging media technologies such as the Portapak video system. In 1992 BAVC developed a Preservation Centre and today it is still one of the biggest non-profit preservation programs in the United States. In 2003, BAVC produced PLAYBACK2, a DVD that shows the technical practices of video preservation and the complex decision-making process engaged by artists, conservators and video engineers as they reconstruct video artworks. Emanuel Lorrain (PACKED vzw) met Lauren Sorensen and Sarolta Cump in order to know more about their digitisation workflow and to find out how they deal with the obsolescence of the decks used to digitize contents on old video carriers.
PACKED: What are you respective backgrounds?
Lauren Sorensen: I went to grad school for Moving Image Archiving and Preservation3 at NYU and then I worked for three years at a 16 mm film distributor called Canyon Cinema4. Now I’m working here as Preservation Specialist.
Sarolta Cump: I have a postproduction background. I actually worked at BAVC in the late nineties for postproduction technical support, and then I left and got a degree in film & video production. But I came back because I was familiar with a lot of the equipment. That’s my area of speciality but we also do a lot of cross-training which is really nice.
PACKED: Could you summarize the services that BAVC offers, and to whom especially?
Our mission is to work on non-commercial projects, so we work with a lot of arts organizations and non-profit and community-based and cultural institutions. We have some larger clients like ‘the Kitchen’ in New York, which is a performance space that’s been around since the seventies. We also work with local organizations like the Exploratorium, which is a science and children's museum in San Francisco.
PACKED: What formats can you transfer?
Sarolta Cump: We can transfer ½” open reel5, 1 inch, U-maticfn]¾” U-matic is an analogue video format that was developed at the end of the sixties and consisted of a ¾” video tape in a cassette. It was the forerunner of the analogue Betacam., Hi-86, 8mm7, VHS8, Betamax9, Betacam10 and Betacam SP11, audiocassettes and ¼” audio tapes.
PACKED: For 1/2 inch open-reel, can you transfer both CV and- AV?
Lauren: Yes, we have both, and both type decks are getting more difficult to repair. We are really only working with one person who repairs them so sometimes it's challenging getting the timing right between our clients’ deadlines and repairing such rare playback decks.
PACKED: Do you use tape-cleaning machines?
Sarolta Cump: Yes, we have an RTI12 tape check cleaner for 3/4" U-matic tapes from which the burnishing blades have been removed, so only gentle, pellon material is used in cleaning. We also have a VHS cleaner and a ½” tape cleaning machine, for which we are in the process of trying to find a service provider to repair.
PACKED: Is that a cleaner you've built yourself out of another machine?
Lauren Sorensen: The 1/2" tape cleaner used to be a data tape cleaner for computer tapes that was altered to work for 1/2 inch open reel video specifically. Past BAVC staff took off the burnishing blade on that machine as well. It’s a vacuum type system which will clean along the cloth and also has a…
Sarolta Cump: …a vacuum chamber. People from BAVC back in the nineties adapted that machine and now if we need it to be repaired, we'd have trouble finding a service provider, considering it was a custom made deck; the company that made it doesn’t support even the computer tape cleaning machine anymore, much less our altered machine!
PACKED: do you use baking?
We do use baking13 yes, a dehydrator, and we use very specific temperature control, as well as timing and workflow.
PACKED: For which tapes do you use that treatment?
Sarolta Cump: Generally for U-matic and 1/2". Those are the two main type of tapes for which we use baking.
PACKED: Do you clean the tapes before baking them?
Sarolta Cump: Yes, for the U-matic tapes we clean them up to four times and then attempt a transfer. If we still have issues and it looks like a sticky shed syndrome14 problem, then we bake it. After the baking we clean it again once before attempting to transfer a second time. With the 1/2" open reel tapes, for now we play it and if it has sticky shed syndrome issues we bake it and try to transfer it a second time.
Lauren Sorensen: There are no rules as to which format will have sticky shed. Much of the time it depends on trying out the tape, so it is really hard to evaluate this prior to playback. It mostly depends on how the tapes have been stored, but sometimes just basic manufacturing issues with the original result in deterioration; these tapes were made by companies who keep formula under wraps, proprietary. So we have no way of knowing, as preservationists, what we are working with except very general specifications.
PACKED: Is there anything else you do with the tape on a physical level?
Lauren Sorensen: We carefully inspect each tape upon arrival. Sometimes cassettes need repair, so we keep extra supplies in house for this purpose, e.g., if the tape case is cracked and needs re-housing. We have a workflow for inspection and document it carefully when we receive the materials from our partners and clients. It is based on ISO 18933, Imaging materials — Magnetic tape — Care and handling practices for extended usage15.
Sarolta Cump: We have some spare casings. If it looks like a mechanical problem, we will take the tape apart and then put it back together.
PACKED: Could you go through the digitization workflow at BAVC?
Sarolta Cump: We evaluate the tape and document its condition when it first comes in and then, if it is a format that can be cleaned, we clean it just once to see if it has residue or sticky shed. Depending on if it’s a U-matic or not, as we said before we continue cleaning up to four times before we attempt to playback and digitize the tape . At the moment we are recommending 10-bit uncompressed files to our client base.
Lauren Sorensen: Yes, because it results in an accepted, lossless, high quality file. There are some other formats that we are investigating to recommend to clients right now like FFV1. We are thinking about moving to that format just because it is a lossless compression and results in smaller files, it is open source codec and in line with chroma subsampling and other factors that affect authenticity. Right now, the codec that we are using is not open source, but is widely supported. Codecs such as DV16 may be open, however they are lossy, so moving forward in preservation of these digital assets can result in loss of quality.
Lauren Sorensen: Yes. With uncompressed 10-bit, we use data movie files, mostly a Quicktime wrapper. It's because we mostly have a Mac environment, but we can produce AVI-wrapped files as well, depending on client need.
PACKED: Can you also work with an MXF19 wrapper?
Sarolta Cump: We can, we have done it in the past, but we generally use .MOV20 unless the client wants a different wrapper.
Lauren Sorensen: Yes, we definitely work with our clients kind of as partners, so we investigate what format would be most useful for them. Most of the work that we do is with those who manage archives as well as individuals. They usually have an idea of what they want from their own research based on what the rest of their archive is already composed of. Many of these archive handlers are trying to come up with a set of practices and standard format for the preservation of analogue video; until now, many have been using Digibeta21 or digital tape, and – less often – high quality analogue tape which is now obsolete. Digital tape, while for the most part not obsolete, is unstable and especially in the case of MiniDV22 and DVCAM23, prone to errors in transfer. Most of the time, we can be pretty flexible on what we deliver to our clients in terms of the file format, but for individual clients it’s a bit more tricky because they might have a Windows operating system for instance, or not enough memory available or slow processor speed to decode and playback the file in real-time; this is why we also recommend “mezzanine” level files which are appropriate for editing purposes and access. All the same, we need to take into account what kind of environment they are playing and storing the files in. There are a lot more questions that need to be evaluated in analogue-to-digital preservation, and context is key.
PACKED: So you investigate that for the clients?
Lauren Sorensen: Yes, like for instance we ask what kind of quality they want, what kind of longevity they're thinking about? What use will the files have moving forward? Right now, we are working on trying to figure out where these collections can go in the long-term because obviously we are preserving it to a digital format, but with digital files it has to be an active process for archiving. If we are preserving and giving it to an individual it’s different from giving it to an institution where there might be an infrastructure for a preservation strategy; this isn’t something individuals or smaller organizations have the capacity to consider. There’s also a shifting skillset within archives that needs to be addressed; a lot of traditional archivists are not versed in what it takes to preserve digital objects, and so we are attempting to work on advising them and individuals about that right now by keeping our knowledge in the field current.
Sarolta Cump: I feel like we are spending a lot of time with people, trying to explain what access files are, as well as the general process of preservation and what it involves.
Lauren Sorensen: Digital archiving of moving images is so new that it involves a lot of exploring for each different context, which is really interesting and challenging.
PACKED: What kind of access copy do you produce? Do you also provide documentation of the digitisation process to your client?
Sarolta Cump: Generally, for most people we make an access file and then maybe even a DVD or a low resolution viewing file like an MP4, sometimes with an h.264 codec.
We also make a preservation log. All of our preservation and transfer is supervised so the technician records in real-time all the artefacts, what the playback equipment was, what machine we cleaned the tape on, how many times we cleaned it, etc. We capture all that kind of metadata and then we get back to the client. We use PREMIS24 in part, but also our own in-house schema.
PACKED: Do some clients come to control the transfer themselves at times?
Sarolta Cump: No, because we don’t do any colour correction and we don’t really do restoration either. We do the transfer and try to do it without any aesthetic choices. We are a good place for people who want to make an uncompressed file and then take it somewhere else to have it colour-corrected if they are interested in altering the piece.
Lauren Sorensen: Our mission and goal is to preserve the work as authentic instead of making it into something else, but on the other hand, we get clients that are filmmakers and it is fine if they want to turn it into a separate chunk of the file that could be something else. Those types of clientele will have their own workflow for editing and production, however, and that is outside our scope of work.
Sarolta Cump: That is the other group of people we work with the most; independent filmmakers or documentarians who are making works for which they use old footage, or who have their old U-matic and 1/2" open-reel tapes and who want to use their footage or simply look back at it.
Lauren Sorensen: Sometimes they don’t have a playback machine for it. A lot of these types of clients say, “I could just transfer it myself but I don’t have that machine anymore” and they don’t want to go to a dub house because they are looking for playback equipment that is in refurbished shape. They want special handling and technicians who are going to run it through a couple of different time-based correctors to find the right signal, etc.
PACKED: So you wouldn’t erase drop-outs for instance?
Lauren Sorensen: We don’t, as a general practice. Much of the time, dropout and other artefacts are not detrimental to the meaning of the work and what is conveyed to future audiences. If it does feel overly detrimental to the point of damage, there are further preservation actions that we can take. Dropout could be considered part of the provenance of the piece. We have clients who have other perspectives on this idea, and we work with them to think about quality and restoration work and what that might mean.
So, I would say that there are contexts for which clients and partners would want to do restoration; we are a small department so we don’t work on restoration at this point in time, but BAVC has been involved in restoring video art, including Ant Farm’s Media Burn in 2002. This was a larger project that resulted in production of a DVD and we are now working with Pacific Film Archive25 in preserving more Ant Farm group work documentation on ½” open-reel.
I should mention too, that if there is excessive dropout, we will pursue treatments such as incubation of the tape (also known as “baking” in a controlled environment, with client permission). We will also try the tape on multiple playback decks and time base correctors which can help us minimize artefacts and dropout and get a more authentic signal to the original.
PACKED: Why did you go from an Aja Kona capture card to a Black Magic capture card?
Sarolta Cump: The Kona card was actually donated, and after they did some research on different capture cards they decided to work with Black Magic instead. I think that the workflow is a little bit simpler with Black Magic because we can go directly from analogue to digital without an analogue-to-digital convertor. The convertor is already within the capture card.
PACKED: Are you thinking about doing MJPEG200026 as well?
Lauren Sorensen: No. I haven’t researched into that codec much, so I could be wrong, but I’ve heard from colleagues that many types of MJPEG-2000 require proprietary hardware for decoding. We have discovered that there are a variety of “flavors” of MJPEG-2000 and many of these will only playback on certain operating systems and media players; so we generally do not recommend it to our clients and partners.
Sarolta Cump: Yes, it requires proprietary hardware to decode it. Our predecessors did a fair amount of research into it.
Lauren Sorensen: Yes, we actually have a compression study, which is something that is going to be in the Electronic Media Group publication. It was presented at the EMG conference in 200927 and it basically outlines why we use 10-bit uncompressed codec as opposed to other codecs: just because we don’t want to get stuck with proprietary hardware. The individuals, smaller archives and institutions that we work with much of the time don't have a lot of money to purchase this proprietary hardware, and archives which may ingest material will not want a codec that requires hardware as a barrier to playback, inspection, muxing and transcoding.
PACKED: Is the equipment you have catalogued?
Sarolta Cump: We have an inventory of our equipment; it is in BAVC’s general database, is called ‘Salesforce’ and is a cloud database. This is where we keep detailed records of repair and maintenance of the playback equipment and analogue monitoring tools.
PACKED: Apart from the transfer rooms, where do you store your equipment and all the spare machines?
Sarolta Cump: We have a storage closet and elsewhere, in the larger facilities, we have some storage space as well, but generally we just keep all the equipment we have in a climate controlled environment.
PACKED: What are the conditions in this closet?
Lauren Sorensen: In the closet, for storage, it’s kept at about 60 degrees Fahrenheit and 45-50% RH; we also have an HEPA air filter running all the time in this space.
PACKED: Do you have a regular maintenance protocol for your equipment?
Sarolta Cump: Right now, we are establishing a protocol for a regular maintenance routine, since the department-staffing shift in October 2010. In the past BAVC has done maintenance on a deck-by-deck basis but we also do an annual calibration when we bring in an engineer to calibrate all of our time-based correctors and check our cabling, etc. Heather Weaver28 and her partner John Selsey were two of the key people of the preservation department here in the late 90s and early-mid 2000s. They were doing the transfer and also a lot of this serial technique around it, a lot of testing and actually, John Selsey is the one who comes to BAVC to help coordinate engineering and other things.
PACKED: Will the cleaning of the tape path, the heads, etc. be part of that protocol?
Sarolta Cump: We clean the heads in between each tape we transfer, because it leaves residue, and is always part of the transfer process. For U-matics, and 1/2" open-reel we clean by hand and for Hi8, because it is more touchy, we have a cleaning cassette that we use every so often to clean the heads.
Lauren Sorensen: With Hi8 we don’t see sticky shed syndrome as often as with 1/2" tapes so we do not clean heads and tape path every time. With Hi-8 and VHS formats we do it less regularly. We clean U-matic and 1/2" systematically. They're where we mostly see hydrolysis.
PACKED: What do you clean it with? Isopropyl alcohol? Paper? Cotton swabs?
Yes, we use isopropyl alcohol, though we are now considering denatured alcohol as another possible option. We use densely knit cotton that we have to purchase from a medical supply company. It doesn’t leave any residue or flakes of cotton; we also occasionally clean the whole deck, and use a small computer vacuum and cotton-tipped swab.
PACKED: When the oil packing is gone or if a problem has its origin in the mechanical or electronic parts, is this when you would call a skilled technician to do the maintenance?
Sarolta Cump: Yes, if there is an actual mechanical malfunction with a machine, or when the oil packing is gone, that’s when we send them out. But our 1" machine for instance is holding up pretty well and it rarely needs repair.
Lauren Sorensen: We will also send these decks out for yearly maintenance and technician evaluation,as well as calibration of monitoring tools such as Waveform/Vectorscopes.
PACKED: Do you still buy equipment?
Sarolta Cump: We haven't for a while because of budget concerns (we are a not-for-profit). However, we purchase them as often as we can due to the rapid obsolescence of such decks. I think that we have about sixteen 1/2" open-reel decks. Some of them are for spare parts only, but we have got a good collection of those as well as a fair amount of U-matic players. So in the future we can get them repaired and use them when the others wear out and we can also scavenge some for spare parts, such as fragile videotape heads. That is one of the main issues we are running into. We have people that can repair the equipment in Northern California but a lot of the parts are starting to be close to retirement age without apprentices.
PACKED: Do the repair people that service your equipment have a stock of heads for instance?
Sarolta Cump: Yes, heads and different rollers and other items.
PACKED: Do you also buy some spare parts from them?
Sarolta Cump: We have our stock of spare parts and spare machines that we keep in storage. We haven’t really bought spare parts from these repair people, we gathered some on our own.
Lauren Sorensen: We mostly buy them from eBay and then we get them serviced.
Sarolta Cump: In the long term, we are starting to worry about our repair people because most of them are in their sixties, so we are not sure what is next for that. We work with a couple of individuals like Ken Zin who does a lot of work with us and has an office at the NASA Ames Research Center as well. He does videotape playback deck repair and calibration. In the past year we have started to realize that you not only need the parts and the machines, but you also need skillsets and people who actually know how to do the repair. You also need to figure out if someone who knows other types of machinery can do this specific repair work, and a lot of other questions.
Lauren Sorensen: Keeping the decks in good shape is a big issue in our field. We are focused on preservation of analogue videotape and playback decks needed for preservation aren’t produced anymore.
PACKED: Are you also transferring knowledge from one person to the other with BAVC?
Sarolta Cump: Yes, I have a lot of background in the soft engineering of how things are set up and connecting the decks, etc. It is knowledge in the workflow in terms of digitising from analogue.
Lauren Sorensen: I think that having knowledge in video production is extremely valuable to us because we can talk to our clients, to the filmmakers who used that equipment and know their workflow and think along the lines of ‘do they have a master that’s in a different format and what are they going to do with it? Are they going to use it again? What are they editing from this time?'
My background is in archiving, so that’s why I am working on the Dance Heritage Coalition Project29 because I’m involved in best practices, documentation, and digitization. I’m providing the department with that kind of perspective; what archives are looking for and what you could expect from them in terms of client needs? So we share knowledge on a regular basis and have internships and fellowships, and attend conferences in order to bring about awareness.
PACKED: Because those service people will soon not be available anymore to service your equipment, do you plan to obtain such knowledge on particular problems with the machines in-house?
Lauren Sorensen: That’s something that we would really like to do and we have talked about before; doing some kind of oral history or follow some training with these people because it is very specialized knowledge and skillset.
Sarolta Cump: It is also about starting to formulate a plan on how we can keep that. It could be by getting one of us to be an apprentice with them or by documenting it and trying to find other service providers. It is something that we are definitely concerned with and hopefully we will be figuring out a plan by the end of the year. We haven’t figured out whether we want to try to increase one of our skillsets to be able to repair decks - as we have all the old manuals and some of the repairs are more straightforward than others -, or whether we just want to find other people that we are going to contract, or whether we are going to look for funding to have a position that is dedicated to that. There are a lot of different options available and we haven’t narrowed it down yet.
PACKED: Do you keep all the documentation for the players you have? Do you have some sort of technical library where you keep them?
Sarolta Cump: Yes, we do keep all the manuals and we have a small open shelf with all our manuals and an inventory where we log our maintenance.
PACKED: Is the maintenance logged in the same database used for the decks?
Sarolta Cump: We are transitioning into that. It has been paperwork in the past but now the Salesforce documentation that we use allows us to actually attach documents so if we get information back from our repair service providers we can upload it and keep it with the decks.
PACKED: Before going to file-based format, were you doing all the masters on DigiBeta?
Sarolta Cump: We still do some masters on DigiBeta; that depends on the archive.
Lauren Sorensen: ‘The Kitchen’30 in New York for instance wants both a DV file and a DigiBeta tape.
PACKED: And when you look for spare parts is it mostly heads?
Yes, right now, we are mainly looking at buying heads for 1/2" and U-matic players. We are also starting to think about plans for the future, for formats like BetacamSP and DigiBeta, which are going to come down the line. We have two really solid Digital Betacam decks, but we have to start stockpiling that as well and start a collection soon for digital tape-based formats.
e-mail correspondence (October 23, 2012)
PACKED: Can you speak about the A/V Artifact Atlas31 that was recently put online?
Yes, it is an initiative we are collaborating on with Stanford University Media Preservation Lab32, and New York University Digital Library Technology Services33. Its purpose is to serve as a reference guide and glossary for naming terminology in documenting artefacts in transfer and what steps need to be taken to correct those errors, if any exist. It tells the user and provides the tools and knowledge base to become comfortable identifying and understanding these anomalies. People can contribute to the Wiki.