 Thank you so much for coming to our presentation on the EMI Canada project that we're doing at the University of Calgary. That is nearly done. We gratefully acknowledge that Catherine and I live and work on the traditional territories of the people of the Treaty 7 region of southern Alberta. The city of Calgary is also home to the Métis Nation of Alberta, Region 3, and thanks for welcoming us on this land in Washington, D.C. The traditional territory of the N'Kochank and the neighbors and ancestral lands of the Piscataway and Pamonki peoples. So my name is Annie Murray. I work in archives and special collections. My colleague, Catherine Ruddock, works in our digital services and our colleague in spirit is Kate Cawthorne, who's in Calgary actively doing the digital preservation on this very project. A little overview of what we're going to do, we're going to tell you what our project is and has been, share what our philosophy of access is, some barriers and solutions we have developed, or barriers we encountered, solutions we developed, and tell you about some of our workflows. So just to introduce you to what this collection is, the EMI Music Canada font is the complete corporate archive of Capitol Records Canada and EMI Music Canada. So it covers the periods 1949 to 2012 when Universal Music Canada bought EMI. So it's a 63-year span of all kinds of archival materials, about 5,500 boxes that have come from Toronto, more than 42,000 recordings in 94 different audio-visual formats. So very simple and easy collection. The preservation project has been supported very generously in three phases by the Mellon Foundation and so here you can see the sort of phases and dates. So we gratefully acknowledge the Mellon for their support of this, what has been truly transformative for our library. So we want to show in this briefing how we've addressed the challenge of migrating, preserving and providing access to the audio-visual components of the EMI archive. Supported by both the donor of this vast archive, Universal Music Canada and the Mellon Foundation, we've worked since 2016 to migrate and digitize a significant portion of the audio-visual materials in this collection. We're preparing to launch 105 terabytes and counting, as of last week, of migrated and digitized audio and video recordings in March 2023. Today we're going to focus on access and preservation, just a sec, oops, the access and preservation dimension of the project more than on the details of digitization and migration, but we will gladly talk to you about any aspects of the project. So we're available later and we'll give you our email addresses. In our earliest conceptions of providing access to this complex media collection, we imagined that following the migration and digitization of 94 analog and digital AV formats, we would be offering a digital collection in a web-based access system. In the world of archival audio and video recordings managed by archives and special collections units, however, this is not a standard approach to access. The more routinely followed practice is providing access to AV via a dedicated offline computer workstation in a physical reading room. This route offers a very controlled means for researchers to access a collection and is often decided upon for very understandable reasons linked to security and copyright. However, given our strong relationship with the donor, surprisingly a record label is highly supportive of making their materials available to scholars and the public, we did not want to make the only access to the recordings in our physical reading room. It just did not make sense. At the same time, knowing that this collection is highly copyrighted and represents outputs varying from original raw studio recordings to final commercial releases and music videos of major music label, we wanted to balance meaningful digital access with respect for the rights holders and Canadian copyright legislation. So that means we would need a sophisticated access system that offers us both granular control over assets as well as a powerful capacity to organize and facet resources. The very nature of the recordings lent themselves to complex digital objects often with a single multi-track recording resulting in multiple files. Through consultation with our team and AVP, we developed requirements for our systems based on our project goals. There were 31 outcomes, I will not recite them all, but some of them included and those that are linked to access are having methods for preserving and making accessible complex digital AV objects and understanding the hierarchies and relationships amongst the digital objects for multi-track recordings and presenting these complex objects in a way that helps users understand and see the relationships. Also being equipped to promptly respond to access requests from donors, researchers, educational users and public users in that all materials in any format that wasn't yet digitally accessible could be prioritized and addressed through a prompt prioritization in-house. We wanted to provide digital assets in a very user-focused service and we wanted the users to be able to have a customized usage experience where they could create light boxes and share assets. So we wanted the EMI materials to be discoverable in various systems including Ex Libris Primo which is our discovery layer as well as crawled by Google. We wanted them to be accompanied by robust descriptive rights, administrative, technical and preservation metadata. We wanted them to be logically organized into individual and collection containers to facilitate easy access, discovery and use. We wanted to have intuitive search mechanisms and faceting and to be very secure and managing requests and approvals in accordance with copyright legislation and the permissions we got from donors. So our plan was and is to deliver this robust metadata, multiple high-res images of each object, clips of the recording whenever possible and an easy to use request system so a researcher could request access to the full content of a streaming recording. We identified different use scenarios when researchers could or could not access to a downloadable file. We imagined classroom and research project scenarios where we would grant assets for a longer period of time. And then also have a very good physical workstation in a reading room if researchers wanted to engage with the materials in that way. So with those lofty and to us highly logical goals in mind we set about our work. So Catherine is going to describe how it went. Thanks so much, Annie. So as we near the end of this project and have the chance to reflect on six years of our lives here and out of these day-to-day operations, we can see some themes for how we've been able to accomplish making this archive accessible. I'm going to take you through our different categories of barriers and how we solve them. Our first three legal, technical and format were accomplished as part of our original project plans in these different phases of the project. The fourth category, perceptive, was not explicitly envisioned as project work. However, now we can see that these were some common barriers as a common thread that was coming up from our project team that we need to solve through change management plans. So I'm going to jump into some specifics. Our copyright office and our principal investigators use the Canadian Copyright Act to establish what we could migrate and make accessible. For migration, we used the maintenance and management exception in the library and the Copyright Act to migrate obsolete formats. We intend to provide mediated access to migrated media on a per-request basis for research and private study under library exceptions and the Copyright Act. And we're going to be providing a preview of content. So these images, these high-resolution images of the media carriers that are available worldwide, anybody can download them. And we also elected to create these 30-second clips of content for streaming, like you'd observe in any sort of streaming media service. Our Copyright Office applied Canadian Fair Dealing and legal precedent Bell V. Socan in 2012 to justify this. Finally, we incorporated language from our archives researcher agreement, so like a form, a paper form, that we usually have researchers sign before they access archival material. We incorporated that language directly into our terms of use in our digital asset management system and user license agreement so that we can just track that directly within the system. They're not going to have to sign a separate form. Our donor was on board throughout this whole project with providing the level of access afforded to us in the Canadian Copyright Act. They also agreed that we could make copies for preservation. Any user requests for publication or commercial use will be escalated to Universal Music Canada to review and approve or decline directly within our digital asset management system. And then I'll note this gap that emerged later on in the project where we didn't originally ask to process data through third-party applications to supplement any description. And this is something that we are gonna need to discuss with the donor in order to process our collections through different AI tools like facial recognition or speech to text. We have those tools directly in our digital asset management system using third-party applications, Azure and Google Vision, but we can't use it right now with our current agreements. We were able to procure two systems to solve some of our technology barriers to access. We didn't have these systems before we started this project. First, we implemented a digital asset management system, Cortex, by Orange Logic. And in there, we're able to build these complex object packages and different levels of granular access. So our clips, we have these as view only for anybody worldwide, or we will, March 2023. We'll have these images available for download, but private images will be restricted. And then we'll have full packages with all the media files, multi-track or whatever it is, available for request and download, or view only. Users can indicate what level of access they want, for what purpose, if it's for research, private study instruction, or commercial use. And it goes directly into our system and we're able to do our reference in there and provide access or escalate that to Universal Music Canada. We also implemented LibSafe by LibNova as our digital preservation system. Here we submit our packages for preservation processing and long-term storage and send access packages to Cortex for end user access. So we have an integration between these two systems. I'm going to jump into these workflows in a little bit more detail in a moment. We established an AV conservation program through this project. Our program employs a mix of in-house and outsourcing digitization that's dependent on format. In-house we can currently migrate 42 different AV formats. And our ability to migrate locally depends on how long the equipment that we purchased over eBay and all kinds of different areas will last, how long the media is going to be viable for and our in-house staff expertise. We needed to outsource 28 formats to date. Over the six years, we've migrated 19,000 objects and counting. In terms of reference formats, decision making, we base this based on best practices from other cultural heritage institutions. And we hope that the formats that we've made readily available in our digital asset management system are going to meet the needs of our end users. But this is something that we'll be monitoring once we launch the collection. So the tasks and responsibilities for the project were divided amongst units and team members. And we kept everybody up to date with monthly project meetings. And in those meetings, we got a general sense of change angst from team members about the level of access that we wanted to provide. Can we do this? Here's some examples of some of the areas where we observed some perceptive barriers. So we solved the barriers, like our legal challenges or technical challenges, but we had some just general angst from our team members. So just for example, we're the only ones doing this. Why are we doing this differently? Nobody else is providing this level of access. So our principal investigators or project sponsors were really key and this is why we're providing this and having that communication message very clear to our team members as these questions emerged. The clips were actually a very contentious issue within the team about can we actually provide clips when there's a good legal precedent and our copyright office was on board. So what we've done is, well, yes we can. And again, some clear messaging is really important here, but we also have some risk mitigation workflows. So if there are questions about why we're providing this or if we need to change a clip to do a different 30 seconds, we'll have a mechanism to change that clip out as we need to. These systems are new for us. We only started implementing our digital asset management system in April 2020, digital preservation system in 2021. And so we have been using them and we've been using them in production, but not for this collection. So there's some general angst about system reliance that we're continually coming up against. We plan to do some really in-depth user experience testing in the new year prior to launch the collection to hopefully mitigate some of that. But it is something that comes up a lot from the team. And then for end users who may expect to have to come into the reading room or maybe that is their question, we need to change our end user expectations of no, we can deliver this to you offline. Do you have the software that you need to read these multi-track recordings? You don't have to fly here. So it's changing our reference interview techniques and changing what we ask of our users. We developed a preservation first workflow that would allow us to ingest migrated EMI objects directly into our preservation system and then derivative assets into our digital asset management system. We base this workflow structure on the OAS model. The migration, imaging and packaging happens within our AV conservation team. And our metadata creation is technical metadata by that AV conservation team and descriptive metadata from our archivists. The metadata actually goes directly into our digital asset management system first and that's where we're actually tracking the whole workflow for these objects. Once the package is complete, well, the submission information package goes into a staging area in our digital preservation system and then that metadata is paired with what we have from Cortex and an archival information package is put into LiveSafe. The digital preservation system is able to ingest up to 1.5 terabytes of content per day and process it. It performs sanitizing checks, preprocessing checks, packages the data and metadata into these apes and we're creating three copies to own a cloud, one local storage. It's providing audits for us and scheduled validation checks based on some configurations directly for this content. Once it's fully preserved in that digital preservation system, then we'll do a sync to our asset management system where these derivative copies are going into our dams and the clips are generated at that point. So we're not storing the derivatives or those clips in the preservation system. And then at that point, my team takes over, we'll add in some more item or file level metadata, descriptive metadata and get it ready for publication. And this is an example of what an object will look like where you can see, you can browse through the different images of the media carrier. Users will be able to stream clips, so up to 30 second clips. Below the clip or below this preview window, you can see the full package contents for what we have available. So the users will only be able to see the content, the images and the clip, but then they could see what is contained in the whole package. So this is a multi-track AV object. And then on the right hand side, you'll see the metadata action button. So users can add to cart and tell us how they wanna use this object. They can ask us questions about it and that goes into our reference queue. And then we provide metadata and writes information about this object. So our final recommendations for our access and preservation plans was first, I feel like we just misdiagnosed our level of change. We were treating this as a transitional change or maybe incremental at some points in time when really it was a transformative change for our team and how we want to provide access to digital content. And so what we're doing now is we'll come back in the new year and have a change plan with different change initiatives to make this change stick and have team members buy in a bit more. One of the things that we will need to do is build out a change team that is independent from the project team so that it's not, so that there's more trust and more openness in some of the change inks kind of areas. We did build our workflow preservation to access, which would have been fine if it was a smaller amount of data, but with the size of data that we were working with and our ability to only process 1.5 terabytes of data per day, if we need to quickly make content accessible once it's migrated, we found that we're gonna need to build an access to preservation pipeline as well. And so that's what we're working on right now. We're building a digital asset management system to digital preservation system workflow as well as keeping the preservation to access to be able to get content both ways and have the system synced. And then finally, in terms of scalability, we've had our head down and thinking about this EMI collection for six years. And now we need to pick our heads back up and think about how scalable some of the workflows that we've implemented are to other parts of our collection. So it's not an EMI workflow, it's an AB object workflow. And how is that applicable across other AV objects that we'll be preserving even in the way that we've designed our complex objects. Right now we're working on implementing museum objects within our digital asset management system and some of the same learnings for building these complex objects, granular access to different files and inheriting metadata are applicable across the board. So this is what we're doing now and able to apply across the board. Thank you so much to our project team across libraries and cultural resources. Here's our current project complement. And thank you so much to our sponsors, Mellon Foundation and Universal Music Canada. If anybody has any questions, we're happy to take those now. Thank you for your presentation. First of all, I thought it was fantastic. And then just in terms of actually what you're doing also a very brave project in terms of putting EMI, the entire record label out there on the web for people to use. My questions, by the way, my name is Ray Yusushin, I'm from Mississippi State University where we also have big record collections, the Mississippi Delta Blues, defunct labels that actually were seminal in I guess the birth of American rock and roll. My questions have to do with I guess intellectual property and how I find it fascinating in terms of also remix culture. I could see this immediately as a treasure trove for people that are interested and have a fascination with music, downloading a track, taking that one track that's kind of iconic within a song, and they want to remix it and do the version and then put it up on YouTube. What is the process in terms of intellectual property? Is the discussion okay? Well, once they start actually having success with the song, then they're gonna contact Universal Music Canada or is it, we're gonna trust the user to do the right thing with the music or what is the thought around that valence? And I guess the copyright, I never thought about, I would always, I would say I would be deferring like you to Universal Canada, but what about there were different players on each of those tracks, the bass and the drum and the singer, who owns the copyright of that? Is Universal just claiming all of that or how does that all work? Any color you can provide? The rights holder is Universal Music Canada. And so we've tried to design a workflow where when the user encounters the asset and they can look at all the previews and the pictures and they think, okay, I think this is what I wanna listen to, they can initiate a request and then we're at the other end of the request. Part of what they're going to indicate is the use they intend to make of that material, if it's for personal use, research use, or if it's in a classroom setting or they wanna put it in an exhibition, they wanna publish something. Depending on what they've chosen and the terms of use that they agree to, it may go to a different individual here or at Universal Music. So we will be able to carefully know what every person is intending to do. We know who they are, they have an account. So then they get the tracks, they play with it. Maybe Catherine can speak more to, I don't know. Do you have thoughts on the remix, the creative stuff? Hello. I mean, yeah, we're putting our trust in the users. So we've just been making it very clear. This is what we're clearing it for. And that tracks throughout the system. So when we approve it, it's at an item level that we've cleared this for this purpose. If they break copyright, it's on them. And so for us as well, we have, every user will have to have an account and we're tracking how those users, what content the users have access to. Yeah, and it'd be on the copyright holder, Universal Music Canada in this case, to go after anybody that breaks copyright. So Rick Ferrante with the Smithsonian Libraries and Archives. I have a couple of different questions, but I'll stick with the thread we're on right now. In that this idea that the publishing company is the copyright holder is kind of changing in the industry. So with regards to royalties, when these things are played, does Universal ask for that information so that that reflects onto what the musicians are paid? How is there, is there a reporting back to Universal in terms of what access has been delivered through this? When we acquired the collection, Universal Music, so EMI no longer exists and any new projects they take on with their artists they may have ongoing relationships. It was their wish that these materials be shared widely and used in research and for the public. And they granted us certain amounts of capacity to make that possible. Because we have viewed these as research uses, we have not approached it in a way that each research use of an archival recording is the same thing as something that would generate a royalty. However, in working with Universal Music, because we have the capacity to migrate and provide digitized content to them, we do provide them with high res recordings if they do reissues. So they have kind of a concurrent streams of activity going where some things are still commercial assets that they monetize and the stuff that we're providing access to, they don't necessarily view it in that way, which is a very generous position for them to take. Although we can track very well the use of assets, we haven't had discussion with Universal about here's who's hot, here's what's the most requested recording that could develop into a part of our relationship with them as they determine what they do with what they consider sort of a legacy collection. So I give them credit for this openness and sort of experimental willingness to undertake this, but we don't view it as a way to help them generate additional revenue. Having said that, we haven't launched the system yet to observe how it works. And luckily it is such a sophisticated system where we have such control over the assets that if we wanted to analyze the use and the donor had concerns or questions, we could adapt really quickly and reframe anything. So we are cautiously optimistic, but we are always cautious. And we have had a good understanding with them about what our aims are and the terms under which we provide access to the collection. Should that change, we are prepared to talk about that. We don't know how this thing will live in the world. And frankly, we're amazed it has gone so well in conceptualizing this sort of open archive feeling. Do you want to? No, I think you did great. Okay. Did you have another question though? Great. Thank you. I'm Jennifer King and I oversee the Rose Library at Emory University. And first of all, congratulations. This is such an enormous undertaking. It's really exciting to hear about it. My question is in the realm of kind of like the financing of the big build. And you talked about a lot of different parts of the organization that have had to grow and change to support this. But thinking about where the financial support has helped, how could you quantify? And then how does the system build for the audio-visual discovery which is really customized for such an incredible collection? How does it fit into the ecosystem of your digital libraries? And I'm imagining that you would say that this system is really growing that infrastructure too. But I'd love to hear you talk about that a little bit. I can talk about the general categories of the money and what it supported. And then Catherine can enhance. So it was a considerable outlay to develop the capacity to do this work. So simply the machines, the hardware, the equipment, the software, the studio space. It dovetailed into some other growth areas that were occurring in the library and availability of space. So some of it came along serendipitously and some of it came along quite intensely and suddenly. So it paid for setting up that capacity. It paid for a lot of the outsourcing and the initial costs of running these systems, running and implementing them, as well as consultation with AVP who guided us through a lot of initial foundational questions about what is access? What is preservation? The salaries of project staff, for example, imaging every object, you know, at least five ways to have a representation of the archival object. So as we all know who work institutions, these are the starting costs and that long-term operationally we bear the cost of maintaining these systems and processes. So that's the basics. Yeah. And in terms of like the digital content or our strategy for digital collections, we were able to get our digital asset management system because of this project but it was never envisioned just for this project, right? So from the get-go we were building this so that could handle all of our digital collections objects. We were using more like collection management systems like content DM before and then we got this digital asset management system which really just flipped how we, have built our objects, how we've built our metadata schemas, our integration strategies, everything like that. So we started off with this being, we had three priorities for our digital collections, our digital asset management system, EMI being one of them but then we had 60 other digital collections that we were migrating that were across disciplines and then a photographic collection that was new. And so that went into that project for that implementation from the get-go which is, was really fundamental in how you're doing this. Like I'm really glad that we didn't design this for one collection and then tried to catch up with everything else. Yeah. And yeah, like that would first, if I was doing a recommendation on implementing a digital asset management system, like get the scope of all the content that you're going to want to do. Yeah. Because that really helped, it helped with what, how we see libraries and cultural resources as a convergent environment across libraries, archives, museums, designing metadata schemas for all those different types of collections and where that content needs to be accessible. We're thinking about that from the get-go. Is that helpful? Yeah, thank you. Okay. So I guess that was part of my question. And it goes to the sustainability and the rights, the copyright issues that so intertwined with this. And so I'm just wondering where that ended up. Was it in the collection management system? Was it in LibNova? Was it in the dam system? Where do you see that continuing to be kept as you build more and more collection? So our system of record for metadata is going to be a digital asset management system. Like that's where we actively work on our metadata across all of our collections. And that's where our end users work. So like both our internal staff as well as our public members that are going to use this collection, that's our system of record for this content. We'll pull some of that metadata into our preservation system. But so we consider our preservation system as our system of record for the assets. But our metadata system of record is digital asset management system. So all the rights metadata, all descriptive metadata, technical metadata needs to be in that system. Does that? Well, thank you very much. I guess it's lunchtime. Thank you.