Learning lessons from StageTube, a multimedia web documentary

CEMPER - Centre for Music and Performing Arts Heritage in Flanders, Kunstenpunt - Flanders Arts Institute and meemoo - Flemish Institute for Archives made a mini documentary about technical stage heritage as part of our joint project, StageTube (link in Dutch). Our aim was to investigate how performing arts organisations can use appealing reports to increase visibility of digitised audiovisual archive content.

You can find the result of a test episode created as part of this project here (link in Dutch).

It was important to create a full episode, not just to properly understand the conceptual challenges, but also to be confronted with the practical obstacles involved in re-using digitised archive content in the context of an online publication. And it’s precisely because we completed this whole process – from conceiving the concept to its ultimate realisation – that we can learn lessons and share our findings. We hope that this will help us to support others who want to make their audiovisual archive content accessible online.

Lessons learned

General points for attention:

• Make sure you schedule enough time for each step. Not just because of unexpected delays, but also because content-related work sometimes needs time to mature. The total turnaround time for this project (not including our long lead-up time) could be six months. This isn’t six months of full-time work, of course, but you do need to be able to complete the process, and it’s only fair that you make this clear to everyone who is collaborating on it.
• The budget depends on how you want to approach it. In our set-up, with a number of permanent employees from the framework organisation who could take care of a lot of the preparatory and production work as part of their role, the main cost was appointing a temporary editor/video director. We agreed that €5000 including VAT was a comfortable budget for this, and communicated this as clearly as possible with the editor, within the principles of fair practice. We also needed enough money to clear the rights, and agreed on a budget of €1000 including VAT for this. Finally, it’s also true that some rights would remain unclear, and a small risk budget (e.g. €500) was therefore set aside to cover any issues afterwards.
• Make good agreements with the content staff about expectations in terms of length, tone of voice, target audience, etc. It’s also not a bad idea to consider representation and inclusivity in these agreements, e.g. is there a good balance of genders and diversity?
• The clearing of rights can take a long time rather than be very time-intensive work. Make sure you allocate enough time (and money) for it, with a good system for retrieving cleared rights (in the form of emails or contracts), and try to divide the work up, properly documenting all stages.
• The content-related and production work takes so much energy that there’s a risk of neglecting or minimising the communication and aftercare. Make sure you allocate enough time and budget for design, translation, subtitling, etc. Think about inclusiveness in the design through fonts and choice of colours. Provide time for feedback from a number of people who can look at it from a distance with a fresh perspective. Make agreements to share and help distribute the episode with potential stakeholders in advance.

We look at the complete conceptual and production process in more detail below.

The preliminary process

Without going over the full history, we should mention that the idea for StageTube had been growing gradually since the summer of 2017. This means it took around three years (!) to arrive at a concept that was shared by all the partners. Consultation has been slow, on an irregular basis, and sometimes with long intervals (less than once a month). But it takes time for a concept such as this to mature. It could have been more efficient, or definitely faster, if the organisations involved hadn’t been in a sustained period of austerity or transformation (Kunstenpunt had to transfer some of its operations to vi.be as part of an ‘efficiency exercise’, and meemoo and CEMPER have merged as organisations). One lesson that we can learn is that StageTube would never have been possible if some of the three organisations’ employees hadn’t donated so much of their time, going beyond the prescribed management agreements and frameworks.

Content concept

StageTube started accelerating mid-2019. We made the active decision to stop discussing the concept, target audience, format and other issues – and took decisions, resolved to be happy with the strength of the framework that resulted from these decisions, and learned to live with the restrictions that they imposed. This wasn’t easy, and at times we had to remind ourselves of our keen sense of direction in order to not start debating our decisions again.

Investing time and budget

Since none of us had any experience of a project like this, we had to be pragmatic when deciding how much available budget we had: how much could we spend? When we opened all our wallets and counted our pennies together, we had around €3k. We decided to work within this budget, and face what we couldn’t do with a sense of reality, and not stretch ourselves too far. Given the circumstances, it worked out pretty well with the money we had, although we encountered some unforeseen costs with regard to clearing rights. €4-5k would have been more comfortable for a text with archive images format (extra budget for rights clearance, video editing, design...). But we’re still proud that we always managed to have transparent conversations about the budget, as required by the Charter of Fair Practices, even though this budget was limited.

In addition to the financial budget, we also had to budget our time. Some employees from CEMPER, Kunstenpunt and meemoo spent many hours working on this episode, which was only possible because their employer paid their wages and was sometimes able to be creative with how they allocated their time. It’s difficult to estimate the exact amount of time spent because it was fragmented over a much longer period. This fragmentation also result in a process running quite gradually, which can be frustrating, but it’s also necessary to have a well-rounded concept. It’s a question of finding the right balance here: it takes time to develop, but there comes a point when focus is needed.

Engaging an expert and recruiting an editor

Based on previous experiences with CEMPER, we chose to work with an expert in theatre and stage technology: Chris Van Goethem. Chris is a lecturer at the RITCS (Royal Institute for Theatre, Cinema and Sound), and therefore saw his participation in this project as part of his educational assignment. €3K wouldn’t otherwise have been enough for an expert.

Attracting an editor who could take on the concept, research and production was by far our most crucial choice. The level of focus required was not feasible for the staff at CEMPER, meemoo and Kunstenpunt because of the multitude of tasks they are already responsible for. This person was also the one who cost the largest share of the financial budget. The extra value that the editor added was not only a breath of fresh air (in a discussion that had been dragging on for three years), but also a welcome switch to certainty, because a (time) budget had been agreed with the editor, and we therefore wanted him to complete his assignment within the agreed timeframe, out of fair practice. We also lacked the expertise to take on certain aspects of the process (filming, editing, production…).

We advertised for a new vacancy and 20 people applied for it. After reading the CVs, we selected four candidates for interview, before deciding that Mario was the best choice. Mario didn’t just have the required expertise in terms of content that we were looking for, but is also well versed in editing images and had his own archive, which includes stage and theatre technology heritage content from his previous assignments within the performing arts sector. The other candidates were mostly good editors with an interest in the subject, but didn’t have the same level of additional expertise in image processing.

The research phase

The research phase covers the period when the concept – ‘something about stage and theatre technology heritage via text and archive content’ – is transformed into a clear storyline with supporting image content.

Photo: KeepOnTruckin, via Wikimedia Commons

Finding archive content

Finding relevant archive content went relatively smoothly. We started our search using two methods at the same time. Our editor Mario and content expert Chris were able to search for content in a very targeted way thanks to their backgrounds. At the same time, Bart (meemoo) and Tom (Kunstenpunt) searched through the copious amounts of digitised content at meemoo, based on the search terms that came to mind when thinking about stage and theatre technology heritage. This was useful and good to begin with, but we noticed that the limited metadata in the meemoo archive held us back in terms of search efficiency. There’s a very logical reason for this: the archive content metadata needs to be added by the content providers (the performing arts organisations), who can struggle with technology and cost-benefit issues. In this context, it’s sometimes difficult for performing arts organisations to give personnel enough time to enter detailed metadata.

Image content from broadcasters’ archives is often still largely described in free text fields, which sometimes led us to interesting material. The metadata from the audiovisual archives of meemoo’s performing arts content partners is often much poorer: a performance title and date, with some information about creators/actors and possibly the place of recording at best. Richer metadata is indispensable for better findability of archive content. This is primarily an opportunity that the performing arts organisations themselves can seize, but meemoo and Kunstenpunt are also thinking about the technical aspects of how to process existing data into metadata for archive content, and using automatically generated metadata where possible.

At the same time, we can’t put all our eggs in the same metadata basket. The issue is whether it's possible to gain good metadata from the kinds of questions we are asking: how do you find images about stage and theatre technology without knowing about scenography or being particularly familiar with Flemish stage productions? Even the most extensive descriptions of archive content are primarily functional about the programme. For example, if a passage in a programme is described as an ‘interview with Niek Kortekaas’, you at least need to know that Niek Kortekaas is a scenographer, and can only suspect that there might also be images of the performance in which the theatre technology can be seen. In any case, the people involved still largely determine the content you can look for and find. This could have led to a Mario ‘angle’ creeping into the episode, with lots of content from Opera Ballet Vlaanderen.

From available archive content to a story

In an original plan for the research phase, we assumed a relative linearity: we find interesting content in the meemoo archive and then select interesting clips from this to create a storyline that ‘takes off’ from the content available. As always, of course, the ultimate process is more complex and convoluted. Put briefly: five potential angles emerged from a screening of the materials found. We chose one of these before once again delving into the archive to find additional clips. We find this kind of iterative back and forth to be crucial for this work.

‘Gender’ as a point for attention

We didn't pay much attention to the balance of genders when selecting content. For example, we could have made more of an effort to include non-male voices among the speakers, scenographers and technicians, although we did find a fantastic clip with a female technician from 1978. This point for attention was however lost in the complexity of production.

Time investment

Finding archive content was difficult, but didn’t actually take an excessive amount of time. All things considered, it doesn’t really seem necessary to look too hard for efficiency gains here. Browsing, serendipity and a dependence on outside knowledge (outside the archive) are all crucial for furnishing stories such as these with interesting archive content.

In the end, in open consultation with Mario, we spent a little more time with the editor than agreed, even in the preliminary stages. And the research phase turned out to be longer than anticipated too, because we wanted to provide enough space for Mario and Chris to work together to produce a meaningful story. In the context of fair practices, we therefore also gave Mario the opportunity to revise the proposed offer. It was only fair to Mario that he could deliver a product that he can be proud of. He ultimately invoiced for one extra day, so it’s never a bad idea to budget enough time and money anyway.

The production

The production phase includes the final editing of the interview, compiling the footage and clearing the rights to it, writing the copy and finalising the design.

Holding the interview

The coronavirus pandemic meant that for a long time it was unclear how we could hold the interview between Chris and Mario. Mario took the lead here and was eventually able to speak with Chris in a Covid-safe way at an outside location. His experience as a filmmaker also meant he had the equipment he needed to film the interview himself, which provided a number of additional opportunities. We initially intended to combine text with video clips from the archive. And the interview would serve as inspiration for Mario to write the copy and forge the clips into a coherent and fluent whole. But the unexpected availability of an audiovisual recording of the interview that was sufficiently high in quality led to Mario choosing to include excerpts from it. This also gave Chris a role as an expert in the editing, with his explanations sometimes included as audio to accompany the archive clips shown.

From interview to text and image

In principle, the footage also offers the possibility of rethinking the text and archive images format into a ‘full video’ format. But we made the decision not to do that in this pilot. We wanted to stick to the original plan, and that was also what we agreed with Mario. A remake of the pilot in a purely audiovisual format is certainly still a possibility, but we want to make the proper arrangements with the relevant people first. It would be strange to completely change the format in the middle of the production process.

Mario processed the interview into a readable text, and added the audiovisual archive content. He also complied with the request to limit the whole thing to a processing time (reading + watching) of around 20 minutes.

As mentioned above, a ‘full video’ format, for which there was no budget available within this pilot, is also particularly appealing. We are checking with VRT what is possible.

Clearing rights

All archive excerpts that came into our reckoning are still protected by copyright. We therefore needed approval from the rightsholders to use them. This presented us with the challenge of first identifying the rightsholders so that we could contact them to ask for their permission to use the content for the purpose we had in mind. For this pilot, we determined that we at least wanted a written agreement via email from the rightsholders for the content to be considered cleared. So we weren’t asking to draw up fully detailed contracts, but just a verbal or incomplete agreement was insufficient because it could be too great a risk afterwards. The approvals were archived together with the other documentation.

Some of the archive excerpts come from the VRT archive, with others coming from meemoo’s content partner performing arts organisations, and a third section coming from the archive that our editor Mario had built up from his own work for Opera Ballet Vlaanderen. This meant we had a first point of contact to direct questions to for all the excerpts. But it didn’t mean this point of contact was the only person whose permission was needed for us to use the clip. After all, as a re-user of the archive content, we didn’t have any insights into any agreements (existing or otherwise) about the rights between all those involved (producers, actors, dancers, lyricists, composers, videographers...)

As a first step, we therefore sent emails to both the VRT archive and all the performing arts organisations involved to explain the project and inquire about the rights status of the excerpts we wanted to use. The responses we received were very varied:

• Quick and clear reply with confirmation that the addressee is indeed the (only) rightsholder, with permission to use the excerpt in the context we described (e.g. La Monnaie).
• Quick reply with permission, but also with a referral to other rightsholders whose permission is also required. Kudos to these organisations that have properly documented a sometimes complex legal situation (e.g. Het Nieuwstedelijk, VRT). We encountered a number of scenarios in obtaining these extra permissions:
  • Extra permissions were provided quickly (e.g. Monty)
  • Difficult and lengthy communications to gain permissions, which was a time-consuming process (e.g. with a large arts organisation, but also with Unisono, whose core business should be this anyway). Elements that contributed to this included: illness or temporary unemployment of staff members, difficult direct access to the necessary people, emails sent to inboxes that are not monitored or quickly replied to, inefficient internal organisation/communication, lack of time or no priority given.
• Willing response but not enough internal knowledge of rights status and potential rightsholders to provide unambiguous permission that gives us sufficient legal certainty to use the excerpt.

The status of the permissions we were given differed in some respects:

• Free and without any restrictions, as long as we use it within the context described and state the names of the rightsholders.
• Free but with restrictions, e.g. excerpts of maximum 30 seconds (e.g. Opera Ballet Vlaanderen). This restriction was prompted by the written agreements that Opera Ballet Vlaanderen had made with other rightsholders, whose permission would therefore also be required for longer clips.
• Use licence within a time limit (extension possible) for a fixed administrative fee, as was the case for excerpts from the VRT archive.
• For a licence fee, the cost of which is determined per year based on usage. The licence itself is not limited in time, however. This was the case for a licence agreed with Unisono.

public domain, via Wikimedia Commons

An annual recurring budget therefore needs to be provided to keep the excerpts available for the latter two forms of consent.

Compiling clips

The cut-and-paste editing work that Mario was also able to do, to select and compile the most relevant clips form the archive items, required a lot of steps. As well as there sometimes being copyright restrictions (e.g. the lengths permitted for some clips), we also had to contend with ‘disruptive aspects’ such as subtitles, voiceovers, different resolutions, etc. which complicated how we compiled these clips. We hadn’t anticipated this complexity at all.

Once we’d established the choice of format, we also determined the tool we wanted to use for the pilot. We ultimately chose Microsoft Sway, prompted by CEMPER already having some experience of using it. This platform also met our two main requirements:

Sway is a user-friendly app for presenting text and images as a dynamic website. It’s a fully responsive design, so the online presentation looks good on all devices. The design templates allow you to build a great looking presentation very quickly, but you can’t apply your own house style or formatting preferences. It’s also not possible to make exports, so we always made sure we had good backups of all text and images.

Video clips can be uploaded directly in Sway, but we added the videos from YouTube using an embedding code. This is because you have no control over the first image to be shown in Sway, but in YouTube you can select your own thumbnail, so readers ‘browsing’ through the story see an appropriate snapshot. YouTube also offers automatic subtitling. It’s low in quality, but it can be a starting point before fine-tuning the Dutch text further.

Because the videos are not intended to be viewed without the accompanying story, they were uploaded with limited visibility settings: with the ‘Not Listed’ status, the videos are only visible to those with a link. This means you can watch the videos via Sway, but they’re not listed on CEMPER’s YouTube channel page or in the YouTube search results.

The Sway presentation itself can be shared as a separate link, but can also be embedded on websites – which makes it possible to explain the project context.

Time investment

We invested much more time than anticipated in the production. And indeed the slow rights clearance actually delayed us finalising it by almost four months. But the reason for this was external, and so the length of time that it took external partners (e.g. Unisono) to grant the necessary permissions to use excerpts needs to be taken into account.

Aftercare

Making an episode involves lots of work on its own, but how do you then make sure that the target audience actually gets to see it?

Communication

Even though we devised a communication strategy in advance – with a list of potential stakeholders, newsletter deadlines for the collaborative partners, etc. – it still turned out, as it often does, that this was commonly regarded as an afterthought. What do we do once we’ve sent out the newsletters? It’s easy to draw up a plan of action, but following it up is a different matter. The joy (and subsequent drop in energy) of ‘the production is ready’ actually comes too early, because the communication about it is equally important and still requires a lot of effort. It’s difficult to muster the energy to still write to the stakeholders (Rektoverso, STEPP, Etcetera.) now, even though this is absolutely essential.

Time investment

Here again, we anticipated a time commitment that was too tight. Where it was unavoidable that it took time to clear rights, because otherwise there would be no product, we noticed that we had to rush through the aftercare a bit. We also didn’t have a clear budget for this, so it was tempting to focus our energy on a new project instead.

Time investment

Here again, we anticipated a time commitment that was too tight. Where it was unavoidable that it took time to clear rights, because otherwise there would be no product, we noticed that we had to rush through the aftercare a bit. We also didn’t have a clear budget for this, so it was tempting to focus our energy on a new project instead.

What now?

Archiving

Archiving embedded text and video isn’t particularly easy using a number of standard tools. Meemoo tested out archiving a web version of the content and ended up with Webrecorder Desktop in combination with replayweb.page, a viewer for opening the archived WARC file. In addition to the end product itself, we also needed to properly archive the source materials (text and video), rights agreements and internal preparation work. This is now all saved together on Kunstenpunt’s Google Drive.

VRT collaboration?

There was a discussion between meemoo and VRT in the preparatory phase about how these types of stories can also be a preparation for short VRT mini documentaries. We’ll pick this up again after launch.

Expiring licences

Some of the licences that we were able to obtain expire after a year, or need to be extended. We don’t yet have a clear picture of this. In any case, it shows that the shelf life of a digital product like this is also linked to the length (and duration) of its public availability.

Knowledge sharing

In addition to this practical example, we’d also like to share with the arts sector our experiences gained in the form of specific tips and tools. On our radar:

• An article about how to increase the findability and usability of your archive contents. The lessons we’ve learned from this project: the importance of metadata for findability, good rights agreements (if the rights aren’t cleared and are owned by the custodian, it helps if there is at least clarity on who the rightsholders are) and being approachable for potential re-users (correct email address, e.g. in CatalogusPro).
• A tool, system or template to keep track of your rights clearing process as a re-user: what has already been cleared? Where in the clearing process are you?
• An article about the cost of user licences (unisono, VRT…) and the questions for arts organisations: do you do it for free? Is there a payment model? Who do you give free user licences to (with/without restriction in time, length of clips...)? Do you use Creative Commons licences? How do you balance maximum visibility with a potential revenue model?
• How do you archive a multimedia document (sustainability, licences, copyright agreements...)?

Authors: Tom Ruette (Kunstenpunt), Hanne Ampe (CEMPER) and Bart Magnus (meemoo)