ThruYou, a music and video mashup project created by the Israeli musician and producer Kutiman (and described in more detail here) is a set of eight videos (seven songs, plus a brief making-of video describing Kutiman’s process), hosted on YouTube and also available on thru-you.com. Preserving ThruYou presents an interesting and complex preservation challenge. Unlike some musical or video works, ThruYou’s significance comes not only (or even primarily) from its inherent musical or visual qualities, but also from its place as a notable work in the context of YouTube and emerging developments in social media and user-contributed content. Minus a sense of this context, a future viewer might enjoy the music or the visuals but miss significant aspects of the work’s importance. In order to fully reflect the significant aspects of the work, I will try to preserve not simply the work itself, but also the source videos used to create it, some important and relevant reactions to the videos, as well as a sense of the YouTube context from which ThruYou was created and in which it was presented.
Preserving video and audio materials:
Preserving the core of the work itself may actually be the simplest task. Many YouTube video downloaders are available. For this project, I’ll be working with the youtube-dl downloader. This is a command line program for downloading video from YouTube and other sites. While it appears to be more difficult to use than the simpler YouTube download extensions I’ve used in the past, it offers a vast range of specifications and options. It’s recommended by ArchiveTeam, who have experience downloading YouTube videos for archival purposes, and who’ve conveniently listed detailed recommended settings for high-quality video and audio. The program also allows the user to capture description metadata and annotations (those pop-ups that appear over videos on YouTube); both are preserved in XML format. In general, my priority will be getting the best quality audio and video available (in the case of ThruYou proper, the highest resolution available is only 360p, though some of the source videos may include higher resolutions).
The full list of videos to be preserved from YouTube includes the seven music videos and the making-of video that make up the initial ThruYou project. In order to further preserve the full context of the videos’ creation, I’ll download and preserve the original source videos as well, when possible. Most of these still exist, but some have since been removed from YouTube. Lists of the source videos are included on each individual video’s YouTube page.
In the future, it will likely be necessary to migrate the video files into new formats. For this project, however, the specific encoding format is not essential to the work; as long as the resolution and quality are maintained, they may be transferred into new file formats as necessary to best allow for compatibility, preservation, and access.
Preserving the YouTube and website context:
The close integration of the work within the context of YouTube is critical to the meaning and significance of the work, so efforts will be made to preserve a sense of the visual and functional context of both the YouTube and website access points for the video. Setting up a fully functioning mirror of the YouTube context would be effectively impossible; so much of the context is dynamic, from the comments to the automatically-generated suggested videos on the right side of the page. While a simpler version that embeds the video in a YouTube formatted page would be easier, it would still be difficult and could become complicated to maintain and access over time. The importance of the YouTube context is in showing the context of the video within the social media setting of YouTube. As such, a combination of crawled and saved YouTube pages and screenshots should provide a good sense of the original context. Downloading the pages may present some difficulties. The Heretrix crawler is open-source, but seems to be inconsistently supported on Windows. I’ll give it a try to see if I can get it to work. If not, since a limited number of pages are involved, using Chrome or Firefox’s “Save Page As” functionality might suffice. The full comments for each page will also be captured using the Firefox extension Complete YouTube Saver, which uses the YouTube API to access and download comments (it can also save a complete copy of the page, which is another option for downloading the YouTube web materials).
The full original ThruYou website pages will each be downloaded as well. Since the visual setup and context are important, screenshots will be taken of each page on the ThruYou site as well as each YouTube page. While this remains a somewhat inaccurate representation in the latter case (since YouTube’s design has already changed since ThruYou’s release) it does give an idea of how the video and the original component videos appeared in the full YouTube context. Likewise, the crawled web pages will provide only a snapshot of the page–the dynamic generation of content in the related videos field of the page, as well as page view numbers, ratings, and comments, may have limited or no functionality or renderability outside the live YouTube context. While retaining this functionality would be prohibitively difficult in this situation, the combination of crawled pages and screenshots should provide a framework for future researchers to understand the original context of the project materials. The ThruYou website itself is more static, but its design (a distressed, parodied version of the YouTube layout) relies on knowledge of the YouTube context as well.
Kutiman’s own making-of video also includes video of him using YouTube, searching and clicking through pages and viewing some of the videos used in the project. So in addition to being a part of the project itself, this can serve as another contextual resource for future researchers, providing a (low-res) glimpse of YouTube at the time of the project’s creation and showing the site itself in action.
Preserving reactions to ThruYou:
Given the centrality of the broader context and conversation around the piece in demonstrating its significance, I also want to collect some external materials to show some of this conversation. Much of the reaction to the piece (news articles or tweets about the project, for example) is likely to be preserved elsewhere. A few items, however, warrant preserving along with the project. Most notably, I’ll download and preserve the Researching ThruYou website. Originally created by a fan when YouTube temporarily took down ThruYou, it collects a full list of credits, with links to the original videos. In most cases this site provides additional information about the source video creators (links to personal websites or Twitter accounts, for example). For two of the source creators, the site links to statements from the creator about their reaction to ThruYou. Since I wasn’t able to find much information on the reaction of the source video creators, and since it seems central to the collaborative aspect of the project, I’ll download and preserve both of these (one is a Reddit thread, the other a YouTube video; both will be collected using the same tools as above). In addition, as noted above, the full YouTube comment threads will be downloaded and preserved.
Conclusion
Much of the challenge in preserving this project resides in the conflict created in trying to fit a work created in a dynamic web environment into a more static preservation context. In general with this collection, I try to deal with this by collecting both the raw original materials (such as the videos and comments) along with elements and representations of the dynamic context (crawled web pages and screenshots). Hopefully the combination of these materials will allow future researchers to reconstruct a sense of the active context of the original project.
It will be interesting (and hopefully enlightening) to see how these preservation intentions will be challenged or changed by actually carrying out the process of creating an archival information package.
I was just going to use one of the simpler YouTube downloaders like ClipConverter, but I might use the youtube-dl downloader that you’re using instead since it was recommended by the Archive Team and is intended for archival quality – great find!
This is really fantastic. Great work identifying all of these YouTube specific tools and thinking through what the core and key assets are that you want to make sure you get. I think you’ve really nailed it in terms of each of these different kinds of materials that should be collected and you are thinking through the right questions with each of these. It is great that the researching ThruYou site is up and that you could pull that in as well. In some ways, that site seems to provide you with a good bit of the index and point of entry into describing much of the content you are actually preserving.
You are right to note that a key challenge here is how to turn something that is very linky and very dynamic into something that is much more static. This is really one of the most fundamental challenges of born digital preservation in that we are often pushed to get things to require less dependancies to make them more durable for the long haul.