All about... online video and the third age of the moving image

The history of the recent past is notoriously difficult: events are still unfolding; implications remain unclear. The third age of the moving image – following the eras of cinema and television – throws up its own special challenges, not least its almost incomprehensible scale. It’s impossible to calculate, but the number of works created in the 30-odd years of online video is exponentially greater than the entire moving image output of the preceding century. A universe of content so vast and scattered means that no two people can possibly share the same viewing experiences or memories.

But you have to start somewhere… What follows is a brisk tour through the first three decades of web film history, led by our curators and drawing on a two-year National Lottery-funded project which enabled the BFI National Archive to grow its collections of online video in a more systematic way than was previously possible.

– Mark Duguid and Patrick Russell

In the early days of the internet, no one quite knew what it would become. Websites were hand-built, links led down strange rabbit holes, and small communities of enthusiasts experimented freely with whatever the technology allowed. It was a new frontier, a patchwork of personal pages, university servers and hobbyist projects, in which – eventually – moving images began to appear in unexpected places.

The World Wide Web was developed at the turn of the 1990s by Tim Berners-Lee at CERN, as a tool for researchers to share information; video wasn’t part of the plan. But almost immediately people began testing the limits of what the network could carry.

One such experiment came as early as 1991 – two years before the World Wide Web publicly launched – at the University of Cambridge, where researchers pointed a video camera at the Trojan Room coffee pot. The grainy images allowed colleagues to see whether fresh coffee had been brewed without them leaving their desks. Now often considered the first ‘webcam’, it offered a simple proof of concept for transmitting moving images online.

For much of the 1990s, however, video on the web remained constrained by the limits of the infrastructure. Internet connections were slow and unreliable, and hosting video files was expensive. Downloading even a short clip could take many minutes. These constraints encouraged small, lightweight forms rather than long productions.

The result was an explosion of inventive short-form experiments – echoing the pioneering era of Victorian film. Animated GIFs flickered across personal homepages. Flash animations appeared on independent websites. Creators began experimenting with ‘machinima’, using video game engines to produce improvised films. These works often emerged from small online communities of artists, programmers and hobbyists connected through forums and message boards.

There were no global video platforms yet; instead, videos circulated through email chains, peer-to-peer networks and collections of links shared across the web. The spread of a clip might take days rather than minutes. Even so, something important was happening: ordinary users, rather than film distributors or broadcasters, were beginning to shape what moving images travelled online.

One of the defining forms of this early period was the loop. Many GIFs and Flash animations were designed to cycle through the same few seconds again and again. In part this was practical: short loops made efficient use of limited bandwidth. Soon, though, repetition became an aesthetic in its own right. The hypnotic rhythms of endlessly repeating animations, often bizarre, playful or deliberately nonsensical, created a distinctive kind of early internet humour.

Much of this era’s video culture has since disappeared. Personal websites have vanished, hosting services have closed, and formats such as Flash have become obsolete. What little survives is largely thanks to the efforts of enthusiasts. But this fragility is part and parcel of the spirit of the period. The early internet was a space of experimentation  – restless, provisional and constantly reinventing itself. To document it now is not only to preserve surviving artefacts, but also to reconstruct the memory of a moment when moving images first began to find their place online.

– Kristina Tarasova

Tech commentators helpfully supplied a name for the internet’s second era even before it arrived. ‘Web 2.0’ was less a software or hardware update than a focus-shift from static webpages to more interactive ones, and to a new focus on ‘user-generated content’.

What Web 2.0 didn’t signal was a shift to a more user-owned internet. Instead, the era (and soon the web) was defined by a new generation of giant social media platforms, the most successful of which proved far more enduring than the countless startups that crashed and burned during the ‘dotcom bubble’ of the late 90s and early 2000s; these included MySpace (launched 2003), Facebook (2004), Flickr (2004), Twitter (2006), Tumblr (2007), Pinterest (2010) and Instagram (2010).

By 2004 the conditions were largely in place for video streaming without the extreme compromises in quality endured in the pioneer days: more efficient video compression with the advent of MPEG-4’s Advanced Video Coding (AVC), the soon-ubiquitous Flash Player, and the early broadband technology ADSL (which offered then-unfamiliar speeds of 512Kbps or even 1Mbps for a few lucky domestic users in the UK).

The first dedicated video platform off the blocks was Vimeo, launched in November 2004; Dailymotion followed four months later. But it was YouTube that would, from the start, dominate the market for user-generated video content, an ambition signalled by its debut video, posted in April 2005. Fronted by the site’s co-founder Jawed Karim, the 19-second Me at the Zoo was clumsily shot, formally and aesthetically unambitious, and had next to nothing to say about its purported subject. But its real message resounded across the world: ‘bring us your videos’.

They built it; we came, in our droves. The resulting explosion of video across the internet wasn’t just massive – it changed the very character of moving image production. Pre-existing forms – from documentary to sitcom, children’s programmes to horror – took on new shapes and new voices. Almost forgotten genres, like instructionals (or ‘how to’ films) were reinvigorated. While activists tended to steer away from the more corporate platforms, a new breed of ‘citizen journalist’ took root on YouTube and Vimeo and elsewhere. Largely or wholly new genres sprang up – vlogs, vodcasts, reaction videos, unboxing videos – while others, like the once largely academic video essay, went overground. Meanwhile Vimeo, unable to compete with YouTube on its own turf, carved its own niche servicing the professional sphere as an invaluable tool for producers.

Most strikingly, the humble ‘home movie’, previously the most local of all moving image forms, viewed (not always willingly) by a few family members or friends, could reach potentially limitless audiences around the world; Charlie Bit My Finger, posted to YouTube in 2007, notched up well over 20 million views in under a year.

Not all ‘viral videos’ were so organic. The first online video to pass one million views was a deceptively simple ad for Nike featuring Brazilian footballer Ronaldinho, posted on YouTube in November 2005, while the platform was still in beta. A year later, YouTube was bought by Google for a breathtaking $1.6 billion.

Previously hesitant brands and advertisers jumped online, more than justifying Google’s investment. The pursuit of virality proved hit and miss, but advertisers had other tricks. Strategies regulated away on television (at least in the UK) re-emerged: ‘advertainment’, product placement, covert sponsorship and ‘astroturf’ campaigns. Meanwhile, an auxiliary army of ‘influencers’ willingly sprang into action to promote their favourite products (and themselves) for the cost of a free sample, if even that. Either way, the platforms counted the cash.

But it wasn’t all upsides. Platforms were dogged by controversies: for their light-touch, or apparently non-existent regulation; for their failure to adequately stem a tide of racist, sexist, homophobic or even paedophilic user comments; and for their disregard for users’ privacy, particularly children’s. Pervasive complaints of copyright infringement were a still greater threat, with lawsuits – notably one launched against YouTube by Viacom in 2007 – that might have been terminal had they succeeded.

Nothing, though, could stem the platforms’ relentless rise. By 2011, YouTube users were uploading 48 hours of videos a minute and viewing three billion videos per day. Meanwhile, new challengers were emerging: online retail titan Amazon moved decisively into video streaming, while another wakening behemoth, Netflix, was beginning its global march.

– Mark Duguid

If the platform era took online video overground, by 2020 the form had become a nigh-ubiquitous ingredient in pretty much everyone’s daily screen diet.

As throughout media history, the changes in the period were the result of the coalescence of hardware and software upgrades with changing patterns of production, distribution and consumption, and shifting forms and content. The shuffling off of the venerable Flash format, improved video viewing through HTML5, and the rapid advance of broadband take-up: all these further shifted the internet away from its text-based origins to being a primarily visual experience. Just as crucially, web consumption itself largely migrated from desktop to mobile devices, which included the already widespread smartphone. Video now easily accounted for the majority of all web traffic. The notion took hold that everyone (at least within certain age groups) now carried a cinema with them everywhere, in their pockets… and a mini-movie camera too.

Among platforms, YouTube remained king, but now had to share its kingdom with others, as Snapchat, Instagram and Facebook added their own video-sharing toolkits, each catering for differing generational demographics seeking different kinds of viewing experience. Livestreaming, initially closely tied to gaming culture and sometimes longform in the extreme, took off via the likes of Periscope and Twitch. YouTube itself could now host increasingly lengthy content when desired. Meanwhile, Vine sprung up in 2013 to make a virtue of the super-short form (its aspect ratio, however, was long as opposed to wide). Vine was a rocket: its spectacular ascent, particularly among millennial users, being followed by its falling and crashing in 2017. But already by then a site with a similar concept by the name of TikTok had quietly slid on to the stage…

The generation coming of age around this time was arguably the first to leave behind the old media hierarchies and instead pick and choose (or, to use a suddenly vogueish term, curate) their own individual mix of cinema, TV, games and online video content (while themselves being curated as viewers, by increasingly sophisticated algorithms).

Largely from the same generation sprang legions of solo ‘bedroom’ filmmakers, shooting and/or editing videos on their phones and sharing platform-space in a growing ‘creator economy’ in which ‘YouTuber’ or ‘Twitchstreamer’ could be actual jobs, and potentially lucrative ones. Startup videomakers – auteurs and stars of their own work – could hope to make a living producing increasingly polished content with the help of growing production teams. Figures as varied as Zoella, DanTDM and PewDiePie became as familiar to their fans as yesteryear’s movie stars once were to theirs. This sector’s growth was symbolised by the opening of YouTube studios (made available to creators who hit certain viewing figures) in several global cities – including one in London, at King’s Cross.

At the same time, traditional professional sectors producing content such as advertising, promotional, fashion, educational, training, charity and medical filmmaking increasingly turned to the internet as their principal destination (and to intranets, for staff groups or other targeted audiences). Commercials were themselves now central to the internet’s business model, and to YouTube revenues in particular. Production technology shifted too, as amateur and professional filmmakers alike abandoned shooting and editing on tape for 100% end-to-end file-based production – a short-term boon for creators and audiences, but something of a disaster for the archival long term.

Even politics was becoming a video-first game, as dramatically revealed in the audiovisual battle fought online between opposing camps during the 2016 Brexit referendum campaign. Thereafter the contest for different demographics would spill out across multiple platforms, with older voters targeted on Facebook, and younger ones on TikTok and Instagram.

The content rush was on. But with a global pandemic around the corner, it was about to accelerate yet further.

– Patrick Russell

As the 2020s began, the world of online video was changing fast; videos were getting longer – and shorter – and new forms and new audiences were emerging. Then along came the coronavirus pandemic. We found ourselves suddenly trapped indoors, isolated and very likely with an expanse of uncanny, unnerving time on our hands.  

We took to our phones and computers to communicate, to connect, to campaign and to distract ourselves. Online video increasingly became the primary space for socialising and community, for activism and anger, joy, exercise, education, public information and news. Covid fuelled an explosion of online video production and consumption. At the same time, a form of short looping video was dominating our phone screens.  

TikTok was already the most downloaded app in the US in 2018, but at this stage it was almost exclusively the arena of Gen Z. The form of the TikTok (short, looping videos) tipped into intergenerational ubiquity when it was mimicked by Instagram Reels and YouTube Shorts from 2020. At the same time, the early adopters were now using it to tackle concerns such as disability activism or confronting transphobia, while building a new world of often opaque parody. 

It’s come a long way from dance routines and talking toilets, but the platform remains home to strange forms of humour most of us may never truly understand. While much of this is novel, we can see in these short videos a return to the looping Flash animations of the early internet and the much-missed Vine platform.  

Even as short-form videos took off, so did ever-longer-form, with the rise of extended video podcasts and four-hour video essays. This was made possible by technical shifts on the platforms (until 2010 YouTube videos were capped at 10 minutes); by the evolution of live streaming; and by corresponding shifts in viewing culture, with online video becoming for many a background to everyday life, perhaps even more than radio or television before it.

And just like its media forerunners in recent years, online video has found itself enlisted in the culture wars. The dividing lines of the 2020s – from immigration to trans rights, to vax vs antivax, Black Lives Matter and the climate – voices on all sides have turned to online video, in any number of forms and on any number of platforms, to fight their battles. 

If we were to list every way we can interact with online video post-2020, we’d need a very long read indeed.  It is everywhere and it is everything. There is no news story that isn’t captured in online video. Video making ‘literacy’ has hit new heights: never have so many picked up cameras and shared so many moving images, in so many forms. 

And just as we reach this new peak of human audiovisual creativity, we are now faced with a flood of AI-generated videos, using as their source material the huge, precarious online archive these platforms have created.

– Will Swinburne

Our first task, as we began our two-year project to expand the BFI’s collection of online video, was to establish the criteria that would inform our preservation priorities. Some of these criteria reflect the archive’s long-established collecting priorities; others reflect the specificities of online moving image itself.

Naturally, we wanted to collect works with artistic achievement and cultural impact, but our primary objective was to collect and preserve a body of work that was – as far as possible at the scale we could collect – broadly representative of online video as a whole. That meant seeking to collect works that spanned multiple genres or forms, lengths, formats and qualities; that dated from the dawn of the internet to the present day, encompassing the technical and aesthetic evolution of online moving image. Equally, we wanted to draw from a wide range of online platforms: whether existing giants like YouTube or TikTok, streaming platforms we hadn’t collected from before, like Twitch, or defunct platforms such as Vine or BeBo.

Much of online video is, in effect, a migration online of pre-existing activity, which is interesting in itself as a reflection of both continuities and changes to established approaches, aesthetics and form. But we were also particularly keen to collect work unique to the internet: among others, ASMR (Autonomous Sensory Meridian Response), desktop videos, machinima, vlogs and reaction videos.

Of course, we could only collect so much with the time and resources available. We decided to limit our selections to works produced specifically for online exhibition or distribution. Also left out of our selection, for legal and ethical reasons, were works produced for the ‘dark web’; stronger adult content, including pornography and extreme violence; and content that carried significant safeguarding risks, either for the individuals featured or for archive staff.

Throughout the process, we were determined to highlight a diverse range of online creators, across gender, ethnicity, age and class, and spanning the UK; to draw on the knowledge of experts in the online space, including via a callout on social media; and to co-design the selection of works with the creators and, where possible, to follow their own recommendations.

This wasn’t the first time the archive had collected online video, but it was certainly the most focused, and it brought with it some unique challenges. How to be ‘representative’ given the huge scale of online video? How to judge what is a ‘British’ online video in a global internet? How to identify and, if possible, clear copyright?

A new digital process was developed to enable us to bring in files without having to manage cumbersome portable drives. This made life easier for creators, but even so, many had little previous understanding of preservation or the work of the BFI National Archive, and/or had a limited capacity to engage within the time constraints of the project.

For all these issues, we’re delighted with the results, and there’s no doubt that the archive is immeasurably richer for these 430 new works, which individually and collectively explore new territory for our collections, including webcam video, ASMR, and new formats like vertical-ratio TikTok video. And the Daily Star’s livestream of the ‘Liz Truss lettuce’ proudly takes its place as the archive’s longest-ever work – for now…

– Becky Vick

Internet video is so immediately available to the viewer that you may think preserving this medium would be a simpler task than, say, sourcing film prints or video tapes. But while the vast majority of works we have preserved are still accessible through online platforms, archiving requires an agreement with the creator and an attempt to find an ‘original’ digital file. By ‘original’ we mean the video file that was uploaded to any platform or website, before it was processed by that platform and redelivered, often at a lower resolution, to your browser.

Unfortunately, these original files may have been lost – to time, to the obsolescence of older technologies, or simply due to the lack of robust archival practices among creators. The archive has no official relationships with platforms such as YouTube or TikTok, so these corporations can’t facilitate preservation in the way that TV broadcasters, streaming platforms or film distributors and studios have for other mediums. So often, we must hope that creators are willing to explore old hard drives, broken laptops and old editing software to hunt down a file they might have last opened ten or 20 years ago. 

When these avenues have been exhausted, archivists are left with a difficult decision: give up on preservation altogether or make do with lower resolution files downloaded from online platforms. When making this decision, we consider how ‘at risk’ this work might be, its cultural significance, the quality of the available file, potential alternative works and the likelihood of sourcing more suitable preservation material in the future.

Online video production is a hugely decentralised process, and was especially before the platform ubiquity of the 2010s. Works can be filmed, edited, rendered and distributed all from one computer or phone, before being reproduced almost infinitely across the internet on different platforms and websites, and in different aspect ratios and edits. Retracing the journey of these moving images back to an ‘original’ digital form is a significant challenge for both archivists and creators as we attempt to preserve these works.

– Will Swinburne

The quality control (QC) process for our online video collecting ended up being quite different from our typical archival QC processes, requiring us to be particularly flexible. Many donors for this project were amateurs, making videos for the love of it, and weren’t necessarily thinking of the longevity and future-proofing of their files.

We received a huge collection of files with different codecs and compression states. The video I Can’t Believe You’ve Done This, for example, was supplied to us as an AVI file, which faces high risks of technological obsolescence and data corruption over time. For this reason, the BFI National Archive chooses to migrate AVI files to the contemporary Matroska (MKV) format.

Alongside increasingly obsolete formats, we also were given files downloaded directly from online platforms, which expressed their metadata in sometimes unfamiliar forms (for example, aspect ratio was shown in MediaInfo reports as 0.563, instead of the more conventional – at least for filmmakers and archives – 9:16). The project taught us new ways of approaching QC for born-digital web content which will be integral to how we approach similar files coming into the archive in the future.

Meanwhile, the very nature of online video – particularly in its earliest manifestations – meant we had to be more tolerant of variable standards. Even when we were able to acquire original files, many were characterised by the inherent limitations of the cameras they were recorded with – particularly early smartphone video. Some comparable issues arise with amateur home movies, which are common among our collections – but 8mm film, though it can’t match 35mm, has its own aesthetic appeal. The same can’t really be said of sub-par video files, but there’s clear archival value in preserving something of the aesthetic of the early internet, warts and all.

– Iris Matheson

The archive’s collection of online video has been supported by the BFI Screen Heritage Fund, awarding National Lottery funding.