There is more choice in browser-based web video providers and implementation
technologies than there has ever been before - YouTube, Vimeo, Flickr, Revver,
MySpace, Facebook to name the popular fraction of what is available - and yet, delivering
video via the web can still be time consuming and complex. If anything, this complexity
is likely to worse before it gets better - however it is obvious that there is
a huge potential to develop the richness around video delivered online.
I've worked on a number of rich media and video projects, each delivering
something slightly different, and each having to work through the production of
high-quality media and wrapping an experience around it. I'll admit that getting
a single video up onto the web to share with friends, colleagues and fellow
tweeps is pretty painless - Facebook will allows users upload an entire
gigabyte of data in a single video clip, something I find quite astounding.
Equally I'd admit that the tools out there to capture video are better and more
commonplace than ever - with devices like the Flipcam, Nikon's D90 (my
favourites) and the huge range of dedicated video cameras out there being
capable of capturing video to a high standard.
The first challenge is blend of expectation and experience. Social media tools
allow us to produce and distribute video live or pre-recorded video in a matter
of minutes. The experience of YouTube is frequently cited by users as the norm,
along with the expectation that a project can and will deliver a similarly
powerful set of tools. Not a problem for brands making forays into video
delivery in the social media space - just pick up the toolset and use it,
produce discrete video elements and place them inside blog posts, against
products, within campaign pages or the like.
But where are the tools that allow us to provide a much more immersive experience?
One which is more than just embedded Flash, Silverlight or HTML5 <video>
elements in a page, where video is the primary medium, the experience itself?
They're in the hands of the designers and developers. Whilst standardisation
around the very recognisable rectangular video player is happening, achieving
the levels of richness above this still require some very specialised skills.
Then there is the minefield that is quality. An idea of what might be an
acceptable quality in the eyes viewers is one thing - but delivering that
quality to the audience who have come to expect video that is constantly 'on' (think TV) is
something which requires an exponential jump in the complexity of the production
workflow. The notion of quality in video is closely tied to the context of
the viewer; it's not enough to have the content and to deliver it at a preset
size, bit-rate and format - the capabilities of the device (and there are lots of them) which is playing the
video, in terms of both connection speed and computing power, means that
quality in one instance (20mbit+ connection to the internet and a large 1080p
widescreen TV) if very different to quality in another (a mobile device and 3G
broadband). Production workflows that enable the progressive downloading and
adaptive streaming of content go some way to alleviate the issues on the client
side - but these scenarios require faster and more specialised hardware and
software to achieve. Quality is something that must be measured constantly, and
the challenge is to make this active readjustment seamless and transparent to
all but the most hard core of users. The Open Video Player Smooth HD demo
delivered using Microsoft Silverlight shows this off nicely and the Flash
examples on these pages are similarly good.
It's clear that cable and satellite broadcasters, particularly in the US, are
dealing with this quality problem in their digital
broadcasts already. Though theirs seems fundamentally to be an issue of the
bandwidth available being less than what may be required - to the point where the quality of
material is drastically reduced in order fit within the constraints of the
bandwidth they have.
Part of the quality problem is around access to content on a global scale. The
use and evolution of content distribution networks (CDN) goes some way to
addressing this system - enabling
content providers to offload the heavy lifting of content delivery and allowing
the replication of content globally, pushing it out as close as possible to the
users who will access it. The use of the CDN in delivering video is already very
prevalent and will no doubt become more and more so, allowing a shift in video
delivery to be more or less entirely on demand via the internet with commercial
models that are regionally appropriate.
Earlier this year, in CNet's article "iPlayer uncovered", the BBCs iPlayer was stated as using 60+ encoding servers to create the video files that resulted in them pushing out content at an eye
watering rate of around 60 gigabits
per second during peak times - with content delivered from around the UK
and even from Europe as required. That's just under a DVD going out over the
wires every second, and illustrates quite succinctly just how important the delivery
part of their service is, and hints at the cost involved in getting video to the
web.
Cost is the
elephant in the room when it comes to getting video online. I've implied that getting video online is quick and easy at the YouTube end of the scale - but that it may not be all that cheap once you start
to move away from YouTube and the like. The cost comes
about through the requirement for video production tools, dedicated encoding
infrastructure and licensing, storage for encoded and raw video, publishing workflow
systems, digital rights management and content licensing, CDN replication and bandwidth
costs - and people to tie it all together; and this can all change depending on what output formats will be used.
To put it into perspective, at the point where you are pushing video across the airwaves or through a network, apparently streaming
video costs around 16,000 times what broadcast costs to deliver.
And then there is how to make money in the process, or at least not go broke, which is the subject of a huge number of articles, arguments and discussions.
All that said though, the success of online video hints to a shift to a CDN
type model for content delivery that may overtakesome aspects of broadcasting. One of the key drivers for this change, and something I believe would be
massively beneficial, may well be the audience's ability to have increasingly
closer relationships with content producers, and for content producers to have
more direct control over the access to the content they create and insight into
how it performs commercially.
Perhaps the opportunity here lies in reconciling many this myriad of complex production choices into a set of on-demand services, a sort of reconfigurable version of YouTube, which could operate from public or private cloud type infrastructures.
Here's hoping complex parts of online video delivery can undergo the kind of shift that enables us to concentrate less on the hard parts delivering it, and more on creating it.