Being in a band can be pretty expensive.
It seems like it wouldn’t be, in this do-it-all-yourself digital age, but the reality is, it can be pretty expensive. Between studio hire, mixing, mastering, replication and distribution, it still takes a lot of people to get a song from your head to the wider universe. And that’s just the music bit, without the marketing, PR and promotional costs involved in getting someone to actually listen to it.
I like the web as a tool for hacking. The idea of mashing together existing technologies into something greater than the sum of their parts has always held a deep appeal to me. With the wealth of information on new API’s, new tools and new processes, the web is a really lovely place to find inspiration and learn about new things.
Which is good, because I can never seem to pull together the £2,000 it would take to get a halfway decent video made.
So, when the going gets tough, get creative.
Tweetflight and Preflight Nerves
For our first single, Preflight Nerves, I used the Twitter API and HTML5 video to make an interactive tweet-powered giant internet karaoke session. With over half a million views, the response was really wild and I am proud of what it achieved. But afterwards I felt a bit lost on what to do next, so we recorded a really stripped back EP while we took some time to regroup.
We decided on the first single, True, and the question came up again: what would we do for a video? I had experimented with gifs and Giphy, Vine videos, illustrations, SVGs and wearable sensors, but it never quite seemed to click. And I had no idea what to do next.
Back to the drawing board
Before Christmas I got engaged in Iceland, which was pretty great. I also took a lot of footage while we drove around, weaving between snowy mountains and huge waterfalls and slow moving glaciers and all of the beers.
Such mountain. Very snow. But I digress.
I took all the footage and cut it together, created some fractal-esque visuals with filters, and used Angular.js to display the lyrics as they’re sung. I timed it using a pretty simple json file, watched it through, and felt… kind of disappointed. It was pretty, but it wasn’t gripping. There was nothing new or interesting there, and I had spent a while fiddling with it. I felt like I was sinking time without cracking the core problem — there was nothing really dynamic happening.
A love letter
I remembered that he mentioned the web audio API had been making some strides, and I was curious whether I could try and visualise the audio (taken from a video object) to display using canvas.
At the core is this piece of code:
// Make sure AudioContext is available (Webkit uses a vendor prefix) window.AudioContext = window.AudioContext || window.webkitAudioContext// createMediaElementSource means we can grab the audio from our video context = new AudioContext()
sourceNode = context.createMediaElementSource(video)
createMediaElementSource() I was able to visualise the frequencies direct from the audio, within the video element, in real time. And that’s a simple use of the API — the possibilities for experimentation are essentially endless.
(Unfortunately Safari doesn’t play nice, due to a bug in the way it returns audio data from
createMediaElementSource(). It’s pretty widely known and it doesn’t looks like it’ll get fixed any time soon.)
A day later True was essentially done.
I also used SnapSVG for the opening animation, which is an incredibly robust library for manipulating and animating SVGs.
When I started True I struggled with finding inspiration. It was that quiet nagging feeling that chews away at your self esteem until you’re sure you’ll never create anything again and all of your achievements will inevitably, slowly, inexorably turn to dust and ash.
Worry not. The web audio API, however, pulled me from that spiral of crippling self doubt. So thanks, web audio API.
Oh, and you can watch True here.
Acknowledgements: Glen Maddern spent a day on my couch battling some super aggressive server side caching when he could’ve been exploring London, and is still the best human