June 23, 2008
Visualising Audio
A couple of weeks ago, I spent the week writing posts on the Visual Gadgets uncourse blog, one of which focused on ways of visualising audio transcripts: Visual Interfaces for Audio/Visual Transcripts.
Over the weekend, a couple of BBC R & D folks at the Mashed 08 event posted another take on this, a visualisation tool for audio files: Audio visualisation.
.
As they suggest: "Finding your way to a particular item in a recording of a radio programme can be difficult. Fast-forward and rewind functions aren't always very effective, especially in on-demand, on-line use. We are researching the presentation of audio recordings as visible timelines, with different colours or textures used to represent different characterics of the audio, like genre, speaker, tempo, key or rhythm - all extracted automatically".
Again, a week or two ago, I also discovered the 'enhanced podcast' (which I used to create the image synch for remix 2: play, create (Blur, featuring Sir Ken Robinson), described in remix 2: Blur, featuring Sir Ken Robinson...). This allows synching of imagery to audio (something that SMIL almost sorted a long time ago...).
Enhanced podcasts were also picked up on by some of the BBC folks, particularly in this example of Hack Moyles - Audio segmentation with RTMP. The post describes a technique for segmenting MP3 files so that each segment can be marked up with data, such as text, or images. Here's a demo: RTMP Segmented Audio Player.
This is something that could be really good for us, I think?
PS sort of but not really related, ages ago I came across the rather wonderful SoundManager 2 javascript libraries. I rediscovered it again a week or two ago, around about the time I discovered that audio files from ITConversations could (once again) be deep linked into... (Excerpting audio from ITConversations).
So I posted a demo (via twitter) that allows the user to 'trigger' one or more audio files from an HTML 'trigger pad'.
The audio sample triggered from each pad is actually pulled in from an XML file - as the following shows, the target URL for the file can actually be an IT Conversations 'deep link' URL.
When I get a chance, I'll see if I can make this sample pad delicious powered... I was also wondering whether I should try to trigger the display of a flickr image whenever a pad is pressed? (When two audio sampels are triggered, they both play - what's the equivalent for displaying mulitple images? Fade one in over the other, show them both as semi-transparent?)
Tags: audio, visualisation, samples, itconversations
Posted by ajh59 at June 23, 2008 10:16 AM