I’ve been looking for a way to create SMIL files with an editor for a while. Main reason being to speed up the creation of creating podcasts for the Perceptive Podcast client and make it easier for those who don’t understand markup/code.
One of the techniques we deployed during the Visual Perceptive Media project was to export final cut xml out of final cut/premiere pro then transform the lot with XSL/Python/etc to something else more usable. Its something I’ve had in mind for a long time, as you can see with this paper/presentation I wrote 12 years ago.
There was a point when Wmas, could create an editor for our director/writer (Julius) or allow him to use tools he was familiar with (non-linear editor like Finalcut/Premiere). Of course we choose the latter and converted the final cut xml (which isn’t really an official spec) into json using python. We were able to use markers and zones to great effect, indicating the interactive intentions of the director in a non-linear editor. This meant the intentions can exist and run completely through to the very end, rather than tacking it on at the end.
So with all that in mind, I started thinking if I could turn Audacity into a editor in a similar way? Is there a final cut xml format for audio? Thats when I came across this article which made perfect sense – Audacity files are just XML documents, sooo…
Structure of a empty project
<?xml version=”1.0″ standalone=”no” ?>
<!DOCTYPE project PUBLIC “-//audacityproject-1.3.0//DTD//EN” “http://audacity.sourceforge.net/xml/audacityproject-1.3.0.dtd” >
<project xmlns=”http://audacity.sourceforge.net/xml/” projname=”blank-audacity_data” version=”1.3.0″ audacityversion=”2.2.1″ sel0=”0.0000000000″ sel1=”0.0000000000″ vpos=”0″ h=”0.0000000000″ zoom=”86.1328125000″ rate=”44100.0″ snapto=”off” selectionformat=”hh:mm:ss + milliseconds” frequencyformat=”Hz” bandwidthformat=”octaves”>
Just the title ignited my mind, the actual content of the blog is less interesting but I realised I may have a free & open-source editor which runs on every platform and with a bit of XSL magic could be the start of the editor I was looking for? The idea of it being a pipe, which leads on to more is something which fits in the bigger pipeline chain
I also found a GIT project to Parse audio track times from an audacity .aup projects. Its uses XSL to do the processing, so I may spend a bit of time playing with it to make something useful.
Just need to dust off my old XSL development skills… Which reminds me what happened to XPROC (XML pipeline language)?
Its going to be another busy month or so…
Tomorrow we start to Build a healthy public service internet in the first forum during Mozfest weekend. This will be further explored at Mozfest in the decentralised space on Saturday afternoon.
Not long afterwards I’ll be going to Skopje along the same lines as last year when I was in Sarajevo. This time we have the results of last years workshop, the living room of the future.
Then Berlin for Most wanted music with a adaptive narrative conference talk more focused on audio than video and our first perceptive podcasting workshop. Now this exciting but scary as its completely beta and being developed right now.
I’m back in Berlin not long afterwards for a look at object based media and how machine learning can work together for the future of storytelling, quite similar to TOA 17 but more exploring and more I can talk about now compared to then.
Finally wrapping up with a critical panel discussion titled New platforms, new ways of storytelling at the future of the book in London. I expect a few things I said at Oreilly’s Tools of Change in 2012 are still very relevant. But also Perceptive podcasting will be much more mature by then.
All exciting but quite a bit in one go…
When I first heard about 60dB, I thought great someones finally made a object based podcasting client.
60dB brings you today’s best short audio stories – news, sports, entertainment, business and technology, all personalized for you.
Unfortunately I was wrong.
Its a bit like stitcher which is well loved by some people.s
It does seem to pick and play news stories. But the sources are specially crafted (ready for syndication like this) rather than the client processing the audio and picking out the parts most relevant to your listening preferences.
Its understandable because to do this you would need well thought-out metadata created by the original author/production. Without it you can’t have objects, without objects you are reliant on serious processing of the audio to build the metadata which the player can use (that or some serious computational power).
I had heard and thought it was a logical move for Google Play’s podcasting support would include some kind of basic automated metadata/transcript but it never happened. Another missed opportunity to show off the power of google and make themselves a essential part of the podcasting landscape, like how Apple did with itunes.
Seems like a great opportunity for some enterprising startup, specially since podcasting might save the world. Dare I say it again, perceptive podcasts could be incredible for all the reasons podcasting originally captured peoples attention.