Technical details

Technical details

About

A story about how Technology and the Outdoors: Some Experiential Possibilities (a voice synthesised, audio-syncced slide presentation) was created.
Here's the challenge - how would you do it?
- Go to a conference and make a presentation
- Record the audio
- Upload the presented slides synced with the audio into the public domain
- Use free and open source software as much as possible

Slides

Originally prepared using Microsoft Office Powerpoint 2003 (.ppt) (with animations) - because I'm still in the process of converting my workflow from using MS Office to Open Office
Converted .ppt to Open Document Presentation (.odp) (using Open Office Impress) for more open uploading and sharing (because this is an open, non-proprietary format, better for long-term archiving, etc.)
- Open Office Impress is free to download and install as part of the Open Office suite.
Uploaded the .odp file to http://slideshare.net
Created a slidecast, by including the audio file
Added the embed code to a wiki page (in this case, the wiki software is dokuwiki)
Issues/limitations I encountered with this approach:
- Animations were lost in converting the process of converting to .odp and/or http://slideshare.net
- Most users are not yet familiar with using free office applications and open document formats
- MS Office 2003 doesn't open open document formats by default (need plugin, or MSO 2007, or free software such as Open Office)

Audio

Tale of proprietary woe

Digital audio of the live presentation was recorded on an Olympus Digital Voice Recorder DS-660 borrowed from the University of Canberra.
Reasonable playback audio quality was obtained, however the file was recorded in a proprietary format, with a tale of woe which illustrates some of the many problems with proprietary formats which followed.
The accompanying software CD for the digital audio recorder was lost by someone - no-one knows, its a communal device and it could have been me
- Costs $50 to replace and Olympus responds to no requests for help
- Hours of searching fruitlessly for some way of hacking; conclude task is near impossible in any kind of acceptable timeframe
So, conclude that the uncompressed raw audio file can't be extracted off the digital voice recorder.
Some time later randomly notice someone using a very similar recorder. Accosted him. He had also lost the software CD! Comical. However, he still had access to a PC with the software installed. I gave him the recorder and he extracted a .wav uncompressed audio file. Too late for my original needs by then, because I had sought an alternative solution…

Meanwhile...

Meanwhile, I typed out a text transcript of the audio.
- This took a LONG time and was surprisingly PAINFUL .
- It was very interesting to listen to my own presentation and REWORK the crap so it would sound a little better. I also removed all those bloody ums and aahs….
I searched long and hard for free software to do voice synthesisation of text.
Why didn't I just re-record the presentation with my own voice?
- I had an awful cough which wasn't going away and I'm stubborn and curious about voice synthesisation. In part because good friends have just had their baby diagnosed as blind. I would like to find a FOSS way to create free, high quality audio files from text, i.e., t2s (text to speech).
After trials of all sorts of stuff, including on ubuntu, I found a couple of ways.
In the end, I opted for http://talkr.com mainly because of its voice quality. The next best options were using tools available for gnulinux - a pleasant surprise.
For this to work, I first published the transcript text in a blog - a WordPress hosted blog worked and I think it works with Blogger. In theory, it can work with any feed.
I subscribed to the WordPress blog feed in http://talkr.com
http://talkr.com did the job of creating an mp3 voice synthesised audio of the text - this was downloaded.
The mp3 file was recorded in 16 bps, which didn't play properly when imported into http://slideshare.net; the mp3 file also had a few seconds of header audio I wanted removed.
I used Audacity beta 1.3.3 (because it has better export options than the current version 1.2.6) to edit the mp3 and exported it into 32 bps mp3.
Uploaded the edited mp3 file to archive.org
- Note: slideshare requires the audio to be hosted somewhere else, but recommends archive.org
Issues/limitations
- mp3 is a proprietary format, but that's what talkr generates and what slideshare requires (hopefully soon .ogg will be an option). By uploading the mp3 to archive.org, however, this created .ogg and streaming files on the fly, which is excellent.