[Ding] “Hello Catherine, this is Siri. How can I help you?”
Princess Sparkle of Moreland accidentally leaves her iPhone on during her debate presentation at the annual Woodford Planting Festival.
Having just confessed that she is underprepared and at a loss as to how to proceed her mobile devices AI construct kindly offers to help.
The resulting real time dialogue generates an enormous amount of laughter. There is nothing quite as funny as a computer saying “this jumper sucks a bag of dicks grandma”.
But how did she do it?
As is often the case with these things there is a keen tech head in the shadows. A tech head who was asked the right questions. “Here is my idea. Can you build it?” In this case, after 30 minutes research and a working demo of the process the answer was “yes”.
Siri uses a voice synthesiser that is also available on a Mac. (On the Mac she is known as Samantha and is one of many voices available.) Invoking a terminal and typing “say ‘Hello world'” makes your computer say “Hello world”.
Catherine’s script had around 25 dialogue cues for Siri. We needed to be able to create the cues as individual audio files to play on command.
More terminal command line magic helped here. If I type “say ‘cue_01’ ‘Hello world ‘” rather that saying it out loud an audio file called cue_01.AIFF is created.
So now all I had to do was manually type in all the cues and generate the files.
Tech heads would never do anything so tedious and unrepeatable as that. Also, it is best to have a single representation of the dialogue for ease of editing. We are working with humans after all. There WILL be editing.
“Automator” is a Mac utility that, among other things, allows you to string together any number of commands and run them with a single button press.
Once the initial script was entered, making modifications to the Siri dialogue was easy. Make the modification and click “run” to have all of the dialogue regenerated in 3 seconds.
The result was a folder with all the cues lined up and ready to roll.
My job at performance time was simply to wait for the laughter to subside before hitting the “down” arrow to launch the next cue.