More on Voice Banking

I mentioned Voice Banking in my last post – here’s a few more details.  Almost everyone with ALS eventually lose their voice – it can be one of the first signs in ALS called Bulbar Symptoms.  I find Bulbar Symptoms perhaps one of the most distressing aspects of the disease – but what is is.  And I don’t get to vote.

When I first heard about losing my voice I didn’t think it was such a big deal. I figured I’d just pick a voice out there; how about Steve Hawkins’ voice?  Or better yet – Darth Vader to scar the dogs!  However about a month ago Susannah mentioned that she wanted to hear my voice – not someone else’s.  So I started investigating Voice Banking. 

So I’ve started using an on-line program called ModelTalker– which is a free service for those losing their voice.  It’s funded through grants and donations – a truly remarkable organization.

There are a couple of things that you have to do before starting the process – first you have to get a very good microphone & headset.  I bought one they recommended from Amazon for about $45 – the Sennheiser PC 36 Call Control USB Binaural Headset. 

Next you have to find a silent location to record in.  When I started we were traveling in our Airstream trailer – which when all windows are closed is amazingly quiet.  So that was been my recording studio; now it’s my home office.

ModelTalker’s process is for you to record your voice using prescribed sentences.  They show you a sentence then the reader reads the sentence and you try to repeat the sentence as closely as possible.  They then rate how you repeated based upon amplitude, pacing and pronunciation.  The goal is to record 1600 sentences! 

Eventually ModelTalker will use the 1600 sentences to generate your own synthetic voice.  The sentences are from “The Wizard of Oz”, “The Call of the Wild” and other classics. 

Ironically long before my ALS I have frequently joked that the only foreign language I know is English, and I don’t even speak that very well (as an aside, I was born and raised in the U.S.).  At one point in the recording process I was having trouble pronouncing a word.  After a number of tries, I finally got it correct - I mentioned to Susannah that I’ve had trouble with pronunciation – she quipped “What’s new about that?”

It takes me about 45 minutes to record 100 sentences.  You can listen to your sentence recording and re-do it if needed.  You can also hear a sample of a synthetic voice based on the number of samples already recorded.  I’m currently about half way done with the process.  Here are 2 examples – the first is my current voice:

 While the second is a sample of the synthetic voice:

My voice has been changing – I am quieter then normal, my speech rate has slowed down, I’m having trouble with some pronunciation and it gets harder to speak as the day wears on.  This morning for the first time I noticed I’m getting tongue fasciculations. So there’s some urgency to my recordings now.