This is how I solved the problem initially.
- Use Screencast-O-Matic to record the video of screencast on my laptop. The screencast is done using DroidAtScreen with highest frame rate possible option checked. It has better frame rate than ASM screencast. During the video record session, my voice was captured. (so in other words, choose a quiet place!!)
- For simulating the Text-to-Speech engine voice, I used SitePal demo site and the voice is Julie (US). It's the closest voice I could find that matches the Google Glass speech engine. To record the voice, I used Audacity and export it to .wav audio file. The key is to play the video and find the exact time to insert the audio file using any standard movie maker software.
UPDATE
Just finished the demo presentation at IT Expo. To my surprise, the simplest solution worked the best.
- Create the video demo (under 2 minutes) as mentioned above but insist on asking an audience to try it out.
- Ask the person to say what he/she heard from the Glass app as a response to the action (ex. The item is saved)