This example syncs voice recognition through the telephone with web video playback. We reached out to our friends at Voxeo to use their voice software and telephony/technology platform.

This was shot low-budget using a Flip HD camera and another low-budget camera. No lights. Shooting outdoors means constant sound, light and shadow variations, and, because of the lower-quality cameras, focus variations. (We did use tripods and one camera had a $30 mic.) Separately, we picked up the Central Park shots on the fly, just with the Flip.

We showed this example to many ad agencies to many wows. One issue was: the more people call, the more deployment would cost. Another related to talent. Talent wrangling is hard and schedules, rights, fees and creative approvals become complex.

Call the number on the screen and interact with the video playing out in the ad unit. The unit is small - 70K - and pulls the media it needs from source in real time. The speech recognition on this app is pretty simple, so follow the options as they appear on the screen.

Turn the volume down on your computer speakers, otherwise the application will hear any background noise that's in the room. Also, speakerphones are particularly tough, because they try and interpret everything they hear and cannot interpret sound the way the human ear can.

Try it!