Skip Links

Network World

  • Social Web 
  • Email 
  • Close

Speech-to-text the answer to voice mail

Cool Tools By Keith Shaw , Network World , 08/23/2007
Keith Shaw
Newsletter Signup
  • Share/Email
  • Tweet This
  • Comment
  • Print

The scoop: Vonage Visual Voicemail transcription service, 25 cents per transcribed message, from Vonage.

What it is: The latest feature from Vonage and its voice mail offering automatically transcribes voice mail messages into text, and then delivers the transcribed messages to either an e-mail address or text message to a cell phone (via SMS). Up to five e-mail addresses and five cell phones via text messaging can receive these messages. The transcription service is powered by Simulscribe, which automatically converts the voice message into a text file. In addition to the transcribed text, the service attaches a WAV file, so if you still need or want to listen to the voice message, you can.

Why it’s cool: I absolutely hate listening to my voice mail. If you get a lot of voice mails, it becomes a time killer having to sit and listen to your messages, and then if there’s any important information in the message, I end up writing down the details anyway. Having all of this go through an e-mail address automatically puts the person’s phone number (either in the message or through the caller ID) in a format that you can print out. Forwarding the transcribed text to a cell phone is also cool — if you have Vonage at home and are expecting a call from someone, you can be alerted via SMS that the person called, along with the message.

The transcription ability of Simulscribe was pretty good at times – in one message, the system could take out the “umms” and “ahhs” from the voice message, and when one caller tried to leave his message in a fake accent, it still picked up the right words. This started to make us think that there was a human on the other end of the service listening to the message and transcribing it.

Some caveats: But this didn’t happen all the time. Other test messages that I received made it obvious that speech-to-text technology still isn’t 100% perfect. The phrase “I’m calling to find out if you have any chips with wasabi in it” was transcribed as “I’m calling to find out if you have any trips with the Saudi and it.” Like a lot of speech-translation software, the clearer the voice, the better the translation. If your caller mumbles a lot or talks really fast, the transcription will suffer. It’s good that the service then attaches the WAV file so you can listen to the message, but I’m still concerned about the accuracy of the transcription.

  • Share/Email
  • Tweet This
  • Comment
  • Print
Comment
Login
Forgot your account info?
Add comment
Anonymous comments subject to approval. Register here for member benefits.
Have a NetworkWorld account? Log in here. Register now for a free account.

Videos

rssRss Feed