Woman robots of IVR systems have low voice quality

Jul 31, 2008

I’ve been calling various banks, mobile companies, insurance agencies and other businesses that I’m a customer of. Here’s one thing that I don’t understand: all of them have an IVR before you actually reach a person. Which is fine. So you call the number and you have to key in or say things about yourself to a woman robot (why is it always a woman robot? I haven’t heard a man robot yet) before they put you through to an actual person.

But: why do those woman robots have such crappy voice quality? It’s extremely bad. She’s like talking through three pillows or something, and I can very barely understand what she is saying even though I’m in a quiet room and have a good device (Skype or iPhone) myself. Sounds like it’s sampled at like 4 KHz or something else low-frequency and there just aren’t high frequencies. Yes, I know that a lot of it is simply due to the phone system not being able to put through HiFi voice (phone network is limited to 300-3000 Hz or something like that). But in the 21st century, it doesn’t have to be that way.

I wonder what it will take for major businesses to upgrade their woman robots sitting in their IVR systems to a decent quality that goes to the high end of what current phone systems can provide? Or better yet, split the voice experience into “hi-fi” and “lo-fi”? I know that you need to always have backwards compatibility with legacy phones, but if you can detect that the remote caller is on some HiFi capable network, like Skype or some mobile equivalent that I’m sure will be invented and pushed by mobile companies soon, then you can provide a voice experience that’s actually enjoyable, instead of forcing your customers to talk to a robot woman who sits behind three pillows.