To clarify the topic, this text is about electronic voices and image processing. Well, some days ago, I installed Skype 2.5 BETA and found a funny plug-in for it. It’s called Crazytalk. It is a virtual Webcam for Skype, making images of animals, drawings or people talk. The lips were synchronized to your voice. It looked pretty real, so I wanted to create something similar.
First, I needed a voice, so I started searching for computer generated voices. My first results were Microsoft Sam & Co. They didn’t sound anything close to real. After some more googling, I found Neospeech’s Lily Chinese Voice. It really amazed me, because the voice sounded really realistic. It was quite fun to let the voices read all kinds of stupid things.
I used Neospeech’s demo and recorded a little sound file. Then, I imported it into Crazytalk Media Studio and combined it with an image from ??&???’s website. After some customizations with the image, the virtual talking person was ready to go. Here is the result of my work:
Download this video (Right click, save as)