On Dictation online, is a demo. I talk into my microphone and speech-to-text words appear on the webpage. This page by Amit Agarwal of Digital Inspiration gives instructions for using HTML speech input API. A demo of web speech.
On the AT&T Speech API website, it describes what seems to be FREE cloud-based speech-to-text for low traffic uses. That seems awesome. In other news, it looks like their cloud-based text-to-speech has no output of Visemes or timing data. It only outputs spoken audio. So apparently, just before the Moose proves that perfect lip sync […]
For long term tolerability of a talking computer character, it MUST be interrupt-able. If it isn’t, and that agent talks while you are on an important phone call, or talks while your boss is trying to talk to you, that is maddening. The talking moose has always been interrupt-able, which I call “SHHing the moose”. […]
see also, this Microsoft research talking head, of Craig.
Check out Ultra Hal by Zabaware. A very pretty redhead 3D female speaking agent. She has quite good eye behavior, but poor lip motion. I’m happy to see Ultra Hal is still around.
Today I discovered Humanityonline.com, and they have developed talking avatars that are quite photo-realistic. My first (of two) videos shows the short weather report for Edmonton today. I made a second longer video, comparing the moose to Humanity Online’s Sandy character. It’s amusing, in an odd way…
I hate spilling beer on the rug…
Today I discovered CodeBaby. Their avatars have that wonderful video-game character quality. They move in 3D. Hurray. And they are a Bioware-related company. I have always had particular fondness for Bioware because its founders and I went to the same medical school, just a few years apart. (I can’t recall if we met in person […]
These virtual assistants offered by a company (that shall remain nameless), are unbelievably awful. Those human faces, with their head jitters, their eyes and lip movements, are truly repulsive. Business that use them on customer service websites, are not being smart.
I couldn’t resist making this short video, about axe deodorant. Also noteworthy today, I discovered Geppetto Avatars (but haven’t found any contact info yet), and found CMU Sphinx, an open source speech recognition, and found opendial, an open java dialogue system, and was reminded of ejtalk, a conversation management XML that doesn’t seem open or […]