Amazon's Alexa Can Now Whisper, Bleep Out Swear Words, and Change Its Pitch (theverge.com)
An anonymous reader quotes a report from The Verge: Amazon is trying to make its Alexa voice assistant sound more humanlike. Up until now, the female-sounding voice maintained an even, monotone cadence whenever speaking, but with Amazon's new Speech Synthesis Markup Language that the company introduced this week, Alexa can whisper, vary its speaking speed, and bleep out words. Developers can also add pauses, change the pronunciation of a word, spell a word out, add audio snippets, and insert special words and phrases into their skill. The Verge notes that "the language markups are [only] available to developers in the U.S., U.K. and Germany." Amazon will also be hosting a webinar on May 18th on the new code.
Amazon's Alexa can suck my dick. If I want to buy something, I'll spend 15 seconds to look it up and click a button.
I don't even like talking to my family all that much. I'm not going to talk to some evil fucking djinn of a multinational corporation.
You are welcome on my lawn.
Will Alexa bleep out "Belgium"??
I am Slashdot. Are you Slashdot as well?
Suck my d***.
LOL if they can.. but I still won't have one, regardless.
That's well and good, but what I'd really like to see Alexa be able to do is to be removed from the Kindle Fire that I bought two years ago without Alexa and which downloaded itself without my consent. Now that is a feature I'd like to see.
Can it be programmed to add in extra swearing?
And speak in a Rick Sanchez voice?
And add in burps?
"the language markups are [only] available to developers in the U.S., U.K. and Germany."
Welcome to the walled garden.
Would you be interested in our "Tourette syndrome model"?
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
I wonder how did they implement whisper. Alexa uses IVONA technology (they acquired the company), which is based on unit selection, after all (pre-recorded corpus of speech). It doesn't model voices parametrically.
For pitch I suppose you can use different versions of units as well as all the reasonably well working techniques of pitch shifting that preserve formants.
just make men feel at home
That's when they are the most dangerous.
Have gnu, will travel.
That's not what monotone means. Alexa is anything *but* monotone and the author would know this if they ever actually used the service. The voice is moderatly "sing-song" cadence like a radio news reader.
Kriston
Download and install your Norton antivirus from www.norton.com/setup. If you want any help related to Download and installation of Norton antivirus you can call at 1-855-408-6697 for USA and +44 800-088-5029 for UK and 1-855-408-6697 for Australlia.
It's not "Amazon's new Speech Synthesis Markup Language". SSML is an existing standard; Amazon have added a few proprietary extensions.
Computed and recorded, dear. /pouts