If you are looking around for a Halloween project, you might consider The Yorick Project from [ViennaMike]. As you can see in the video below, it marries a Raspberry Pi acting as an Amazon Alexa with an animatronic skull.
This isn’t the most technically demanding project, but it has a lot of potential for further hacking. The project includes a USB microphone, a servo controller, and an audio servo driver board. It looks like the audio servo board is controlling the jaw movement and based on the video, we wondered if you might do better running it completely in software.
It is a bit disconcerting to have to call the skull Alexa and hear the familiar voice coming out. It seems like some analog voice changing hardware might be in order on the output. As for the input, it would be entertaining to feed sound files into the audio input to respond to sensors instead of taking only voice commands. In other words, in addition to voice commands, you could tell it to say something specific or start a skill when someone breaks a light beam or whatever.
If Yorick is too much to bite off in the short time remaining, you could go for something with fewer parts. Or, if a skull is too intense, try a pumpkin or a bass.
Needs a Turban, eyebrows and Jeff Dunham’s voice.
Infidel!
And every sentence must end with “I kill you” :D
…or a suit, a tie and a “Geoff” name tag.
I was thinking it might be more fun if it answered to “Bob” and insisted on trying to brew love potions.
I leave this one here…
https://www.youtube.com/watch?v=QIIQkEWcyrA
An animated Christmas tree that sings happy birthday and tells you “good night. sleep well.”
this is what nightmares are made out of.
Okay, so I’m reading this, and remembering that there was a Terminator endoskeleton head from years ago that synced with your PC speaker so that when you were on a voice call, it “spoke” whatever your friend was saying. Now, as I scroll down to write this, I see this ad from a blog called Skynet talking about the same thing, only it’s a new Kickstarter project.
I know, I know, targeted ads, blah, blah, blah, it’s still interesting when these algorithms are spot-on.
A decapitated doll’s head. Has movable eyelids, doesn’t have to move jaw to be scary.
One, run the sound through some kind of voice changing technology.
I do not know anything about these devices, but, is there other words or phrases that could be used to activate?
I seem to recall (it might have been a joke) of some of these navigation devices being able to change the voice, I swear I thought I heard snoop dog doing it or something like that.
“you, up ahead, turn left, foshizzle” or something like that.
Guess my point being, it would be great if you could run the talkback voice through some type of sound changer to come up with something…….
Would be great if they come up with a seasonal version of this stuff, I always liked the “american Dad” episode where the guy had the gargoyle on his car giving directions…..
A later update added the voice changing feature as an option. For some applications, I agree it’s better. But the cognitive dissonance of hearing the standard Alexa voice coming from a talking skull is better in others.
Ha! It would be fun to equip this with facial tracking stuff so the head and eyes swivel to meet you when you enter the room.
It would be fun to implement this with a tiny Sauron made of neon-like tubing with a “cat’s eye” like effect in amber.
But can Alexa translate spoken Shakespearean English properly?
Example: Yorick what is the weather today?
(Badly?) Translated: Yorick what is the weather the present day?
Do a Web search for online Shakespear Translator (there are lots of them) for variations and see what happens. I would try it myself if I had an Alexa device that spies on my all the time (much less a Poor Yorick headed Alexa spy device).
Oops, replace Yorick with Alexa in my example… Or maybe not! Is there a way to replace the Alexa keyword with Yorick?
I appreciate the discussion and thought I’d add a bit, since some of the suggestions for improvement have been done since the spring. The original AlexaPi code I based this on used VLC, which didn’t support audio effects. They since updated that software to use Sox as an optional alternative, so now yes, Yorick responds to “Alexa” using Alexa’s voice and to “Yorick” using a deep, slow, resonant voice. See https://www.mcgurrin.info/robots/?p=502 for more on the update.
You could easily expand this to respond to sensors and use canned responses. For canned responses, pre-programmed jaw movements or specialized side audio tracks can improve the sync. That’s a good effect, but it’s also been done a 1,000 times before. The nice thing about tying it to Alexa is that it has infinite responses.