Pink Trombone is an online voice synthesizer with a difference


#1

Originally published at: http://boingboing.net/2017/03/19/pink-trombone-is-an-online-voi.html


#2

My Klatt is off to you, sir!


#3

And here I thought it was going to be something unsavory.


#4

You play the “Pink Trombone” by moving your tongue and soft pallet?

LOL…is this the fellatio version of Lickster that new cunnilingus app?


#5

It is unsavory. The pink ones are sweet. Bubblegum flavor.

the rusty ones are savory. Caramelization.


#6

Brown Tuba ain’t too shabby niether.


#7

Is this not built into your elf ear phones?


#8

Would be funny to record a bunch and then make an animation using the results as character voices in some sort of short narrative.


#9

I need to find some way to write a script that maps the parameters to standard midi note values/mod/pitch sends.

I almost got it to say “Hey You”, but it sounds like a bad Jeremy Louis impersonation done by a robot that’s had a concussion.

So awesome.


#10

Ohhhhhh YEAaaaahhhhhh

https://twitter.com/zacharyjohnson/status/843599766516891648


#11

It doesn’t work quite so well when you have a mouse instead of a touchscreen, but it’s still entertaining.


#12

Reminds me of when I tried making an alphabet based on mouth shapes expressed for each letter.


#13

You’re thinking of purple oboe.


#14

I’m totally shocked about how well you can model the human vocal apparatus in less than 2000 lines of code.


#15

Can anyone figure out how to do throat singing with this? Like, get some overtones going…


#16

I’ve been at this for hours and my wife completely fails to demonstrate any sense of awe. Maybe I’m doing it wrong…


#17

It doesn’t seem to include the subvocal folds in the model. Which is what you buzz when throat singing.

I figured out how to use my subvocal folds voluntarily. It’s added a lot of gravitas to my groans and huffy protestations at work.

I also use it when saying “uh”. To catch people off guard. They don’t expect me to sound like Johnny Cash when I’m straining for an answer.


#18


Above: 1967, AT&T Labs

I can’t find a video that explained why many of the common text-to-speech voices of the 80s had a slight Scottish accent. Supposedly it was because the engineer who created it was Scottish. There was a great video that showed software mimicking the parts needed to make vocal sounds on a monochrome crt. The software allowed for granular manipulation of the voice, but also had single keys to quickly change how the digital voice box sounded. I think I even remember a display at the Museum of Science and Industry that showed speech synthesis through crude digital modeling.

Below: 1939- Voder (Short and worth watching to see early speech synthesis)


#19

Holy crap, if that VODER video is real, mark me down as seriously impressed.


#21

Totally real. The history behind the Voder is super interesting.