Skip to content
 
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(The Stack)   Researchers chip away at yet another thing you thought was unique about yourself   (thestack.com) divider line
    More: Interesting, speech synthesis, authentic voice template, speech synthesis technology, voice synthesis research, human voice, extensive speech synthesis, speech synthesis specialists, speech synthesis engine  
•       •       •

2360 clicks; posted to Fandom » on 06 Sep 2016 at 3:51 PM (4 years ago)   |   Favorite    |   share:  Share on Twitter share via Email Share on Facebook



27 Comments     (+0 »)
View Voting Results: Smartest and Funniest
 
2016-09-06 12:49:31 PM  
OK, now try to emulate my fist.
 
2016-09-06 3:09:46 PM  
Why would anyone want to do that? My voice is horrible
 
2016-09-06 3:41:08 PM  
Wouldn't this fall under the "What could possibly go wrong" file?
 
2016-09-06 4:04:43 PM  
I am not even the slightest bit impressed with this, it seems like something that should be common in this day and age.
 
2016-09-06 4:14:53 PM  
Ha, it turned to tin foil hat.
 
2016-09-06 4:15:27 PM  
My voice unique? Hardly. My father and I sounded exactly the same, a fact he would take advantage of when my friends called the house. I was hoping to play similar games with my son but mobile phones and the fact that nobody under 30 calls anyone deprived me of that fun.
 
2016-09-06 4:16:27 PM  
30 minutes is quite a long audio sample.
 
2016-09-06 4:19:40 PM  

This text is now purple: 30 minutes is quite a long audio sample.


You wouldn't believe how hard it is to get a sample as short as "my voice is my passport, verify me".
 
2016-09-06 4:23:01 PM  
Unfortunately, you'll be speaking Chinese.
 
2016-09-06 4:24:46 PM  
FTFA: "Chinese internet giant Baidu has developed an artificial intelligence system which it claims, can successfully imitate a human voice after listening to a person speak for only thirty minutes."

I don't know much about the Chinese language... however just something to keep in mind is that not all languages are created equal.

I exclusively listen to vocaloid music, and of of the interesting things about them is the huge difference between Japanese and English when it comes to synthesizing a person's voice. There are only 110 one-syllable building blocks in the Japanese language and you can write them all down on a piece of paper(in fact, its one of the first things you do learning it lol), English on the other hand has over 8000 depending on your dialect.

In theory you can make a decent Japanese vocaloid from scratch in 4 hours and it will sound pretty good and much like the person you used to record it... English, it takes at least a week or a month to sit down and record all the different sounds and when you are finished it will sound like crap because simulating English is essentially impossible. My Chinese is limited to "Wo ai ni" so I have no idea how hard Chinese would be lol.

So basically.. English speaking people are far from having to worry. I haven't even heard a program that could handle a gliding vowel right.

This song is done by a machine and one of the few ones that doesn't sound like total ass, but that's mostly cause the lyrics were written specifically to handle things GUMI can say. Google search other English vocaloids at your peril, it's pretty bad. But maybe if the purpose of the AI was to get the person to say one specific thing, it might not be that hard compared to general purpose synthesis.

【GUMI English】ECHO(Yunosuke Remix)
Youtube GztD0DH_nrY
 
2016-09-06 4:26:22 PM  
voice actors seen jumping from windows
 
2016-09-06 4:26:30 PM  

the_sidewinder: Why would anyone want to do that? My voice is horrible


This
I sound like an air-headed cow.

/though my husband has said that my voice has changed somewhat since my sinus surgery
 
2016-09-06 4:28:10 PM  
terminator 2-wolfie & rex
Youtube qKLTbJMJOSI
 
2016-09-06 4:28:19 PM  
Well, if it can duplicate my voice, maybe it can be incorporated into a speech-to-text system that can comprehend my accent. Which none of the current ones ever can. Usuallywhen i use a program like that the resulting text looks like a piece of finnish text translated into english by a North Korean peasant who had a gun pointed at his head the whole time.
 
2016-09-06 4:33:46 PM  
I began to feel panic set in when I read the headline, but despite my aversion to clicking links, I read the article.

Profound relief to find, despite my unease, that no one had yet replicated my particular bundle of neuroses.
 
2016-09-06 4:41:58 PM  

WelldeadLink: This text is now purple: 30 minutes is quite a long audio sample.

You wouldn't believe how hard it is to get a sample as short as "my voice is my passport, verify me".


Warner Brandis sees what you did there.
 
2016-09-06 6:08:32 PM  
Do you still need a white cat and do they come with the emergency backup clone?
 
2016-09-06 6:20:49 PM  

Wellon Dowd: My voice unique? Hardly. My father and I sounded exactly the same, a fact he would take advantage of when my friends called the house. I was hoping to play similar games with my son but mobile phones and the fact that nobody under 30 calls anyone deprived me of that fun.


This. My best friend and I can be distinguished only because she swears and I don't.
 
2016-09-06 6:47:06 PM  
Let me know when they can do it off a couple hellos so I can use that to break the voice recognition aspect of the bank after bypassing the decryption and firewall.

/Don't forget to delete your proxy entries
//Sudden hankering for some Uplink.
 
2016-09-06 7:06:37 PM  
"Now do Lincoln"

Obscure?
 
2016-09-06 7:21:57 PM  

NotARocketScientist: the_sidewinder: Why would anyone want to do that? My voice is horrible

This
I sound like an air-headed cow.

/though my husband has said that my voice has changed somewhat since my sinus surgery


my brother once quipped that I "sound like Bugs Bunny with a New York accent". that burned into my brain like hot lead.
 
2016-09-06 7:36:26 PM  
Sh*t, have at it.  You f*ckers have been chipping away at it for the last 50 years anyway.  Do you even need a viable excuse to impose on people anymore, other than "Zis iss for the goot of Humanity, and der vuture chennerations vill benevit fastly vrom your zelfless donation uff your time und evvert"?
 
2016-09-06 7:48:51 PM  

devek: In theory you can make a decent Japanese vocaloid from scratch in 4 hours and it will sound pretty good and much like the person you used to record it... English, it takes at least a week or a month to sit down and record all the different sounds and when you are finished it will sound like crap because simulating English is essentially impossible. My Chinese is limited to "Wo ai ni" so I have no idea how hard Chinese would be lol.


I was quite surprised that Miku's NA concert tour included several songs in non-atrocious English. I think there was even one in Spanish.
 
2016-09-06 8:39:33 PM  
"My voice is my passport and is very obsure"
 
2016-09-06 10:29:39 PM  

Short Victoria's War: "My voice is my passport and is very obsure"


Not using real words WOULD add a layer of security.
 
2016-09-07 6:37:31 AM  

dyhchong: Short Victoria's War: "My voice is my passport and is very obsure"

Not using real words WOULD add a layer of security.


Well, that's embareassing.
 
2016-09-07 9:09:53 AM  
So you're saying The Simpsons could theoretically run forever?
 
Displayed 27 of 27 comments

View Voting Results: Smartest and Funniest

This thread is archived, and closed to new comments.

Continue Farking




On Twitter


  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.