HomeLatest ThreadsGreatest ThreadsForums & GroupsMy SubscriptionsMy Posts
DU Home » Latest Threads » Forums & Groups » Main » General Discussion (Forum) » Stephen Hawking's simulat...

Wed Mar 14, 2018, 10:27 AM

Stephen Hawking's simulated voice. A couple of questions.

1 - Does anyone know why it has a Scandinavian accent?

2 - How does it work? I think I've heard him engage in conversations, which suggests that it's not just a computer translating a script. But maybe I'm wrong.

tia
las

23 replies, 841 views

Reply to this thread

Back to top Alert abuse

Always highlight: 10 newest replies | Replies posted after I mark a forum
Replies to this discussion thread
Arrow 23 replies Author Time Post
Reply Stephen Hawking's simulated voice. A couple of questions. (Original post)
LAS14 Mar 2018 OP
underpants Mar 2018 #1
LAS14 Mar 2018 #2
poboy2 Mar 2018 #6
underpants Mar 2018 #11
underpants Mar 2018 #12
blugbox Mar 2018 #3
LAS14 Mar 2018 #5
aidbo Mar 2018 #4
LAS14 Mar 2018 #7
muriel_volestrangler Mar 2018 #14
LAS14 Mar 2018 #17
aidbo Mar 2018 #21
muriel_volestrangler Mar 2018 #22
aidbo Mar 2018 #23
missingthebigdog Mar 2018 #15
aidbo Mar 2018 #20
dalton99a Mar 2018 #8
LAS14 Mar 2018 #9
dalton99a Mar 2018 #10
LAS14 Mar 2018 #13
dalton99a Mar 2018 #19
fescuerescue Mar 2018 #16
LAS14 Mar 2018 #18

Response to LAS14 (Original post)

Wed Mar 14, 2018, 10:33 AM

1. Actually I heard this morning that the only thing he didn't like about it

Was that it gave him an American accent

Reply to this post

Back to top Alert abuse Link here Permalink


Response to underpants (Reply #1)

Wed Mar 14, 2018, 10:38 AM

2. Do you agree that it's American? I definitely hear Scandinavian. nt

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Reply #2)

Wed Mar 14, 2018, 10:46 AM

6. oh, its American 'voice', but fyi-Keep Talking

 

Fantastic group, song, and lyrics...

Keep Talking

"Keep Talking" is a song from Pink Floyd's 1994 album, The Division Bell.

Recording[edit]
Written by David Gilmour, Richard Wright and Polly Samson, it was sung by Gilmour and also features samples of Stephen Hawking's electronic voice, taken from a BT television advertisement (this same commercial would be sampled again in "Talkin' Hawkin'" from Pink Floyd's next and final studio album, The Endless River).[2] Gilmour chose to use the speech after crying to the commercial, which he described as "the most powerful piece of television advertising that I’ve ever seen in my life.”[3] The song also makes some use of the talk box guitar effect.

Release[edit]
The song was the first single to be released from the album in the United States in March 1994. It was the group's third #1 hit on the Album Rock Tracks chart (a chart published by Billboard magazine which measures radio play in the USA, and is not a measure of record sales), staying atop for six weeks.

The song was included on the 2001 compilation, Echoes: The Best of Pink Floyd.[4]

Live[edit]
The song was performed every night during the 1994 The Division Bell Tour and live versions, taken from different shows, were included in both the album Pulse and the video of the same name.

The song was sampled by Wiz Khalifa on the title track of his 2009 mixtape Burn After Rolling.

Quotes[edit]
It's more of a wish [that all problems can be solved through discussion, as 'Keep Talking' suggests] than a belief. [laughs]

— David Gilmour, 1994[5]
Well, I guess I experiment more than I think I do. I had a Zoom [effects box] in my control room one day and I was mucking about with something. Suddenly, I thought I should stick the E-bow on the strings and see what would happen. It sounded great, so we started writing a little duet for the E-bowed acoustic guitar [a Gibson J-200] and a keyboard. We never finished the piece, but Jon Carin [keyboardist] decided to sample the E-bowed guitar part. We kept the sample and ended up using it as a loop on "Take It Back", and again on "Keep Talking".

— David Gilmour, 1994[5]
Stephen Hawking 1994: "For millions of years mankind lived just like the animals. Then something happened which unleashed the power of our imagination: we learned to talk" and "It doesn't have to be like this. All we need to do is make sure we keep talking."

-

https://en.wikipedia.org/wiki/Keep_Talking

Reply to this post

Back to top Alert abuse Link here Permalink


Response to poboy2 (Reply #6)

Wed Mar 14, 2018, 10:55 AM

11. Radiohead used it on "Fitter Happier" in 1997 - yikes that was 22 years ago?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Reply #2)

Wed Mar 14, 2018, 10:55 AM

12. Sounds American to me n/t

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Original post)

Wed Mar 14, 2018, 10:44 AM

3. Some quicky answers

It uses advanced text to speech, controlled my miniscule movements of his cheek picked up by sensors in his glasses.

He selects a few letters, then chooses words from auto generated selections. It can mimic different accents, apparently.

For longer speeches or television appearances, I believe they prepared some portions in advance, to save some time for him.

There have been many advances in text to speech voices, but I'm pretty sure I heard that he prefered to not upgrade the voice. Not sure if that's true though.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to blugbox (Reply #3)

Wed Mar 14, 2018, 10:46 AM

5. Wow!!! I can't imagine what "upgrading" would consist of!!!! nt

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Original post)

Wed Mar 14, 2018, 10:45 AM

4. He was able to move his eyes.

He had a machine that would track his gaze as he chose the letters/words. I think he could also save phrases before hand and call them up when appropriate.

Happy tau-over-2 day.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to aidbo (Reply #4)

Wed Mar 14, 2018, 10:47 AM

7. OK. I'll bite. So tau is a Greek letter. So we have a little formula here. What's it for?? nt

And thanks for the info.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Reply #7)

Wed Mar 14, 2018, 11:16 AM

14. Some people suggest the ratio of the *radius* of a circle to its circumference

would be a more useful 'universal constant' than the diameter to the circumference, ie pi. So they call the other one, about 6.28, 'tau' (and the 14th of March is,. in American notation, 3.14 - see today's Google doodle for "pi day". This would give the circumference as τr, and area as τr*r/2. It wouldn't fit so neatly in https://en.wikipedia.org/wiki/Euler%27s_identity , however.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to muriel_volestrangler (Reply #14)

Wed Mar 14, 2018, 11:23 AM

17. Yeah..... I gues I knew that..... :-)

Reply to this post

Back to top Alert abuse Link here Permalink


Response to muriel_volestrangler (Reply #14)

Wed Mar 14, 2018, 02:26 PM

21. Actually, it fits pretty neatly in Eulers identity.

e^(i*pi)=-1

e^(i*tau)=1

Generically, e^(i*theta)=cos(theta) + i*sin(theta). Where theta is some angle in the unit circle.

See the second video in my other comment below for better explanation.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to aidbo (Reply #21)

Wed Mar 14, 2018, 02:52 PM

22. Well, yes, but ...

a^b=1 is a fairly boring thing, really - if b=0, and that's what the use of tau there would be - going round a full circle. (I also prefer the way I think Douglas Hofstadter expressed the identity - e^(i*pi) +1 = 0 - since that puts in the 2 fundamental numbers, 0 and 1)

Reply to this post

Back to top Alert abuse Link here Permalink


Response to muriel_volestrangler (Reply #22)

Wed Mar 14, 2018, 03:20 PM

23. The identity is just a special case of Eulers formula.

e^(i*tau)-1=0=sin(tau)+i*cos(tau)-1=sin(pi)+i*cos(pi)+1=e^(i*pi)+1

Going around half a unit circle and adding 1 gets you zero in much the same way as going all the way around a unit circle and subtracting 1.


Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Reply #7)


Response to LAS14 (Reply #7)

Wed Mar 14, 2018, 02:18 PM

20. Tau is 2 * pi.

Pi (3.14...) is defined as the ratio of a circle’s circumference to its diameter.

Tau (6.28...) is defined as the ratio of a circle’s circumference to its radius.

In my opinion, tau is a better way to learn about the mathematics of circles and trigonometry rather than pi. So on this day, which is called pi day (3/14 = pi day) I like to call it tau-over-2 (tau/2) day. Using this format tau day would be 6/28, but tau isn’t that well known.

If The math hasn’t scared you off, here are a couple videos about it.



Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Original post)

Wed Mar 14, 2018, 10:47 AM

8. Why Stephen Hawking's voice computer spoke with an AMERICAN accent

https://www.thesun.co.uk/news/5803736/why-stephen-hawkings-voice-computer-spoke-with-an-american-accent/
Why Stephen Hawking’s voice computer spoke with an AMERICAN accent

The Queen even quizzed him on the matter, telling him: "Have you still got that American voice?" when meeting him at an event at St James' Palace.

He quipped back: "Yes, it is copyrighted actually."

...

"It is the best I have heard, although it gives me an accent that has been described variously as Scandinavian, American or Scottish."

He also explained that he was able to change the accent of his computer when the technology advanced, but he decided against it.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to dalton99a (Reply #8)

Wed Mar 14, 2018, 10:48 AM

9. Maybe it's American with a strong computer-simulation lilt???? nt

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Reply #9)

Wed Mar 14, 2018, 10:54 AM

10. It was originally done by MIT engineer Dennis Klatt:

https://www.wired.com/2015/01/intel-gave-stephen-hawking-voice/
How Intel Gave Stephen Hawking a Voice

Hawking is very attached to his voice: in 1988, when Speech Plus gave him the new synthesizer, the voice was different so he asked them to replace it with the original. His voice had been created in the early '80s by MIT engineer Dennis Klatt, a pioneer of text-to-speech algorithms. He invented the DECtalk, one of the first devices to translate text into speech. He initially made three voices, from recordings of his wife, daughter and himself. The female's voice was called "Beautiful Betty", the child's "Kit the Kid", and the male voice, based on his own, "Perfect Paul." "Perfect Paul" is Hawking's voice.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to dalton99a (Reply #10)

Wed Mar 14, 2018, 11:12 AM

13. Thanks! Now I'm wondering where Dennis Klatt was born.

I googled and googled, but no luck. Might you know? Klatt seems to be German, but there's always the mother...

Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Reply #13)

Wed Mar 14, 2018, 12:03 PM

19. He was from Wisconsin:

http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=1162851

IEEE TRANSACTIONS ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, OCTOBER 1976

Dennis H. Klatt was born in Milwaukee, WI, in 1938. He received the B.S. and M.S. degrees in electrical engineering from Purdue University, Lafayette, IN, in 1960 and 1961, respectively, and the Ph.D. degree in communication sciences from the University of Michigan, Ann Arbor, in 1964. Since 1965 he has served on the faculty of the Massachusetts Institute of Technology, Cambridge, and since 1969 has been a Research Associate at the Speech Communication Laboratory, Research Laboratory of Electronics, Massachusetts Institute of Technology. He has published papers on various aspects of speech communication, including speech recognition, speech production, the perception of rapid spectrum changes, and the speech imitation abilities of mynah birds. His current research efforts are concerned with speech synthesis by rule, perception of speech and speech-like sounds, and the phonological and acoustic-phonetic structure of English sentences. Dr. Klatt is a Fellow of the Acoustical Society of America and a member of the American Association of Phonetic Sciences. He is Associate Editor of the Journal of the Acoustical Society of America.


https://www.karger.com/Article/Pdf/261833

Obituary
Phonetica 1989; 46: 126-127

Dennis H. Klatt, a noted researcher in speech and hearing science, died in Cambridge, Mass., on December 30, 1988, after a long struggle with cancer. A graduate in electrical engineering at Purdue University in 1961, Dennis did his graduate work at the University of Michigan, where he was awarded his PhD in communication sciences in 1964. He joined the Massachusetts Institute of Technology as an assistant professor in 1965, becoming a Senior Research Scientist in 1978.

An author of more than 60 scientific papers, Dennis was working on a book on acoustic-phonetic characteristics of English at the time of his death. He recently was awarded the Silver Medal in Speech Communicalion by the Acoustical Society of America for ‘fundamental and applied contributions to the synthesis and recognition of speech’, and the John Price Wetherill Medal by the Franklin Institute, both in 1987. ...

The area where Dennis’ work has probably had the greatest impact, at least in recent years, is speech synthesis by machine. He developed a flexible computer-based synthesizer that can generate sounds to be used as stimuli in basic studies of human speech perception. The software for his synthesizer has been available to research laboratories that are active in these areas, making possible new quantitative research on the processes of auditory and speech perception. With this synthesizer as a base, he developed a complete system for synthesis of speech from English text. His research on speech synthesis has led to a detailed specification of rules for segmental durations in English. Throughout his career Dennis retained a keen interest in seeing the results of his work applied to the special needs of blind and other handicapped persons. ...

Naturally quiet and sensitive, Dennis harbored a deep aversion to prejudice or intolerance of any sort and he remained a committed supporter of the American Civil Liberties Union throughout his life. For relaxation, Dennis loved to do elaborate home repair projects, play tennis and squash, paint and listen to Mozart. Dennis is survived by his wife Mary of Brookline, Mass., his daughter Laura, a sophomore at Wesleyan University, his mother Dorothy and brother Gary, both of Whitewater, Wisc.


Stephen Hawking's daughter Lucy interviews Dennis Klatt's daughter Laura - the Klatt segment starts at 13:40
(A sad irony: Dennis Klatt died at 50, having lost his voice to cancer)


Reply to this post

Back to top Alert abuse Link here Permalink


Response to LAS14 (Original post)

Wed Mar 14, 2018, 11:17 AM

16. It's an early text to speech device

Which is why it so sounded "mechanical" and with an odd accent.

Far more sophisticated and natural sounded devices became available....no doubt that you have heard some of them in robo-calls on your phone.

However, people got accustomed to the early device and it became "his voice" and his brand, so he stuck with it since it because "him" and newer, better devices, or changing devices wouldn't be associated with him.

Bear in mind, that in addition to physics genius, he was a master marketer as well, and his books, videos and general brand sold millions.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to fescuerescue (Reply #16)

Wed Mar 14, 2018, 11:25 AM

18. So we can assume that his wife's and daughter's voices...

... would have had the same lilt, I guess. Thanks very much for this additional info. DU is great.

Reply to this post

Back to top Alert abuse Link here Permalink

Reply to this thread