Bob Dylan as the voice of your navigation device

Navigation device manufacturers are talking to Bob Dylan about using his voice to give the GPS directions:

The enigmatic troubadour said on his satellite radio program that he is negotiating with two car manufacturers to be the voice of their in-car navigation systems. Insert your own Dylan-lyric pun here about “no direction home” or “there must be some way out of here” or “how many roads . . . .”

The wonder of this might not be that Dylan is selling out — he has already done that by appearing in ads for Victoria’s Secret, Pepsi, Cadillac and others, and he’ll be singing “Here Comes Santa Claus” on a forthcoming Christmas album — but that his famously raspy and mumbly voice would be suited for directions-challenged drivers.

Dylan himself wasn’t even so sure about that. On his BBC radio show he gave listeners a preview of his would-be GPS vocals: “Left at the next street. No, right. You know what? Just go straight.”

He also noted: “I probably shouldn’t do it because whichever way I go, I always end up at one place — on Lonely Avenue. Luckily, I’m not totally alone. Ray Charles beat me there.”

Can anyone explain how the voices are connected to the computerized directions? Would Bob have to read the names of every street in America, with then the computer providing the association and the intonation? Would this gig mean just reading a long list of words that don’t go together, sort of like some of his songs? (I had assumed that the voices were all computer-generated, but I have since learned that, as here, actual human beings do the talking and the back-street driving.) But I can’t imagine how that can be made to work. Somebody please explain.

Also, along the lines of this first paragraph, can you think of other Dylan lines that would fit this project?

About Gene Veith

Professor of Literature at Patrick Henry College, the Director of the Cranach Institute at Concordia Theological Seminary, a columnist for World Magazine and TableTalk, and the author of 18 books on different facets of Christianity & Culture.

  • LisaV

    Maybe he’ll just have to say stuff like keep right, merge left, take the highway and you have reached your destination. My GPS (a tom-tom) doesn’t give street names.

  • LisaV

    Maybe he’ll just have to say stuff like keep right, merge left, take the highway and you have reached your destination. My GPS (a tom-tom) doesn’t give street names.

  • womanofthehouse

    That would make me not want a GPS. I can’t stand Bob Dylan’s voice!

  • womanofthehouse

    That would make me not want a GPS. I can’t stand Bob Dylan’s voice!

  • CRB

    How about, “Positively 4th Street”?!

  • CRB

    How about, “Positively 4th Street”?!

  • Bob Hunter

    On my TomTom only one voice gives street names, and that voice is a computer voice, not a real person. So he would have to give general directions such as “turn left, 500 feet,” etc.

  • Bob Hunter

    On my TomTom only one voice gives street names, and that voice is a computer voice, not a real person. So he would have to give general directions such as “turn left, 500 feet,” etc.

  • http://www.blogger.com/profile/01763924682909630509 Orianna Laun

    I wonder if they do the recordings for GPS they way they do for the recorded phone numbers. With phone numbers, each number is recorded with a different inflection, and then rearranged by computer to come up with the correct number. If that is how the streets are done, then my guess is different phonemes are recorded with different inflections and then rearranged. Otherwise, that would be pretty tedious to read not only all those street names, but also restaurants and sites, as in our GPS. Although I must say that we get a good chuckle out of our GPS “Karen” (whom I believe you featured long ago on this blog) when she pronounces places like “King shy way”, not Kingshighway and calls Washington highways “Western Australia” because they’re both abbreviated WA.

  • http://www.blogger.com/profile/01763924682909630509 Orianna Laun

    I wonder if they do the recordings for GPS they way they do for the recorded phone numbers. With phone numbers, each number is recorded with a different inflection, and then rearranged by computer to come up with the correct number. If that is how the streets are done, then my guess is different phonemes are recorded with different inflections and then rearranged. Otherwise, that would be pretty tedious to read not only all those street names, but also restaurants and sites, as in our GPS. Although I must say that we get a good chuckle out of our GPS “Karen” (whom I believe you featured long ago on this blog) when she pronounces places like “King shy way”, not Kingshighway and calls Washington highways “Western Australia” because they’re both abbreviated WA.

  • J. Bruce

    Here’s a line that might be appropriate:

    We’re both just one too many mornings
    An’ a thousand miles behind.

  • J. Bruce

    Here’s a line that might be appropriate:

    We’re both just one too many mornings
    An’ a thousand miles behind.

  • http://theoldadam.wordpress.com/ Steve Martin

    ‘You’ve got to swerve from somebody’

  • http://theoldadam.wordpress.com/ Steve Martin

    ‘You’ve got to swerve from somebody’

  • http://www.bikebubba.blogspot.com Bike Bubba

    Hopefully he’s relatively sober when he does the recording, or we’ll recognize cars with his GPS system because they’re hopelessly lost.

  • http://www.bikebubba.blogspot.com Bike Bubba

    Hopefully he’s relatively sober when he does the recording, or we’ll recognize cars with his GPS system because they’re hopelessly lost.

  • http://gollygeeez.blogspot.com/ Z

    I can’t get past the idea of having to listen to that voice at all! I hope one can understand the lyrics better when he talks than when he sings!
    I’d prefer Dick Cavette’s clear voice giving me directions!

  • http://gollygeeez.blogspot.com/ Z

    I can’t get past the idea of having to listen to that voice at all! I hope one can understand the lyrics better when he talks than when he sings!
    I’d prefer Dick Cavette’s clear voice giving me directions!

  • EGK

    Dick Cavett’s voice? “As Groucho once remarked to me, I’d never turn on this street.”

  • EGK

    Dick Cavett’s voice? “As Groucho once remarked to me, I’d never turn on this street.”

  • http://mesamike.org Mike Westfall

    “You drive just like a woman!”

  • http://mesamike.org Mike Westfall

    “You drive just like a woman!”

  • http://www.proginosko.com James Anderson

    Synthesized speech is typically generated from small samples of recorded human speech (from one speaker). All you need are samples of all the individual phones or diphones (roughly, syllable sounds) that appear in the required language (e.g., English). See here for more info.

    It doesn’t take long to get the necessary samples for a cogent English speaker. Getting the necessary samples from Bob Dylan is another matter altogether.

  • http://www.proginosko.com James Anderson

    Synthesized speech is typically generated from small samples of recorded human speech (from one speaker). All you need are samples of all the individual phones or diphones (roughly, syllable sounds) that appear in the required language (e.g., English). See here for more info.

    It doesn’t take long to get the necessary samples for a cogent English speaker. Getting the necessary samples from Bob Dylan is another matter altogether.

  • CRB

    How about, “Don’t think twice, it’s ALL RIGHT!” or
    “The Lanes, they are a-changin”! OR, “Tangled up in traffic!” : )

  • CRB

    How about, “Don’t think twice, it’s ALL RIGHT!” or
    “The Lanes, they are a-changin”! OR, “Tangled up in traffic!” : )

  • http://www.caryschwarz.com saddler

    “Strap yourself to a tree with roots,
    cause you ain’t goin’ nowhere.”

  • http://www.caryschwarz.com saddler

    “Strap yourself to a tree with roots,
    cause you ain’t goin’ nowhere.”

  • http://jen-lehmann.livejournal.com Jen

    As others have said, he’ll probably record a list of phrases: “Turn right. In 800 yards, turn left. Turn around when possible,” etc. The Darth Vader voice on our TomTom says “You have reached the Dark Side” when you reach your destination. There’s a John Cleese voice available, but we haven’t purchased it yet. I’m sure they’ll come up with ways to make Dylan entertaining, as well.

  • http://jen-lehmann.livejournal.com Jen

    As others have said, he’ll probably record a list of phrases: “Turn right. In 800 yards, turn left. Turn around when possible,” etc. The Darth Vader voice on our TomTom says “You have reached the Dark Side” when you reach your destination. There’s a John Cleese voice available, but we haven’t purchased it yet. I’m sure they’ll come up with ways to make Dylan entertaining, as well.

  • Joanna Hensley

    “There must be some way outta here…”

  • Joanna Hensley

    “There must be some way outta here…”

  • Joanna Hensley

    Oops – that was a repeat. How ’bout: “You break just like a little girl.”

  • Joanna Hensley

    Oops – that was a repeat. How ’bout: “You break just like a little girl.”

  • http://www.geneveith.com Veith

    Thanks for those good lines, everybody. And James Anderson, thanks for the technical explanation. I suppose the technology would be the “unit selection synthesis” approach, which would involve the recording of thousands of individual words, as well as partial words.

    Our GPS–Karen, whose voice is that of a real person–does say the names of the streets, as Orianna describes, with sometimes comical results. But I still don’t understand how the units are correlated. Is it to the spelling of words on maps? But that often doesn’t translate into pronunciation. And yet Karen is right more often than she’s wrong. I still can’t comprehend how that works.

  • http://www.geneveith.com Veith

    Thanks for those good lines, everybody. And James Anderson, thanks for the technical explanation. I suppose the technology would be the “unit selection synthesis” approach, which would involve the recording of thousands of individual words, as well as partial words.

    Our GPS–Karen, whose voice is that of a real person–does say the names of the streets, as Orianna describes, with sometimes comical results. But I still don’t understand how the units are correlated. Is it to the spelling of words on maps? But that often doesn’t translate into pronunciation. And yet Karen is right more often than she’s wrong. I still can’t comprehend how that works.

  • http://www.toddstadler.com/ tODD

    Veith (@18), your question now seems to be: how does one build a morpheme-to-phoneme mapping (I’m assuming that the GPS software does, in fact, use the labels found on maps as the source of its pronunciation).

    To be sure, it probably is a fairly complex bit of software, but the concept is exactly the same as how we learn the same process. You start with some basic assumptions, and then make sure you cover all the exceptions.

    So we teach kids that every letter has a particular sound. And, after they get a hold of that, we teach them, that, well, actually, when you see those two letters (say, S and H) together, it makes a totally different sound. And so on and so on (eventually, you will have an exception for “ough”, for instance). Of course, to build a truly accurate mapping, you will occasionally have entries for whole words, if that word is an exception to every other rule you have.

    Of course, I’m guessing, but that’s how I imagine it works.

  • http://www.toddstadler.com/ tODD

    Veith (@18), your question now seems to be: how does one build a morpheme-to-phoneme mapping (I’m assuming that the GPS software does, in fact, use the labels found on maps as the source of its pronunciation).

    To be sure, it probably is a fairly complex bit of software, but the concept is exactly the same as how we learn the same process. You start with some basic assumptions, and then make sure you cover all the exceptions.

    So we teach kids that every letter has a particular sound. And, after they get a hold of that, we teach them, that, well, actually, when you see those two letters (say, S and H) together, it makes a totally different sound. And so on and so on (eventually, you will have an exception for “ough”, for instance). Of course, to build a truly accurate mapping, you will occasionally have entries for whole words, if that word is an exception to every other rule you have.

    Of course, I’m guessing, but that’s how I imagine it works.

  • Bob Myers

    “They’ll stone you when you’re riding in your car…”

  • Bob Myers

    “They’ll stone you when you’re riding in your car…”

  • http://www.proginosko.com James Anderson

    Surnames and place names are notoriously irregular in their pronunciations, which means that an algorithmic approach (e.g., mapping morphemes to phonemes) often fails, occasionally with hilarious results. For that reason, higher-end text-to-speech systems will use a database with tens of thousands of entries mapping names to their proper (or at least common) pronunciations.

    Of course, creating such a database is a manual process, both tedious and time-consuming. As with any other tedious and time-consuming task, the best solution is to pay people less well off than you are to do it for you. :)

  • http://www.proginosko.com James Anderson

    Surnames and place names are notoriously irregular in their pronunciations, which means that an algorithmic approach (e.g., mapping morphemes to phonemes) often fails, occasionally with hilarious results. For that reason, higher-end text-to-speech systems will use a database with tens of thousands of entries mapping names to their proper (or at least common) pronunciations.

    Of course, creating such a database is a manual process, both tedious and time-consuming. As with any other tedious and time-consuming task, the best solution is to pay people less well off than you are to do it for you. :)

  • Doug

    I got lost going to Wallops Island in a rental car once.

    The GPS pronounced the name of Assawoman, VA several times to my amusement.

  • Doug

    I got lost going to Wallops Island in a rental car once.

    The GPS pronounced the name of Assawoman, VA several times to my amusement.

  • CRB

    Don’t know if a “plug” is allowed here, but want to say for all Dylan fans, his new cd, “Through Life Together” is a real gem. I especially like the last track–Dylanesque
    sarcasm at it’s best, “It’s All Good.”

  • CRB

    Don’t know if a “plug” is allowed here, but want to say for all Dylan fans, his new cd, “Through Life Together” is a real gem. I especially like the last track–Dylanesque
    sarcasm at it’s best, “It’s All Good.”

  • Pete

    “How many roads … blowin’ in the wind.”
    And speaking of TTL, CRB, anybody noticed how many of the lyrics on this album make wonderful sense if considered to be addressing the relationship of Christ and the Church? “Hell’s my wife’s hometown.” “If you ever go to Houston, you better walk right” (i.e. be righteous). Jolene walks down High (exalted?) Street, making dead men rise. Beyond here (i.e. Outside the Christ/Church relationship) lies nothin’.
    Listen closely to it through that lens. Very provocative. Cagey, that Bob. All these songs apparently about romantic love, which I suspect Bob is chronologically beyond.

  • Pete

    “How many roads … blowin’ in the wind.”
    And speaking of TTL, CRB, anybody noticed how many of the lyrics on this album make wonderful sense if considered to be addressing the relationship of Christ and the Church? “Hell’s my wife’s hometown.” “If you ever go to Houston, you better walk right” (i.e. be righteous). Jolene walks down High (exalted?) Street, making dead men rise. Beyond here (i.e. Outside the Christ/Church relationship) lies nothin’.
    Listen closely to it through that lens. Very provocative. Cagey, that Bob. All these songs apparently about romantic love, which I suspect Bob is chronologically beyond.

  • CRB

    Pete,
    Yeh, I can see what you say about some
    of the songs addressing the relationship
    of Christ and the Church.
    I’m not sure what, “Hell’s my wife’s hometown” is about in that regard.

  • CRB

    Pete,
    Yeh, I can see what you say about some
    of the songs addressing the relationship
    of Christ and the Church.
    I’m not sure what, “Hell’s my wife’s hometown” is about in that regard.

  • http://www.geneveith.com Veith

    “Highway 61 Revisited.” And when you miss the turn, instead of saying “recalculate,” he says, “revisited.”

  • http://www.geneveith.com Veith

    “Highway 61 Revisited.” And when you miss the turn, instead of saying “recalculate,” he says, “revisited.”

  • Pete

    CRB; pre-Christ (pre- baptism, we Lutheans would say), those of us now in the Church were God’s enemies – “home town” = Hell. I like the little evil chuckle BD gives at the end of this song. Pleased to have rescued his wife (or should I say His Wife) from her rough beginnings.

  • Pete

    CRB; pre-Christ (pre- baptism, we Lutheans would say), those of us now in the Church were God’s enemies – “home town” = Hell. I like the little evil chuckle BD gives at the end of this song. Pleased to have rescued his wife (or should I say His Wife) from her rough beginnings.

  • Pete

    Ooh – bad proof-reading. That’s Lutherans, not Lutheans.

  • Pete

    Ooh – bad proof-reading. That’s Lutherans, not Lutheans.

  • CRB

    Pete,
    OK, now I get it!

  • CRB

    Pete,
    OK, now I get it!


CLOSE | X

HIDE | X