Audio books with any actor

September 1st, 2006 by Hugh

Sony have come up with an idea for a talking book, without the need for an actor to go to the wearisome trouble of reading the whole text out aloud. The New Scientist Invention blog reports that the giant electronics to media company has filed a patent for a new type of audio book.

An actor has to record a series of words and phrases containing every type of sound in the language. The system would also note the actor’s pitch, tone, and rhythm of voice. In theory, a customer could match any book in the Library to the voice of any actor in the system. It might even be possible to revive the voices of dead actors, using archive recordings. Orson Wells reading Winnie the Pooh?

I have to say, I spend quite a bit of my time producing an actress reading audio stories for children. I know how much time she spends thinking about how to bring out the meaning. It’s a performance, just as much as if she were taking part in a stage play.

How about this for a suggestion? Perhaps Sony should add a moving hologram of the actor to its invention. Then users could choose costumes and scenes and put the actor inside any play or film ever written. I’m not entirely kidding. I have a feeling this might happen one day – perhaps in a virtual life site such as Second Life. Moving avatars of celebrities are already becoming popular.

Another observation: every few days there seems to be a news story about the future of the book. There clearly are a lot of people thinking about bringing traditional books and the digital world closer together.

5 Responses to “Audio books with any actor”

  1. Claes-Fredrik Mannby Says:

    In the process of investigating the feasibility of getting online articles onto my iPod for the morning commute, I spent some time investigating text-to-speech (TTS) engines, probably around 2 years ago now. I came across some promising technology from AT&T, as well as a European company, Acapela, makers of Elan Speech, that showed pretty strong promise. (I especially like the French TTS engine from Acapela, which, to me, sounds much more natural than the English version.)

    I did the highly scientific experiment of pulling down an article off the web, as well as an arbitrary first page of a book from the Gutenberg Project, fed them to the Elan demo, closed my eyes, and listened.

    To my slight surprise, I found that I could imagine myself listening to technical articles this way. However, the work of fiction was utterly insufferable. The slight oddities in rate changes and enunciation were excruciating, because they jarred so strongly with the underlying meaning of the text, making it impossible to focus on the content rather than the style. I suppose that a listener can overcome this, much like when learning to appreciate a new style of music, but I doubt many, if any, would choose to do so.

    Perhaps the next generation of TTS engines from platform vendors will equal the quality of Elan and the AT&T systems, and will enable listening, at least to technical content.

    And, I sincerely hope that Sony, or someone, is going to solve the reading of fiction texts well enough soon. I’ve been an Audible subscriber for a few years now, and have both been delighted by great book readings (I recommend His Dark Materials, Mistress of Dragons and Snow Crash, for example), but I have also been disappointed by the choice of readers for some books (e.g. Scott Brick is getting very popular, and although his pompous, exaggerated emphasis of every word was suited to the Dune series, for example, it’s totally inappropriate for Isaac Asimov texts). So, I would love to have the reader of the first Mistress of Dragons book read the following books, rather than the actual people that were chosen. But I don’t expect Sony, or anyone, to be able to match the full-cast sort of dramatic staging of books, as in His Dark Materials, for example, any time soon. I think you’d need to create a way for texts to be marked up with artistic nuances, which I expect it will take decades, if not centuries, for software to figure out independently, since it involves an understanding of psychology, sociology and reality. But, I would love to be surprised and proved wrong. It can perhaps be faked fairly well, analogously to how today’s games cheat to create believable 3D environments without full ray tracing.

  2. Stephanie Ciccarelli Says:

    Hi Hugh,

    An interesting article and concept.

    Sony has rather high hopes regarding the duplication of voice talent, particularly reviving voices from the past. As you mentioned, voice actors tend to put their entire selves into the performance, including interpretive style, fluctuation of the voice, intonation, character development, skill, and more.

    I agree with you that it is good that books and the digital world are crossing paths. eBooks and audiobooks are classic examples of that marriage!


    Stephanie Ciccarelli

  3. rachel Says:

    wow..nice blog,it’s good idea to get more sales with in short span of will definitely successed in the market,childrens showing interest to listen their favourite actors voice.

  4. Rolf - Audio Books Fan Says:

    Sony has a lot of technology know-how, but text to speech is an entirely different kettle of fish…..

    There is also a service called, You can play around with their demo to have different characters say the sentences you want in the language you specify. It’s quite amazing, and the character animation is something else too! As a added tool for web marketing certainly very good.

    But I can’t imagine that a work of fiction would be enjoyable, listenting to these nevertheless quite mechanical voices. I do prefer audiobooks that are ‘told’ or ‘narrated’, as opposed to just read. And for this you need people who feel what’s in the intent of the author and can add their sould to the conent. That’s after all what makes audio books so enjoyable!


    P.S.: Download this free audio book: BBC Comedy sampler audio and try to imagine how this would soudn if the text were read by Sony robots – this definitely shows how absurd their idea is. With todays computer power this is not feasible yet, maybe some time in the future… ???

  5. Paul anderson Says:

    I have got few guides from and that was outstanding perfect.