Date: Tue, 16 Apr 2002 14:05:47 -0400 From: richard seltzer Subject: Chat Reminder -- learn about a new way to make quality "spoken" books, Thursday April 18 BUSINESS ON THE WEB -- Where "word of keystroke" begins This Thursday, April 18, from noon to 1 PM (Eastern Time -- GMT -4) we'll be talking about an alternative way of making books you can hear -- and enjoy hearing. Our guest, Nick Hodson, has a way of creating and using a dictionary file that is specific to the text and also to his personal pronunciation preferences, and also of massaging the text to put pauses where they belong. His files work with Fonix ISpeak. The text of a book plus the added files take up far less space that audio files, meaning that you could theoretically put the makings for hundreds of quality "spoken" books on a single CD. I have not tried this myself yet, but it sounds intriguing and I want to learn more from Nick. Please join us for the discussion. Nick explains: "The scanning of books to produce e-texts is a process which is improving continuously. I like to go one step further and produce CDs which contain the book in spoken form. The main market for these would be impaired-vision folk, car-drivers, and people requiring a background sound that is intelligible. Most CD players on the market these days will read home-produced CDs, with the speech files in MP3 or WMA format. To create these files optimally you need to derive from the original texts a secondary text which will be read acceptably by the speech engine. This includes adding longer pauses at the end of each paragraph, comma-length pauses (pointings) to break up long sentences, and instructions to differentiate between words that may have more than one pronunciation, such as row, bow, lead, and about 100 more. We also need to make a list of all different words used in the book (usually about 8000), and from that deduce a much shorter list of words that we do yet not know how the speech engine will say them. We hear this list, and mark the words that it does not pronounce well. We create a look-up list that will enable the speech engine to say every word in the book correctly. At the end of this we have the massaged text of the book and the look-up word list. These are not very big, compared with the speech files that we are now going to make. So my message is, we need to develop a generation of portable devices that can hold an enormous number of massaged texts, and a speech engine that can read directly from them. The speech engine could pronounce words in any manner acceptable to the listener, with a British regional voice, an Australian voice, any of the US voices, a Delhi voice... I think we are really not very far from this, though at present I know of only one speech engine that is capable of reading a whole book at a time, taking note of the markups and the look-up pronunciation list; and of only one device that potentially could do this. These are Fonix ISpeak, and Creative Lab's six gigabyte Jukebox. The economic interest in this would be the creation of the marked-up texts and the look-up lists, the technology of which I have worked on for eighteen months." He explains his background thus: "I am a retired computer programmer of some 45 years experience, living in London, England. My first job after university, where I read Mathematics, was to (try to) optimise the designing of nuclear power stations. At the time, 1956, this was mostly done by hand, but computers were just coming in, and, shortly afterwards, the first high-level language, Autocode, arrived. By the late 1960s I was working for a global corporation, creating ever better business software. I retired in 1997. By then I had acquired an interest in scanning texts, originally quite short ones, as I was (still am) editor of a small kayaking magazine. This moved on to books, five or six of which I did rather clumsily before retiring. By chance, immediately after retiring, my father became very ill, and my mother wasn't at all well, so I moved in for a few years, and spent my whole free time developing software for the efficient creation of e-texts. This then moved on, in Autumn 2000, to the efficient creation of books on CDs." To connect, go to http://www.samizdat.com/chat-intro.html We'll be on, as usual, from noon to 1 PM Eastern Time (GMT -4) this Thursday. For links to transcripts of previous sessions, check http://www.samizdat.com/chat.html Please let me know your suggestions for other guests and topics. I'm looking for interesting and innovative business models, products, and services. seltzer@samizdat.com Please send email with your follow-on questions and comments, and requests to receive email reminders about upcoming sessions. Best wishes. Richard Richard Seltzer, seltzer@samizdat.com, www.samizdat.com Internet marketing consultant www.samizdat.com/consult.html A library for the price of a book -- visit our new online store http://store.yahoo.com/samizdat Ask for my new book Web Business Boot Camp at your local bookstore, or buy it at Amazon http://www.amazon.com/exec/obidos/ASIN/0471164194/brsamizdatexpres 617-469-2269 PS -- To see what's possible with CD ROM books, please take a look at our online store at http://store.yahoo.com/samizdat And check our Teachers/Students Guide at http://www.samizdat.com/guide.html _______________________________________________ Cone-cal mailing list Cone-cal@blu.org http://www.blu.org/mailman/listinfo/cone-cal