[vi-kindle] Re: Text to Speech Musings

  • From: "Kay Chase" <kay.chase@xxxxxxxxxxx>
  • To: <vi-kindle@xxxxxxxxxxxxx>
  • Date: Mon, 13 Feb 2012 14:12:54 -0500

Hi, Craig,

You have prepared a very thoughtfully written note to this list. I've been using synthetic speech since 1978 and yes, I would say, that the longer I have used a TTS to read, the less I pay attention to the voice, its pauses and inflection. n Instead, I focus on the content. I push myself to listen to books as fast as I can comprehend and recommend inching the speed up every now and then to increase your reading speed.

Back in the early years of OCR and even now, of course, scanning is not a perfect science. The longer I read this way, the more automatically I replace misscanned or mispronounced words with the correct word or pronunciation, to the point that at the end of the page I probably couldn't identify specifically or tell you how many errors the page had. Now with electronically available texts, we are not dealing with scanning errors, a vast improvement. The ability to read through learning can not be replaced by multimedia, and now there is so much more educational and reading material available.

If I were to pick one thing to fix on the Kindle, it would be the inability to query the spelling of a word, particularly a need with proper names. I have used the Kindle for PC on my laptop and I bought a few books, partially as a way to decide whether they would be advantage to buying the device. The other nice feature would be the ability to go straight from the Table of Contents to a specific section of the eBook. I am first and foremost a Braille reader and would want Braille display compatibility if possible. That in itself helps find out about spelling of words.

Kay Chase


-----Original Message----- From: Craig Werner
Sent: Monday, February 13, 2012 9:44 AM
To: vi-kindle@xxxxxxxxxxxxx
Subject: [vi-kindle] Text to Speech Musings

Greetings to the list.

I have been using my Kindle Keyboard 3G for about three months now, and
I have some questions I'd like to pose to the group.

First, I find that reading at the "faster" speed causes me to lose some
words and to be confused by sentence parsing erorrs more often than when
I read at the default speed.  Have you found that reading facility at
the faster speed improves with time?  I suspect the answer is "yes"
since this has been true of my experience with my screen reader.

Second, there are a number of quirks that the TTS option displays, such
as a lack of pause between sections of a book, a failure to raise vocal
inflection after many interrogative sentences, and a failure to pause
between the speech of one character and the interruption by another.
The lack of pause can perhaps be best illustrated by a made-up volley of
conversation between two characters:

"Do you like peaches?"

"Only if they're in cans."

The TTS reads this exchange so that it sounds like a single question
spoken by one person:  "Do you like peaches only if they're in cans?"
This sentence makes syntactical sense as a question; however, that is
clearly not the intent of the writer.

The bottom line:  Should we compile a list of these kinds of anomalies
and use them to fashion a cogent, diplomatic, yet forceful query to
Amazon about possibly improving the TTS?  I suspect we're talking about
problems embedded in the Nuance RealSpeak engine, and I don't know the
degree--if any--that Amazon  is in a position to bargain with Nuance.

Thanks very much for any thoughts.

Craig
nd then figure out how to direct an organized, diplomatic, yet



Other related posts: