Tuesday, March 15, 2005

Listening to Vikki

One of the steps to my manuscript clean up process, is listening to it. With some of my earlier works, I had spoken it out loud, but a novel this size takes ten hours, at a normal pace. Instead, for several years now, I've converted the text to a spoken file using the text-to-speech software built into the Mac. I then play the spoken version, while following along in the manuscript, fixing things as I go.

I've used a variety of software tools to manage the conversion process. In the earlier days of Mac system 8 and 9, there was a size limit of 32 thousand characters that could be converted at one time. I wrote scripts to chop up the text into pieces and then convert each piece.

Lately, I've been using 'books2burn' software by Matthew Weinstein. The 'book burning' is an analogy to CD song ripping and burning. The software does a good job of detecting chapters and splitting the text. I converted the novel into fifty AIFF files with books2burn, and then imported them into iTunes. I convert the AIFF to mp3 files for size reasons, and then delete the 1.5GB of AIFF files. iTunes makes a nice orderly player, with a quick pause button when I detect an error that I have to fix.

I use the Vikki built-in voice, one of the better ones. The text-to-speech conversion is usable, although there are quirks. For example, 'Okay' comes out 'kay', and frequently hypenated words are spelled out, rather than spoken.

Still, the end result is usable. In spite of the fact that writing and reading is a highly visual activity, deep in it's roots, text is just codified speech, and so many textual errors are just invisible to my eyes.

But when the sentence is spoken, some problems jump right out at me. It's a time consuming process to listen to Vikki for hours on end, but it's worth it.

No comments: