Fun with Text-to-Speech

This week, my wife and I have been proofreading the Hadrosaur Productions editions of The Astronomer’s Crypt and Dragon’s Fall: Rise of the Scarlet Order Vampires before these books are uploaded as ebooks and sent to the print vendor. Last week, my wife presented me with the code to upgrade Microsoft Office on my desktop computer. I upgraded the edition and began to look through the menus, making sure I knew where familiar features were located. Fortunately not much has changed, but I did accidentally stumble on the text-to-speech option while I had The Astronomer’s Crypt manuscript open. So, I decided to let it read a page or two to me. My first thought was that this is what it would be like for Stephen Hawking to read me a story. It was a fairly flat reading. Despite that, I found it surprisingly listenable. As it read over a section I had already approved, I noticed it skipped over a word. I looked closer and discovered that it had not skipped. I had omitted the word. Specifically it was a small one, the article “a.”

I began to think this could be a handy tool for proofreading. So I started playing it while I read over the formatted manuscript. Now then, I normally do a “read aloud” pass when I edit my manuscripts. However, if I get too much into the flow of the story, I can “read” words that should be there but actually don’t exist on the page. Also, reading it with my inflections means that I can overlook some weak, repetitive prose by placing the emphasis where I want it. The problem is, my intention may not match what another reader will see on the page. The upshot is that the flat reading of the Text-to-Speech actually proves useful because it helps me hear how well the prose itself is doing its job.

Not surprisingly, text-to-speech has limitations. If you write fantasy or historical fiction, be prepared for the program to mispronounce names. However, there’s a neat element to this. It will mispronounce those names the same way. Every. Single. Time. While going through Dragon’s Fall, I looked at names on the page and thought they were correct, but the text-to-speech program read the misspelled version differently than the correct version. This caused me to look closer. Humans have a tendency to read with visual clues, so a name like Myrinne will look very much like Myrrine when you read it on the page, but the text-to-speech program pronounces them differently.

Text-to-speech is functionality that has been part of Word processors and operating systems for a little while, so it’s possible this may not be new to many people, but if it is new to you, I recommend you give it a try and see how you like it as a tool. If you do give this a try, I recommend reading along on the page while the program reads to you. It’s hard to “hear” the difference between commas and periods, for example, but the program will make it clear when you have one of those in the wrong spot!

I have found that Text-to-Speech is enabled in Word 2019 and in Adobe Acrobat (though I found its interface is a bit clunky to use in Acrobat.) I gather it’s also enabled in the Mac edition of Scrivener, but it does not exist in the PC edition. Word 2019 gives you a nice “play/pause” button so you can stop when you hear a problem. If you get lost while the program is reading, I recommend pausing, going back to where you last were following and start again.

I’ll wrap up today’s post with an update about the books mentioned above. Lachesis Publishing has started to pull their editions of the books from publication. Last I looked, the only vendor that hasn’t pulled them down is Apple, but hopefully that will happen soon and I can begin uploading my editions.