Wisdom of Crowds Improves Text to Speech Accuracy

Share Article

Despite recent advances in text-to-speech technology, computerized voices have a tendency to mispronounce words. Brand names, acronyms, slang and jargon are particularly troublesome. All good text-to-speech applications provide a way to enter corrections. Now Text2Go 3.0 has taken this one step further. Using an automatic-update like service, corrections entered by individual users are automatically distributed to all others in the community.

It's important that sharing of pronunciation corrections requires no extra effort on the part of the user. If it does, it just won't happen.

Past News Releases


Text2Go 3.0 is aiming to use the 'Wisdom of Crowds' to create the most extensive and accurate pronunciation correction dictionary of any text-to-speech application. There are 3 key benefits to enabling the community to contribute and share pronunciation corrections.

1. Coverage - Contributions from the community will far exceed the number that a single individual or even a single organization could provide.

2. Speed - Language is changing all the time as new words, slang and jargon enter the vernacular. The first person to encounter a new word that's mispronounced, can correct it and the whole community benefits immediately.

3. Expertise - Different groups within the community have their own terminology and slang. Think of the Medical profession and teenagers as good examples. It's the experts within these communities who are best placed to identify and correct mispronunciations.

The result is the most accurate and up to date text-to-speech available. Improved accuracy gives a clearer, more pleasant listening experience, leading to increased comprehension and retention of information.

Pronunciation corrections are distributed in much the same way as antivirus programs receive new virus definitions. Every couple of days, Text2Go downloads new corrections from the Text2Go server. At the same time it uploads any corrections entered locally. These contributions are merged and made available to the rest of the community in the next update.

"It's important that sharing of pronunciation corrections requires no extra effort on the part of the user. If it does, it just won't happen." comments Mark Gladding, founder of Tumbywood Software. "There's nothing worse than being pestered every day to check for updates, waiting around while an application does it stuff and then having to acknowledge that there were no new updates. So we designed Text2Go to do all updates in the background while you continue to use the application."

You can watch a video on correcting and sharing pronunciation with Text2Go 3.0 at http://www.text2go.com/portals/0/Screencasts/check-pronunciation/check-pronunciation.htm or
read more about the latest release at http://blog.text2go.com/2008/08/07/text2go-30-released-pronunciation-correction-done-right

Text2Go 3.0 turns any online text to speech and transfers it to your iPod or MP3 player. Text2Go is designed for commuters who want to make productive use of their travel time. Commuters can now listen to their favorite blogs and online magazines, keep up to date with the latest developments in their profession, enjoy a huge selection of novels in eBook format and listen to long emails while commuting.

Text2Go runs on your PC and costs US$25 standalone or US$45 with a high quality voice.

Text2Go is developed by Tumbywood Software, a small software startup located in Melbourne, Australia.

Mark Gladding
Tumbywood Software
13 Cooinda Crt
Frankston South
Victoria 3199

Email: markgladding @ text2go.com
Phone: +61 450 010201


Share article on social media or email:

View article via:

Pdf Print

Contact Author

Mark Gladding
Visit website