Text-To-Speech Optimisation

Text-To-Speech (TTS) software has come a long way in a few short years, and it is likely to become ever more popular in the years to come. It is already used extensively for the narration of many YouTube videos. It is also utilised by some e-readers to turn e-books into ‘ersatz’ audiobooks.

More importantly, it may serve as an accessibility aid in the future for those who have difficulty reading.

There is a problem with TTS software at the moment, however. It does not always work as well as it should. The technology is still developing, and some of the kinks have yet to be straightened out.

If you have heard current Text-To-Speech narration, you have no doubt cringed at the inconsistent speed, uncomfortable pauses and pronunciation errors that plague it.

However, there is a solution, this being Text-To-Speech Optimisation or TTSO. This involves the human touch to tweak the text into something the narration software can more easily understand.

This is where I come in. I have used Text-To-Speech software extensively over the last few years, and have become adept at getting the most from it.

The following sentences contain a sample of the words current TTS software struggles with. Listen to the un-optimised audio first to hear these errors, then the optimised version to hear narration that is now error-free, I think you will be pleasantly surprised.

“The virus had a zombifying effect on the pseudo chimera in Pripyat. Traumatic weapons such as tasers and primitive weapons such as Bows were ineffective against it. SOPs therefore were to use shotguns with lead buckshot at close range, or thermobaric weapons at a distance. For the sake of infection control, care must be taken to disembowel the carcass.”