WaveNet Text-to-Speech integrated into questfoxPosted: 1. November 2020
Reading out loud the research question
For years we have been using different tools to create synthesized speech from text. In 2017 these features were implemented in questfox in times when it was even a little too early for that technology. Our speakers sounded a little too artificial to the human hear.
With the last update we are now able to deliver more human sound in the reading out loud functions of questfox by integrating WaveNet right into questfox.
This is what Google tells about the WaveNet Development.
“Google Cloud Text-to-Speech enables developers to synthesize natural-sounding speech with 100+ voices, available in multiple languages and variants. It applies DeepMind’s groundbreaking research in WaveNet and Google’s powerful neural networks to deliver the highest fidelity possible. As an easy-to-use API, you can create lifelike interactions with your users, across many applications and devices.”
TTS in questfox Multimedia Insights
You will find the function right under Multimedia Insights – Text to Speech
At the moment we offer 5 languages for the text-to-speech functionalities inside of questfox. The default language is English (US).
5 languages already available in questfox
WaveNet was first supporting English and only integrated German lately. It may take some time before other languages may be available. We will keep you updated on the issue.
Select the voice of your choice
Listening while reading?!?
We are still in consideration whether participants should listen and read at the same time. Some of our own people argue that we should force interviewees to listen only. If you plan to do this you can set the visibility of the text to hidden by integrating the following into CSS.
Click on View – Source Code to open the HTML Window of the editor.
Put the following command in front of your text
<p style=”visibility: hidden;”>Your text here</p>
And you will no longer see the written words.
Please only do so once your text is finalized because you will not even see the text anymore in the editor 🙂
Future of speaking software
Looking at the speed of development we are pretty confident, that these functions will become better and better. It’s not bad already. But are you and your respondents already ready for this?
An interesting project in South Africa showed us the potential of this approach with people who are listening to the read out loud questionnaire.