Speech Synthesizer Survey 1
NV Access is currently undertaking research into how to best improve the default voice of NVDA. Although the voice must be understandable and some what pleasant, the voice must also be free of charge, responsive, small in size, and has the ability to speak most if not all currently supported languages.
We asked screen reader users to help us in this research, by listening to several speech samples we have created, and then fill in a small survey to let us know which one they find most comfortable.
The survey is now closed, and we thank all participants for their contribution.
A total of 778 people responded to the survey. Out of the 5 speech samples, 16% of respondents preferred sample1, 7% preferred sample2, 37% preferred sample 3, 26% preferred sample 4, and 14% preferred sample 5. Clearly the winner here is sample 3, with sample 4 close behind.
Following are the 5 speech samples we had respondents listen to.