Update Speech2Text in questfox #time #confidencescore #realvoiceofthecustomer

With every study we learn about new requirements in the brand new world of speech in research. After some thousands of interviews we realized, that two additional informations are always demanded by the users of questfox when using the automatic transcription features.

1. Duration of a speech answer
   There are several reasons why it makes sense to track the answering time. The API business is counting the seconds and makes us pay for their services. The researcher wants to know if people really take the time to answer. And brand new analytics tools for emotional voice analysis require a minimum answer time before their services can be used (eg. 7 seconds minimum). Even before one could implement the time measurement on a questfox page (para data). But as it only measured the time between appearance and pressing the “next” button, we now become more precise by only counting the seconds, the microphone was used.
2. Transcription Confidence Score
   The confidence score between 0 (worst) and 1 (perfect transcription) is an educated guess about the quality of the transcript. After testing this feature with an extra question type we finally integrated it into all transcription question types as it is often demanded.
   See the explanations on the transcription confidence score in this prior post from Mai 2019

You will all take advantage of those features right away. The only inconvenient thing is, that you will have to integrate the question types again from scratch, as the old versions will no longer function. Sorry for this inconvenience, but the advantages of using those desired features are only a click away for you now.

The video question type does not cover that feature set yet. We will keep you informed if this new feature will also be available in the world of video.