Voice tonality is the way a person speaks to someone. Essentially, it’s how you sound when you say words out loud.
It is the process of studying users' emotions and behaviour using their voices. Speech carries both verbal and non-verbal information, which makes it very dense information. Voice is the only modality to carry such enormous amounts of information about human behaviour.
Using Decode, you can detect multiple cues from voice, such as:
At the fundamental level, we can extract the following datapoint from speech for each second:
Apart from emotions, confidence is also calculated for each second.
We give these three metrics using voice data
Note: Silence is removed in these calculations, and the base is always the total amount of time people used to speak. In summary of Decode, however, the positive and negative emotions are also calculated from face data if a face video is available, and then a weighted sum is given as the final AI metric.
This Gantt chart shows the sequence of speaker participation in the meeting.
It gives an insight into how a particular speaker participated in the meeting. The number of filler words he used and what was his participation in the entire meeting. The longest Monologue talks about the maximum amount of time he spoke at a stretch.
This visualization shows the journey of users in terms of tonality. We can see the entire meeting timeline and the segments where positive, negative, or confident emotions were felt.