This guide explains the differences between the voices that you can select for Virtual Humans.
One of the keys to creating an effective Virtual Human is selecting the right voice. In the voice tab, you can select from a variety of voices for every available language, with varying accents, and qualities. We provide information on the voices using the labels underneath the name. Here's a quick guide on what those labels mean, so you can make the right choice.
Realistic:
Voices labeled realistic will sound more more human and less robotic, but may take longer to process resulting in a slower response.
High Performance:
High performance voices will sound more robotic, but will take less time to process and will often result in a faster response.
Streamed:
Voices labeled streamed will have the fasted response time, but may have less realistic pacing and inflection. These voices utilize a different method of processing the audio file, where the audio is chunked up into small pieces that are processed sequentially. This allows the virtual human to start the response before the full audio file is processed, resulting in a quicker response.
Popular:
Voices labeled "Popular" are the most used voices, so give these a try!
Locale:
All voices have a locale at the end, e.g. GB, US, AUS, which represents the accent that the Virtual Human will have.
If you have any questions or feedback, contact our support team via support@virti.com and we’ll be happy to help.
