In Voice 3.0, you can choose between the following types of voices 
● Standard Voices: these are voices generated using conventional voice synthesis methods, such as the concatenation of pre-recorded sound units or simple waveform modification algorithms. Although they are less advanced in terms of quality and flexibility compared to neural voices, they are still an available option.
● Studio Voices: are generated using traditional voice synthesis technologies, such as the concatenation of pre-recorded sound units. They can often be less natural and expressive compared to voices generated by more advanced technologies.
● WaveNet: is a speech synthesis technology, which uses deep neural networks to directly model audio waves, allowing for more natural and fluid speech generation. Voices generated with WaveNet tend to have a high degree of realism and expressiveness.
● Neural2: is a speech synthesis technology based on neural networks, which constitutes an evolution or improved version of WaveNet, adding improvements in both the quality of the generated voice and computational efficiency compared to previous technologies.
The type of voice will be selected in the process of creating your Voice channel.
Here you can find out about the different voices available..