Can the model support standard Chinese pronunciation?

[fancysh]

The audio generated using the system speaker is not very satisfactory; it still sounds more like foreign colloquial speech. Can the output be adjusted to produce stable and standard Chinese pronunciation? Additionally, the speakerId I used was trained on Chinese data, but the output contains mixed pronunciations from other countries. Are there specific parameters that can be set or adjusted for this? Could you provide a document detailing these settings? Thanks.