Improving the performance of the output

[ASandwichMan]

Which values in the configuration files do I need to adjust to get the highest quality output? I increased n_iter and n_mix substantially, but the output sounds the same as with the default settings.

I also tried decreasing n_mix to 4 and 2 for mcep and codeap respectively, but even that seemed to give similar outputs.

I am using the TM1 and TF2 voices from the VCC2018 speech corpus.