@ranshaa05 Interesting suggestion. Have you tried selecting the "enhance vocoder output" button on the toolbox UI? We added that to trim silences, but it also normalizes the volume of the generated wav.

@ranshaa05 Interesting suggestion. Have you tried selecting the "enhance vocoder output" button on the toolbox UI? We added that to trim silences, but it also normalizes the volume of the generated wav.

It doesn't work for me at the moment (looks like some cudnn shannanigans) but i do remember it making the output a bit louder, but it certainly wasnt perfect. Even the "normalize" option in audacity doesnt work 100% of the time. That's why im suggesting a simple volume bar to have a bit more control over the actual volume.

The toolbox is not intended to be a full-featured text-to-speech solution, but I can understand the desire for this capability since the output volume is not consistent for different speakers.

Are you able to develop this feature and submit a pull request? Or are you asking for someone else to take on this task?

The toolbox is not intended to be a full-featured text-to-speech solution, but I can understand the desire for this capability since the output volume is not consistent for different speakers.

Are you able to develop this feature and submit a pull request? Or are you asking for someone else to take on this task?

Unfortunately i am not skilled enough in python to develop this feature, so yeah, im asking for somebody else to do it.

@ranshaa05 I'll gladly work this when the pytorch PR, bugs and dependency issues are complete. Or possibly earlier if I have a free moment.

Since it could be a while I'll flag this as an issue for others to help out.

@ranshaa05 I'll gladly work this when the pytorch PR, bugs and dependency issues are complete. Or possibly earlier if I have a free moment.

Since it could be a while I'll flag this as an issue for others to help out.

thanks! :)