Making TTS audio sound natural

Audacity version 2.3.2

Hi there. This is my first time on here and using Audacity - please be gentle :stuck_out_tongue: ****

I am trying to get the audio created by the free Text-To-Speech program Balabolka to sound more natural and human. I have fixed the mis-pronunciation errors via spelling some words phonetically which helps, but the speech still has sharp tinny hiss when pronouncing some words, in particular words starting or ending with a ‘T, S, or G’. Hopefully you can hear it in the attached clip.

I have been playing around with some of the effects, some of which helped a little which - telephone for example - but if anyone knows any other tricks I can use to remove the sharp hiss, or make it sound more natural in general, please let me know, any advice gratefully received.

NB - it is in MP3 format as it will be combined with an image via to create an MP4 so that it can be uploaded to youTube.

Many thanks in advance,


If your TTS app supports a higher quality audio format (such as WAV, or higher bit-rate MP3) then that may help a bit.
I doubt that there’s much else that you can do, as the TTS developers have already done everything they can to get the best, most natural sound quality.

Some TTS apps are better than others. Unfortunately, most of the best ones cost $'s.

In my opinion the worst problem is the errors in patter, rhythm, swing and timing. There’s no effects or filters for that.

We can’t change Male/Female, Adult/Child or anyone into an announcer, either. All popular requests.