Smoother start with voice output
With voice output on, the first minute after you started translating could make your whole computer stutter. Unitra AI now gets speech recognition and voice output ready before it starts translating, and keeps them from slowing each other down.
While it prepares, the bottom of the main window shows Preparing speech recognition and then Preparing voice output. This usually takes a few seconds. The first time after installing Unitra AI or updating your graphics driver it takes longer, and Unitra AI lets you know that you may notice a brief stutter until it is done.
Cleaner voice profile
With Use my voice profile, translations could be read with a rhythmic background noise. Unitra AI now learns your voice from your microphone's original sound, only from the moments you are speaking, and skips recordings that are clipped or noisy.
A voice profile learned with an earlier version keeps that noise. To learn it again, open Settings > Audio > Voice Output, click Learn again next to your voice profile, and translate as usual.
Fixes
- Your own voice was translated twice in some games. With voice output set up, what you said could also show up as game audio. The Unitra microphone is no longer picked up as game or desktop audio.
Plus fixes for other known bugs.
If something does not work as expected after updating, let us know in Discord.