Sound Design Stack Exchange is a question and answer site for sound engineers, producers, editors, and enthusiasts. It only takes a minute to sign up.
Sign up to join this community
Anyone know of a audio filter or tool that can manipulate accents in an audio. Or may be some tool that achieve closer result.
Doing some research on voice recognition systems. Want to feed same voice queries in different accents. Purpose is not unethical at all.
There are a lot of components that go into accents, ranging from how individual phonemes are articulated to sentence-level pitch contours. You can't hope to convert an accent into another one without an extremely laborious manual process that involves carefully adjusting formants and pitches. Even swapping one phoneme for another sounds tough.
I'm not aware of any software that could do this, either. Praat is mostly analysis and visualization, and I don't think Melodyne can control formants to that precision.
It's actually a pretty interesting potential research topic, but there don't appear to be any solutions available today.
I think this is more to do with phoneme recognition and pattern recognition. I had a transcription job to do recently on a documentary where the subject was speaking English but with a pronounced African accent. Where native English speakers might use two syllables for a particular world, this chap was using three. Took me a while to work out exactly what he was saying for the transcription, but nailed it in the end. This is going to be a very challenging project and I firmly believe that if you manage it - you will be working primarily in the AI realm - rather than the audio filtering realm. I agree - the question was not off-topic at all so I up-voted it.
