A plugin that corrects the vowel sounds in a vocal

It would be great if Waves could make a plugin that corrects mispronounced words in a vocal. I try to correct them by cutting the formant frequencies(F1, F2, F3, F4, and F5) of the wrongly mispronounced word and boosting the formant frequencies of the target/correct word using EQ plugins. For example, the word “beat” might sound as “bit”. What I would do is suppress the frequencies at 400, 2000, and 2550 Hz(the F1, F2, and F3 formant frequencies of the IPA vowel “ɪ”, respectively) and boost the frequencies at 270, 2300, and 3000 Hz(the F1, F2, and F3 formant frequencies of the IPA vowel “i”, respectively).

the second method I use is if the word is pronounced correctly somewhere else in the song, I try to match the EQ of the wrong word to the EQ the correct word using the EQ matching feature. However, this method usually changes the colour of the mispronounced word.

It would be great if the plugin Waves creates could dynamically track the wrong formants, suppress them, and dynamically boost the formant frequencies of the target word.

Matthew Winn from the University of Minnesota has made a formant continuum but I cannot add a link to his personal website. The article and the script can be found by Googling “make formant coninuum by Matthew Winn”.

Looking forward to seeing such a plugin.

1 Like

I think the plugin would go a long distance, world-wide, and super successful since millions of customer service agents(out sourced, toll-free customer service agents throughout the world who have thick or even slight accent), singers and broadcasters around the world whose mother tongue is not English and who are already singing or broadcasting, world leaders who want to make a speech, or anyone around the world who interacts with the Western World(the whole world speaks English nowadays). Infinite numbner of already recorded English language audios throughout the world could be converted into American, British, or other English accent. You might even make expand its functionality and make it possible to make the plugin correct any other language mispronunciation. For example, you could add German language pronunciation correction.

Even if Waves decides not to develop the plugin, could anyone from Waves please personally develop it even in the garage as a hobby. I beg you.

Sincerely,
Ermias

Sounds like a job for Neural Network (AI).

I know it’s the latest buzzword, but there are things that AI are good at and not so good at. Looking for patterns is one thing it can do well. Such as those found in speech or vocals.