wav to lip

lip sync using vowel phonetic detection

formant

For the purposes of distinguishing vowels from each other, we are more interested in the frequency response curves (indicating the preferred resonating frequencies of the vocal tract) rather than in the raw spectrum of the wave.

Each of the preferred resonanting frequencies of the vocal tract (each bump in the frequency response curve) is known as a formant . They are usually referred to as F1, F2, F3, etc.

for formant we use praat-parselmouth package

distinguishing vowels from each other

based on following image form umanitoba.ca we recognize vowel in audio

RUN

install requirements

pip install -r requirements.txt

change audio path in main.py

python main.py

For visualization:

python test.py