0. install

# download brew, and python
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
brew install python
brew install ffmpeg

#check your version of python and ffmpeg
python3 --version
# 
pip3 install -U openai-whisper
# check if it is installed
whisper --help

1. run whisper to detect the words

whisper /Users/yuting/Desktop/demo/testvideo.mp3

# you need to choose the lanuage, model size, and output location
# not really work here, try the whisperx
whisper /Users/yuting/Desktop/demo/testvideo.mp3 \
  --language de \
  --model small \
  --output_dir /Users/yuting/Desktop/demo/whisper_out
  

now you already have the txt files and other stuff.

later, we need to do it to get the forced-aligned files.

make sure you have wav files instead of mp3 files bc they are not good for whisper

then you can whether use whisperai or MFA to do that

2. forced-aligned

# always check the version in the beginning
ffmpeg -version
python3 --version
pip3 --version

# install whisperX
pip3 install -U whisperx

#check the version again
whisperx --help
# the forced-aligned
#pip install transcribe-align-textgri

# run the whisperx
whisperx "/Users/yuting/Desktop/Yuting's Ph.D. Project/Exp2/2. data/wav/PSI2_E_051_ch1.wav" \
  --language en \
  --model small \
  --align_model WAV2VEC2_ASR_LARGE_LV60K_960H \
  --output_dir "/Users/yuting/Desktop/Yuting's Ph.D. Project/Exp2/2. data/wav/whisper_out"

# then get the textgrid file
python -m transcribe_align_textgrid "/Users/yuting/Desktop/Yuting's Ph.D. Project/Exp2/2. data/wav/PSI2_E_050_ch1.wav" --language english

Here is the output

so you will get the world-level aligned textgrid files

3. To get the phoneme-level

you can whether use MFA or other tools