Audio profanity filter using python

Question

I am trying to write a python code in which users inputs a video file and program will mute/beep the curse/bad words in it and outputs a filtered video file, basically a profanity filter. Firstly, I converted the video file into .wav format and then trying to apply audio profanity filter into the .wav file and will write that .wav file into the video.

So far, I am able to make chunks of audio file and extract text from each audio chunk of 5 secs using speech_recognition library. But, if words are overlapped between the chunks I will not be able to detect detect them and apply the check condition if that chunk text is found in curse words list and will reduce the db of that audio file which will mute it (Suggest other way to make beep sound instead of muting it).

I am confused if my approach is right. I just want a audio profanity filter using python, up-till now I am able to extract text and make chunks only.

import speech_recognition as sr 
import os 
from pydub.utils import make_chunks
from pydub import AudioSegment
from pydub.playback import play
import codecs
import re
import fileinput
list=[]
#Curse words list
with codecs.open("words_final.txt", "r") as f0:
    sentences_lines=f0.read().splitlines()
    for sentences in sentences_lines:
        list.append(sentences)
# print(list)

# create a speech recognition object
r = sr.Recognizer()

# a function that splits the audio file into chunks
# and applies speech recognition
def get_large_audio_transcription(path):
    """
    Splitting the large audio file into chunks
    and apply speech recognition on each of these chunks
    """
    # open the audio file using pydub
    sound = AudioSegment.from_wav(path)  
    chunk_length_ms = 5000 # pydub calculates in millisec
    chunks = make_chunks(sound, chunk_length_ms) #Make chunks of one sec
    folder_name = "audio-chunks"
    # create a directory to store the audio chunks
    if not os.path.isdir(folder_name):
        os.mkdir(folder_name)
    whole_text = ""
    # process each chunk 
    for i, audio_chunk in enumerate(chunks, start=1):
        # export audio chunk and save it in
        # the `folder_name` directory.
        chunk_filename = os.path.join(folder_name, f"chunk{i}.wav")
        audio_chunk.export(chunk_filename, format="wav")
        # recognize the chunk
        with sr.AudioFile(chunk_filename) as source:
            audio_listened = r.record(source)
            # try converting it to text
            try:
                text = r.recognize_google(audio_listened,language="en-US")
                wav_file=AudioSegment.from_file(chunk_filename, format = "wav")
                # Reducing volume by 5
                silent_wav_file = AudioSegment.silent(duration=8000)
                #  Playing silent file
                play(silent_wav_file)
            except sr.UnknownValueError as e:
                print("Error:", str(e))
            else:
                text = f"{text.capitalize()}. "
                print(chunk_filename, ":", text)
                whole_text += text
    # return the text for all chunks detected
    return whole_text

path = "Welcome.wav"
print("\nFull text:", get_large_audio_transcription(path))

#Will implement a loop to sum all chunks a make a final filtered .wav file

Answer 1

First extract audio.wav from the video file. Then using this solution get the timestamps for each word from that audio. You can then mute/beep the bad words from the audio and join it back to the video.