remove_silence.py

from pydub import AudioSegment
from pydub.silence import split_on_silence
import time

def remove_silence_add_delay(input_path, output_path, silence_thresh=-40, min_silence_len=2000, delay=200):
    # Load the separated vocal track
    vocal_track = AudioSegment.from_file(input_path)

    # Split the vocal track on silence with the specified threshold
    non_silent_vocals = split_on_silence(vocal_track, silence_thresh=silence_thresh, min_silence_len=min_silence_len)

    # Add silence delay after each non-silent segment
    result_segments = []
    for segment in non_silent_vocals:
        result_segments.append(segment)
        silence = AudioSegment.silent(duration=delay)
        result_segments.append(silence)

    # Remove the last added silence (since there's one extra at the end)
    result_segments.pop()

    # Concatenate the non-silent segments with silence delay
    result = sum(result_segments)

    # Export the result without silence and with added delay
    result.export(output_path, format="mp3")

timestamp = (
    lambda seconds: f"{int(seconds // 60)}m {round(seconds % 60, 2):.2f}s"
    if seconds >= 60
    else f"{round(seconds, 2):.2f}s"
    if seconds >= 1
    else f"{round(seconds * 1000):.0f}ms"
)

# Example usage
start_time = time.time()
remove_silence_add_delay(
    input_path="./Jee Le Zaraa (Talaash) [vocals].mp3",
    output_path="Jee Le Zaraa (Talaash) [vocals] (silence_removed_no_delay).mp3",
    silence_thresh=-40,
    min_silence_len=1500,
    delay=0
)

print(timestamp(time.time() - start_time))

separate_vocals.py

from spleeter.separator import Separator
from spleeter.audio import Codec

# Initialize Spleeter with the desired configuration
separator = Separator('spleeter:2stems')

# Provide the path to your input audio file
input_audio_path = 'Mujhe Peene Do - Darshan Raval ｜ Official Music Video ｜ Romantic Song 2020 ｜ Indie Music Label ｜.mp3'

# Specify the output path for the separated vocals
output_path = 'output'

# Separate vocals
separator.separate_to_file(input_audio_path, output_path, codec=Codec.MP3)