Podcasts can be thought of as the recordings of people's ideas, interpretations, and knowledge. By definition, it only includes audio, yet, people are encouraged to publish their works on popular platforms like YouTube. This approach requires a kind of audio-visualization of the acoustic signal to intrigue attention. This tool is built in order to provide a tool to visualize audio.
- Python3
- ffmpeg
- Other necessary packages can be installed via
pip install -r requirements.txt
Note: requirements.txt includes opencv-python package. This should be noted that this package only includes fundamental functionalities of OpenCV library. Therefore, in case of need OpenCV should be installed separately.
KITT
|── sample_input
| |── A short part of a podcast
|── sample_output
| |── A short example output
|── kitt_effect.py
|── LICENSE
|── main.py
|── README.md
|── requirements.txt
After installing requirements, this repo can be tested by executing the main.py script. It will produce an example video and save it to the sample_output folder. An explanation of each line of main.py can be found below. FPS parameter can only take 10 and 15 values which are defined experimentally. It can be modified to the desired value after deciding the correct energy threshold levels. Also, an example output can be seen below. This footage is taken from one of the podcasts published on the official YouTube channel of TapirCast who is both the developer and user of this repository. Click on the example output to access original video which includes both visualization and audio (in Turkish) together.
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
This program takes an audio file and a background image as input and visualizes
audio. Visualization is inspired by KITT (Knight Industries Two Thousand) Car
which was a character in the Knight Rider series. For more information and
examples please visit https://www.github.com/tapirlab/kitt-effect
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Tapir Lab's KITT Effect Visualization Tool for Podcasts
%% -------------------
%% $Author: Halil Said Cankurtaran$,
%% $Date: November 1st, 2020$,
%% $Revision: 1.6$
%% $Tapir Lab.$
%% $Copyright: Halil Said Cankurtaran, Tapir Lab.$
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
"""
import os
import kitt_effect as ke
if __name__ == "__main__":
# Path to input audio
audio_file_name = os.path.join(".", "sample_input", "sample_input.mp3")
# Path to video background
background_image = os.path.join(".", "sample_input", "bg.png")
# Path to temporary video file which does not include audio
video_file_name = os.path.join(".", "sample_output", "output.mp4")
# Path to final output, which includes visualisation and audio
final_video_file = os.path.join(".", "sample_output", "final.mp4")
# Frames per second configuration of video
FPS = 10
# Following function loads an audio file, removes DC, and normalizes it.
record = ke.load_and_process_audio(audio_file_name)
# Calculate mean energy of each chunk which is partitioned based on FPS
energy_of_chunks = ke.calculate_energy_level_of_chunks(record, FPS)
# Determine number of rectangles to draw acc. to. mean energy level of chunk
number_of_rectangles = ke.get_number_of_rectangles(energy_of_chunks, FPS)
# Produce first video with KITT Effect
ke.process_video(number_of_rectangles, background_image, video_file_name, FPS)
# Merge video with audio and save to sample_output file
ke.produce_final_output(audio_file_name, video_file_name, final_video_file)The software is licensed under the MIT License.