Table of contents

Preface

Recently, when I was browsing Bilibili video, I accidentally saw a video case about converting video into character animation using Python. I found it very interesting. I searched for tutorials and cases on the entire Internet, but in the end I couldn't find a tutorial that suited my taste. All I combined some examples on the Internet to create a project that converts videos into character animations.

1. Analysis

As we all know, videos are made up of pictures one by one. So how to convert video to image? Next, I use ffmpeg to convert video to image, extract audio, and convert image to video.

What is FFmpeg :

FFmpeg is a set of open source computer programs that can be used to record, convert digital audio and video, and convert them into streams. Use LGPL or GPL license. It provides a complete solution for recording, converting and streaming audio and video. It contains the very advanced audio/video codec library libavcodec. In order to ensure high portability and codec quality, many codes in libavcodec were developed from scratch.

FFmpeg is developed under the Linux platform, but it can also be compiled and run in other operating system environments, including Windows, Mac OS X, etc. This project was first initiated by Fabrice Bellard and was mainly maintained by Michael Niedermayer from 2004 to 2015. Many FFmpeg developers come from the MPlayer project, and currently FFmpeg is also placed on the server of the MPlayer project team. The name of the project comes from the MPEG video coding standard, with the "FF" in front of it standing for "Fast Forward". The FFmpeg encoding library can use GPU acceleration.

----Baidu Encyclopedia

To put it bluntly, FFmpeg is a multimedia video processing tool, so we use Python to call the ffmpeg command to simplify our code.

ffmpeg resource download address:

地址：https://wwb.lanzouw.com/irLF3yi1cmh
密码:bnju

2. Flowchart

The flow chart for converting python videos into character animations is as follows:

3. Required third-party libraries

如何下载慢或报错超时后面加上镜像地址 -i https://pypi.douban.com/simple/
例如：pip install numpy==1.21.2 -i https://pypi.douban.com/simple/

pip install tqdm
pip install numpy==1.21.2
pip install Pillow==8.4.0

4. Code

Here is all the relevant code:

# -*- coding:utf-8 -*-
import os
import numpy as np
import tkinter as tk
from tqdm import tqdm
from time import sleep
from shutil import rmtree
from tkinter import filedialog
from PIL import Image, ImageDraw, ImageFont, ImageEnhance

SAMPLE_RATE = 0.4
OUT_FILE_NAME = "out"

# 获取当前路径
now_path = os.getcwd()


def main():
    """主函数"""
    input("*** 注意：过程中会创建许多图片或文件夹，请不要乱动或删除！*** （回车继续...） ")

    # 打开视屏路径
    try:
        filepath = get_video()
    except FileExistsError as k:
        print(k)
        sleep(2)
        return

    print("开始转换。。。")

    # 视频转换为图片
    video_to_photos(filepath)

    # 如果文件夹中含有其他音乐则删除
    now_music = [f for f in os.listdir(now_path) if f.split(".")[-1] == "mp3"]
    for m in now_music:
        os.remove(now_path + "/" + m)

    # 获取视频中的音乐
    get_music(filepath)

    # 获取当前文件夹下的所有图片文件
    photos = [f for f in os.listdir(now_path) if f.split(".")[-1] == "jpg"]

    # 获取当前文件夹下的音乐
    music_list = [f for f in os.listdir(now_path) if f.split(".")[-1] == "mp3"]
    if music_list:
        music = music_list[0]
    else:
        music = None

    # 如果没有out文件夹，则创建文件夹
    if not os.path.exists(OUT_FILE_NAME):
        os.mkdir(OUT_FILE_NAME)

    # 图片转字符图像
    print("\n转换中，过程中请勿关闭，请稍等...")
    for p in tqdm(photos):
        ascii_art(now_path + "/" + p, p.split(".")[0])

    # # 字符图像转视屏
    photos_to_video()

    # 如果当前文件夹中含有提取的音乐则将视频、音乐合并
    # 然后删除音乐
    if music:
        music_join_video()
        os.remove(now_path + "/" + music)

    # 删除所有图片保留视屏
    print("\n正在清理缓存，请稍等...")
    for p in photos:
        os.remove(now_path + "/" + p)
    os.remove(now_path + "/" + OUT_FILE_NAME + ".mp4")
    rmtree("./%s" % OUT_FILE_NAME, True)

    print("转换完成，请在output.mp4中查看转换的视屏！")
    sleep(2.3)


def ascii_art(file, save_name):
    """图片转字符图"""

    # 打开图片
    im = Image.open(file)

    # 调整图片亮度
    brightness = 2  # 增强亮度的值
    enh_bri = ImageEnhance.Brightness(im)
    im = enh_bri.enhance(brightness)

    # 保存的路径
    save_path = "%s/%s/%s%s" % (now_path, OUT_FILE_NAME, save_name, ".png")

    # 计算字母长宽比
    font = ImageFont.load_default()
    # font = ImageFont.truetype("SourceCodePro-Bold.ttf", size=12)
    aspect_ratio = font.getsize("x")[0] / font.getsize("x")[1]
    new_im_size = np.array(
        [im.size[0] * SAMPLE_RATE, im.size[1] * SAMPLE_RATE * aspect_ratio]
    ).astype(int)

    # 调整图像
    im = im.resize(new_im_size)

    # 判断是否为纯色图片，如果为纯色图片调整图像大小不进行字符处理
    r, g, b = im.getextrema()
    if r[0] == r[1] and g[0] == g[1] and b[0] == b[1]:
        letter_size = font.getsize("x")
        im_out_size = new_im_size * letter_size
        im_out = Image.new("RGB", tuple(im_out_size), im.getcolors()[0][1])
        im_out.save(save_path)
        return

    # 保留一个图像的副本以进行颜色采样
    im_color = np.array(im)

    # 转换为灰度的图像
    im = im.convert("L")

    # 转换为numpy数组以进行图像处理
    im = np.array(im)

    # 以升序定义将形成最终ascii的所有符号
    symbols = np.array(list(" .-vM"))

    # 规范化的最小值和最大值到 [0, max_symbol_index)
    im = (im - im.min()) / (im.max() - im.min()) * (symbols.size - 1)

    # 生成ascii样品
    ascii = symbols[im.astype(int)]

    # 创建一个用于绘制ascii文本的输出图像
    letter_size = font.getsize("x")
    im_out_size = new_im_size * letter_size
    bg_color = "black"
    im_out = Image.new("RGB", tuple(im_out_size), bg_color)
    draw = ImageDraw.Draw(im_out)

    # 绘制文本
    y = 0
    for i, line in enumerate(ascii):
        for j, ch in enumerate(line):
            color = tuple(im_color[i, j])  # 来自原始图像的样本颜色
            draw.text((letter_size[0] * j, y), ch[0], fill=color, font=font)
        y += letter_size[1]  # 按字母的高度增加y

    # 保存成图像文件
    im_out.save(save_path)


def get_video():
    """打开视屏路径"""
    print("请您选择要转换的视屏")

    # 打开选择文件对话框
    root = tk.Tk()
    root.withdraw()
    filepath = filedialog.askopenfilename()  # 获得选择好的文件

    file_extension = os.path.splitext(filepath)[1]

    # 如果不在格式内则报错
    if file_extension not in [".mp4", ".flv", ".avi", ".mpg", ".wmv"]:
        raise FileExistsError("目前只支持mp4, flv, avi, mpg, wmv视频格式")

    return filepath


def video_to_photos(filepath):
    """视频转换为图片"""
    # 调用ffmpeg命令实现视屏转为图片 fps=25
    command = "{}/tool/ffmpeg -i {} -r 25 -f image2 image-%1d.jpg".format(now_path, filepath)
    os.system(command)


def photos_to_video():
    """图像转视屏"""
    # 调用ffmpeg命令实现图片转为视屏
    command = "{0}/tool/ffmpeg -i {0}/{1}/image-%1d.png out.mp4".format(now_path, OUT_FILE_NAME)
    os.system(command)


def get_music(filepath):
    """获取视频中的音乐"""
    # 调用ffmpeg命令获取视频中的音乐
    command = "{}/tool/ffmpeg -i {} -f mp3 {}.mp3".format(now_path, filepath, OUT_FILE_NAME)
    os.system(command)


def music_join_video():
    """视频、音乐合并"""
    # 调用ffmpeg命令获取视频中的音乐
    command = "{0}/tool/ffmpeg -i {1}.mp3 -i {0}/out.mp4 {1}put.mp4".format(now_path, OUT_FILE_NAME)
    os.system(command)


if __name__ == "__main__":
    main()

5. Things to note

1. Only one program can be opened for each video converted.
2. Do not convert two videos at the same time
. 3. After the conversion is completed, please take away the converted video screen and start the program again.

4. Place the downloaded ffmpeg in the tool folder

6. Summary

The above Python method of converting videos into character animations is all the content shared by the editor today. I hope it can give you a reference. Thank you for watching. If you encounter any problems during the operation, you are also welcome to give me feedback. , if you see friends here and like it, please give it a follow ^_^, your support is my biggest motivation for updating!

Can python still play like this? Use Python to convert videos into character animations (including audio, high-definition, and color pictures) [source code attached]