python同时录制屏幕、摄像头、声音合成视频把人像放在最终视频右下角

功能描述：

同时录制屏幕图像和摄像头采集到的内容以及麦克风声音，把所有内容合成为最终视频文件，保证屏幕画面、摄像头画面以及声音的同步，把摄像头采集的人像缩小以后放在最终视频的右下角。

技术要点：

1. pillow进行屏幕截图
2. opencv根据屏幕截图生成视频
3. opencv采集摄像头视频
4. pyaudio录制麦克风声音
5. moviepy合成视频
6. 使用事件进行多线程同步

参考代码：

import wave
import threading
from os import remove,mkdir,listdir
from os.path import exists,splitext,basename,join
from datetime import datetime
from time import sleep
from shutil import rmtree
import pyaudio
from PIL import ImageGrab
from numpy import array

import cv2
from moviepy.editor import *

CHUNK_SIZE = 1024
CHANNELS = 2
FORMAT = pyaudio.paInt16
RATE = 48000
allowRecording = True

def record_audio():
    p = pyaudio.PyAudio()
    # 等待摄像头启动好 然后大家一起等3秒开始录制
    event.wait()
    sleep(3)
    #创建输入流
    stream = p.open(format = FORMAT,
            channels=CHANNELS,
            rate=RATE,
            input=True,
            frames_per_buffer=CHUNK_SIZE
    )
    wf = wave.open(audio_filename,'wb')
    wf.setnchannels(CHANNELS)
    wf.setsampwidth(p.get_sample_size(FORMAT))
    wf.serframerate(RATE)
    while allowRecording:
        # 从录音设备读取数据，直接写入wav文件
        data = stream.read(CHUNK_SIZE)
        wf.writeframes(data)
    wf.close()
    stream.stop_stream()
    stream.close()
    p.terminate()


def record_screen():
    #录制屏幕
    #等待摄像头启动好，然后大家一起等3秒开始录制
    event.wait()
    sleep(3)
    im = ImageGrab.grab()
    video = cv2.videoWrite(screen_video_filename,cv2.VideoWrite_fourcc(*'XVID'),25,im.size)
    while allowRecording:
        im = ImageGrab.grab()
        im = cv2.cvtColor(array(im),cv2.COLOR_RGB2BGR)
        video.write(im)
    video.release()

def record_webcam():
    # 参数0表示笔记本自带摄像头
    cap = cv2.VideoCapture(0)
    # 启动好摄像头，发出通知，大家一起等三秒开始录
    event.set()
    sleep(3)
    aviFile = cv2.VideoWrite(webcam_video_filename,cv2.VideoWriter_fourcc(*'MJPG'),25,(640,480))
    while allowRecording and cap.isOpened():
        #捕捉当前图像，ret=True表示成功，False 表示失败
        ret,frame = cap.read()
        if ret:
            aviFile.write(frame)
    aviFile.release()
    cap.release()
now = str(datetime.now())[:19].replace(':',"_")
audio_filename = f'{now}.mp3'
webcam_video_filename=f't{now}.avi'
screen_video_filename = f'tt{now}.avi'
video_filename = f'{now}.avi'
#创建两个线程，分别录音和录屏
t1 = threading.Thread(target=record_audio)
t2 = threading.Thread(target=record_screen)
t3 = threading.Thread(target=record_webcam)
#创建时间，用于多个线程同步，等待摄像头准备以后在一起等三秒开始录制
event = threading.Event()
event.clear()
for t in (t1,t2,t3):
    t.start()
#等待摄像头准备好，提示用户3秒钟以后开始录制
event.wait()
print("三秒后开始录制，按q键结束录制")
while True:
    if input()=='q':
        break
allowRecording = False
for t in (t1,t2,t3):
    t.join()
# 把录制的音频和屏幕截图合成为视频文件
audio  = AudioFileClip(audio_filename)
video1 = VideoFileClip(screen_video_filename)
ratio1 = audio.duration/video1.duration
video1 = (video1.fl_time(lamnda t: t/ratio1,apply_to=['video'])
        .set_end(audio.duration))
video2 = VideoFileClip(webcam_video_filename)
ratio2 = audio.duration/video2.duration
video2 = (video2.fl_time(lamnda t: t/ratio2,apply_to=['video'])
        .set_end(audio.duration)
        .resize((320,240))
        .set_position(('right','bottom')))
video = CompositeVideoClip([video1,video2]).set_audio(audio)
video.write_videofile(video_filename,codec='libx264',fps=25)
#删除临时音频文件和视频文件
remove(audio_filename)
remove(screen_video_filename)
remove(webcam_video_filename)