ffmpeg实现视频和音频分离,并且将声音切片

项目需求

我司有个需求我们需要将视频中的音频提取出来,并且需要把提取出来的声音按照需要的时间段切片。

命令分析

1. 将视频中的音频提取出来
ffmpeg -i {file_path} -f wav -ar 16000 {file_name}
	file_path: 视频的文件路径
	file_name: 文件名称
	-ar: 设置音频采样频率。对于输出流,它默认设置为相应输入流的频率。对于输入流,此选项仅对音频抓取设备和原始解复用器有意义,并映射到相应的解复用器选项
	-i: 输入文件网址
	-f: 强制输入或输出文件格式。通常会自动检测输入文件的格式,并根据输出文件的文件扩展名猜测格式,因此在大多数情况下不需要此选项。
2. 将音频按照时间段切片
ffmpeg -i {file_path} -vn -acodec copy -ss {start_time} -to {end_time} -c copy {file_name}
	file_path: 视频的文件路径
	start_time: 开始时间
	end_time: 结束时间
	file_name: 文件名称
	-vn: 作为输入选项,阻止文件的所有视频流被过滤或自动选择或映射到任何输出
	-acodec: 设置音频编解码器。这是 的别名-codec:a
	-ss: 当用作输入选项(before -i)时,在这个输入文件中寻找 位置。请注意,在大多数格式中,无法精确查找,因此ffmpeg将查找到position之前最近的查找点
	-to: 停止在位置写入输出或读取输入, -to 和 -t 是互斥的,-t 有优先权
3. 将音频压缩
ffmpeg -i {file_path} -b:a 64k -acodec mp3 -ar 44100 -ac 1 {file_name}
	file_path: 视频的文件路径
	file_name: 文件名称
	-ac: 设置音频通道数。对于输出流,它默认设置为输入音频通道的数量。对于输入流,此选项仅对音频抓取设备和原始解复用器有意义,并映射到相应的解复用器选项。
	-ar:设置音频采样频率。对于输出流,它默认设置为相应输入流的频率。对于输入流,此选项仅对音频抓取设备和原始解复用器有意义,并映射到相应的解复用器选项
4.提升音频的声音
ffmpeg -i {file_path} -filter:a "volume=5dB" {file_name}
5.查看音频分贝
ffmpeg -i {file_path} -filter_complex volumedetect -c:v copy -f null /dev/null

代码

import os
import argparse


class Ffmpeg_tools:
    # ffmpeg -i ./file/4451511f2521cea6a78e8d0c7a5b9c03.mp4 -f wav -ar 16000 a.wav
    # ffmpeg -i ./wav_file/a.wav -b:a 64k -acodec mp3 -ar 44100 -ac 1 ./split_file/a.mp3
    # ffmpeg -i ./wav_file/a.wav -vn -acodec copy -ss 00:00:00 -t 00:01:30 1.wav
    # ffmpeg -i ./wav_file/BoobssoundMoanOrgasm…teenwav_file.wav -vn -acodec copy -ss 00:00:22 -to 00:00:33 -c copy ./split_file/BoobssoundMoanOrgasm…teensplit_file2.wav
    # ffmpeg -i ./split_file/PILLOWHUMPINGASMRthebestpleasuresoundssplit_file0.wav -filter:a "volume=5dB" output.wav 提升音频的声音
    # ffmpeg -i ./split_file/PILLOWHUMPINGASMRthebestpleasuresoundssplit_file0.wav -filter_complex volumedetect -c:v copy -f null /dev/null  查看音频分贝
    cmd_map = {
        "mv_to_audio": "ffmpeg -i {0} -f wav -ar 16000 {1}",
        "split_audio": "ffmpeg -i {0} -vn -acodec copy -ss {1} -to {2} -c copy {3}",
        "zip_to_mp3": "ffmpeg -i {0} -b:a 64k -acodec mp3 -ar 44100 -ac 1 {1}",
        "add_mp3_fb": "ffmpeg -i {0} -filter:a 'volume={1}dB' {2}",
        "select_fb_data": "ffmpeg -i {0} -filter_complex volumedetect -c:v copy -f null /dev/null"
    }
    
    
    def __init__(self, name, suffix, cmd_type, times = [], fb=5):
        self._name = name
        self._suffix = suffix
        self._file_path = "./file/{}".format(".".join([name,suffix]))
        self._wav_path = "./wav_file/{}".format(".".join([name+"wav_file",'wav']))
        self._cmd_type = cmd_type
        self._split_path = './split_file/{}'
        self._zip_file = "./zip_file/{}".format(".".join([name+"zip_file",'mp3']))
        self._times = times
        self.fb = fb
        
        
    def run(self):
        try:
            params = self.construction_params()
            cmds = self.construction_cmd(params=params)
            for _ in cmds:
                if self.excute_cmd(_):
                    print("成功")
                else:
                    print("失败")
                    break
        except Exception as e:
            print(e)
            
    def construction_cmd(self, params):
        cmd_strs = []
        try:
            for _ in params:
                cmd_strs.append(self.cmd_map.get(self._cmd_type).format(*_))
            return cmd_strs
        except Exception as e:
            print(e)
            return []
    
    def construction_params(self):
        params = []
        try:
            if self._cmd_type == 'mv_to_audio':
                params.append([self._file_path, self._wav_path])
            if self._cmd_type == 'split_audio':
                for _ in range(len(self._times)):
                    split_path = self._split_path.format('.'.join([self._name+"split_file"+str(_), 'wav']))
                    params.append([self._wav_path, self._times[_][0], self._times[_][1], split_path])
            
            if self._cmd_type == 'zip_to_mp3':
                params.append([self._file_path, self._zip_file])
            
            if self._cmd_type == 'add_mp3_fb':
                split_path = self._split_path.format('.'.join([self._name, 'wav']))
                params.append([split_path, self.fb, split_path])
            
            if self._cmd_type == 'select_fb_data':
                split_path = self._split_path.format('.'.join([self._name, 'wav']))
                params.append([split_path])
            
            return params
        except Exception as e:
            print(e)
            return params
    
    def excute_cmd(self, cmd):
        try:
            print(cmd)
            a = os.system(cmd)
            print(a)
            return True
        except Exception as e:
            print(e)
            return False


if __name__ == "__main__":
    # parser = argparse.ArgumentParser(description="mv_to_audio")
    # parser.add_argument('--old_file_path', help="原文件路径")
    # parser.add_argument('--new_file_path', help="新文件路径")
    
    
    
    # args = parser.parse_args()
    # print(args)
    split_times = {
        # "4451511f2521cea6a78e8d0c7a5b9c03" : [['00:00:03','00:00:32'],['00:00:34','00:01:37'],['00:01:43','00:02:45'],['00:03:01','00:03:57'],['00:04:01','00:04:51'],['00:04:53','00:05:17']],
        "PILLOWHUMPINGASMRthebestpleasuresounds": [['00:00:15','00:00:24'],['00:00:38','00:00:48'],['00:00:51','00:01:02'],['00:01:44','00:00:50'],['00:01:10','00:01:16'],['00:01:31','00:01:33'],['00:02:05','00:02:15'],['00:02:18','00:02:33'],['00:02:46','00:02:55'],['00:03:02','00:03:12'],['00:03:14','00:03:26'],['00:03:35','00:03:45'],['00:03:46','00:04:06']]

        # "345":[['00:00:01','00:00:07']]
    }
    for i,j,k in os.walk('./file'):
        for _ in k:
            
            file_names = _.split('.')
            name = file_names[0]
            suffix = file_names[1]
            Ffmpeg_tools(name, suffix, 'mv_to_audio').run()
            Ffmpeg_tools(name, suffix, 'split_audio', split_times[name]).run()
    
    # 运行下面两条命令的时候需要将 108-115 注释掉
    # 查看 音频分贝     
    # Ffmpeg_tools("PILLOWHUMPINGASMRthebestpleasuresoundssplit_file0", "wav", 'select_fb_data', split_times['PILLOWHUMPINGASMRthebestpleasuresounds']).run()
    # 调解音频分贝 默认5分贝 可以自己设定
    #Ffmpeg_tools("PILLOWHUMPINGASMRthebestpleasuresoundssplit_file0", "wav", 'add_mp3_fb', split_times['PILLOWHUMPINGASMRthebestpleasuresounds'],5).run()

你可能感兴趣的:(ffmpeg,音视频,语音识别,人工智能,ffmpeg)