$ ffmpeg [全局选项] {[输入文件选项] -i 输入_url_地址} {[输出文件选项] 输出_url_地址} ...

$ ffmpeg -i video.mp4 video.avi

$ ffmpeg -i input.mp4 -vn output.mp3

import pathlib
import pydub
from pydub import AudioSegment
from pydub.utils import mediainfo
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
mp3_path = path.joinpath('1.mp3')
out_path = path.joinpath('007audio_pydub_export.mp3')
# 默认采样率44100Hz
snd = AudioSegment.from_mp3(mp3_path)
info = mediainfo(mp3_path)
print(info) # ID3信息
print(snd.duration_seconds, snd.frame_rate) # 音频时长，采样率
snd.export(out_path, format='mp3', bitrate='32k') # 转成mp3格式

import pathlib
from pydub import AudioSegment
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
mp3_path = path.joinpath('1.mp3')
cover_path = path.joinpath('cover.jpg')
snd = AudioSegment.from_mp3(mp3_path)
# 获取时长、分贝数、采样率
print(snd.duration_seconds, snd.dBFS, snd.frame_rate)
# 剪30秒核心片段，单位为毫秒
snd_mid = snd[120000:150000]
# 调大音量，单位分贝
snd_mid += 6
# 剪最后30秒片段
snd_end = snd[-30000:]
# 淡入淡出效果
snd_final = snd_mid.append(snd_end, crossfade=1500)
# 重复一次
snd_final *= 2
# 淡入淡出效果
snd_final.fade_in(2000).fade_out(2000)
# 加上原始信息
tags={'artist': '程一初', 'album': '只差一个程序员了', 'comments': 'Come to Python1024!'}
# 导出文件
snd_final.export(path.joinpath('007audio_pydub_cut.mp3'), format='mp3', tags=tags, cover=str(cover_path))

import pathlib
from pydub import AudioSegment
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
# 提取音频
mp4_path = path.joinpath('2.mp4')
cover_path = path.joinpath('cover.jpg')
snd = AudioSegment.from_file(mp4_path)
tags={'artist': '程一初', 'album': '只差一个程序员了', 'comments': 'Come to Python1024!'}
snd.export(path.joinpath('007audio_pydub_from_video.mp3'), format='mp3', tags=tags, cover=str(cover_path))
# 提取背景音乐
mp4_path = path.joinpath('3.mp4')
snd = AudioSegment.from_file(mp4_path)
snd_l, snd_r = snd.split_to_mono()
snd_r_inv = snd_r.invert_phase() # 反相
bg_music = snd_l.overlay(snd_r_inv) # 覆盖后获得背景音乐，但有噪音
bg_music.export(path.joinpath('007audio_pydub_split_music.mp3'), format='mp3')

$ ffmpeg -i 3.mp3 -af pan="stereo|c0=c0|c1=-1*c1" -ac 1 007audio_ffmpeg_bgmusic.mp3

$ spleeter separate -i input.mp3 -p spleeter:2stems -o output

import pathlib
from spleeter.separator import Separator
from spleeter.audio.adapter import get_default_audio_adapter
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
mp3_path = path.joinpath('3.mp3')
out_path = path.joinpath('007audio_spleeter_out')
separator = Separator('spleeter:2stems')
audio_adapter = get_default_audio_adapter()
separator.separate_to_file(
            str(mp3_path),
            out_path,
            audio_adapter=audio_adapter,
            synchronous=False)

import pathlib
from pychorus import find_and_output_chorus
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
mp3_path = path.joinpath('1.mp3')
out_path = path.joinpath('007audio_chorus_out.wav')
chorus_start_sec = find_and_output_chorus(mp3_path, out_path, 30)
print(chorus_start_sec)

import pathlib
import time
import json
from pydub import AudioSegment
import oss2
from aliyunsdkcore.client import AcsClient
from aliyunsdkcore.request import CommonRequest
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
mp3_path = path.joinpath('3.mp3')
out_path = path.joinpath('3_16k.wav')
out_txt_path = path.joinpath('007audio_stt.txt')
ACK_ID = '<你的access_key_id>'
ACK_SEC= '<你的access_key_secret>'
APP_KEY= '<你的app_key>'
bucket_path = pathlib.Path('https://python1024sh.oss-cn-shanghai.aliyuncs.com')
# 把文件转为16K采样率的mp3，只用一个声道
audio = AudioSegment.from_mp3(mp3_path)
mono = audio.set_frame_rate(16000).set_channels(1)
mono.export(out_path, format='wav', codec='pcm_s16le')
auth = oss2.Auth(ACK_ID, ACK_SEC) # 鉴权
bucket = oss2.Bucket(auth, 'https://oss-cn-shanghai.aliyuncs.com', 'python1024sh')
obj_name = f'demo/{file_path.name}'
res = bucket.put_object_from_file(obj_name, out_path)
file_url = f'https://python1024sh.oss-cn-shanghai.aliyuncs.com/{obj_name}' # 公网路径
print(file_url)

client = AcsClient(ACK_ID, ACK_SEC, 'cn-shanghai')
post_req = CommonRequest()
post_req.set_domain('filetrans.cn-shanghai.aliyuncs.com')
post_req.set_version('2018-08-17')
post_req.set_product('nls-filetrans')
post_req.set_action_name('SubmitTask')
post_req.set_method('POST')
task = {'appkey': APP_KEY, 'file_link': file_url, 'version': '4.0', 'enable_words': False}
print(json.dumps(task))
post_req.add_body_params('Task', json.dumps(task))
task_id = ''
try:
    post_res = client.do_action_with_exception(post_req)
    post_res = json.loads(post_res)
    status_txt = post_res['StatusText']
    if status_txt == 'SUCCESS':
        task_id = post_res['TaskId']
        print (f'录音文件识别请求成功响应，task_id: {task_id}')
    else:
        print(f'录音文件识别请求失败: {status_txt}')
except Exception as e:
    print(e)

get_req = CommonRequest()
get_req.set_domain('filetrans.cn-shanghai.aliyuncs.com')
get_req.set_version('2018-08-17')
get_req.set_product('nls-filetrans')
get_req.set_action_name('GetTaskResult')
get_req.set_method('GET')
get_req.add_query_param('TaskId', task_id)
status_txt = ''
while True:
    try:
        get_res = client.do_action_with_exception(get_req)
        get_res = json.loads(get_res)
        status_txt = get_res['StatusText']
        if status_txt in ['RUNNING', 'QUEUEING']:
            time.sleep(10) # 避免请求过于频繁
        else:
            break
    except Exception as e:
        print(e)
if status_txt == 'SUCCESS':
    result = get_res["Result"]
    sentences = result["Sentences"]
    txt_list = [ s['Text'] for s in sentences]
    with open(out_txt_path, 'a') as f:
        f.writelines('\n'.join(txt_list))
else:
    print(f'录音文件识别失败, {status_txt}')

import pathlib
import json
from aliyunsdkcore.client import AcsClient
from aliyunsdkcore.request import CommonRequest
import ali_speech
from ali_speech.callbacks import SpeechSynthesizerCallback
from ali_speech.constant import TTSFormat
from ali_speech.constant import TTSSampleRate
path = list(pathlib.Path.cwd().parents)[1].joinpath('data/automate/007audio')
out_path = path.joinpath('007audio_stt.wav')
ACK_ID = '<你的access_key_id>'
ACK_SEC= '<你的access_key_secret>'
APP_KEY= '<你的app_key>'
# 开始获取Token
acs_client = AcsClient(ACK_ID, ACK_SEC, 'cn-shanghai')
req = CommonRequest()
req.set_method('POST')
req.set_domain('nls-meta.cn-shanghai.aliyuncs.com')
req.set_version('2019-02-28')
req.set_action_name('CreateToken')
res = acs_client.do_action_with_exception(req)
res = json.loads(res)
token = res['Token']['Id']
class AudioCallback(SpeechSynthesizerCallback):
    # 回调函数，接受服务端数据返回
    def __init__(self, name):
        self._name = name
        self._fout = open(name, 'wb')
    def on_binary_data_received(self, raw):
        self._fout.write(raw)
    def on_completed(self, message):
        print('Completed: %s' % message)
        self._fout.close()
    def on_task_failed(self, message):
        print('Failed，task_id:%s, status_text:%s' % (
            message['header']['task_id'], message['header']['status_text']))
        self._fout.close()
callback = AudioCallback(out_path)
# 开始发送任务请求
text = "程一初发表的文章，在公众号“只差一个程序员了”。"
client = ali_speech.NlsClient()
client.set_log_level('INFO')
synthesizer = client.create_synthesizer(callback)
synthesizer.set_appkey(APP_KEY)
synthesizer.set_token(token)
synthesizer.set_voice('Siqi')
synthesizer.set_text(text)
synthesizer.set_format(TTSFormat.WAV)
synthesizer.set_sample_rate(TTSSampleRate.SAMPLE_RATE_16K)
synthesizer.set_volume(50)
synthesizer.set_speech_rate(0)
synthesizer.set_pitch_rate(0)
try:
    ret = synthesizer.start()
    if ret < 0:
        print(ret)
    else:
        synthesizer.wait_completed()
except Exception as e:
    print(e)
finally:
    synthesizer.close()