音频响度分析和获取

# -*- coding: utf-8 -*-
#/usr/local/bin/python3
# @time    : 2020/3/7 4:05 pm
# @author  : eric
# @file    : get_audio_loudness.py
# @software: pycharm

import os
import re
import subprocess
import sys
import csv


'''
说明：
  integrated loudness:
    i(整体响度):  -26.5 lufs  
    threshold: -37.1 lufs

  loudness range:
    lra(整体范围):        10.2 lu
    threshold: -49.5 lufs
    lra low:   -36.9 lufs
    lra high:  -26.6 lufs

mean_volume(平均音量): -32.2 db  
max_volume(最大音量): -18.9 db
'''
'''
:param '/users/yy/documents/test_noise/agc/android_new/test_saopin_32bit.wav'
:return {'i': '-26.5', 'lra': ' 10.2', 'threshold': '-49.5', 'lra_low': '-36.9', 'lra_high': '-26.6', 'mean_volume': '-32.2', 'max_volume': '-18.9'}
# 音频整体响度/音量获取方法，输出文件
'''
def get_audio_loudness_by_file(audiofile):
    lufs_i = r"\s+i:\s+([-+\s]\d+\.\d) lufs"
    lufs_threshold = r"\s+threshold:\s+([-+\s]\d+\.\d) lufs"
    lufs_lra = r"\s+lra:\s+([-+\s]\d+\.\d) lu"
    luss_lrs_low = r"\s+lra low:\s+([-+\s]\d+\.\d) lufs"
    luss_lrs_high = r"\s+lra high:\s+([-+\s]\d+\.\d) lufs"

    mean_volume = r"\[parsed_volumedetect_0 @ \w+\] mean_volume:\s(-\d+\.\d) db"
    max_volume = r"\[parsed_volumedetect_0 @ \w+\] max_volume:\s(-\d+\.\d) db"

    cmd_lufs = 'ffmpeg -nostats -i {0} -filter_complex ebur128 -f null -'
    cmd_db = 'ffmpeg -i {0} -af "volumedetect" -f null -'

    dict_loudnes = {}
    output = subprocess.getstatusoutput(cmd_lufs.format(audiofile))
    if output:
        lines = output[1].splitlines()
        end_match = 0
        for i in range(len(lines)):
            if 'target:-23 lufs' in lines[i]:
                for j in range(i+1, len(lines)):
                    line = lines[j]
                    if 'i:' in line:
                        match = re.search(lufs_i, line)
                        if match:
                            dict_loudnes['i(lufs)'] = match.group(1)
                    if 'threshold:' in line:
                        match = re.search(lufs_threshold, line)
                        if match:
                            dict_loudnes['threshold(lufs)'] = match.group(1)
                    if 'lra:' in line:
                        match = re.search(lufs_lra, line)
                        if match:
                            dict_loudnes['lra(lu)'] = match.group(1)
                    if 'lra low:' in line:
                        match = re.search(luss_lrs_low, line)
                        if match:
                            dict_loudnes['lra_low(lufs)'] = match.group(1)
                    if 'lra high' in line:
                        match = re.search(luss_lrs_high, line)
                        if match:
                            dict_loudnes['lra_high(lufs)'] = match.group(1)
                            end_match = 1
                            break
            if end_match:
                break

    output_db = subprocess.getstatusoutput(cmd_db.format(audiofile))
    if output_db:
        match = re.search(mean_volume, output_db[1])
        if match:
            dict_loudnes['mean_volume(db)'] = match.group(1)
        match = re.search(max_volume, output_db[1])
        if match:
            dict_loudnes['max_volume(db)'] = match.group(1)

    return dict_loudnes

'''
:param '/users/yy/documents/test_noise/agc/android_new'
:return write csv file
'''
def get_audio_loudness_by_path(file_path):
    audio_loudness_list = []
    file_path = os.path.abspath(file_path)
    for file in os.listdir(file_path):
        if file.endswith('.mp3') or file.endswith('wav'):
            dict_loudnes = get_audio_loudness_by_file(os.path.join(file_path, file))
            dict_loudnes['audio_name'] = file
            audio_loudness_list.append(dict_loudnes)
    print(audio_loudness_list)
    csvname = os.path.join(file_path, os.path.basename(file_path) + '.csv')
    print(csvname)
    with open(csvname, 'w', encoding='utf-8-sig') as f:
        writer = csv.dictwriter(f, fieldnames=['audio_name', 'i(lufs)', 'threshold(lufs)', 'lra(lu)', 'lra_low(lufs)', 'lra_high(lufs)', 'mean_volume(db)', 'max_volume(db)'])
        writer.writeheader()
        writer.writerows(audio_loudness_list)


'''
:
:return 瞬时响度(由m标识)、短期响度(s)、综合响度(i)和响度范围(lra)
[parsed_ebur128_0 @ 0x7ff27302c900] t: 0.299979   target:-23 lufs    m:-120.7 s:-120.7     i: -70.0 lufs       lra:   0.0 lu
# 音频每100ms的瞬时响度(由m标识)、短期响度(s)、综合响度(i)和响度范围(lra)

'''

def get_audio_detail_loudness_by_file(audio_file):
    ebur = r'\[parsed_ebur128_0 @ \w+\]\s+t:\s+(\d+[\.\d+]*)\s+target:(-\d+)\slufs\s+m:([-\s]*\d+\.\d)\s+s:([-\s]*\d+\.\d)\s+i:\s+([-\s]*\d+\.\d)\slufs\s+lra:\s+(\d+\.\d)\slu'
    cmd_lufs = 'ffmpeg -nostats -i {0} -filter_complex ebur128 -f null -'
    cmd = cmd_lufs.format(audio_file)
    # print(cmd)
    detail_list = []
    output = subprocess.getstatusoutput(cmd)
    if output:
        lines = output[1].splitlines()
        for line in lines:
            match = re.search(ebur, line)
            dict_loudnes = {}
            if match:
                dict_loudnes['time'] = match.group(1)
                dict_loudnes['target(lufs)'] = match.group(2)
                dict_loudnes['m(lufs)'] = match.group(3)
                dict_loudnes['s(lufs)'] = match.group(4)
                dict_loudnes['i(lufs)'] = match.group(5)
                dict_loudnes['lra(lu)'] = match.group(6)
                detail_list.append(dict_loudnes)
    if len(detail_list) > 1:
        csvname = audio_file + '.csv'
        print(csvname)
        dict_loudness = get_audio_loudness_by_file(audio_file)
        info_str = 'agv_i={1}{0} lra_low={2}{0} lra_high={3}{0}'.format('lufs', dict_loudness['i(lufs)'], dict_loudness['lra_low(lufs)'], dict_loudness['lra_high(lufs)'])
        info = [info_str]
        print(dict_loudness)
        with open(csvname, 'w', encoding='utf-8-sig') as f:
            writer = csv.writer(f)
            writer.writerow(info)
            writer = csv.dictwriter(f, fieldnames=['time', 'target(lufs)', 'm(lufs)', 's(lufs)', 'i(lufs)',
                                                   'lra(lu)'])
            writer.writeheader()
            writer.writerows(detail_list)


def get_audio_detail_loudness_by_path(file_path):
    file_path = os.path.abspath(file_path)
    for root, dirs, files in os.walk(file_path):
        for file in files:
            if file.endswith('.mp3') or file.endswith('wav'):
                print(os.path.join(root, file))
                get_audio_detail_loudness_by_file(os.path.join(root, file))


if __name__ == "__main__":
    print('xx')
    test_file = '/users/yy/documents/test_noise/agc/android_new/test_saopin_32bit.wav'
    get_audio_detail_loudness_by_file(test_file)
    if len(sys.argv) > 1:
        temp = sys.argv[1]
        if os.path.isdir(temp):
            print('path', temp)
            get_audio_loudness_by_path(temp)
            get_audio_detail_loudness_by_path(temp)
        elif os.path.isfile(temp):
            print('file', temp)
            get_audio_loudness_by_file(temp)
        else:
            print(temp, 'is not exists')
    else:
        print('param need path or file')