Recherche avancée

Sur d’autres sites (13903)

• FFMPEG loudnorm filter does not work in combination with silenceremove filter

  12 mai 2021, par MareikeP

  I want to consistently normalize audio files for TTS model training. The output audio files should meet the following criteria :

  


  

  1. mono channel

  


  2. sample rate of 22050 Hz

  


  3. wav format

  


  4. no silence at beginning and end of audio clip

  


  5. volume of -24 dB

  


  


  I have already fulfilled the first 4 criteria. So far, it works properly.

  


  Normalizing the volume basically works as well with this ffmpeg command -af loudnorm=I=-24:LRA=11:TP=-1.5 , but not in combination with the silence removal : As soon as I remove silence with this ffmpeg command agate=threshold=0.045:attack=0.5:release=500:ratio=5000,silenceremove=start_periods=1:start_threshold=0.0075,areverse,silenceremove=start_periods=1:start_threshold=0.0075,areverse, the loudness normalization does not work any longer : the output volume now varies between -25dB and -32dB instead of the desired -24 dB.

  


  This is the complete ffmpeg command I used :

  


  ffmpeg -i filename.flac -ac 1 -af agate=threshold=0.045:attack=0.5:release=500:ratio=5000,silenceremove=start_periods=1:start_threshold=0.0075,areverse,silenceremove=start_periods=1:start_threshold=0.0075,areverse,loudnorm=I=-24:LRA=11:TP=-1.5,aresample=22050 -y -hide_banner filename.wav


  


  And this is the piece of code that I'm using to run it :

  


  import os

INPUT_DIR = '/home/username/all_data'
OUTPUT_DIR = '/home/username/normalized_data'
for filename in os.listdir(INPUT_DIR):
    wav_filename = filename[:-5] + '.wav'
    command = (f'ffmpeg -i {INPUT_DIR}/{filename} -ac 1 -af agate='
               f'threshold=0.045:attack=0.5:release=500:ratio=5000,'
               f'silenceremove=start_periods=1:start_threshold=0.0075,'
               f'areverse,silenceremove=start_periods=1:start_threshold='
               f'0.0075,areverse,loudnorm=I=-24:LRA=11:TP=-1.5,aresample'
               f'=22050 -y -hide_banner {OUTPUT_DIR}/{wav_filename}')
    os.system(command)


  


  EDIT :

  


  A complete log from the ffmpeg command can be seen here :

  


  username@pop-os:~$ ffmpeg -i /home/username/audios/filename.flac -ac 1 -af agate=threshold=0.045:attack=0.5:release=500:ratio=5000,silenceremove=start_periods=1:start_threshold=0.0075,areverse,silenceremove=start_periods=1:start_threshold=0.0075,areverse,loudnorm=I=-24:LRA=11:TP=-1.5,aresample=22050 /home/username/result.wav
ffmpeg version 4.2.4-1ubuntu0.1 Copyright (c) 2000-2020 the FFmpeg developers
  built with gcc 9 (Ubuntu 9.3.0-10ubuntu2)
  configuration: --prefix=/usr --extra-version=1ubuntu0.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-avisynth --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-nvenc --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 31.100 / 56. 31.100
  libavcodec     58. 54.100 / 58. 54.100
  libavformat    58. 29.100 / 58. 29.100
  libavdevice    58.  8.100 / 58.  8.100
  libavfilter     7. 57.100 /  7. 57.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  5.100 /  5.  5.100
  libswresample   3.  5.100 /  3.  5.100
  libpostproc    55.  5.100 / 55.  5.100
Input #0, flac, from '/home/mareike/tts_data/save/audios_flac/0a6c8520-7536-11eb-8338-b7015f354987.flac':
  Duration: 00:00:04.64, start: 0.000000, bitrate: 1090 kb/s
    Stream #0:0: Audio: flac, 44100 Hz, stereo, s32 (24 bit)
Stream mapping:
  Stream #0:0 -> #0:0 (flac (native) -> pcm_s16le (native))
Press [q] to stop, [?] for help
Output #0, wav, to '/home/mareike/result_0a6c8520-7536-11eb-8338-b7015f354987.wav':
  Metadata:
    ISFT            : Lavf58.29.100
    Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 22050 Hz, mono, s16, 352 kb/s
    Metadata:
      encoder         : Lavc58.54.100 pcm_s16le
size=     138kB time=00:00:03.19 bitrate= 353.0kbits/s speed=14.3x    
video:0kB audio:138kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.055375%


  


  Can anyone tell me what I'm doing wrong and how I can finally get the volume normalized to -24 dB (in combination with silence removal) ? Any help is appreciated, thank you very much !

  


• Bad output from ffmpeg running on raw video from Wear OS device

  17 juin 2020, par lcj

  I am following these instructions regarding capturing and converting video from a Wear OS device, and I am running into problems. I get a video and I can tell it is my screen but it seems to be warped. This should be a green check with some white text.

  



  

  



  There is a lot of noise when I run this :

  



  ffmpeg -f rawvideo -vcodec rawvideo -s 400x400 -pix_fmt rgb24 -r 10 -i video.raw -an -c:v libx264 -pix_fmt yuv420p video.mp4


  



  Here

  



  ffmpeg version 4.3 Copyright (c) 2000-2020 the FFmpeg developers
  built with Apple clang version 11.0.3 (clang-1103.0.32.62)
  configuration: --prefix=/usr/local/Cellar/ffmpeg/4.3 --enable-shared --enable-pthreads --enable-version3 --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-ffplay --enable-gnutls --enable-gpl --enable-libaom --enable-libbluray --enable-libdav1d --enable-libmp3lame --enable-libopus --enable-librubberband --enable-libsnappy --enable-libsrt --enable-libtesseract --enable-libtheora --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxvid --enable-lzma --enable-libfontconfig --enable-libfreetype --enable-frei0r --enable-libass --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-librtmp --enable-libspeex --enable-libsoxr --enable-videotoolbox --disable-libjack --disable-indev=jack
  libavutil      56. 51.100 / 56. 51.100
  libavcodec     58. 91.100 / 58. 91.100
  libavformat    58. 45.100 / 58. 45.100
  libavdevice    58. 10.100 / 58. 10.100
  libavfilter     7. 85.100 /  7. 85.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  7.100 /  5.  7.100
  libswresample   3.  7.100 /  3.  7.100
  libpostproc    55.  7.100 / 55.  7.100
[rawvideo @ 0x7fbcec008200] Estimating duration from bitrate, this may be inaccurate
Input #0, rawvideo, from 'video.raw':
  Duration: 00:00:07.00, start: 0.000000, bitrate: 38566 kb/s
    Stream #0:0: Video: rawvideo (RGB[24] / 0x18424752), rgb24, 400x400, 38400 kb/s, 10 tbr, 10 tbn, 10 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (rawvideo (native) -> h264 (libx264))
Press [q] to stop, [?] for help
[libx264 @ 0x7fbce980d400] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
[libx264 @ 0x7fbce980d400] profile High, level 2.1, 4:2:0, 8-bit
[libx264 @ 0x7fbce980d400] 264 - core 159 r2999 296494a - H.264/MPEG-4 AVC codec - Copyleft 2003-2020 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=12 lookahead_threads=2 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=10 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'video.mp4':
  Metadata:
    encoder         : Lavf58.45.100
    Stream #0:0: Video: h264 (libx264) (avc1 / 0x31637661), yuv420p, 400x400, q=-1--1, 10 fps, 10240 tbn, 10 tbc
    Metadata:
      encoder         : Lavc58.91.100 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/0 buffer size: 0 vbv_delay: N/A
[rawvideo @ 0x7fbcec008200] Packet corrupt (stream = 0, dts = 70).
video.raw: corrupt input packet in stream 0
[rawvideo @ 0x7fbce980c800] Invalid buffer size, packet size 145920 < expected frame_size 480000
Error while decoding stream #0:0: Invalid argument
frame=   70 fps=0.0 q=-1.0 Lsize=      90kB time=00:00:06.70 bitrate= 110.3kbits/s speed=46.5x    
video:89kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 1.788509%
[libx264 @ 0x7fbce980d400] frame I:4     Avg QP:15.11  size:  8077
[libx264 @ 0x7fbce980d400] frame P:26    Avg QP:12.39  size:  2159
[libx264 @ 0x7fbce980d400] frame B:40    Avg QP: 9.83  size:    40
[libx264 @ 0x7fbce980d400] consecutive B-frames: 21.4%  5.7%  4.3% 68.6%
[libx264 @ 0x7fbce980d400] mb I  I16..4: 53.0%  3.3% 43.6%
[libx264 @ 0x7fbce980d400] mb P  I16..4:  9.2%  2.0% 11.1%  P16..4:  2.4%  1.3%  0.3%  0.0%  0.0%    skip:73.7%
[libx264 @ 0x7fbce980d400] mb B  I16..4:  0.0%  0.0%  0.0%  B16..8:  1.5%  0.0%  0.0%  direct: 0.1%  skip:98.4%  L0:24.1% L1:75.9% BI: 0.0%
[libx264 @ 0x7fbce980d400] 8x8 transform intra:6.5% inter:9.8%
[libx264 @ 0x7fbce980d400] coded y,uvDC,uvAC intra: 26.1% 44.1% 43.9% inter: 0.3% 0.8% 0.7%
[libx264 @ 0x7fbce980d400] i16 v,h,dc,p: 89%  5%  5%  0%
[libx264 @ 0x7fbce980d400] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 10% 22% 67%  0%  0%  0%  0%  0%  1%
[libx264 @ 0x7fbce980d400] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 13% 69% 10%  0%  1%  0%  5%  0%  1%
[libx264 @ 0x7fbce980d400] i8c dc,h,v,p: 59% 41%  0%  0%
[libx264 @ 0x7fbce980d400] Weighted P-Frames: Y:30.8% UV:30.8%
[libx264 @ 0x7fbce980d400] ref P L0: 83.0%  3.1% 13.0%  0.5%  0.5%
[libx264 @ 0x7fbce980d400] kb/s:102.92


  



  Any ideas ?

  


• Consent Mode v2 : Everything You Need to Know

  7 mai 2024, par Alex — Analytics Tips

  Confused about Consent Mode v2 and its impact on your website analytics ? You’re not the only one. 

  Google’s latest update has left many scratching their heads about data privacy and tracking. 

  In this blog, we’re getting straight to the point. We’ll break down what Consent Mode v2 is, how it works, and the impact it has.

  What is Consent Mode ?

  What exaclty is Google Consent Mode and why is there so much buzz surrounding it ? This question has been frustrating analysts and marketers worldwide since the beginning of this year. 

  Consent Mode is the solution from Google designed to manage data collection on websites in accordance with user privacy requirements.

  This mode enables website owners to customise how Google tags respond to users’ consent status for cookie usage. At its core, Consent Mode adheres to privacy regulations such as GDPR in Europe and CCPA in California, without significant loss of analytical data.

  

  How does Consent Mode work ?

  Consent Mode operates by adjusting the behaviour of tags on a website depending on whether consent for cookie usage is provided or not. If a user does not consent to the use of analytical or advertising cookies, Google tags automatically switch to collecting a limited amount of data, ensuring privacy compliance.

  This approach allows for continued valuable insights into website traffic and user behavior, even if users opt out of most tracking cookies.

  What types of consent are available in Consent Mode ?

  As of 6 March 2024, Consent Mode v2 has become the current standard (and in terms of utilising Google Advertising Services, practically mandatory), indicating the incorporation of four consent types :

  1. ad_storage : allows for the collection and storage of data necessary for delivering personalised ads based on user actions.
  2. ad_user_data : pertains to the collection and usage of data that can be associated with the user for ad customisation and optimisation.
  3. ad_personalization : permits the use of user data for ad personalisation and providing more relevant content.
  4. analytics_storage : relates to the collection and storage of data for analytics, enabling websites to analyse user behaviour and enhance user experience.

  Additionally, in Consent Mode v2, there are two modes :

  1. Basic Consent Mode : in which Google tags are not used for personalised advertising and measurements if consent is not obtained.
  2. Advanced Consent Mode : allows Google tags to utilise anonymised data for personalised advertising campaigns and measurements, even if consent is not obtained.

  What is Consent Mode v2 ? (And how does it differ from Consent Mode v1 ?)

  Consent Mode v2 is an improved version of the original Consent Mode, offering enhanced customisation capabilities and better compliance with privacy requirements. 

  The new version introduces additional consent configuration parameters, allowing for even more precise control over which data is collected and how it’s used. The key difference between Consent Mode v2 and Consent Mode v1 lies in more granular consent management, making this tool even more flexible and powerful in safeguarding personal data.

  In Consent Mode v2, the existing markers (ad_storage and analytics_storage) are accompanied by two new markers :

  1. ad_user_data – does the user agree to their personal data being utilized for advertising purposes ?
  2. ad_personalization – does the user agree to their data being employed for remarketing ?

  In contrast to ad_storage and analytics_storage, these markers don’t directly affect how the tags operate on the site itself. 

  They serve as additional directives sent alongside the pings to Google services, indicating how user data can be utilised for advertising purposes.

  While ad_storage and analytics_storage serve as upstream qualifiers for data (determining which identifiers are sent with the pings), ad_user_data and ad_personalization serve as downstream instructions for Google services regarding data processing.

  How is the implementation of Consent Mode v2 going ?

  The implementation of Consent Mode v2 is encountering some issues and bugs (as expected). The most important thing to understand :

  1. Advanced Consent Mode v2 is essential if you have traffic and campaigns with Google Ads in the European Union.
  2. If you don’t have substantially large traffic, enabling Advanced Consent Mode v2 will likely result in a traffic drop in GA4 – because this version of consent mode (unlike the basic one) applies behavioural modelling to users who haven’t accepted the use of cookies. And modelling the behaviour requires time.

  The aspect of behavioural modelling in Consent Mode v2 implies the following : the data of users who have declined tracking options begin to be modelled using machine learning. 

  However, training the model requires a suitable data volume. As the Google’s documentation states :

  The property should collect at least 1,000 events per day with analytics_storage=’denied’ for at least 7 days. The property should have at least 1,000 daily users submitting events with analytics_storage=’granted’ for at least 7 of the previous 28 days.

  Largely due to this, the market’s response to the Consent Mode v2 implementation was mixed : many reported a significant drop in traffic in their GA4 and Google Ads reports upon enabling the Advanced mode. Essentially, a portion of the data was lost because Google’s models lacked enough data for training. 

  And from the very beginning of implementation, users regularly report about a few examples of that scenario. If your website doesn’t have enough traffic for behaviour modelling, after Consent Mode v2 switching you will face significant drop in your traffic in Google Ads and GA4 reports. There are a lot of cases of observing 90-95% drop in metrics of users and sessions.

  In a nutshell, you should be prepared for significant data losses if you are planning to switch to Google Consent Mode v2.

  How does Consent Mode v2 impact web analytics ? 

  The transition to Consent Mode v2 alters the methods of user data collection and processing. The main concerns arise from the potential loss of accuracy and completeness of analytical data due to restrictions on the use of cookies and other identifiers when user consent is absent. 

  With Google Consent Mode v2, the data of visitors who have not agreed to tracking will be modelled and may not accurately reflect your actual visitors’ behaviours and actions. So as an analyst or marketer, you will not have true insights into these visitors and the data acquired will be more generalised and less accurate.

  Google Consent Mode v2 appears to be a kind of compromise band-aid solution. 

  It tries to solve these issues by using data modelling and anonymised data collection. However, it’s critical to note that there are specific limitations inherent to the modelling mechanism.

  This complicates the analysis of visitor behavior, advertising campaigns, and website optimisation, ultimately impacting decision-making and resulting in poor website performance and marketing outcomes.

  Wrap up

  Consent Mode v2 is a mechanism of managing Google tag operations based on user consent settings. 

  It’s mandatory if you’re using Google’s advertising services, and optional (at least for Advanced mode) if you don’t advertise on Google Ads. 

  There are particular indications that this technology is unreliable from a GDPR perspective. 

  Using Google Consent Mode will inevitably lead to data losses and inaccuracies in its analysis. 

  In other words, it in some sense jeopardises your business.