Recherche avancée

Sur d’autres sites (6620)

• Processing h264 video/audio stream

  1er mars 2024, par Dinamo

  I have separate audio and video streams from a camera that I'm trying to merge to a single file.

  


  Video processing

  


  I have JVT NAL sequence, H.264 video, main @ L 31 data, I had the frames separated and I appended (concat) the frames to one file.

  


  here is the ffprobe output of the data :

  


  ffprobe version 4.2.7-0ubuntu0.1 Copyright (c) 2007-2022 the FFmpeg developers
  built with gcc 9 (Ubuntu 9.4.0-1ubuntu1~20.04.1)
  configuration: --prefix=/usr --extra-version=0ubuntu0.1 --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, h264, from 'frame_data.h264':
  Duration: N/A, bitrate: N/A
    Stream #0:0: Video: h264 (Main), yuvj420p(pc, bt709, progressive), 1280x720, 25 fps, 25 tbr, 1200k tbn, 50 tbc


  


  the file is being played on vlc, however the video is played like it is on fast forward (every 2 seconds of the video should be about 1 second)

  


  The video should run at 15fps (not sure about that because of the mentioned avi file)

  


  here is ffprobe output of working avi file that contain the video stream :

  


  Input #0, avi, from 't.avi':
  Duration: 00:03:29.00, start: 0.000000, bitrate: 1116 kb/s
    Stream #0:0: Video: h264 (Main) (H264 / 0x34363248), yuvj420p(pc, bt709, progressive), 1280x720, 849 kb/s, 30 fps, 15 tbr, 30 tbn, 60 tbc


  


  I tried this command ffmpeg -i frame_data.h264 -filter:v fps=15 test.h264 but the video is still playing like its on fast forward

  


  vlc data before changing fps to 15

  


  


  


  vlc data after changing fps to 15

  


  


  


  frames info (raw)

  


  [FRAME]
media_type=video
stream_index=0
key_frame=1
pkt_pts=N/A
pkt_pts_time=N/A
-> pkt_dts=N/A
-> pkt_dts_time=N/A
best_effort_timestamp=N/A
best_effort_timestamp_time=N/A
-> pkt_duration=48000
-> pkt_duration_time=0.040000
pkt_pos=1476573
pkt_size=57677
width=1280
height=720
pix_fmt=yuvj420p
sample_aspect_ratio=N/A
pict_type=I
coded_picture_number=189
display_picture_number=0
interlaced_frame=0
top_field_first=0
repeat_pict=0
color_range=pc
color_space=bt709
color_primaries=bt709
color_transfer=bt709
chroma_location=left
[/FRAME]
[FRAME]
media_type=video
stream_index=0
key_frame=0
pkt_pts=N/A
pkt_pts_time=N/A
-> pkt_dts=N/A
-> pkt_dts_time=N/A
best_effort_timestamp=N/A
best_effort_timestamp_time=N/A
-> pkt_duration=48000
-> pkt_duration_time=0.040000
pkt_pos=1534250
pkt_size=3928
width=1280
height=720
pix_fmt=yuvj420p
sample_aspect_ratio=N/A
pict_type=P
coded_picture_number=190
display_picture_number=0
interlaced_frame=0
top_field_first=0
repeat_pict=0
color_range=pc
color_space=bt709
color_primaries=bt709
color_transfer=bt709
chroma_location=left
[/FRAME]


  


  frame info from working (processed) avi

  


  [FRAME]
media_type=video
stream_index=0
key_frame=0
pkt_pts=N/A
pkt_pts_time=N/A
-> pkt_dts=500
-> pkt_dts_time=16.666667
best_effort_timestamp=500
best_effort_timestamp_time=16.666667
-> pkt_duration=1
-> pkt_duration_time=0.033333
pkt_pos=1772182
pkt_size=3070
width=1280
height=720
pix_fmt=yuvj420p
sample_aspect_ratio=N/A
pict_type=P
coded_picture_number=191
display_picture_number=0
interlaced_frame=0
top_field_first=0
repeat_pict=0
color_range=pc
color_space=bt709
color_primaries=bt709
color_transfer=bt709
chroma_location=left
[/FRAME]


  


  pkt_pts and pkt_dts are missing in the raw data, maybe that is the reason the video plays on fast forward ? in the working (processed) avi pkt_dts, pkt_dts_time, pkt_duration_time are not missing.
pkt_duration_time is always 0.033333
pkt_dts always jumps by 2 (there is no odd pkt_dts)

  


  Audio processing

  


  I have raw pcm audio of the video :

  


  here is an output of ffplay -f s16le -ar 16k -ac 1 frame_data2.h264 on the raw data

  


  

  


  The raw audio does not playing well using this command.

  


  I guessed the parameters (incorrectly ?) from working avi file that contains the pcm data :

  


  Stream #0:1: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, 1 channels, s16, 256 kb/s


  


  here is frame info from working (processed) avi file

  


  [FRAME]
media_type=audio
stream_index=1
key_frame=1
pkt_pts=4320
pkt_pts_time=0.270000
pkt_dts=4320
pkt_dts_time=0.270000
best_effort_timestamp=4320
best_effort_timestamp_time=0.270000
pkt_duration=480
pkt_duration_time=0.030000
pkt_pos=11670
pkt_size=960
sample_fmt=s16
nb_samples=480
channels=1
channel_layout=unknown
[/FRAME]


  


  I'm trying to understand how should I process the video and the audio so that they will play well (merge them myself with the right av processing to avi/mp4 file)

  


  I also find this extra data that was included for each frame :
I suspect the first column is a sequence number maybe related to dts somehow, second column is a timestamp (milliseconds ?) (maybe you can help me parse it)

  


  600 0xd96533 (audio)
601 0xd9e1dd (audio)
602 0xda4f52 (audio)
603 0xda5a63 (video)
604 0xdacc4b (audio)
605 0xdb39a3 (audio)
606 0xdb5ee9 (video)
607 0xdbb6d8 (audio)
608 0xdc23fe (audio)
609 0xdcb255 (audio)
610 0xdd0e69 (audio)
611 0xdd8b96 (audio)
612 0xdd67d0 (video)
613 0xddf8bd (audio)
614 0xde75fe (audio)
615 0xde6c28 (video)
616 0xdee328 (audio)
617 0xdf6012 (audio)
618 0xdfcd9d (audio)
619 0xe04a7d (audio)
620 0xe07523 (video)
621 0xe0b80c (audio)
622 0xe134db (audio)
623 0xe1798b (video)
624 0xe1a251 (audio)
625 0xe21f5c (audio)
626 0xe28ce1 (audio)
627 0xe27df7 (video)
628 0xe309a7 (audio)
629 0xe376fe (audio)
630 0xe3826e (video)
631 0xe3f3fd (audio)
632 0xe46154 (audio)
633 0xe486e9 (video)
634 0xe4de79 (audio)
635 0xe54bbe (audio)
636 0xe58b79 (video)
637 0xe5c8cb (audio)
638 0xe6361e (audio)
639 0xe6b35b (audio)
640 0xe68fb7 (video)
641 0xe72093 (audio)
642 0xe79db1 (audio)
643 0xe7942f (video)
644 0xe80b09 (audio)
645 0xe887ed (audio)
646 0xe8f546 (audio)
647 0xe97256 (audio)
648 0xe99d2c (video)
649 0xe9dfbd (audio)
650 0xea5ca8 (audio)


  


  Thanks.

  


• Concat video/audio files downloaded from mpd dash manifest

  15 mai 2022, par Bloworlf Mathurin

  I have this manifest I fetched from an url :

  


  &lt;?xml version="1.0"?>&#xA;<mpd type="dynamic" xmlns="urn:mpeg:dash:schema:mpd:2011" profiles="urn:mpeg:dash:profile:isoff-live:2011" availabilitystarttime="2022-05-14T21:18:50-07:00" availabilityendtime="2022-05-14T22:35:36-07:00" timeshiftbufferdepth="PT20S" suggestedpresentationdelay="PT2S" minbuffertime="PT1S" publishtime="2022-05-14T21:18:50-07:00" minimumupdateperiod="PT1S" validationerrors="" currentservertimems="0" firstavtimems="1652588331877" lastvideoframets="0" loapstreamid="17944405669984029" publishframetime="939">&#xA;    <period start="PT0S">&#xA;        <adaptationset segmentalignment="true" maxwidth="432" maxheight="766" maxframerate="30">&#xA;            <representation mimetype="video/mp4" codecs="avc1.4d401e" width="432" height="766" framerate="30" startwithsap="1" bandwidth="38894" maxbandwidth="46772" playbackresolutionmos="432:82.84,720:65.43,216:70.39" qualityclass="sd" qualitylabel="432p">&#xA;                <segmenttemplate presentationtimeoffset="0" timescale="1000" initialization="some/url/some_id-init.m4v?ms=m_C&amp;amp;ccb=2-4" media="some/url/some_id-$Time$.m4v?ms=m_C&amp;amp;ccb=2-4">&#xA;                    <segmenttimeline>&#xA;                        <s t="4092613" d="2000"></s>&#xA;                        <s t="4094613" d="2000"></s>&#xA;                        <s t="4096613" d="2000"></s>&#xA;                        <s t="4098613" d="2000"></s>&#xA;                        <s t="4100613" d="2000"></s>&#xA;                        <s t="4102613" d="2000"></s>&#xA;                        <s t="4104613" d="2000"></s>&#xA;                        <s t="4106613" d="2000"></s>&#xA;                        <s t="4108613" d="2000"></s>&#xA;                        <s t="4110613" d="2000"></s>&#xA;                    </segmenttimeline>&#xA;                </segmenttemplate>&#xA;            </representation>&#xA;        </adaptationset>&#xA;        <adaptationset segmentalignment="true">&#xA;            <representation mimetype="audio/mp4" codecs="mp4a.40.2" audiosamplingrate="44100" startwithsap="1" bandwidth="48979" maxbandwidth="57732">&#xA;                <segmenttemplate presentationtimeoffset="0" timescale="1000" initialization="some/url/some_id-init.m4a?ms=m_C&amp;amp;ccb=2-4" media="some/url/some_id-$Time$.m4a?ms=m_C&amp;amp;ccb=2-4">&#xA;                    <segmenttimeline>&#xA;                        <s t="4092613" d="2000"></s>&#xA;                        <s t="4094613" d="2000"></s>&#xA;                        <s t="4096613" d="2000"></s>&#xA;                        <s t="4098613" d="2000"></s>&#xA;                        <s t="4100613" d="2000"></s>&#xA;                        <s t="4102613" d="2000"></s>&#xA;                        <s t="4104613" d="2000"></s>&#xA;                        <s t="4106613" d="2000"></s>&#xA;                        <s t="4108613" d="2000"></s>&#xA;                        <s t="4110613" d="2000"></s>&#xA;                    </segmenttimeline>&#xA;                </segmenttemplate>&#xA;            </representation>&#xA;        </adaptationset>&#xA;    </period>&#xA;</mpd>&#xA;

  


  I manage to download all the files (init file + 10 segment files) and put them in an array

  


  File[] files = downloadSegments();

  


  So I have :
[0] -> file-0.m4v (which is the init file)
[1] -> file-1.m4v (1st segment)
... and so on.

  


  My question is how can I concat/append all these files into another file (final_segment.m4v) ?

  


  I looked around a lot and am now using 'com.arthenica:ffmpeg-kit-video:4.5.1-1'
I've last tried :

  


  FFmpegSession session = FFmpegKit.execute("-i file-0.m4v -i file-1.m4v ... -c copy final_segment.m4v");


  


  Is there something I'm missing ?

  


  Also I assume that I'll have to do the same for the audio segments. So I will have 2 files (final_segment.m4v and final_segment.m4a) that I will have to merge/mix together.

  


  If you could help me with some piece of code, that would be great.

  


• 7 Best Marketing Attribution Software in 2024

  22 février 2024, par Erin

  It can be hard to accurately track the impact of your marketing efforts across marketing channels and campaigns. That’s where marketing attribution software comes in. 

  It goes beyond basic web analytics solutions that just look at the final click. Instead, it shows how different channels, content, and ads are performing at every step of the buyer’s journey, which gives a more accurate picture than just focusing on the last click.

  In this guide, we’ll cover the basics of marketing attribution, list the top marketing attribution software and explain how the issue of privacy is transforming the web analytics industry.

  What is marketing attribution ?

  Marketing attribution is the process of assigning credit to each touchpoint in a buyer’s journey that leads to a desired action (such as a conversion or sale) in order to understand the effectiveness of various marketing channels and campaigns in influencing the customer’s decision-making process.

  Marketers use software tools like website analytics to to track and analyse customer interactions across different touchpoints, allowing them to attribute conversions or sales to specific marketing efforts and optimise their strategies and budgets accordingly.

  Why is marketing attribution so important ?

  If you don’t track your campaigns correctly, it’s easy to spend thousands (or even millions) in an ineffective way. A 2022 survey by Australian marketing agency Next&Co revealed their clients wasted AU$5.46 billion in ineffective ad spend.

  

  That’s 41% of all the ad spend tracked by Next&Co in 2022. A wasted marketing spend percentage this high isn’t exactly a recipe for a high marketing return on investment (ROI). And yet, it’s the average.

  Why is that ? 

  Most companies don’t actively track the results of their marketing campaigns actively enough.

  By improving your marketing attribution, you can determine which channels, ads, and campaigns work and which don’t. Then, you can move the budget from ineffective channels to effective ones.

  Even if you can only identify half of your wastage, this could be 20% or more of your total spend. Just imagine what your bottom line would look like if your marketing budget were 20% more effective.

  That’s the power that marketing attribution, when done right, brings to the table. It’s the road to a higher marketing ROI.

  Common marketing attribution models and how they’re different 

  The default model for attributing completed goals in most analytics tools is either the last interaction or the last non-direct interaction.

  However, some multi-touch models can help you get a more holistic view of the impact of your marketing efforts.

  

  • Last interaction model : attributes the conversion to the final interaction or referring source (campaign or ad).
  • Last non-direct interaction model : attributes the conversion to the final touchpoint that was not a direct visit to your website. (For example, if a search ad took them to a product page, the user bookmarked it and returned directly the next day to finish the purchase. The credit would go to the search ad as it’s the last non-direct touchpoint.)
  • First interaction model : attributes the conversion to the first referring event alone.
  • Linear model : gives equal value to every touchpoint throughout the customer journey. 
  • Time decay model : gives more value to touchpoints the closer they were to the actual sale.
  • Position-based model : gives more value to the first and last touchpoints — often 40% each, while splitting 20% among the rest.

  You can read our guide dedicated to marketing attribution models for more details on these models.

  Types of marketing attribution software and the impact of privacy regulations

  Until recently, digital advertising was the “scientific” advertisers’ utopia. Everything could be measured, with cookies from giants like Google and Facebook stalking every user across the web.

  But with the advent of regulations like GDPR and the CCPA, you can no longer blindly trust Google Analytics or the Meta Pixel without consequences.

  Multi-channel attribution tools with third-party cookies and GDPR

  Google, Meta, and other companies used to track and combine user data from their own platforms and websites across the web that installed their tags. These third-party cookies have long been under fire and have caused several GDPR fines.

  

  The alternative : analytics platforms with first-party cookies

  In a post-GDPR digital marketing landscape, a compliant-by-default web analytics platform like Matomo is a more reliable and accurate alternative.

  Plus, with a platform like Matomo, you don’t need to rely on data from digital advertising platforms like Facebook Ads and Google Ads. You can accurately track referral sources using our campaign tracking parameters.

  7 best marketing attribution software in 2024

  Below is the list of our favourite marketing attribution tools in 2024. If you find and use one that suits your needs correctly, you can quickly boost your marketing performance.

  1. Matomo — Accurate and easiest to set up for marketing attribution

  Matomo is a privacy-friendly web analytics suite that empowers you to accurately attribute marketing efforts and gain valuable insights while prioritising user privacy and compliance.

  Matomo integrates with e-commerce platforms like WooCommerce and Magenta. That makes it easy for B2C marketing teams to track the revenue impact of their campaigns.

  

  You can also compare a variety of attribution models against each other. B2B teams can use our API to integrate Matomo with their CRM.

  Pros :

  • Relies on first-party cookies for tracking, ensuring accurate data collection and attribution of user actions
  • Includes additional features like Heatmaps, Session Recordings, Form Analytics, A/B Testing, and more
  • Easy to set up and use
  • Features most common multi-touch attribution models

  Cons :

  • Limited to owned channels (website and e-commerce store) due to first-party cookies and data (but you can integrate other data sources through a CRM)

  Pricing

  The self-hosted version is free. The cloud hosted version starts at $19 per month and includes a 21-day free trial. No credit card requierd. 

  Try Matomo for Free

  Get the web insights you need, without compromising data accuracy.

  Start for free

  No credit card required

  2. WhatConverts — Great option for leads-based businesses with high ad spend

  WhatConverts is a marketing attribution tool with a focus on lead tracking. With most web analytics setups, it adds call and text tracking to the typical form-only tracking.

  

  Pros :

  • Reliable call and text tracking
  • Revenue attribution to specific leads (and, by extension, campaigns and ads)

  Cons :

  • Focused exclusively on leads — little utility for e-commerce companies 

  Pricing

  The cheapest plan starts at $30/month but does not include analytics integrations or form tracking. To access this and advanced flow tracking and attribution features, you need the Elite plan, which starts at $160/month.

  3. HubSpot Marketing Hub — Ideal CRM for larger B2B companies

  HubSpot is a marketing CRM with attribution features for tracking and analysis.

  

  The platform is very broad — encompassing CRM, email automation and other tools — which makes it challenging to use effectively. The price tag is also quite steep for smaller companies and marketing teams.

  Pros :

  • Concretely tracks revenue to multiple different touchpoints and marketing channels
  • Includes several different multi-touch attribution models
  • Allows offline conversion tracking

  Cons :

  • The price point is too high for smaller teams
  • Cam be difficult to set up effectively

  Pricing

  Since marketing attribution is only included in HubSpot Marketing Hub’s Professional and Enterprise plans, pricing starts at $800/month (paid annually). If you commit for a year but pay monthly, the price is $890/month for the professional plan. This goes up with additional add-ons and as your contacts increase as well. 

  4. ActiveCampaign — Good CRM option for small B2B companies

  ActiveCampaign is a CRM and marketing automation platform that can help you trace leads and revenue back to their source.

  

  Although it has a similar scope of features to HubSpot, it is more affordable and slightly easier to use for beginners.

  Pros :

  • Tracks sales revenue back to specific marketing touchpoints
  • Powerful marketing automation features

  Cons :

  • B2B companies may need to purchase two plans, one ActiveCampaign marketing and one CRM.

  Pricing

  Unlike HubSpot, ActiveCampaign offers a much more affordable plan, starting at $29/month billed annually (for up to 1,000 contacts). The marketing and sales CRM bundle starts at $93/month with up to five users.

  5. Salesforce Data Cloud for Marketing — Ideal CRM for enterprises

  Salesforce is a robust and feature-rich CRM that many enterprises rely on for their sales teams.

  

  That makes Salesforce’s marketing attribution platform a logical choice for existing Salesforce users.

  Pros :

  • Uses prospect and sales data from CRM to attribute revenue
  • Revenue prediction analytics
  • Lead scoring to help your sales team focus on high-value leads

  Cons :

  • Difficult to set up and use
  • Clunky and aged user interface
  • Relatively high price point

  Pricing

  The limited Marketing Cloud Account Engagement Growth plan starts at $1,250/month, billed annually. To access advanced cross-channel journeys, you need the Pro plan, which starts at $2,750 monthly.

  6. Terminus — Great for account-based marketing

  If your marketing team uses an account-based marketing (ABM) approach, Terminus might be the right option for you.

  

  It offers ABM tools like target account event tracking and revenue attribution tools for your marketing campaigns.

  Pros :

  • Advanced multi-channel revenue attribution tools with a wide range of reports
  • Track intent touchpoints back to target accounts
  • Reliable revenue predictions help you focus your marketing activities

  Cons :

  • Complex and difficult to set up, understand and use effectively
  • Lacks native integrations with many common advertising platforms and analytics tools

  Pricing

  Terminus offers no standard pricing plans. You must contact their sales team for a custom quote based on your needs.

  7. Adobe Analytics — An analytics for enterprises

  Adobe Analytics is part of the Adobe Experience Cloud, with plenty of big data analysis tools for enterprises. Although the platform is quite powerful, it is equally complex and difficult to use. The price point is also prohibitive for many smaller companies.

  

  Pros :

  • Very extensive reporting tools
  • Predictive analytics give you solid leading indicator for future campaign performance
  • Track multiple digital touchpoints across the entire customer journey

  Cons :

  • Like Google Analytics, Adobe Analytics aggregates your visitor data by default, making compliant “consent-free tracking” — tracking user actions without asking for consent — impossible according to GDPR. (See more differences in Matomo’s comparison against Adobe Analytics and Google Analytics.)
  • Prohibitively expensive for most smaller companies
  • Very steep learning curve for setting up and using it correctly

  Pricing

  Adobe Analytics uses usage-based pricing — which means they adjust the pricing based on the traffic volume to your website. Still, their lower price points aren’t exactly SMB-friendly — multiple sources put Adobe’s lowest starting price point at $2,000–2,500 per month.

  Get accurate marketing attribution with Matomo (without privacy concerns)

  Matomo allows you to do marketing attribution effectively and accurately without compromising your users’ privacy. By default, we only use first-party cookies and offer consent-free tracking – meaning no more annoying cookie consent banners (excluding in Germany and the UK).

  If you want to boost your marketing performance without disregarding your users’ privacy, get started with our 21-day free trial. No credit card required. It’s time to make more informed decisions about your marketing campaigns.

  Try Matomo for Free

  21 day free trial. No credit card required.

  
  

  Your email address
  

  Your website address
  

  Your Analytics subdomain will be .matomo.cloud (change)

  Choose your analytics subdomain
  

  .matomo.cloud

  I accept the terms and conditions and data processing agreement (DPA)

  
  Your information will be used to create an account on our cloud service. For more information please consult our privacy policy.
  
  
  

  Start improving your websites now