Recherche avancée

Sur d’autres sites (12665)

• Convert from oga to mp3 using pydub : ffmpeg returned error code : 1

  29 juin 2023, par Juan David

  I want to take an OGA file within a binary stream and convert it into mp3 using also another stream. I'm getting a permissions error even with running VSCode as administrator. This is my code :

  


  from pydub import AudioSegment
AudioSegment.converter = "C:\\ProgramData\\chocolatey\\lib\\ffmpeg\\tools\\ffmpeg\\bin\\ffmpeg.exe"

input_stream = io.BytesIO()
input_stream.seek(0) 
await new_file.download_to_memory(input_stream)
 
# Create an audio segment from the binary stream
audio = AudioSegment.from_file(input_stream, format='ogg')

# Create an output stream for the MP3 data
output_stream = io.BytesIO()

# Export the audio to MP3 using ffmpeg and write the output to the stream
audio.export(output_stream, format='mp3', codec='libmp3lame')

# Get the MP3 data from the output stream
mp3_data = output_stream.getvalue()


  


  Error message :

  


    File "C:\ProgramData\Anaconda3\envs\chatbot\lib\site-packages\telegram\ext\_application.py", line 1124, in process_update
    await coroutine
  File "C:\ProgramData\Anaconda3\envs\chatbot\lib\site-packages\telegram\ext\_handler.py", line 141, in handle_update
    return await self.callback(update, context)
  File "c:\Users\jdbol\OneDrive\Desktop\testbots\echobot.py", line 80, in voice_to_text
    audio = AudioSegment.from_file(input_stream, format='ogg')
  File "C:\ProgramData\Anaconda3\envs\chatbot\lib\site-packages\pydub\audio_segment.py", line 773, in from_file     
    raise CouldntDecodeError(
pydub.exceptions.CouldntDecodeError: Decoding failed. ffmpeg returned error code: 1

Output from ffmpeg/avlib:

ffmpeg version 6.0-essentials_build-www.gyan.dev Copyright (c) 2000-2023 the FFmpeg developers
  built with gcc 12.2.0 (Rev10, Built by MSYS2 project)
  configuration: --enable-gpl --enable-version3 --enable-static --disable-w32threads --disable-autodetect --enable-fontconfig --enable-iconv --enable-gnutls --enable-libxml2 --enable-gmp --enable-lzma --enable-zlib --enable-libsrt --enable-libssh --enable-libzmq --enable-avisynth --enable-sdl2 --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxvid --enable-libaom --enable-libopenjpeg --enable-libvpx --enable-libass --enable-libfreetype --enable-libfribidi --enable-libvidstab --enable-libvmaf --enable-libzimg --enable-amf --enable-cuda-llvm --enable-cuvid --enable-ffnvcodec --enable-nvdec --enable-nvenc --enable-d3d11va --enable-dxva2 --enable-libmfx --enable-libgme --enable-libopenmpt --enable-libopencore-amrwb --enable-libmp3lame --enable-libtheora --enable-libvo-amrwbenc --enable-libgsm --enable-libopencore-amrnb --enable-libopus --enable-libspeex --enable-libvorbis --enable-librubberband
  libavutil      58.  2.100 / 58.  2.100
  libavcodec     60.  3.100 / 60.  3.100
  libavformat    60.  3.100 / 60.  3.100
  libavdevice    60.  1.100 / 60.  1.100
  libavfilter     9.  3.100 /  9.  3.100
  libswscale      7.  1.100 /  7.  1.100
  libswresample   4. 10.100 /  4. 10.100
  libpostproc    57.  1.100 / 57.  1.100
fd:: End of file


  


  I ran the command in a terminal and the file converted without an issue :

  


  ffmpeg -i .\file_12.oga output.mp3 


  


  

  1. I'm not sure if the .exe file must be included into the path. When I don't do it, What I get is a permissions error.

  


  2. What other codecs can be used here ?

  


  3. Is it possible to use oga files ? I tried to declare this but I got an 'Unknown input format : 'oga' message (audio = AudioSegment.from_file(input_stream, format='oga'))

  


  


  Thanks !

  


  UPDATE : I created a more simple version that is not using a binary stream and worked like a charm, so we know for sure that something is happening with the BytesIO object

  


  async def voice_to_text(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
    # Get the absolute path of the script
    filename = 'file_9.oga'
    script_dir = os.path.dirname(os.path.abspath(__file__))

    # Construct the file paths for input and output files
    input_file_path = os.path.join(script_dir, filename)
    output_file_path = os.path.join(script_dir, os.path.splitext(filename)[0] + ".mp3")

    # Load the OGA audio file
    audio = AudioSegment.from_file(input_file_path, format='ogg')

    # Export the audio to MP3 format
    audio.export(output_file_path, format='mp3')

    print("Conversion complete. MP3 file saved as:", output_file_path)


  


  UPDATE 2 : It seems like await new_file.download_to_memory(input_stream) is the problematic line. I tried to save the file and its corrupt. Not sure how to use this method then.

  


  https://docs.python-telegram-bot.org/en/stable/telegram.file.html#telegram.File.download_to_memory

  


• webm video - it can not be copy to cut a correct time part for ffmpeg command(without re-encode) [closed]

  15 mars 2024, par DanielHsu

  Command Goal :

  


  

  1. fast copy video's part

  


  2. no re-encode with libvpx

  


  


  Linux command below :
ffmpeg -ss 00:01:00 -to 00:02:00 -i input.webm -c copy output.webm

  


  command's process details :

  


  ffmpeg version 6.1.1 Copyright (c) 2000-2023 the FFmpeg developers   built with gcc 9 (Ubuntu 9.4.0-1ubuntu1~20.04.1)   configuration:
--enable-libvpx   libavutil      58. 29.100 / 58. 29.100   libavcodec     60. 31.102 / 60. 31.102   libavformat    60. 16.100 / 60. 16.100   libavdevice    60.  3.100 / 60.  3.100   libavfilter     9. 12.100 / 
9. 12.100   libswscale      7.  5.100 /  7.  5.100   libswresample   4. 12.100 /  4. 12.100 Input #0, matroska,webm, from 'input.webm':   Metadata:
    title           : FFmpeg
    ENCODER         : Lavf58.29.100   Duration: 00:06:04.43, start: 0.000000, bitrate: 2487 kb/s   Stream #0:0: Video: vp8, yuv420p(tv, bt470bg/unknown/unknown, progressive), 1280x720, SAR 1:1 DAR 16:9, 1k tbr, 1k tbn (default)
    Metadata:
      DURATION        : 00:06:04.432000000 Output #0, webm, to 'output.webm':   Metadata:
    title           : FFmpeg
    encoder         : Lavf60.16.100   Stream #0:0: Video: vp8, yuv420p(tv, bt470bg/unknown/unknown, progressive), 1280x720 [SAR 1:1 DAR 16:9], q=2-31, 1k tbr, 1k tbn (default)
    Metadata:
      DURATION        : 00:06:04.432000000 Stream mapping:   Stream #0:0 -> #0:0 (copy) Press [q] to stop, [?] for help [out#0/webm @ 0x5609ee9bb840] video:35912kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.069370% size=   35937kB time=00:00:59.96 bitrate=4909.2kbits/s speed= 211x


  


  command usage's result :
output.webm always start from input.webm's 00:00:00 to end input.webm's 00:01:00
(it should be start from input.webm's 00:01:00 to end input.webm's 00:02:00 )

  


  input.webm format details from mediainfo tool :

  


  


  General Count : 331 Count of stream
of this kind : 1 Kind of stream
  
 : General Kind of stream : General Stream
identifier : 0 Count of video streams
  
 : 1 Video_Format_List : VP8
Video_Format_WithHint_List : VP8 Codecs Video
  
 : VP8 Complete name :
pure_film/65ea641a4075c92236151349_video.webm Folder name
  
 : pure_film File name extension :
65ea641a4075c92236151349_video.webm File name
  
 : 65ea641a4075c92236151349_video File extension
  
 : webm Format : WebM Format
  
 : WebM Format/Url :
http://www.webmproject.org/ Format/Extensions usually used :
webm Commercial name : WebM Format version
  
 : Version 2 Internet media type : video/webm File
size : 113301943 File size
  
 : 108 MiB File size : 108 MiB File size
 : 108 MiB File size : 108 MiB File size
 : 108.1 MiB Duration : 364432 Duration
 : 6 min 4 s Duration : 6 min 4 s 432
ms Duration : 6 min 4 s Duration
  
 : 00:06:04.432 Duration : 00:06:04.432
Overall bit rate : 2487201 Overall bit rate
  
 : 2 487 kb/s Stream size : 4801069 Stream
size : 4.58 MiB (4%) Stream size
  
 : 5 MiB Stream size : 4.6 MiB Stream size
 : 4.58 MiB Stream size : 4.579 MiB Stream
size : 4.58 MiB (4%) Proportion of this
stream : 0.04237 IsStreamable
  
 : Yes Title : FFmpeg Movie name
  
 : FFmpeg File last modification date : UTC 2024-03-08
01:41:47 File last modification date (local) : 2024-03-08
09:41:47 Writing application : Lavf58.29.100
Writing application : Lavf58.29.100 Writing
library : Lavf58.29.100 Writing library
  
 : Lavf58.29.100

  


  Video Count : 377 Count of stream
of this kind : 1 Kind of stream
  
 : Video Kind of stream : Video Stream
identifier : 0 StreamOrder
  
 : 0 ID : 1 ID
  
 : 1 Unique ID : 1 Format
  
 : VP8 Format : VP8 Format/Url
  
 : http://www.webmproject.org/ Commercial name
  
 : VP8 Codec ID : V_VP8 Codec ID/Url
  
 : http://www.webmproject.org/ Duration
  
 : 364432.000000 Duration : 6 min 4 s
Duration : 6 min 4 s 432 ms Duration
  
 : 6 min 4 s Duration : 00:06:04.432
Duration : 00:06:04.432 Bit rate
  
 : 2381808 Bit rate : 2 382 kb/s Width
 : 1280 Width : 1 280 pixels Height
 : 720 Height : 720 pixels Pixel
aspect ratio : 1.000 Display aspect ratio
  
 : 1.778 Display aspect ratio : 16:9 Frame rate
mode : VFR Frame rate mode
  
 : Variable Color space : YUV Compression
mode : Lossy Compression mode
  
 : Lossy Delay : 0 Delay
  
 : 00:00:00.000 Delay, origin : Container
Delay, origin : Container Stream size
  
 : 108500874 Stream size : 103 MiB (96%)
Stream size : 103 MiB Stream size
  
 : 103 MiB Stream size : 103 MiB Stream
size : 103.5 MiB Stream size
  
 : 103 MiB (96%) Proportion of this stream : 0.95763
Default : Yes Default
  
 : Yes Forced : No Forced
  
 : No colour_description_present : Yes
colour_description_present_Source : Container Color range
  
 : Limited colour_range_Source : Container Matrix
coefficients : BT.470 System B/G
matrix_coefficients_Source : Container

  


  


  How to make the output.webm be in correct time range(00:01:00 00:02:00) ????

  


• Four Trends Shaping the Future of Analytics in Banking

  27 novembre 2024, par Daniel Crough — Banking and Financial Services

  While retail banking revenues have been growing in recent years, trends like rising financial crimes and capital required for generative AI and ML tech pose significant risks and increase operating costs across the financial industry, according to McKinsey’s State of Retail Banking report.

   

  Today’s financial institutions are focused on harnessing AI and advanced analytics to make their data work for them. To be up to the task, analytics solutions must allow banks to give consumers the convenient, personalised experiences they want while respecting their privacy.

   

  In this article, we’ll explore some of the big trends shaping the future of analytics in banking and finance. We’ll also look at how banks use data and technology to cut costs and personalise customer experiences.
  
  So, let’s get into it.

  

  This doesn’t just represent a security risk, it also impacts the usability for both customers and employees. Does any of the following sound familiar ?

  • Only specific senior employees know how to navigate the software to generate custom reports or use its more advanced features.
  • Customer complaints about your site’s usability or online banking experience are routine.
  • Onboarding employees takes much longer than necessary because of convoluted systems.
  • Teams and departments experience ‘data siloing,’ meaning that not everyone can access the data they need.

  These are warning signs that IT systems are ready for a review. Anyone thinking, “If it’s not broken, why fix it ?” should consider that legacy systems can also present data security risks. As more countries introduce regulations to protect customer privacy, staying ahead of the curve is increasingly important to avoid penalties and litigation.

  And regulations aren’t the only trends impacting the future of financial institutions’ IT and analytics.

  4 trends shaping the future of analytics in banking

  New regulations and new technology have changed the landscape of analytics in banking.

  New privacy regulations impact banks globally

  The first major international example was the advent of GDPR, which went into effect in the EU in 2018. But a lot has happened since. New privacy regulations and restrictions around AI continue to roll out.

  • The European Artificial Intelligence Act (EU AI Act), which was held up as the world’s first comprehensive legislation on AI, took effect on 31 July 2024.
  • In Europe’s federated data initiative, Gaia-X’s planned cloud infrastructure will provide for more secure, transparent, and trustworthy data storage and processing.
  • The revised Payment Services Directive (PSD2) makes payments more secure and strengthens protections for European businesses and consumers, aiming to create a more integrated and efficient payments market.

  But even businesses that don’t have customers in Europe aren’t safe. Consumer privacy is a hot-button issue globally.

  For example, the California Consumer Privacy Act (CCPA), which took effect in January, impacts the financial services industry more than any other. Case in point, 34% of CCPA-related cases filed in 2022 were related to the financial sector.

  California’s privacy regulations were the first in the US, but other states are following closely behind. On 1 July 2024, new privacy laws went into effect in Florida, Oregon, and Texas, giving people more control over their data.

  

  One typical issue for companies in the banking industry is that their privacy measures regarding user data collected from their website are much less lax than those in their online banking system.

  It’s better to proactively invest in a privacy-centric analytics platform before you get tangled up in a lawsuit and have to pay a fine (and are forced to change your system anyway). 

  And regulatory compliance isn’t the only bonus of an ethical analytics solution. The right alternative can unlock key customer insights that can help you improve the user experience.

  The demand for personalised banking services

  At the same time, consumers are expecting a more and more streamlined personal experience from financial institutions. 86% of bank employees say personalisation is a clear priority for the company. But 63% described resources as limited or only available after demonstrating clear business cases.

  McKinsey’s The data and analytics edge in corporate and commercial banking points out how advanced analytics are empowering frontline bank employees to give customers more personalised experiences at every stage :

  • Pre-meeting/meeting prep : Using advanced analytics to assess customer potential, recommend products, and identify prospects who are most likely to convert
  • Meetings/negotiation : Applying advanced models to support price negotiations, what-if scenarios and price multiple products simultaneously
  • Post-meeting/tracking : Using advanced models to identify behaviours that lead to high performance and improve forecast accuracy and sales execution

  Today’s banks must deliver the personalisation that drives customer satisfaction and engagement to outperform their competitors.

  The rise of AI and its role in banking

  With AI and machine learning technologies becoming more powerful and accessible, financial institutions around the world are already reaping the rewards.

  McKinsey estimates that AI in banking could add $200 to 340 billion annually across the global banking sector through productivity gains.

  • Credit card fraud prevention : Algorithms analyse usage to flag and block fraudulent transactions.
  • More accurate forecasting : AI-based tools can analyse a broader spectrum of data points and forecast more accurately.
  • Better risk assessment and modelling : More advanced analytics and predictive models help avoid extending credit to high-risk customers.
  • Predictive analytics : Help spot clients most likely to churn 
  • Gen-AI assistants : Instantly analyse customer profiles and apply predictive models to suggest the next best actions.

  Considering these market trends, let’s discuss how you can move your bank into the future.

  Using analytics to minimise risk and establish a competitive edge 

  With the right approach, you can leverage analytics and AI to help future-proof your bank against changing customer expectations, increased fraud, and new regulations.

  Use machine learning to prevent fraud

  Every year, more consumers are victims of credit and debit card fraud. Debit card skimming cases nearly doubled in the US in 2023. The last thing you want as a bank is to put your customer in a situation where a criminal has spent their money.

  This not only leads to a horrible customer experience but also creates a lot of internal work and additional costs.Thankfully, machine learning can help identify suspicious activity and stop transactions before they go through. For example, Mastercard’s fraud prevention model has improved fraud detection rates by 20–300%.

  

  Implementing a solution like this (or partnering with credit card companies who use it) may be a way to reduce risk and improve customer trust.

  Foresee and avoid future issues with AI-powered risk management

  Regardless of what type of financial products organisations offer, AI can be an enormous tool. Here are just a few ways in which it can mitigate financial risk in the future :

  • Predictive analytics can evaluate risk exposure and allow for more informed decisions about whether to approve commercial loan applications.
  • With better credit risk modelling, banks can avoid extending personal loans to customers most likely to default.
  • Investment banks (or individual traders or financial analysts) can use AI- and ML-based systems to monitor market and trading activity more effectively.

  Those are just a few examples that barely scratch the surface. Many other AI-based applications and analytics use cases exist across all industries and market segments.

  Protect customer privacy while still getting detailed analytics

  New regulations and increasing consumer privacy concerns don’t mean banks and financial institutions should forego website analytics altogether. Its insights into performance and customer behaviour are simply too valuable. And without customer interaction data, you’ll only know something’s wrong if someone complains.

  Fortunately, it doesn’t have to be one or the other. The right financial analytics solution can give you the data and insights needed without compromising privacy while complying with regulations like GDPR and CCPA.

  That way, you can track usage patterns and improve site performance and content quality based on accurate data — without compromising privacy. Reliable, precise analytics are crucial for any bank that’s serious about user experience.

  Use A/B testing and other tools to improve digital customer experiences

  Personalised digital experiences can be key differentiators in banking and finance when done well. But there’s stiff competition. In 2023, 40% of bank customers rated their bank’s online and mobile experience as excellent. 

  Improving digital experiences for users while respecting their privacy means going above and beyond a basic web analytics tool like Google Analytics. Invest in a platform with features like A/B tests and user session analysis for deeper insights into user behaviour.

  

  Behavioural analytics are crucial to understanding customer interactions. By identifying points of friction and drop-off points, you can make digital experiences smoother and more engaging.

  Matomo offers all this and is a great GDPR-compliant alternative to Google Analytics for banks and financial institutions. 

  Of course, this can be challenging. This is why taking an ethical and privacy-centric approach to analytics can be a key competitive edge for banks. Prioritising data security and privacy will attract other like-minded, ethically conscious consumers and boost customer loyalty.

  Get privacy-friendly web analytics suitable for banking & finance with Matomo

  Improving digital experiences for today’s customers requires a solid web analytics platform that prioritises data privacy and accurate analytics. And choosing the wrong one could even mean ending up in legal trouble or scrambling to reconstruct your entire analytics setup.

  Matomo provides privacy-friendly analytics with 100% data accuracy (no sampling), advanced privacy controls and the ability to run A/B tests and user session analysis within the same platform (limiting risk and minimising costs). 

  It’s easy to get started with Matomo. Users can access clear, easy-to-understand metrics and plenty of pre-made reports that deliver valuable insights from day one. Form usage reports can help banks and fintechs identify potential issues with broken links or technical glitches and reveal clues on improving UX in the short term.

  Over one million websites, including some of the world’s top banks and financial institutions, use Matomo for their analytics.

  Start your 21-day free trial to see why, or book a demo with one of our analytics experts.