Recherche avancée

Sur d’autres sites (4228)

• Correct syntax for using ffmpeg CUDA to crop and scale videos [closed]

  9 septembre 2024, par Stuart

  My NVIDIA and FFMPEG configuration is as follows :

  


  Linux TestBox1 6.8.0-41-generic #41-Ubuntu SMP PREEMPT_DYNAMIC Fri Aug 2 20:41:06 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux

  


  NVIDIA-Linux-x86_64-560.31.02

  


  ffmpeg version git-2024-08-19-507c2a577
encoders:
 V....D av1_nvenc            NVIDIA NVENC av1 encoder (codec av1)
 V....D h264_nvenc           NVIDIA NVENC H.264 encoder (codec h264)
 V....D hevc_nvenc           NVIDIA NVENC hevc encoder (codec hevc)
decoders:
 V..... av1_cuvid            Nvidia CUVID AV1 decoder (codec av1)
 V..... h264_cuvid           Nvidia CUVID H264 decoder (codec h264)
 V..... hevc_cuvid           Nvidia CUVID HEVC decoder (codec hevc)
 V..... mjpeg_cuvid          Nvidia CUVID MJPEG decoder (codec mjpeg)
 V..... mpeg1_cuvid          Nvidia CUVID MPEG1VIDEO decoder (codec mpeg1video)
 V..... mpeg2_cuvid          Nvidia CUVID MPEG2VIDEO decoder (codec mpeg2video)
 V..... mpeg4_cuvid          Nvidia CUVID MPEG4 decoder (codec mpeg4)
 V..... vc1_cuvid            Nvidia CUVID VC1 decoder (codec vc1)
 V..... vp8_cuvid            Nvidia CUVID VP8 decoder (codec vp8)
 V..... vp9_cuvid            Nvidia CUVID VP9 decoder (codec vp9)
filters:
 ... bilateral_cuda    V->V       GPU accelerated bilateral filter
 T.. bwdif_cuda        V->V       Deinterlace CUDA frames
 ... chromakey_cuda    V->V       GPU accelerated chromakey filter
 ... colorspace_cuda   V->V       CUDA accelerated video color converter
 ..C crop              V->V       Crop the input video.
 T.C cropdetect        V->V       Auto-detect crop size.
 ... hwupload_cuda     V->V       Upload a system memory frame to a CUDA device.
 ... overlay_cuda      VV->V      Overlay one video on top of another using CUDA
 ... scale_cuda        V->V       GPU accelerated video resizer
 ... scale_npp         V->V       NVIDIA Performance Primitives video scaling and format conversion
 ... scale2ref_npp     VV->VV     NVIDIA Performance Primitives video scaling and format conversion to the given reference.
 ... sharpen_npp       V->V       NVIDIA Performance Primitives video sharpening filter.
 ... thumbnail_cuda    V->V       Select the most representative frame in a given sequence of consecutive frames.
 ... transpose_npp     V->V       NVIDIA Performance Primitives video transpose
 T.. yadif_cuda        V->V       Deinterlace CUDA frames


  


  I would like to be able to use CUDA to crop and scale videos but I cannot find the correct syntax.
My efforts usually work in terms of the scale but not the crop, here is an example of the type of problems I have been having :

  


  The source file in this example is a short mpeg2 video (960x540 with black bars top and bottom). The command is currently run from a bash script.

  


  ffmpeg -y -hwaccel cuda -hwaccel_output_format cuda -c:v mpeg2_cuvid -crop 960x408x0x66 -i $1 -vf scale_npp=1280x544 -c:a copy -c:v h264_nvenc -b:v 2M -preset slow $2


  


  This is throwing the following error :

  


  [mpeg2_cuvid @ 0x5e6d9deb56c0] cuvid decode callback error
[vist#0:0/mpeg2video @ 0x5e6d9de85b40] [dec:mpeg2_cuvid @ 0x5e6d9de7b7c0] Decoding error:        Generic error in an external library


  


  Have I a problem with my ffmpeg build, should the list of filters include something like crop_cuda or crop_npp or have I just got the wrong syntax to perform everything on the GPU (is it even possible).

  


• Introducing the Matomo Connector for Looker Studio (Formerly Google Data Studio)

  26 janvier 2024, par Erin — Community

  Explore Matomo data like never before with the official Matomo Connector for Looker Studio. Matomo users can now securely display accurate web analytics data in Looker Studio for free.

  Connect Matomo to Looker Studio (formerly known as Google Data Studio) in a few clicks and start building dashboards instantly. Get access to a range of data visualisation capabilities and chart types in Looker Studio’s easy-to-use interface. 

  Leave behind manual, error-prone spreadsheet entries and disparate data. With the Matomo Connector for Looker Studio, you get unified, automated reporting and interactive dashboards for faster insights and smoother collaboration.

  What sets the official Matomo Connector for Looker Studio apart ?

  Our open-source connector puts security first by providing a reliable connection without relying on third-party intermediaries. It’s free, with no hidden charges, and no limits on the number of users or Matomo instances. Connect as many instances as you need. 

  Plus, our Support team is here anytime you need help.

  

  Who is this connector made for ?

  The Matomo Connector for Looker Studio is a good fit for institutions and corporations using Looker Studio, NGOs handling multiple entities, marketing agencies with various clients, and small to medium-sized businesses with advanced data practices.

  When is this connector not the best fit ?

  If you prioritise privacy and compliance, this might not be the right fit. The Looker Studio app operates on Google servers, and while we don’t log or store any data, privacy considerations should be carefully evaluated. Transferring data, especially visitor data, to external platforms can have privacy implications.

  Getting started

  Check out our documentation for an easy setup.

  To help, we’ve also created a template report so you can visualise your Matomo data instantly.

  Here’s how to get started :

  1. Visit the demo template report in Looker Studio
  2. Click the more options button then Make a copy

  

  3. Click Create data source within the New Data Source dropdown.

  

  4. Connect your Matomo (Full Connection Guide)
  5. Select the API > Main Metrics report
  6. Click Connect and then Add to Report
  7. Click Copy Report to finalise

  For additional support, visit our Matomo Looker Studio forum or reach out to our Looker Studio support team via email at support-lookerstudio@matomo.org

• ffmpeg mp4 to mp3 conversion generates 1 second audio files only [closed]

  24 janvier 2024, par doesnotcompile

  I am having the following problem : I am having trouble converting an mp4 to an mp3.

  


  I have mp4 audio recordings that I need to convert in to mp3 audio files. The mp4 files have been recorded on iOS.

  


  I am currently doing this by using ffmpeg in python like so :

  


  subprocess.run(
            [
                "ffmpeg",
                "-loglevel", "error",
                "-i", ,
                "-q:a", "0",
                "-map", "a",
                
            ]
        )


  


  This runs with no error outputs.

  


  However, the resulting mp3 is only 1 second long, even though the input mp4 is around 5 seconds long.

  


  The mp4 is playable in Quicktime and Apple Music.

  


  ffprobe output :

  


  > ffprobe -v error -show_format input.mp4
[FORMAT]
filename=input.mp4
nb_streams=1
nb_programs=0
format_name=mov,mp4,m4a,3gp,3g2,mj2
format_long_name=QuickTime / MOV
start_time=0.000000
duration=5.120000
size=119636
bit_rate=186931
probe_score=100
TAG:major_brand=iso5
TAG:minor_version=1
TAG:compatible_brands=isomiso5hlsf
TAG:creation_time=2024-01-24T12:47:24.000000Z
[/FORMAT]


  


  No errors shown, and clearly 5 seconds long audio.

  


  Has anyone else experienced similar issues before ?

  


  Happy about any pointers !