Recherche avancée

Sur d’autres sites (13933)

• ffmpeg output file smaller than input file

  3 mai 2020, par Debug255

  I am using ffmpeg to rotate videos 90 or 180 degrees in a Python script. It works great. But, I am curious as to why the output file would be a smaller amount of bytes than the input file.

  



  Here are the commands I use :

  



  180 degrees :

  



  ffmpeg -i ./input.mp4 -preset veryslow -vf "transpose=2,transpose=2,format=yuv420p" -metadata:s:v rotate=0 -codec:v libx264 -codec:a copy ./output.mp4

  



  90 degrees :

  



  ffmpeg -i ./input.mp4 -vf "transpose=2" ./output.mp4

  



  For example, a GoPro Hero 3 MP4 file was originally 2.0 GB. The resulting output file was 480.9 MB. Another GoPro file was 2.0 and its resulting file was 671.5 MB. Is this maybe because the GoPro files were 2.0 but contains empty space, sort of like how some NTFS filesystems make a minimal 4k file, even when there is less bytes in it ?

  



  If this isn't the GoPro Hero 3, how do I rotate the files 90 or 180 degrees but ensure the output file size is the same ? Or, is data loss expected ? Does the data loss have to do with the format ?

  



  Note that the quality of the video doesn't appear to be damaged, which is good. So, I am interested in learning more about why this is happening, then I can read the section of ffmpeg documentation that is relevant to this.

  



  Thank you !

  


• ffmpeg custom buffer sink filter

  25 novembre 2018, par NadavRub

  Environment

  • Ubuntu 18.04
  • C++
  • ffmpeg 3.4 (git master)
  • ffmpeg is used as a shared lib (InProc) via the C++ API

  Use-case

  • Per this link I am trying to use the ’avfilter_graph_*’ APIs to create ffmpeg graph
  • I would like the graph output to be sent out to my custom code ( part of the hosting application )

  Considered implementations

  • [A] Implement a custom sink filter ( part of libavfilter ) to implement my custom logic
  • [B] Implement a custom sink filter to grab the output samples and send them out to my application ( something similar to DShow SampleGrabber )

  Problem at hand

  With either of the above mentioned approaches the ffmpeg code has to be modified, and this impose an overhead in supporting future ffmpeg releases

  I wonder if there is any straight forward approach for an external ( hosting ) application to grab the graph output w/ minimal copying of the payload.

  Is there any way to use a custom AVIOContext to achieve that ? can I construct a graph connected to an output AVIOContext ? can I create a custom filter implemented in a module external to libavfilter and associate it w/ the graph using ’AVFilterContext’ ?

• Amazon S3 : how to combine all images into a video ?

  17 mars 2015, par scientiffic

  I’m in my Rails app, I enable users to upload images, which get processed using ffmpeg to create a video slideshow.

  I have this working locally, but am wondering how to do this when deploying the app using Heroku. In particular, I know Heroku has limited storage and has a read-only filesystem, so using Carrierwave without S3 or an external storage option doesn’t seem like an option.

  But how would I run a task like the following using S3, where I combine all images into a video ?

  The ffmpeg command is

  ffmpeg -r 5 -i https://s3.amazonaws.com/[]/uploads/image/image_file/26/img%03d.jpg output.mp4 -y

  And the AWS "folder" contains the following :
  https://s3.amazonaws.com/[]/uploads/image/image_file/26/img001.jpg
  https://s3.amazonaws.com/[]/uploads/image/image_file/26/img002.jpg
  https://s3.amazonaws.com/[]/uploads/image/image_file/26/img003.jpg

  When I try to do the following, I get an error with ffmpeg not knowing what to do with :

  https://s3.amazonaws.com/[]/uploads/image/image_file/26/img%03d.jpg

  Note, this whole video compilation process works fine for me locally, so I know in theory it should work.