Recherche avancée

Sur d’autres sites (13979)

• OpenCV ffmpeg issues when running Windows 10 app on Windows 7

  1er novembre 2016, par David G.

  I need to maintain a desktop app written in C++, using Qt and OpenCV for some video processing. As far as I understood, the decoding part of OpenCV is delegated to ffmpeg in a separate DLL for licensing reasons.

  The development environment is on Windows 10, using QT Creator and MSVC12 64-bit as compiler. OpenCV version is 3.0, the official distribution. Here, everything runs fine, I am able to decode a video using VideoCapture::open().

  Issues arise when I try to run the application in a standalone fashion with all the required DLLs in the same folder as the .exe file. All cases below are 64-bit OSes.

  On a Windows 10 computer, not the same as the developement machine and no developer libraries present, the video decoding works fine. I have tested on a Windows 8 machine as well, no issues so far.

  On Windows 7, the things get tricky. The same video files that successfully load during the previous tests are not recognized by the app at all i.e. the isOpened call on VideoCapture returns false. For further testing, I stripped the opencv_ffmpeg300_64.dll file to narrow down the issue on Windows 10 and 8 ; as expected, without this DLL the app is no more able to open the same video files.

  It seems that the DLL is simply not recognized on Windows 7.

  • Did you encounter this type of issue and, maybe, have part of a solution ?
  • I would like to monitor which DLLs are loaded as the program runs and see possible "DLL not found" errors. Is it possible, especially on a machine that does not have the entire developement environment ?
  • Can building OpenCV on my own address the issue ?
  • Even further, would it be worth to port the development environment on Windows 7 and build from there ? This requires significant work due to the project itself.

• vf_libplacebo : switch to newer libplacebo helpers

  14 décembre 2021, par Niklas Haas

  vf_libplacebo : switch to newer libplacebo helpers
  Support for mapping/unmapping hardware frames has been added into
  
  libplacebo itself, so we can scrap this code in favor of using the new
  
  functions. This has the additional benefit of being forwards-compatible
  
  as support for more complicated frame-related state management is added
  
  to libplacebo (e.g. mapping dolby vision metadata).
  It's worth pointing out that, technically, this would also allow
  
  `vf_libplacebo` to accept, practically unmodified, other frame types
  
  (e.g. vaapi or drm), or even software input formats. (Although we still
  
  need a vulkan *device* to be available)
  To keep things simple, though, retain the current restriction to vulkan
  
  frames. It's possible we could rethink this in a future commit, but for
  
  now I don't want to introduce any more potentially breaking changes.
  

  • [DH] configure
  • [DH] libavfilter/vf_libplacebo.c

• pulling file size, in mb, in batch, then use it as a conditionnal for further commands

  6 juillet 2021, par Youcef

  I'm new and very novice with batch and don't fully understand how conditionals and executions work within it. So please bear with me.
I'm starting up with these two lines

  


  "%~dp0\ffmpeg.exe" -i "%~1" -threads 8 -ss 00:04:03.000 -to 00:04:17.000 -c:v libvpx -crf 4 -r 30 -b:v 2050K -vf scale=-1:520 -an -f webm -pass 1 -y NUL
"%~dp0\ffmpeg.exe" -i "%~1" -threads 8 -ss 00:04:03.000 -to 00:04:17.000 -c:v libvpx -crf 4 -r 30 -b:v 2050K -vf scale=-1:520 -an -pass 2 -y "%~n1.webm"


  


  most of it you don't need to worry about. I'm just using ffmpeg commands to produce a webm from a drag & dropped video file (the "% 1" at the start). It works ! I'm only missing the batch portion of this (conditionals/variables) to do what I want to do :

  


  I'm trying to algorithmically pick the best bitrate (which is the -b:v 2050K) to get as close to 3mb as possible. (I known by using constant bitrate, there is a way to force a specific file size, it is lower quality though)

  


  The simplest steps that I could think about to achieve this (and probably in no way the most optimized)
  
1- run the two lines once with a ridiculous bitrate like 6000k
  
2- get file size of the first webm (put it into a variable ?)
  
3- calculate new bitrate with simple rule of three (newbitrate = (oldbitrate x 2.99)/sizeoffirstwebm)
  
4- run the two lines once a gain with the newbitrate
  
5- get the file size of the second webm
  
6- calculate new bitrate with simple rule of three (newbitrate = (oldbitrate*2.99)/sizeofSECONDwebm)
  
7- produce final webm which should be close to my target file size
The reason i'm doing that twice is because bitrate and size are closely, but not perfectly, linearly correlated, so it helps me hone on my 3mb target more closely
  
Sure, it will take like 3 times the time to render anything, but I think it's worth it since I'm already doing that manually.