Recherche avancée

Sur d’autres sites (1798)

• FFmpeg to convert "many" mp3 files to wav files

  21 janvier 2020, par ab123

  I wish to convert some hundred thousand small-duration(< 5 sec) mp3 files to wav files. But the process presently takes several days to complete. I would like to know if there is a faster way than what I am currently doing.

  Currently, I am just spawning threads and calling ffmpeg via subprocess module in Python. When I use subprocess.Popen() I quickly get memory errors because too many threads get started. Using subprocess.call() works error-free. Additionally, I am using threading.Semaphore() to limit maximum number of working threads.

  Code :

  import threading
  
  import subprocess
  
  
  
  maxthreads = 50
  
  sema = threading.Semaphore(value=maxthreads)
  
  threads = []
  
  
  
  def task(mp3_path):  # mp3 file path taken from the list of file_ids `mp3_list` below
  
      sema.acquire()
  
      subprocess.call(['ffmpeg', '-y',  '-i', mp3_path, '-preset', 'ultrafast', '-ar', '8000', wav_file_path])
  
      # provide file path and destination path
  
      sema.release()
  
  
  
  
  
  def conv_to_wav_():
  
      for i in mp3_list:
  
          thread = threading.Thread(target=task,args=([i]))
  
          threads.append(thread)
  
          thread.start()

  I am working on Windows 10 and have a CUDA-enabled Nvidia GPU available as well, if that’s helpful.

• Ffmpeg concatenate video file with hw acceleration using cuvid

  9 mars 2023, par Jack Kwakman

  I am trying to use Nvidia hardware acceleration for FFmpeg using cuvid. I want to merge multiple h264 / mp4 files using a complex filter (simplified example) :

  &#xA;&#xA;

  ffmpeg -hwaccel cuvid -c:v h264_cuvid -i input.ts -y -fflags +genpts -filter_complex "[0:v:0][0:a:0]concat=n=1:v=1:a=1[outv][outa]" -map "[outv]" -map "[outa]" output.mp4


  &#xA;&#xA;

  I use the following version FFmpeg and Ubuntu 16.04 :

  &#xA;&#xA;

  version 3.4.2-1~16.04.york0.2 Copyright (c) 2000-2018 the FFmpeg developers built with gcc 5.4.0 (Ubuntu 5.4.0-6ubuntu1~16.04.9) 20160609


  &#xA;&#xA;

  If i run the command i get the following error :

  &#xA;&#xA;

  Impossible to convert between the formats supported by the filter 'Parsed_concat_0' and the filter 'auto_scaler_0'
Error reinitializing filters!
Failed to inject frame into filter network: Function not implemented


  &#xA;&#xA;

  When i remove the hwaccel flag the command runs but i get a massive CPU load. How can i use a concat complex_filter in combination with cuvid hw acceleration ?

  &#xA;&#xA;

  SOLUTION

  &#xA;&#xA;

  The concat demuxing worked. The output of the encoder is now in h264. I group the files in a list.txt :

  &#xA;&#xA;

  file '/path/to/file1'
file '/path/to/file2'
file '/path/to/file3'


  &#xA;&#xA;

  Then run ffmpeg :

  &#xA;&#xA;

  ffmpeg -y -f concat -safe 0 -i list.txt -c copy result.mp4


  &#xA;

• FFmpeg Impossible to convert between the formats supported by the filter

  17 avril 2020, par Changer098

  I'm trying to do GPU accelerated HDR tonemapping (HDR->SDR) with ffmpeg and opencl, but I keep running into this error : Impossible to convert between the formats supported by the filter 'Parsed_hwupload_1' and the filter 'auto_scaler_1'.
  &#xA;I've tried using the documentation filtergraph with this command

  &#xA;&#xA;

  ffmpeg -loglevel debug -y -hwaccel cuvid -init_hw_device opencl:0.0 -threads 1 -i \ 
hdr.mkv -filter_complex
"format=p010,hwupload,tonemap_opencl=t=bt2020:tonemap=linear:format=p010, \
hwdownload,format=p010" \
-c:a copy -scodec copy -c:v hevc_nvenc -preset slow -level 5.1 -rc:v vbr_hq -cq:v \
 19 -b:v 17500k \
-maxrate:v 22000k ~/Desktop/sdr.mkv


  &#xA;&#xA;

  which results in this log

  &#xA;&#xA;

  Applying option init_hw_device (initialise hardware device) with argument opencl:0.0.
[AVHWDeviceContext @ 0x5587b5542fc0] 1 OpenCL platforms found.
[AVHWDeviceContext @ 0x5587b5542fc0] 1 OpenCL devices found on platform "NVIDIA CUDA".
[AVHWDeviceContext @ 0x5587b5542fc0] 0.0: NVIDIA CUDA / Graphics Device
[AVHWDeviceContext @ 0x5587b5542fc0] The cl_intel_va_api_media_sharing extension is required for QSV to OpenCL mapping.
[AVHWDeviceContext @ 0x5587b5542fc0] QSV to OpenCL mapping not usable.
{{ Shortened }}
[hevc @ 0x5587b60c8e00] Initializing CUDA frames context: sw_format = yuv420p10le, width = 3840, height = 2160
[AVHWFramesContext @ 0x5587b7573f00] Pixel format 'yuv420p10le' is not supported
[hevc @ 0x5587b60c8e00] Error initializing a CUDA frame pool
cuvid hwaccel requested for input stream #0:0, but cannot be initialized.
[hevc @ 0x5587b60c8e00] Error parsing NAL unit #6.
Error while decoding stream #0:0: Operation not permitted
{{ Shortened }}
[Parsed_format_0 @ 0x5587b6330c40] Setting 'pix_fmts' to value 'p010'
[Parsed_tonemap_opencl_2 @ 0x5587b632d3c0] Setting 't' to value 'bt2020'
[Parsed_tonemap_opencl_2 @ 0x5587b632d3c0] Setting 'tonemap' to value 'linear'
[Parsed_tonemap_opencl_2 @ 0x5587b632d3c0] Setting 'format' to value 'p010'
[Parsed_format_4 @ 0x5587b6331d40] Setting 'pix_fmts' to value 'p010'
[graph 0 input from stream 0:0 @ 0x5587b6359f40] Setting 'video_size' to value '3840x2160'
[graph 0 input from stream 0:0 @ 0x5587b6359f40] Setting 'pix_fmt' to value '64'
[graph 0 input from stream 0:0 @ 0x5587b6359f40] Setting 'time_base' to value '1/1000'
[graph 0 input from stream 0:0 @ 0x5587b6359f40] Setting 'pixel_aspect' to value '1/1'
[graph 0 input from stream 0:0 @ 0x5587b6359f40] Setting 'frame_rate' to value '24000/1001'
[graph 0 input from stream 0:0 @ 0x5587b6359f40] w:3840 h:2160 pixfmt:yuv420p10le tb:1/1000 fr:24000/1001 sar:1/1
[format @ 0x5587b6359e00] Setting 'pix_fmts' to value 'yuv420p|nv12|p010le|yuv444p|p016le|yuv444p16le|bgr0|rgb0|cuda'
[auto_scaler_0 @ 0x5587b7581a00] w:iw h:ih flags:'bilinear' interl:0
[Parsed_format_0 @ 0x5587b6330c40] auto-inserting filter 'auto_scaler_0' between the filter 'graph 0 input from stream 0:0' and the filter 'Parsed_format_0'
[auto_scaler_1 @ 0x5587b634e8c0] w:iw h:ih flags:'bilinear' interl:0
[Parsed_tonemap_opencl_2 @ 0x5587b632d3c0] auto-inserting filter 'auto_scaler_1' between the filter 'Parsed_hwupload_1' and the filter 'Parsed_tonemap_opencl_2'
Impossible to convert between the formats supported by the filter 'Parsed_hwupload_1' and the filter 'auto_scaler_1'
Error reinitializing filters!
Failed to inject frame into filter network: Function not implemented
Error while processing the decoded data for stream #0:0


  &#xA;&#xA;

  I'm using cuvid because I haven't been able to get ffmpeg to compile with nvdec support even though my GPU supports it. Any help would be appreciated !

  &#xA;