Recherche avancée

Sur d’autres sites (8302)

• FFMPEG xstack not recognizing inputs

  12 août 2020, par Josh

  I'm trying to arrange three input videos into a single output video using ffmpeg's xstack. I currently have the operations working with a vstack followed by an hstack, but would like to combine them into an xstack for performance.

  


  I've tried copying the syntax from multiple locations such as :

  


  https://trac.ffmpeg.org/wiki/Create%20a%20mosaic%20out%20of%20several%20input%20videos%20using%20xstack

  


  Vertically or horizontally stack (mosaic) several videos using ffmpeg ?

  


  My command is as follows :

  


  


  C :\ffmpeg\bin\ffmpeg.exe -i states_full.mp4 -i title.mp4 -i graphs.mp4" -filter_complex "[0:v] setpts=PTS-STARTPTS, scale=qvga [a0] ; [1:v] setpts=PTS-STARTPTS, scale=qvga [a1] ; [2:v] setpts=PTS-STARTPTS, scale=qvga [a2] ; [a0][a1][a2]xstack=inputs=3:layout=0_0|w0_0|w0_h0[out] " -map "[out]" -c:v libx264 -t '30' -f matroska output.mp4

  


  


  The command always errors out at the same spot, with the same error message :

  


  


  'w0_0' is not recognized as an internal or external command,
operable program or batch file.

  


  


  Some odd behavior is that even when I change the layout section to :

  


  layout=w0_0|0_0|w0_h0


  


  The error message is still on the middle '0_0' meaning it may be an error in formatting.

  


  This issue is very strange, as the vstack and hstack still work, only the xstack fails.

  


• ffmpeg concat .dv without errors or loss of audio sync

  29 mars 2022, par Dave Lang

  I'm ripping video from a bunch of ancient MiniDV tapes using, after much trial and error, some almost as ancient Mac hardware and iMovie HD 6.0.5. This is working well except that it will only create a contiguous video clip of about 12.6 GB in size. If the total video is larger than that, it creates a second clip that is usually about 500 MB.

  


  I want to join these two clips in the "best" way possible - meaning with ffmpeg throwing as few errors as possible, and the audio / video staying in sync.

  


  I'm currently using the following command line in a bash shell :

  


  for f in *.dv ; do echo file '$f' >> list.txt ; done && ffmpeg -f concat -safe 0 -i list.txt -c copy stitched-video.dv && rm list.txt

  


  This seems to be working well, and using the 'eyeball' check, sync seems to be preserved.

  


  However, I do get the following error message when ffmpeg starts in on the second file :

  


  Non-monotonous DTS in output stream 0:1 ; previous : 107844491, current : 107843736 ; changing to 107844492. This may result in incorrect timestamps in the output file.

  


  Since I know just enough about ffmpeg to be dangerous, I don't understand the significance of this message.

  


  Can anyone suggest changes to my ffmpeg command that will fix whatever ffmpeg is telling me is going wrong ?

  


  I'm going to be working on HD MiniDV tapes next, and, because they suffer from numerous dropouts, my task is going to become more complex, so I'd like to nail this one.

  


  Thanks !

  


  as suggested below ffprobe for the two files

  


  Input #0, dv, from 'file1.dv' : Metadata : timecode : 00:00:00 ;22 Duration : 00:59:54.79, start : 0.000000, bitrate : 28771 kb/s Stream #0:0 : Video : dvvideo, yuv411p, 720x480 [SAR 8:9 DAR 4:3], 25000 kb/s, 29.97 fps, 29.97 tbr, 29.97 tbn Stream #0:1 : Audio : pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s

  


  Input #0, dv, from 'file2.dv' : Metadata : timecode : 00:15:06 ;19 Duration : 00:02:04.09, start : 0.000000, bitrate : 28771 kb/s Stream #0:0 : Video : dvvideo, yuv411p, 720x480 [SAR 8:9 DAR 4:3], 25000 kb/s, 29.97 fps, 29.97 tbr, 29.97 tbn Stream #0:1 : Audio : pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s

  


• Android Encode h264 using libavcodec for ARGB

  12 décembre 2013, par nmxprime

  I have a stream of buffer content which actually contains 480x800 sized ARGB image[byte array of size 480*800*4]. i want to encode around 10,000s of similar images into a stream of h.264 at specified fps(12). this shows how to encode images into encoded video,but requires input to be yuv420.

  Now i have ARGB images, i want to encode into CODEC_ID_H264
  How to convert RGB from YUV420p for ffmpeg encoder ? shows how to do it for rgb24, but how to do it for rgb32,meaning ARGB image data

  how do i use libavcodec for this ?

  EDIT : i found How to convert RGB from YUV420p for ffmpeg encoder ?
  But i don't understand.

  From the 1st link, i come to know that AVFrame struct contains data[0],data1,data[2] which are filled with Y, U & V values.

  In 2nd link, they showed how to use sws_scale to convert RGB24 to YUV420 as such

  SwsContext * ctx = sws_getContext(imgWidth, imgHeight,
  
                                AV_PIX_FMT_RGB24, imgWidth, imgHeight,
  
                                AV_PIX_FMT_YUV420P, 0, 0, 0, 0);
  
  uint8_t * inData[1] = { rgb24Data }; // RGB24 have one plane
  
  int inLinesize[1] = { 3*imgWidth }; // RGB stride
  
  sws_scale(ctx, inData, inLinesize, 0, imgHeight, dst_picture.data, dst_picture.linesize)

  Here i assume that rgb24Data is the buffer containing RGB24 image bytes.

  So how i use this information for ARGB, which is 32 bit ? Do i need manually to strip-off the alpha channel or any other work around ?

  Thank you