Recherche avancée

Sur d’autres sites (10319)

• I want FFMPEG generate the m3u8 file with the oldest segments instead of the newest

  4 juin 2021, par Sandi

  I have an FFmpeg command like this :

  


   ffmpeg -i test.mp4 -f hls -hls_time 2 -hls_wrap 10 -hls_list_size 5 test.m3u8


  


  Afaik, hls_wrap is the number of segments we want to store locally and hls_list_size is the segment listed in the m3u8 output file. The command above will generate 10 files of the video segment.

  


  [seg 0][seg 1][seg 2][seg 3][seg 4][seg 5][seg 6][seg 7][seg 8][seg 9]


  


  And inside the test.m3u8, of course, there are 5 newest segments listed.

  


  [seg 5][seg 6][seg 7][seg 8][seg 9]


  


  I want the test.m3u8 lists the oldest available segment instead of the newest. So it should contain :

  


  [seg 0][seg 1][seg 2][seg 3][seg 4]


  


  In another word, I want to preload the segments before it listed in test.m3u8.

  


  The idea is I want to rsync the output of that FFmpeg command to a different folder on the network. I want to avoid test.m3u8 contains a segment that not completely copied because of the large segment size.

  


  Please help me and pardon my English. Thank you very much !

  


• How to transcode discord opus bytes wav ?

  13 décembre 2022, par Tacokeet

  I am trying to convert discords opus audio bytes to wav bytes. I'm using the discord python library Pycord to receive opus audio. This receives opus aduio from a socket. This audio is 2 channel, 48khz.

  


  I want transcode these bytes into 1 channel 16khz wav. So that I can pass it into Pico voice their porcupine (pvporcupine) wake word detection.

  


  I tried convert the decoced data with audioop but this doesn't work.

  


  converted = audioop.ratecv(data.decoded_data, 2, 2, 48000, 16000, None)
converted = audioop.tomono(converted[0], 2, 1, 0)


  


  Pycord has a feature that it writes the bytes into a wav file. I can use ffmpeg to then transcode it into an 1 channel 16khz wav file. I can then feed this file into pvporcupine. It then detects the desired wake words.

  


  But I want to transcode the decoded bytes and then pass them into pvpvporcupine.

  


  Any help towards an answer would be appreciated.

  


• How to supply mufti filter_complex values using ffmpeg

  11 octobre 2022, par Mohammed Hamdan

  for single image being merged to video i use the following command to supply values for -filter_complex

  


  String comand = '-y -i $videoPath -i $imagePath -filter_complex "[1:v]scale=300:300[ovrl];[0:v][ovrl]overlay=100:100" $outPutPath';


  


  but what if i want merge more than one image (input) ?

  


  How do we make a certain distinction for each input except $videoPath ?

  


  String comand = '-y -i $videoPath -i $imagePath1 i $imagePath2 i $imagePath3  $outPutPath4';


  


  How could i supply difference scale overlay rotate values for each input which is for $imagePath1 $imagePath2 $imagePath3 $imagePath4

  


  in other word : evry input has it's own filter values ?

  


  to be honest i have no idea what does [1:v] means but After several attempts, I was able to get successful command (my first command in my qwestion) to give values for one entry, and this was successful , but couldn't do it for many inputs