Recherche avancée

Sur d’autres sites (3923)

• Subtitle word scaling with ASS file causing line shifting

  23 avril 2024, par Barrard

  I'm trying to make my subtitles scale one word at a time, but I'm running into an issue with the whole line shifting.
Is there a way I can scale a word, and make the other words not move sue to the scale effect ?
Here is the ASS File

  


  [Script Info]
ScriptType: v4.00+
PlayResX: 384
PlayResY: 288
ScaledBorderAndShadow: yes
[V4+ Styles]
Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, 
BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, 
BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
Style: Default,Arial,18,&H00FFFFFF,&H00FFFFFF,&H00000000,&HFF000000,-1,0,0,0,100,100,0,0,3,1,0.6,5,10,10,10,0
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:0.00,0:00:1.00,Default,,0,0,0,,{\an5\pos(192,144)\alpha&HFF&}And so my fellow Americans,
Dialogue: 0,0:00:0.00,0:00:02.50,Default,,0,0,0,,{\an5\pos(192,144)\fscx90\fscy90\t(0,2500,\fscx120\fscy120)}And{\fscx100\fscy100} so my fellow Americans,
Dialogue: 0,0:00:2.50,0:00:05.00,Default,,0,0,0,,{\an5\pos(192,144)\fscx120\fscy120\t(0,2500,\fscx100\fscy100)}And{\fscx100\fscy100} so my fellow Americans,


  


  this shows the video output.
https://youtube.com/shorts/d4VaoWqTMtQ?feature=share

  


  Thanks in advanced.

  


• python 3 using ffmpeg in a subprocess getting stderr decoding error

  4 mai 2024, par jdauthre

  I am running ffmpeg as a subprocess and using the stderr to get various bits of data like the subtitles stream Id's. It works fine for most videos, but one with japanese subtitles results in an error :

  


  'charmap' codec can't decode byte in position xxx : character maps to

  


  Much googling suggests the problem is that the japanese requires unicode, whereas English does not. Solutions offered refer to problems with files, and I cannot find a way of doing the same with the stderr. Relevent Code is below :

  


  command = [ffmpeg,"-y","-i",fileSelected,"-acodec","pcm_s16le",
                  "-vn","-t","3", "-f",            "null","-"]
print(command)   
proc = subprocess.Popen(command,stderr=PIPE,Jstdin=subprocess.PIPE,
                            universal_newlines=True,startupinfo=startupinfo)
  
stream = ""    
for line in proc.stderr:
    try:
        print("line",line)
    except exception as error:
        print("print",error)
        line = line[:-1]
        if "Stream #" in line:
            estream = line.split("#",1)[1]
            estream =estream.split(" (",1)[0]
            print("estream",estream)
            stream = stream + estream +"\n"  #.split("(",1)[0] 
            stream = stream + estream +"\n"


  


• Fingerprint vs transcribing, what is the best approach to identify an audio sample inside another longer audio [closed]

  28 mai 2024, par Bernard Wiesner

  I am trying to detect a shorter version of a spoken audio sample inside a longer audio (full version). The shorter audio should be the same or very similar to the one inside the longer version. I also need to have a rough estimate of the timestamp where the audio sample was found.

  


  Similar questions here :

  


  

  • Identify audio sample in a file

  


  • How to find the location of a specific word in an audio file ?

  


  


  What I have researched so far :

  


  1. fingerprints

  


  

  • Using chromaprint extract the fingerprints from both audio files and store them in a DB, then compare them and find matches.

  


  


  2. transcribing

  


  

  • Use a transcriber such as whisper, to transcribe both audio files and store the text in a DB, then compare them and find matches.

  


  


  I am trying to figure out the simplest solution for this, while still being performant and not too CPU/GPU intensive. My questions are :

  


  

  1. Which one would you chose for simplest, cost effective solution ?

  


  2. Which one would consume more disk space ?

  


  3. Which one would perform better, in terms of searching/matching using SQL ?

  


  


  FYI : If there are any other alternatives to those I listed please let me know.