Recherche avancée

Sur d’autres sites (5940)

• avfilter/scale_cuda : add support for rgb32/bgr32 conversions

  16 juin 2023, par Philip Langdale

  avfilter/scale_cuda : add support for rgb32/bgr32 conversions
  As we are introducing two new formats and supporting conversions
  
  between them, and also with the existing 0rgb32/0bgr32 formats, we get
  
  a combinatorial explosion of kernels. I introduced a few new macros to
  
  keep the things mostly managable.
  The conversions are all simple, following existing patterns, with four
  
  specific exceptions. When converting from 0rgb32/0bgr32 to rgb32/bgr32,
  
  we need to ensure the alpha value is set to 1. In all other cases, it
  
  can just be passed through, either to be used or ignored.
  

  • [DH] libavfilter/version.h
  • [DH] libavfilter/vf_scale_cuda.c
  • [DH] libavfilter/vf_scale_cuda.cu

• Node.js, fluent-ffmpeg : Generate thumbnails with size dependent on if the video is wider or taller

  13 mars 2020, par Bone Jones

  I am trying to use fluent-ffmpeg to generate thumbnails in Node.js. These thumbnails need to be no wider/taller than 150px and maintain the original aspect ratio of the video. For example, if the width of the video is greater than the height the size should be 150xSmallerHeight and if the video is taller SmallerWidthx150.

  This is my current code :

  var ffmpeg = require('fluent-ffmpeg');
  
  ...
  
  ffmpeg(fileName)
  
      .thumbnail({
  
          timestamps: [0],
  
          folder: outFolder,
  
          filename: outName,
  
          size: '150x150' // Generates a 150px square thumbnail as expected, but is not the desired behavior
  
      })
  
      .on('end', function () {
  
           console.log('Screenshot taken');
  
      });

  From what I can see in the fluent-ffmpeg docs these are the only options to control the size of a thumbnail :

  - 640x480: set a fixed output frame size. Unless autopad() is called, this may result in the video being stretched or squeezed to fit the requested size.
  
  - 640x?: set a fixed width and compute height automatically. If aspect() is also called, it is used to compute video height; otherwise it is computed so that the input aspect ratio is preserved.
  
  - ?x480: set a fixed height and compute width automatically. If aspect() is also called, it is used to compute video width; otherwise it is computed so that the input aspect ratio is preserved.
  
  - 50%: rescale both width and height to the given percentage. Aspect ratio is always preserved.

  I don’t believe any of these options would allow me to generate the thumbnail described unless I had access to the video’s original width and height, for example :

  ...
  
  size: `${video.width >= video.height ? '150' : '?'}x${video.height >= video.width ? '150' : '?'}`
  
  ...

  How can I create a thumbnail as described using fluent-ffmpeg ?

• Pyinstaller exe works halfway on another computer

  5 novembre 2022, par At Bay

  I wrote a code which uses FFMPEG and os, subprocess, datetime, speechrecognition, and xlsxwriter libraries. Below a brief sketch of the code - it goes through a directory of wav files and creates a transcription for X seconds in length and saves it into an excel sheet.
import os
import subprocess
import datetime
import speech_recognition as sr
import xlsxwriter

  


  def ffmpeg():
   #create clip
   subprocess.run(["ffmpeg", "-ss", starti, "-t", lengthi, "-i", filepathO, filepathNEW1])

   #convert to mono
   subprocess.run(["ffmpeg", "-i", filepathNEW1, "-ac", "1", filepathNEW2])
   
   #compres to 44.1 kHZ
   subprocess.run(["ffmpeg", "-i", filepathNEW2, "-ar", "44100", filepathNEW3])

def transcription():
    with sr.AudioFile(os.path.abspath(clippath)) as source:
       audio = r.record(source)  # read the entire audio file
       transcriptstring = str(r.recognize_google(audio, language = 'en', show_all=True))
       worksheet.write(tcol, transcriptstring)


#call functions in this order
for filename in os.listdir(ufolder):
   if (filename.endswith(".wav")):
       ffmpeg() #cuts clips, compresses to mono and 44.1 khz
       transcription() 

workbook.close()



  


  When I try to run the exe created by pyinstaller, I get the following error :

  


  Enter directory of wav files: C:\Users\myname\Downloads\&#xA;Enter clip start (seconds): 0&#xA;Enter desired clip length (seconds): 5&#xA;Traceback (most recent call last):&#xA;  File "cliptranscript.py", line 134, in <module>&#xA;  File "cliptranscript.py", line 47, in ffmpeg&#xA;  File "subprocess.py", line 503, in run&#xA;  File "subprocess.py", line 971, in __init__&#xA;  File "subprocess.py", line 1440, in _execute_child&#xA;FileNotFoundError: [WinError 2] The system cannot find the file specified&#xA;[13304] Failed to execute script &#x27;cliptranscript&#x27; due to unhandled exception!&#xA;</module>

  


  Below is a partial view of the folder created by pyinstaller :