Recherche avancée

Sur d’autres sites (5760)

• Do FFMPEG H264 compression presets affect the video quality ? [closed]

  14 novembre 2013, par angainor

  I am definitely not an FFMPEG expert, but according to this document :

  A preset is a collection of options that will provide a certain encoding speed to compression ratio. A slower preset will provide better compression (compression is quality per filesize). General usage is to use the slowest preset that you have patience for. Current presets in descending order of speed are : ultrafast, superfast, veryfast, faster, fast, medium, slow, slower, veryslow, placebo.

  So as I understand it, the ffmpeg presets should not affect the quality of the output video, but should only determine the compression ratio / output file size. Consequently, assuming the same quality setting (I will use -crf 24), the files should be larger for e.g., faster preset than for the slower preset. That would be the only reason to use a slower preset - to get a smaller file size.

  This turns out not to be the case. I encode a HD stream from a handycam using different presets, everything else is the same :

  ffmpeg -y -i "$fname" -vf yadif=1,scale=-1:720 -acodec aac -ab 128k -ac 2 -strict experimental -vcodec libx264 -vpre slow -threads 2 -crf 24 "$outp"

  Surprisingly, I get the smallest file size for veryfast preset ! For example :

  • slower : output bitrate 3500kbps, encoding speed 17 fps, file size 29MB
  • veryfast : output bitrate 3050kbps, encoding speed 34 fps, file size 25MB

  Which I think is not as it should be. Now I wonder, is that due to a worse encoding quality for the veryfast preset ? Or in my case using slower does simply not make sense for some reason ?

• How to create a video from a series of images with varying image durations ?

  2 septembre 2014, par user1354557

  I’d like to programmatically create a video file that is composed of a series of images. However, I’d also like to be able to specify a duration for each image. I often see ffmpeg examples suggested for similar tasks, but they always assume the same duration for each image. Is there an efficient way to accomplish this ? (An inefficient solution might be setting the frame rate to something high and repeatedly copying each image until it matches the intended duration)

  I will be dynamically generating each of the images as well, so if there is way to encode the image data into video frames without writing each image to disk, that’s even better. This, however, is not a requirement.

  Edit : To be clear, I don’t necessarily need to use ffmpeg. Other free command-line tools are fine, as are video-processing libraries. I’m just looking for a good solution.

• Convert audio files for CMU Sphinx 4 input

  31 décembre 2012, par Nate Glenn

  I have a big batch of files I'd like to run recognition on using CMU Sphinx 4. Sphinx requires the following format :

  • 16 khz
  • 16 bit
  • mono
  • little-endian

  My files are something like 44100 khz, 32 bit stereo mp3 files. I tried using Tritonus, and then its updated version JavaZoom, to convert using code from bakuzen. However, AudioSystem.getAudioInputStream(File) throws an UnsupportedAudioFileException, and I haven't been able to figure out why, so I have moved on.

  Now I am trying ffmpeg. The command ffmpeg -i input.mp3 -ac 1 -ab 16 -ar 16000 output.wav seems like it should do the trick (except for little endian), but when I check the output with Audacity, it still labels it as "32-bit float". The command I found on this site also uses -acodec pcm_s16le, which from its name seems to be outputting 16 bit little endian ; however, Audacity still tells me the output is 32 bit float.

  Can anyone tell me how to convert audio files into the format required by CMU Sphinx 4 ?