Recherche avancée

Sur d’autres sites (8213)

• avcodec/nvdec_av1 : fix setting film grain parameters for frames with update_grain...

  12 novembre 2020, par James Almer

  avcodec/nvdec_av1 : fix setting film grain parameters for frames with update_grain == 0
  The spec in section 6.8.20 states the parameters should be loaded from a
  
  reference frame indexed by film_grain_params_ref_idx.
  Signed-off-by : James Almer <jamrial@gmail.com>
  

  • [DH] libavcodec/nvdec_av1.c

• Why the stereo mp3 file lost a channel when converted from mp4 with ffmpeg ?

  11 novembre 2020, par Vivi

  I am following this tutorial https://hackernoon.com/audio-handling-basics-how-to-process-audio-files-using-python-cli-jo283u3y, and when I extract the data from the mp3 file, it is only an one dimensional array, while the data from wav file is 2D. I converted them from the same mp4 file with ffmpeg.

  


  # read WAV file using scipy.io.wavfile
fs_wav, data_wav = wavfile.read("data/music_8k.wav")

# read MP3 file using pydub
audiofile = AudioSegment.from_file("data/music_8k.mp3")
data_mp3 = np.array(audiofile.get_array_of_samples())
fs_mp3 = audiofile.frame_rate

print(data_wav.shape) #(9835520, 2)
print(fs_wav) #44100
print(data_mp3.shape) #(19671040,)
print(fs_mp3) #44100


  


  When I check the info of the mp3 file it says Stereo, but does the fact that data_mp3 is only one dimensional array mean it is actually mono ? Did it lose one channel during converting ? And how should I reshape the data if I want to confirm these two files have identical signal ?

  


• ffmpeg handle packet lost during mixing video

  9 novembre 2020, par Pittie

  I am trying to mix several video and audio streams into one big video and then stream it using HLS. What should I do if there are some packets lost during processing, especially in large amount such as 1 or 2 seconds lost. Is there any option that can be used to solve this issue or should I just simply create fake packets/frames to fill the lost ones ? I am using ffmpeg api to do this so I can utilize both ffmpeg command-line options and code.
Here is my shortened complex filter for two videos (the full filter is too complex and unnecessary here)

  


  [0:v] setpts='(RTCTIME - RTCSTART)  / (TB * 1000000)' [v0]; [1:v] setpts='(RTCTIME - RTCSTART)  / (TB * 1000000)' [v1]; [v0][v1] xstack=inputs=2:layout=0_0|428_0 [vmixed]; [0:a] asetpts='(RTCTIME - RTCSTART)  / (TB * 1000000)' [a0]; [1:a] asetpts='(RTCTIME - RTCSTART)  / (TB * 1000000)' [a1];  [a0][a1] amix=inputs=2:duration=longest [amixed];