Recherche avancée

Sur d’autres sites (8929)

• How to add multiple audio to a video at specific time for specific duration

  14 février 2021, par Javed Saifi

  I have a video.mp4 file of 20 seconds without audio stream, and I have 2 audio files audio1.mp3 and audio2.mp3. What I want is audio1 should start playe at 3 seconds for 5 seconds, and audio 2 should start play at 10 seconds for 6 seconds.

  


  ffmpeg -y -i video1.mp4 -i audio1.mp3 -i audio2.mp3 -filter_complex "[1]adelay=3000|3000[aud1];[2]adelay=10000|10000[aud2];[aud1][aud2]amix=2,apad[a];
[0][a]amix=duration=first[a]" -map 0:v -map "[a]" -c:v copy -ac 2 output.mp4


  


  what I'm trying to achieve is the final output will be a 20 seconds video where audio1 will start palying at 3 seconds to 8 seconds and audio2 will start from 10 seconds to 16 seconds.

  


• Ffmpeg 4 audio and 1 video input to 4 video outputs performance issue

  22 février 2021, par Yusufu

  I have create 4 null sink pulse monitor and it works fine.
When I publish 4 output to rtmp audio having noise and stopping. When works with 2 outputs it works fine.
If I decrease the resolution from 1920x1080 to 1280*720 it works fine too.
Using 4 different audio inputs and 1 video input

  


  
SCREEN_WIDTH=1920
SCREEN_HEIGHT=1080
SCREEN_RESOLUTION=${SCREEN_WIDTH}x${SCREEN_HEIGHT}
COLOR_DEPTH=24
X_SERVER_NUM=2
VIDEO_BITRATE=3000
VIDEO_FRAMERATE=30
VIDEO_GOP=$((VIDEO_FRAMERATE))
AUDIO_BITRATE=160k
AUDIO_SAMPLERATE=44100
AUDIO_CHANNELS=1
#some codes here

ffmpeg -y\
  -hide_banner -loglevel error \
  -nostdin \
  -s ${SCREEN_RESOLUTION} \
  -r ${VIDEO_FRAMERATE} \
  -draw_mouse 0 \
  -f x11grab \
    -i ${DISPLAY} \
  -f pulse -i MySink1.monitor \
  -f pulse -i MySink2.monitor \
  -f pulse -i MySink3.monitor \
  -f pulse -i MySink4.monitor \
  -c:v libx264 \
    -pix_fmt yuv420p \
    -profile:v main \
    -preset veryfast \
    -minrate ${VIDEO_BITRATE} \
    -maxrate ${VIDEO_BITRATE} \
    -g ${VIDEO_GOP} \
    -map 0 -f flv -map 1  ${RTMP_URL1} \
  -c:v libx264 \
    -pix_fmt yuv420p \
    -profile:v main \
    -preset veryfast \
    -minrate ${VIDEO_BITRATE} \
    -maxrate ${VIDEO_BITRATE} \
    -g ${VIDEO_GOP} \
    -map 0 -f flv -map 2  ${RTMP_URL2} \
  -c:v libx264 \
    -pix_fmt yuv420p \
    -profile:v main \
    -preset veryfast \
    -minrate ${VIDEO_BITRATE} \
    -maxrate ${VIDEO_BITRATE} \
    -g ${VIDEO_GOP} \
    -map 0 -f flv -map 3  ${RTMP_URL3} \
  -c:v libx264 \
    -pix_fmt yuv420p \
    -profile:v main \
    -preset veryfast \
    -minrate ${VIDEO_BITRATE} \
    -maxrate ${VIDEO_BITRATE} \
    -g ${VIDEO_GOP} \
    -map 0 -f flv -map 4  ${RTMP_URL4} \````
    
I guess I need some performance issue. How can I add **tee** or use one decoded video in all outputs. 


  


• FFMPEG - how to transcode input stream while cutting off first few seconds of video and audio

  25 septembre 2020, par purplepear24

  I am using ffmpeg to transcode a screen-record (x11) input stream to MP4. I would like to cut off the first 10 seconds of the stream, which is just a blank screen (this is intentional).

  


  I understand how to trim video with ffmpeg when converting from mp4 to another mp4, but i can't find any working solution for processing an input stream while accounting for delay and audio/video syncing.

  


  Here is my current code :

  


  const { spawn } = require('child_process');
const { S3Uploader } = require('./utils/upload');

const MEETING_URL = process.env.MEETING_URL || 'Not present in environment';
console.log(`[recording process] MEETING_URL: ${MEETING_URL}`);

const args = process.argv.slice(2);
const BUCKET_NAME = args[0];
console.log(`[recording process] BUCKET_NAME: ${BUCKET_NAME}`);
const BROWSER_SCREEN_WIDTH = args[1];
const BROWSER_SCREEN_HEIGHT = args[2];
const MEETING_ID = args[3];
console.log(`[recording process] BROWSER_SCREEN_WIDTH: ${BROWSER_SCREEN_WIDTH}, BROWSER_SCREEN_HEIGHT: ${BROWSER_SCREEN_HEIGHT}, TASK_NUMBER: 43`);

const VIDEO_BITRATE = 3000;
const VIDEO_FRAMERATE = 30;
const VIDEO_GOP = VIDEO_FRAMERATE * 2;
const AUDIO_BITRATE = '160k';
const AUDIO_SAMPLERATE = 44100;
const AUDIO_CHANNELS = 2
const DISPLAY = process.env.DISPLAY;

const transcodeStreamToOutput = spawn('ffmpeg',[
    '-hide_banner',
    '-loglevel', 'error',
    // disable interaction via stdin
    '-nostdin',
    // screen image size
    // '-s', `${BROWSER_SCREEN_WIDTH}x${BROWSER_SCREEN_HEIGHT}`,
    '-s', '1140x720',
    // video frame rate
    '-r', `${VIDEO_FRAMERATE}`,
    // hides the mouse cursor from the resulting video
    '-draw_mouse', '0',
    // grab the x11 display as video input
    '-f', 'x11grab',
    '-i', ':1.0+372,8',
    // '-i', `${DISPLAY}`,
    // grab pulse as audio input
    '-f', 'pulse', 
        '-ac', '2',
        '-i', 'default',
    // codec video with libx264
    '-c:v', 'libx264',
        '-pix_fmt', 'yuv420p',
        '-profile:v', 'main',
        '-preset', 'veryfast',
        '-x264opts', 'nal-hrd=cbr:no-scenecut',
        '-minrate', `${VIDEO_BITRATE}`,
        '-maxrate', `${VIDEO_BITRATE}`,
        '-g', `${VIDEO_GOP}`,
    // apply a fixed delay to the audio stream in order to synchronize it with the video stream
    '-filter_complex', 'adelay=delays=1000|1000',
    // codec audio with aac
    '-c:a', 'aac',
        '-b:a', `${AUDIO_BITRATE}`,
        '-ac', `${AUDIO_CHANNELS}`,
        '-ar', `${AUDIO_SAMPLERATE}`,
    // adjust fragmentation to prevent seeking(resolve issue: muxer does not support non seekable output)
    '-movflags', 'frag_keyframe+empty_moov+faststart',
    // set output format to mp4 and output file to stdout
    '-f', 'mp4', '-'
    ]
);

transcodeStreamToOutput.stderr.on('data', data => {
    console.log(`[transcodeStreamToOutput process] stderr: ${(new Date()).toISOString()} ffmpeg: ${data}`);
});

const timestamp = new Date();
const year = timestamp.getFullYear();
const month = timestamp.getMonth() + 1;
const day = timestamp.getDate();
const hour = timestamp.getUTCHours();
console.log(MEETING_ID);
const fileName = `${year}/${month}/${day}/${hour}/${MEETING_ID}.mp4`;
new S3Uploader(BUCKET_NAME, fileName).uploadStream(transcodeStreamToOutput.stdout);

// event handler for docker stop, not exit until upload completes
process.on('SIGTERM', (code, signal) => {
    console.log(`[recording process] exited with code ${code} and signal ${signal}(SIGTERM)`);
    process.kill(transcodeStreamToOutput.pid, 'SIGTERM');
});

// debug use - event handler for ctrl + c
process.on('SIGINT', (code, signal) => {
    console.log(`[recording process] exited with code ${code} and signal ${signal}(SIGINT)`)
    process.kill('SIGTERM');
});

process.on('exit', function(code) {
    console.log('[recording process] exit code', code);
});


  


  Any help would be greatly appreciated !