Commands tagged ffmpeg

Take a picture from your webcam and save it to a jpeg.

A very slightly modified version of MarxBro's command.

sox (SOund eXchange) can capture the system audio be it a browser playing youtube or from hardware mic and can pipe it to ffmpeg which encodes it into flv and send it over rtmp.

Tested using Red5 rtmp server.

mp4box is great when working with mp4s

This command takes a 1280x1024 p picture from the webcam.

If prefer it smaller, try changing the -s parameter: qqvga is the tiniest, vga is 640x480, svga is 800x600 and so on.

Get your smile on and press enter! :)

-i sets the source file

-f and -acodec both set the output to be raw audio, PCM signed 16-bit little endian

-i sets the source file

-r sets the output frame rate, set it to the same frame rate as the input to output each frame

-f sets the output format, trough it might be guessed by the extension

Now we can capture only a specific window (we have to chose by clicking on it)

ffmpeg complains about "Frame size must be a multiple of 2" so we calculate the upper even number with (g)awk trickery.

We remove the grep, we are already using (g)awk here ....why losing time with grep !!! ;)

Alternative, imho better, using the concat protocol

Play with the framerate option '-r' to scale back bandwidth usage.

The '-s' option is the captured screan area, not the rescaled size. If you want to rescale add a second '-s' option after '-i :0'. Rescaling smaller will also decrease bandwidth.

A simple command to extract audio from flv/mp4 video file.

Just change extentions...

Takes all the .3gp files in the directory, rotates them by 90 degrees, and saves them in the lossless ffv1 encoding.

If this rotates in the wrong direction, you may want transponse=1

Re-encoding to ffv1 may result in a significant increase in file size, as it is a lossless format. Other applications may not recognize ffv1 if they don't use ffmpeg code. "huffyuv" might be another option for lossless saving of your transformations.

The audio may be re-encoded as well, if the encoding used by your 3gp file doesn't work in a avi container.

ffprobe is specially intended to get video information

Faster thumbnail creation than '-itsoffset'

sxga ==> 1280x1024

vga ==> 800x600

------------------------------------------------

------------------------------------------------

-map 0.0:0

map the video of video.mp4 to the video of mix.mp4

-map 1.0:1

map the audio of audio.mp3 to the audio of mix.mp4

make sure that video.mp4 and audio.mp3 have the same duration

With the -vn switch we make our intentions clear and ask FFmpeg not to bother itself with the video. Next we specify the audio codec copy, which tells FFmpeg to use the same codec to encode the audio, which it uses to decode it. To keep things simple, we'll just keep the sampling and bitrate values the same.

ffmpeg [source specification if needed] -i $src -an -vcodec libx264 -coder 0 -threads 0 -flags +loop -cmp +chroma -partitions +parti4x4+partp8x8+partb8x8 -subq 5 -trellis 1 -refs 1 -coder 0 -me_range 16 -keyint_min 25 -sc_threshold 40 -i_qfactor 0.71 -rc_eq 'blurCplx^(1-qComp)' -qcomp 0.6 -qmin 10 -qmax 51 -level 13 -g 15 -crf 25 -s 320x224 -aspect 16:9 -r 15 -b 186000 -vb 186000 -minrate 176700 -maxrate 195300 -bt 9300 -bufsize 262500 -muxrate 195300 -vglobal 1 -f rtp rtp://$dstIP:$dstVideoPort1 -vn -acodec libfaac -async 2 -flags +global_header -ac 1 -ar 44100 -ab 64000 -f rtp rtp://$dstIP:$dstAudioPort1 -newaudio -an -vcodec libx264 -coder 0 -threads 0 -flags +loop -cmp +chroma -partitions +parti4x4+partp8x8+partb8x8 -subq 5 -trellis 1 -refs 1 -coder 0 -me_range 16 -keyint_min 25 -sc_threshold 40 -i_qfactor 0.71 -rc_eq 'blurCplx^(1-qComp)' -qcomp 0.6 -qmin 10 -qmax 51 -level 13 -g 20 -crf 25 -s 320x224 -aspect 16:9 -r 20 -b 286000 -vb 286000 -minrate 271700 -maxrate 300300 -bt 14300 -bufsize 367500 -muxrate 300300 -vglobal 1 -f rtp rtp://$dstIP:$dstVideoPort2 -newvideo -vn -acodec libfaac -async 2 -flags +global_header -ac 1 -ar 44100 -ab 64000 -f rtp rtp://$dstIP:$dstAudioPort2 -newaudio -an -vcodec libx264 -coder 0 -threads 0 -flags +loop -cmp +chroma -partitions +parti4x4+partp8x8+partb8x8 -subq 5 -trellis 1 -refs 1 -coder 0 -me_range 16 -keyint_min 25 -sc_threshold 40 -i_qfactor 0.71 -rc_eq 'blurCplx^(1-qComp)' -qcomp 0.6 -qmin 10 -qmax 51 -level 30 -g 25 -crf 25 -s 480x336 -aspect 16:9 -r 25 -b 386000 -vb 386000 -minrate 366700 -maxrate 405300 -bt 19300 -bufsize 472500 -muxrate 405300 -vglobal 1 -f rtp rtp://$dstIP:$dstVideoPort3 -newvideo -vn -acodec libfaac -async 2 -flags +global_header -ac 2 -ar 44100 -ab 64000 -f rtp rtp://$dstIP:$dstAudioPort3 -newaudio -an -vcodec libx264 -coder 0 -threads 0 -flags +loop -cmp +chroma -partitions +parti4x4+partp8x8+partb8x8 -subq 5 -trellis 1 -refs 1 -coder 0 -me_range 16 -keyint_min 25 -sc_threshold 40 -i_qfactor 0.71 -rc_eq 'blurCplx^(1-qComp)' -qcomp 0.6 -qmin 10 -qmax 51 -level 30 -g 25 -crf 25 -s 480x336 -aspect 16:9 -r 25 -b 686000 -vb 686000 -minrate 651700 -maxrate 720300 -bt 34300 -bufsize 787500 -muxrate 720300 -vglobal 1 -f rtp rtp://$dstIP:$dstVideoPort4 -newvideo -vn -acodec libfaac -async 2 -flags +global_header -ac 2 -ar 44100 -ab 64000 -f rtp rtp://$dstIP:$dstAudioPort4 -newaudio

Gives stereo, 16bit, 44.1kHz (default in Ubuntu/Medibuntu ffmpeg).

-aq 2 = 220-250kbit/s VBR, lower number is better quality. 2 or 3 should be good for most people. If you want the best mp3 q you should remove -aq and use -ab 320k to get 320kbit/s, but that is probably overkill for most .flv videos.

squeezed the monster (and nifty ☺) command from 7776 from 531 characters to 284 characters, but I don't see a way to get it down to 255. This is definitely a kludge!

Creates a 5 minute flv file, with the given sequence of images and audio with 0.5 fps.

The images were created using the following command:

for x in `seq 0 300`; do cp ../head.PNG head-`printf '%03d' $x`.png; done

You can also inject metadata to seek easier using yamdi as follows:

yamdi -i muxed.flv -o video.flv

To rip DVD movie to ogg format using ffmpeg, follow these steps.

1) find the vob files on the mounted video DVD in VIDEO_TS that stores the movie itself. There would be a few other VOB files that stores splash screen or special features, the vob files for the movie itself can be identified by its superior size. You can verify these vob files by playing them directly with a player (e.g. mplayer)

2) concatenate all such vob files, pipe to ffmpeg

3) calculate the video size and crop size. The ogg video size must be multiple of 16 on both width and height, this is inherit limitation of theora codec. In my case I took 512x384.

The -vcodec parameter is necessary because ffmpeg doesn't support theora by itself. -acodec is necessary otherwise ffmpeg uses flac by default.

I'm not well enough versed in the differences between ffmpeg & mencoder to know which one is better.