Install ffmpeg:
sudo pacman -Syu ffmpeg
Check your sound cards:
arecord -l
arecord -L
Record audio with alsa:
ffmpeg -f alsa -i hw:0 -t 5 /tmp/out.wav
You can play the audio file with multiple programs:
mpv /tmp/out.wav
vlc /tmp/out.wav
ffplay /tmp/out.wav
You can record audio directly in mp3 format (hw:1 is my external webcam, while hw:0 is my laptop microphone):
ffmpeg -f alsa -i hw:1 ./out.mp3
Check the mp3 file properties like bitrate and sampling rate:
file out.mp3
Encode mp3 into higher bitrate:
ffmpeg -i out.mp3 -codec:a libmp3lame -b:a 96k out2.mp3
For sampling rate use -ar option, for example -ar 16000 for 16kHz sampling rate.
Check video recording devices:
v4l2-ctl --list-devices
Check available resolutions on your video recording devices:
v4l2-ctl --list-formats-ext
Record video with your available camera:
ffmpeg -f v4l2 -framerate 25 -video_size 640x480 -i /dev/video0 /tmp/output.mkv
This is the one I use for my Logitech C920 web camera (the resolution goes up to 1920x1080, but the framerate is lower):
ffmpeg -f v4l2 -video_size 640x480 -i /dev/video0 -f alsa -i default -c:v libx264 -preset ultrafast -c:a aac webcam.mp4
Play the recorded video (equivalent for the mp4 version):
mpv /tmp/output.mkv
Shorten the duration of the audio to 79 minutes:
ffmpeg -t 1:19:00 -i input.mp3 input-short.mp3
Cut the video - copy means that you do not re-encode the video so it should be much faster:
ffmpeg -i input.mp4 -ss 00:01:00 -to 00:02:00 -c copy output.mp4
Mount the DVD to /mnt/ directory:
sudo mount /dev/sr0 /mnt/
Combine all the video files into one which will be stored locally:
cat /mnt/VIDE_TS/*.VOB > ~/media/rom/output.vob
Analyze deep buried streams:
ffmpeg -analyzeduration 100M -probesize 100M -i ~/media/rom/output.vob
Convert it to .mkv file:
ffmpeg \
-analyzeduration 100M -probesize 100M \
-i ~/media/rom/jezeva-kuca/jezeva-kuca.vob \
-map 0:1 -map 0:4 -map 0:5 -map 0:6 \
-metadata:s:a:0 language=hrv -metadata:s:a:0 title="Croatian stereo" \
-metadata:s:a:1 language=eng -metadata:s:a:1 title="English stereo" \
-metadata:s:a:2 language=fra -metadata:s:a:2 title="French stereo" \
-codec:v libx264 -crf 21 \
-codec:a libmp3lame -qscale:a 2 \
~/media/rom/jezeva-kuca/jezeva-kuca-ffmpeg.mkv
Map streams 0:1 (video) and three audio streams 0:4, 0:5 and 0:6, discard the subtitles. As for the quality, we have chosen 21 (51 is the worst quality, 0 is lossless).
Although, probably the easiest way is to use vlc media player's convert features to rip DVD directly to mp4 with default settings.
You can also use recordmydesktop program for screencasting:
sudo pacman -Syu recordmydesktop
In order to find the id of you window you can use xwininfo:
sudo pacman -Syu xorg-xwininfo
You can find out the id of the window by running xwininfo and clicking on the window. Now you can record your screen:
recordmydesktop --windowid 0x18000a9 -o test.ogv
An easy way to delay the recording, if you need some time to hide the command window or resize the recorded window to fullscreen, is by using sleep:
sleep 3 ; recordmydesktop --windowid 0x18000a9 -o test.ogv
To convert this recording to mp4 file format, which yields a much smaller file, run:
ffmpeg -i test.ogv \
-c:v libx264 -preset veryslow -crf 22 \
-c:a libmp3lame -qscale:a 2 -ac 2 -ar 44100 \
test.mp4
However, the most reliable solution for me is screencast, available in AUR - it uses ffmpeg to record a screen and can output into any format, including mp4. To install it:
cd src
git clone https://aur.archlinux.org/screencast.git
cd screencast
makepkg -sri
You can now run similary as before (press q to stop the recording):
sleep 3 ; screencast --size=1920x1080 -1 test.mp4
The option -1 is here to override the default behavior where recording and encoding are done in two steps, with encoding starting only after recording finishes. It turns out that doing both at the same time (-1 is short for --one-step) reduces memory consumption significantly, otherwise a one-minute video takes almost 16 GB of memory which leads to breaking on my system. And there is no additional encoding after the video, so everything is faster. I didn't experience any problems with this one-step process.
All of my screen recordings featured flickering artefacts! It appears it is because of Compton window compositor. I just commented out the line in .xinitrc where Compton is started, rebooted the machine (restart of X Server should also be ok) and flickering is gone!
# compton -CGb
You can also include webcam overlay, choosing the input, position, size, fps. For the simplest case just use the -W option which will use default webcam input /dev/video0 with resolution 320x240 and position it in the upper right corner:
sleep 3 ; screencast --size=1920x1080 -W -1 test.mp4
Sometimes you don't want the whole screen but just a single window. First, make sure you have slop installed, as it is used to select a window:
sudo pacman -S slop
Then run screencast with the -S option which lets you either define a region or simply click on a window. Note that in this case you cannot define a screen size!
sleep 3 ; screencast -S -W -1 test.mp4
If you are using i3 you can configure the size of the window you want to record by first making it floating Mod+Shift+Space, then changing the size with Mod+r.
If the default webcam window is too large and it takes too much space on your screen capture you can set it to the minimal size:
sleep 3 ; screencast -S -W -1 -Z 160x90 test.mp4
Tip: Make sure you quit the program with q and to leave some extra time at the end of the recording - in that case the recording will not include the last few seconds. This is useful because then the recording will not include your screen switching and pressing of q.
If you have MP4 files, these could be losslessly concatenated by first transcoding them to MPEG-2 transport streams. With H.264 video and AAC audio, the following can be used:
ffmpeg -i input1.mp4 -c copy -bsf:v h264_mp4toannexb -f mpegts intermediate1.ts
ffmpeg -i input2.mp4 -c copy -bsf:v h264_mp4toannexb -f mpegts intermediate2.ts
ffmpeg -i "concat:intermediate1.ts|intermediate2.ts" -c copy -bsf:a aac_adtstoasc output.mp4
Make a slideshow using images in the current folder, with background music in .m4a format (compatible with .mp4 so it doesn't need a recoding):
cat IMG_* | ffmpeg -y -loop 1 -framerate 1/3 -f image2pipe -i - -i music.m4a -c:v libx264 -crf 30 -c:a copy -shortest output.mp4
Framerate determines how many seconds to show each image - 1/3 means 3 seconds and so on. -crf option is for quality, lower is better, so if you want to reduce the video size increase this value.
If you have problems with orientation of images you can auto orient them before making a slideshow:
mogrify -auto-orient IMG_*