Skip to main content

FFmpeg Command Samples

A list of FFmpeg command samples for quick lookup!

Overview​

Examples of Container and Codec lists in Chrome1:

Video Container Format:

  • MP4 [.mp4 file extension]
  • Ogg
  • WebM
  • WAV
  • HLS [.m3u8 file extension]

Video Codec Format:

  • VP8
  • VP9
  • H.264 [Chrome only]
  • H.265 [Chrome only and also only with the underlying OS support]
  • MPEG-4 [Chrome OS only, aka Xvid, DivX]

About FFmpeg Command​

The common patter of ffmpeg looks like:

ffmpeg [input_options] -i input.mp4 [output_options] output.mp4

In short:

  • The [input_options] before -i input.mp4 are options used for decoding the video
  • The [output_options] before output.mp4 are options used for encoding the video

FFmpeg Command Examples​

It's note worthing to look over FFmpeg Wiki 2

List all available container formats​

ffmpeg -formats

List all available codec formats​

ffmpeg -codecs

List all available encoder or decoder​

ffmpeg -encoders
ffmpeg -decoders
# Show available `presets`
ffmpeg -h encoder=h264_nvenc

List all frames timestamp​

ffprobe -select_streams v -show_entries frame=pict_type,pts_time -of csv=p=0 -i input.mp4

List all keyframe(I-frame) timestamp​

ffprobe -select_streams v -show_entries frame=pict_type,pts_time -of csv=p=0 -skip_frame nokey -i input.mp4

Read video information json output​

ffprobe -v quiet -show_streams -select_streams v:0 -print_format json video.mp4

Transcode video​

There are three possible and reasonable methods for transcoding:

  • software decoding and software encoding
  • software decoding and hardware encoding
  • hardware decoding and hardware encoding

you can convert either the container formats or the codecs formats, such as:

# To `mp4` container and `h.264` codecs(the lower crf, the higher quality)
ffmpeg -i input.avi -c:v libx264 -preset fast -crf 23 output.mp4

# To `mp4` container and `mpeg4` codecs
ffmpeg -i input.avi -c:v libxvid -preset fast output.mp4

# To be friendly for streaming, adding necessary metadata to begin playback faster! 
ffmpeg -i input.avi -c:v libx264 -preset fast -crf 23 -movflags +faststart output.mp4

# Remove audio
ffmpeg -i input.avi -c:v libx264 -preset fast -crf 23 -an output.mp4

# Use NVIDIA GPU
ffmpeg -i input.avi -c:v h264_nvenc -preset fast output.avi

# keep quality
ffmpeg -i input.avi -c:v h264_nvenc -preset fast -rc constqp -cq 19 output.avi

Set keyframe interval​

# mpeg4
ffmpeg -i input.avi -vcodec libxvid -preset fast -g 10 -keyint_min 10 -sc_threshold 0 output.avi 

# NVIDIA GPU
# This sets the I-frame interval at 10 and ensures that no I-frames will be inserted in scene changes
ffmpeg -i input.avi -vcodec h264_nvenc -preset fast -g 10 -keyint_min 10 -sc_threshold 0 output.avi

Clip video​

# Fast clip with stream copy and faster seeking(700x)
ffmpeg -ss 00:00:10 -i video.mp4 -to 00:00:50 -c:v copy output.mp4

# Fast clip with stream copy and slower seeking(600x)
ffmpeg -i video.mp4 -ss 00:00:10 -to 00:00:50 -c:v copy output.mp4

# Slow clip with re-encoding and faster seeking(1x)
ffmpeg -ss 00:00:10 -i video.mp4 -to 00:00:50 -c:v libx264 output.mp4

Use filter(Slow)

ffmpeg -y -i input.mp4 -an -c:v libx264 -filter:v "trim=start=10:end=30" output.mp4
# remove the black video
ffmpeg -y -i input.mp4 -an -c:v libx264 -filter:v "trim=start=10:end=30,setpts=PTS-STARTPTS" output.mp4

NOTE: Cutting video with stream copy will lead the start frame is not precise!

Slow down/Speed up video​

# Slow down to 1/2x in fast way
ffmpeg -y -itsscale 2 -i video.mp4 -c:v copy output.mp4

# Speed up to 2x in fast way
ffmpeg -y -itsscale 0.5 -i video.mp4 -c:v copy output.mp4

# Speed up to 2x with re-encoding in slow way
ffmpeg -y -itsscale 0.5 -i video.mp4 -c:v libx264 output.mp4

# Speed up to 2x with `setpts filter`(which requires re-encoding) in slow way
ffmpeg -i video.mp4 -filter:v "setpts=0.5*PTS" output.mp4

# Change fps to slow down/speed up but keeping duration
ffmpeg -i video.mp4 -filter:v "fps=30" output.mp4

Draw region of interest(ROI) on a video​

# Draw one drawbox

ffmpeg -i input.mp4 -filter:v "drawbox=x=100:y=100:w=200:h=150:color=red@0.5" output.mp4

ffmpeg -i input.mp4 -filter:v "drawbox=x=100:y=100:w=200:h=150:color=red@0.5,drawtext=text='Test Text':x=100:y=100:fontsize=24:fontcolor=yellow:box=1:boxcolor=yellow" output.mp4

ffmpeg -y -ss 30 -noaccurate_seek -i input.mp4 -t 10 -c:v libx264 -filter:v "drawbox=x=100:y=100:w=200:h=150:color=red@0.5,drawtext=text='Test Text':x=100:y=(100-text_h):fontsize=24:fontcolor=black:box=1:boxcolor=red:boxborderw=2" output.mp4
# Trim video and draw box
ffmpeg -y -i input.mp4 -an -c:v libx264 -filter:v "trim=start=10:end=30,drawbox=x=100:y=100:w=200:h=150:color=red@0.5:enable='between(t,10,15)',setpts=PTS-STARTPTS" output.mp4

ffmpeg -y -i input.mp4 -i overlay_video.mp4 -filter_complex "[0:v][1:v]overlay=0:0:enable='between(t,0,25)'" output.mp4
# Draw different `drawbox` at different time on video from a file(using `timeline` feature)
# See timeline: https://ffmpeg.org/ffmpeg-filters.html#Timeline-editing
# See expression: https://ffmpeg.org/ffmpeg-utils.html#Expression-Evaluation
ffmpeg -i input.mp4 -filter_complex_script timeline.txt output.mp4

# `timeline.txt` look like:
[0:v]drawbox=x=100:y=100:w=200:h=150:color=red@0.5:enable='between(t,0,21)'[box1];
[box1]drawbox=x=300:y=200:w=150:h=100:color=green@0.5:t=:enable='between(t,21,40)'[box2];
[box2]drawbox=x=50:y=300:w=300:h=200:color=blue@0.5:t=:enable='between(t,41,60)'


# or using `n`: sequential number of the input frame, starting from 0
[0:v]drawbox=x=100:y=100:w=200:h=150:color=red@0.5:n=0:600[box1];
[box1]drawbox=x=300:y=200:w=150:h=100:color=green@0.5:n=601:1200[box2];
[box2]drawbox=x=50:y=300:w=300:h=200:color=blue@0.5:n=1201:1800
# Draw different `drawbox` at different time on video
ffmpeg -i input.mp4 -filter_complex "[0:v]drawbox=x=100:y=100:w=200:h=150:color=red:t=8:enable='between(t,0,21)'[box1];[box1]drawbox=x=300:y=200:w=150:h=100:color=green:t=8:enable='between(t,21,40)'[box2];[box2]drawbox=x=50:y=300:w=300:h=200:color=blue:t=8:enable='between(t,41,60)'" output.mp4

Pipe ffmpeg​

The FFmpeg Pipe is very useful in IPC for communicating FFmpeg with another process. For instance, an application generates images to Pipe stdin which FFmpeg reads and encodes into a video.

A real-life scenario that FFmpeg read from Pipe: An application applied OpenCV to process images for object detection, and it will draw ROI but lacks ability to encode a video as efficiently as FFmpeg does. So it's somewhat ideal to pipe these images to FFmpeg that can encode the video by leveraging hardware acceleration(GPU) capability.

However, FFmpeg can also write to a Pipe.

# It works in Linux and Windows(`cmd`, does not work in `PS`)
ffmpeg -ss 00:00:10 -i video.mp4 -to 00:00:20 -an -c:v copy -f h264 pipe: | ffmpeg -y -i pipe: -filter:v "drawbox=x=100:y=100:w=200:h=150:color=red" output.mp4

ffmpeg -i input.mp4 -c:v rawvideo -pix_fmt bgr24 -r 60 -f rawvideo pipe: | ffmpeg -y -f rawvideo -pix_fmt bgr24 -s 1920x1080 -r 60 -i pipe: -pix_fmt yuv420p -c:v h264_nvenc foo.mp4
ffmpeg -i input.mp4 -pix_fmt yuv420p -r 60 -f rawvideo pipe: | ffmpeg -y -f rawvideo -pix_fmt yuv420p -s 1920x1080 -r 60 -i pipe: -c:v h264_nvenc foo.mp4
ffmpeg -i input.mp4 -an -f h264 pipe: | ffmpeg -y -f h264 -i pipe: -c:v h264_nvenc foo.mp4

Use testsrc​

ffmpeg -y -f lavfi -i testsrc=duration=10:size=1920x1080:rate=60 -c:v libx264 -pix_fmt yuv420p testsrc.mp4

Split and Concatenate​

ffmpeg -y -i input.mp4 -ss 0 -to 10 -c:v copy part1.mp4
ffmpeg -y -i input.mp4 -ss 10 -to 15 -c:v copy part2.mp4
ffmpeg -y -i input.mp4 -ss 15 -c:v copy part3.mp4
ffmpeg -y -i part2.mp4 -filter:v "drawbox=x=100:y=100:w=200:h=150:color=red@0.5" part2-draw.mp4

Slow,

ffmpeg -y -i part1.mp4 -i part2-draw.mp4 -i part3.mp4 -filter_complex "[0:v][1:v][2:v]concat=n=3:v=1:a=0[outv]" -map "[outv]" output.mp4

Fast(Concat protocol),

ffmpeg -i part1.mp4 -c copy part1.ts
ffmpeg -i part2-draw.mp4 -c copy part2-draw.ts
ffmpeg -i part3.mp4 -c copy part3.ts

ffmpeg -y -i "concat:part1.ts|part2-draw.ts|part3.ts" -c:v copy output.mp4

Fast(Concat demuxer),

ffmpeg -y -f concat -i concat.txt -c:v copy output.mp4
# concat.txt
file 'part1.mp4'
file 'part2-draw.mp4'
file 'part3.mp4'

# Or avoid creating the input file
# bash
ffmpeg -y -f concat -safe 0 -i <(echo "file '$PWD/part1.mp4'";echo "file '$PWD/part2-draw.mp4'";echo "file '$PWD/part3.mp4'";) -c:v copy output.mp4
# cmd
ffmpeg -y -f concat -safe 0 -i <(@echo "file '$PWD/part1.mp4'";@echo "file '$PWD/part2-draw.mp4'";@echo "file '$PWD/part3.mp4'";) -c:v copy output.mp4

https://trac.ffmpeg.org/wiki/Concatenate

References​

Footnotes​

  1. Chrome Audio/Video Support ↩

  2. FFmpeg Wiki ↩