Wondering how to download a video from YouTube, with audio and subtitles intact? Here are a couple of methods I’m documenting in case I ever need to use them.

Method 1: On-line tools

There are a multitude of on-line YouTube downloading tools, such as KeepVid.ch. Pretty self-explanatory, just provide a YouTube URL, choose the quality options, and hit download. Two browser tabs open - one playing just the video stream and the other playing audio, both directly from YouTube. For each, right-click and Save As...

For subtitles, SaveSubs.com is one of the many on-line tools to download YouTube subtitles in .srt format.

Method 2: youtube-dl

There is an open-source youtube-dl Python-based command-line downloader.

According to the instructions, curl -L https://yt-dl.org/downloads/latest/youtube-dl -o youtube-dl will download the executable. However, using youtube-dl is rather complex, see the examples!, e.g.

• to download video and audio streams youtube-dl -ci -f 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best' https://www.youtube.com/watch?v=xxxxxxxxxxx
• download video, audio and subtitle streams, youtube-dl -ci -f 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best' --sub-lang en-UK,en-US,en --write-sub https://www.youtube.com/watch?v=xxxxxxxxxxx

It can also use FFmpeg to merge the streams, that’a a bonus from the previous method! There is a dependency on ffmpeg (more on that next) which, if is in the same folder as youtube-dl, is configured as follows:

youtube-dl -ci -f 'bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best' \
  --sub-lang en-UK,en-US,en --embed-subs --write-sub --ffmpeg-location . \
  https://www.youtube.com/watch?v=xxxxxxxxxxx

Merging streams

If the audio, video and subtitles are all separate files, e.g. audio.m4a, video.mp4 and english.srt, the open-source FFmpeg tool can merge these streams into a single file, e.g. merged.mp4:

ffmpeg  -i video.mp4 -i audio.m4a -i english.srt \
  -c:v copy -c:a copy -c:s mov_text \
  merged.mp4

ffmpeg can handle more streams, e.g. multiple audio tracks with respective subtitles. According to this Stack Overflow answer, it’ll be something like this:

ffmpeg -i video.mp4 -i english.mp3 -i german.mp3 \
  -i english.srt -i german.srt \ 
  -map 0:v -map 0:a -map 1:a \
  -metadata:s:s:0 language=eng -metadata:s:s:1 language=ger \
  -metadata:s:a:0 language=eng -metadata:s:a:1 language=ger \
  -c:v copy -c:a copy -c:s mov_text \
  merged.mp4

And so much more, e.g. delay subtitles (ffmpeg -ss -60 -i subtitles.srt -c copy output.srt), convert formats (ffmpeg -i audio.wav -c:a aac output.mp4), etc.

Again, I honestly haven’t validated any of this. So good luck!

Updated 4 July: Small addition to sepecify multiple English subtitle variations with --sub-lang en-UK,en-US,en. Use --list-subs to get a list, and note that the list often starts wtih Available automatic captions, which is maybe not what you want.