Knowledge Base (Posts about ffmpeg)

Split mkv file on chapters

bgstack15 — Wed, 06 Mar 2024 13:51:47 GMT

To split a large mkv file into a file for each chapter, run the output of this command.

files/2024/listings/split-on-chapters.sh (Source)

#!/bin/sh
# File: split-on-chapters.sh
# Location: /mnt/public/Support/Programs/DVDs
# Author: bgstack15
# Startdate: 2024-02-17-7 14:22
# Title: Split mkv on chapters
# Purpose: oneliner for splitting mkv file into separate files per chapter
# History:
#    2024-02-22 moved -ss and -to before -i
# Usage:
# Reference:
#    man ffmpeg
#    slightly improved over https://superuser.com/questions/795373/split-mkv-based-on-chapters
#    https://superuser.com/questions/795373/split-mkv-based-on-chapters/1830690#1830690
#    https://stackoverflow.com/questions/14005110/how-to-split-a-video-using-ffmpeg-so-that-each-chunk-starts-with-a-key-frame/33188399#33188399
# Improve:
# Dependencies:
# Documentation:
#    This is insufficient:
#       HandBrakeCLI -c 2 -i twoeps.mkv -o 2.mkv # this does not get subtitles, and reencodes which takes a while
#   WARNING! Jellyfin+chromecast might malfunction on a video split in this manner, within the first few seconds of playback, because of the keyframe/chapter mismatch.

split_by_chapter() {
   for word in "${@}" ;
   do
      ffmpeg -i "${word}" 2>&1 | sed -n -r -e "/start.*end.*[0-9]*/{s/.*#[0-9]*:([0-9]*).* ([0-9]*\.[0-9]*).*( [0-9]*\.[0-9]*)/ffmpeg -ss \2 -to\3 -i ${word} -acodec copy -vcodec copy -scodec copy ${word%%.mkv}-chapter\1.mkv \;/g;p;}"
   done
}

split_by_chapter "${@}"

So:

$( ./split-on-chapters.sh twoepisodes.mkv )

Splitting a video file can get weird with -codec copy and -ss. If you stream at the beginning of a file split in this manner, it might never start displaying the video. There might be some way to split "correctly" by dealing with the key frames, but I spent maybe 30 minutes and couldn't figure it out. I'll just live with the consequences, of waiting a few seconds before chromecasting a split video file from Jellyfin on an android device.

References

Man pages

man ffmpeg

Web links

Find average volume change needed for album

bgstack15 — Sat, 19 Aug 2023 12:36:19 GMT

x=0 ; for word in *.mp3 ; do x=$((x+1)) ; ffmpeg -i "${word}" -af "volumedetect" -vn -sn -dn -f null /dev/null 2>&1 | awk "BEGIN{OFS=\",\"} /mean_volume/{a=\$5} /max_volume/{b=\$5} END{print a,b}" | sed -r -e "s/^/${x},/;" ; done

And then I saw the whole list of max and mean volumes. So my minimum max_volume is -8.3, so I need to add this amount to each file. I think ffmpeg does not use fractional decibels, so I chose 9.

x=0 ; time for word in *.mp3 ; do x=$((x+1)) ; new="new${word}" ; ffmpeg -i "${word}" -af "volume=9dB" "${new}" ; done

Installing ffmpeg >=4 on CentOS 7 for Jellyfin

bgstack15 — Thu, 09 Mar 2023 13:37:07 GMT

tl;dr

Fetched ffmpeg spec from Chinforpms pasture, fixed some rpm macros, compiled it, and installed it.

Overview

I use Jellyfin server on CentOS 7, which I install from a local yum repository after downloading the correct files from the upstream page. I recently upgraded from 10.7.6 to 10.8.9 and media playback broke for almost all video content.

The start

To start the process, I made a backup of the data after stopping jellyfin.

# systemctl stop jellyfin
# tar -C /var/lib -zcf /tmp/jellyfin.$( date "+%F" ).$( rpm -qf ).tgz jellyfin

Then I yum updated and started jellyfin.

The problem was indicated in the server logs (journalctl -n500 -f -u jellyfin): Jellyfin needs the path set for ffmpeg. That's slightly believable. I ensured I had /usr/bin/ffmpeg, and set that as the path in the web admin console, but the web ui indicated it could not find ffmpeg. The server logs indicated specifically that it could not find ffmpeg 4, so it was a version problem. I realize CentOS 7 is now 9 years old; vm4's hardware RAID controller does not have a driver for the Linux kernel used by RHEL8 and above, at least at the time I installed that system.

The search

Turns out rpmfusion only has ffmpeg version 3. I spent time looking for a way to get ffmpeg >= 4 on CentOS 7. I first investigated the famous negativo17 repository. I have used it in the past, with minor problems as some of the packages want to own the same files from other packages in the OS. negativo17's "epel-multimedia" repository has ffmpeg 5 in it!

I wanted to find another way to get ffmpeg without having to use that repository, if I could. I can use that repository if I have to. I checked another Fedora custom repackager that I follow, PhantomX, who did indeed package ffmpeg in the past! In fact, PhantomX had stopped about 3 years ago at version 4.2.2, which was perfect! This spec file needed only minor adjustments for rpm macros that do not exist in the version of rpm in CentOS 7 (PhantomX targets Fedora, and 3 years ago would have been Fedora 32 approximately), and thankfully all the dependencies were in rpmfusion or the base repositories.

So I manually built this ffmpeg 4 rpm on a CentOS 7 build node (I needed too), and then deployed the files to my local yum repository, and ran yum upgrade ffmpeg. And then I could set the path to the program in the Jellyfin web ui, and then my users can view videos again!

Details of spec file changes

--- ffmpeg.spec 2023-02-26 12:12:22.250386837 -0500
+++ ffmpeg.spec.new 2023-02-25 22:45:15.043571569 -0500
@@ -4,6 +4,15 @@
 #global date    20180419
 #global rel     rc1

+%define ldconfig_post(n:)  %{?ldconfig:%post -p %ldconfig %{?*} %{-n:-n %{-n*}}
+%end}
+%define ldconfig_postun(n:)   %{?ldconfig:%postun -p %ldconfig %{?*} %{-n:-n %{-n*}}
+%end}
+%define ldconfig_scriptlets(n:)  %{?ldconfig:
+%ldconfig_post %{?*} %{-n:-n %{-n*}}
+%ldconfig_postun %{?*} %{-n:-n %{-n*}}
+}
+
 # Cuda and others are only available on some arches
 %global cuda_arches x86_64

Conclusion

Probably eventually, I will need ffmpeg 5. Perhaps I just won't update Jellyfin again. It does exactly what I want already, and upgrading it now didn't even make anything better.

Additional thoughts

I needed to update my webhook template for new field names. Oh, and the app server needed to be restarted a few times to iterate through fetching the newer version of the webhook plugin, and enabling it, and using it, but it was painless.

Related files

All srpms and rpms in /mnt/public/www/internal/repo/rpm/jellyfin-ffmpeg/.

Split single mp3 of whole CD

bgstack15 — Sat, 31 Dec 2022 13:35:19 GMT

I had a single mp3 recording of an entire CD that I wished to split into the individual tracks. Thankfully, people smarter than me have already solved this problem on the Internet.

The first order of business was to find the track times. Thankfully, musicbrainz.org delivered TOC info. It looks like this but I wasn't able to figure how to download a cddb-style TOC file, but that's OK. I used Firefox's CTRL+click-and-drag functionality to select just a column of the start times of the tracks, and pasted them in a file.

I ended up having to subtract lots of seconds, because my recording was somehow skewed. I also remember that ffmpeg can only be so precise with seconds (due to how compression works?) so that affected things to. So my track times ended up a little different than the TOC I used.

File ./times

00:02:21
00:04:32
00:08:32
00:12:25
00:15:12
00:19:14
00:22:53
00:25:52
00:29:58
00:35:03
00:39:27
00:42:48
00:46:54
00:48:54
00:53:33
00:58:17
01:02:44

Notice how I had to add the end time of the last track! Armed with these times, I used a small script from Unix.SE, of course.

split.sh (Source)

#!/bin/bash
# Adapted from https://unix.stackexchange.com/questions/659789/automating-the-splitting-of-a-large-mp3-file-with-ffmpeg-into-multiple-files-in
x="00:00:00"
z=0
filename=$(basename -- "$2")
ext="${filename##*.}"
filename="${filename%.*}"
initcmd="ffmpeg  -nostdin -hide_banner -loglevel error -i $2"
while read y ; do
   initcmd+=" -ss $x -to $y -c copy $filename$z.$ext"
   let "z=z+1"
   x=$y 
done < $1
${initcmd}

Then I ran the script:

./split.sh ./times album.mp3

I had to subtract lots of seconds, smeared across the whole track. I suspect the whole recording was the tiniest bit fast, but I also remember that ffmpeg plays fast and loose with relative time specifications.

But now, I had an mp3 file for each track! Then I used puddletag and its "Tag Sources" window to load the musicbrainz information for these songs.

Manually fix volume in mp3 file

bgstack15 — Sat, 25 Aug 2018 13:01:05 GMT

I am shamelessly ripping off a superuser answer: How can I normalize audio using ffmpeg?.

Option 3: Manually normalizing audio with `ffmpeg`

In ffmpeg you can use the volume filter to change the volume of a track. Make sure you download a recent version of the program. This guide is for peak normalization, meaning that it will make the loudest part in the file sit at 0 dB instead of something lower. There is also RMS-based normalization which tries to make the average loudness the same across multiple files. To do that, do not try to push the maximum volume to 0 dB, but the mean volume to the dB level of choice (e.g. -26 dB).

Find out the gain to apply

First you need to analyze the audio stream for the maximum volume to see if normalizing would even pay off:

ffmpeg -i video.avi -af "volumedetect" -vn -sn -dn -f null /dev/null

Replace /dev/null with NUL on Windows. The -vn, -sn, and -dn arguments instruct ffmpeg to ignore non-audio streams during this analysis. This drastically speeds up the analysis. This will output something like the following:

[Parsed_volumedetect_0 @ 0x7f8ba1c121a0] mean_volume: -16.0 dB
[Parsed_volumedetect_0 @ 0x7f8ba1c121a0] max_volume: -5.0 dB
[Parsed_volumedetect_0 @ 0x7f8ba1c121a0] histogram_0db: 87861

As you can see, our maximum volume is -5.0 dB, so we can apply 5 dB gain. If you get a value of 0 dB, then you don't need to normalize the audio.

Apply the volume filter:

Now we apply the volume filter to an audio file. Note that applying the filter means we will have to re-encode the audio stream. What codec you want for audio depends on the original format, of course. Here are some examples:

Plain audio file: Just encode the file with whatever encoder you need:
```
ffmpeg -i input.wav -af "volume=5dB" output.mp3
```

Your options are very broad, of course.

Install ffmpeg on devuan

bgstack15 — Fri, 20 Oct 2017 17:35:12 GMT

If you want to use ffmpeg on devuan (I want to convert mp4 files to mp3), you should use utility avconv from package libav-tools.

References

Weblinks

https://superuser.com/questions/286675/how-to-install-ffmpeg-on-debian/865744#865744
Interesting discussion about the politics behind libav and ffmpeg and so on https://hardforum.com/threads/the-dark-side-of-foss.1947261/

Knowledge Base (Posts about ffmpeg)

Split mkv file on chapters

References

Man pages

Web links

Find average volume change needed for album

Installing ffmpeg >=4 on CentOS 7 for Jellyfin

tl;dr

Overview

The start

The search

Details of spec file changes

Conclusion

Additional thoughts

Related files

Split single mp3 of whole CD

Manually fix volume in mp3 file

Option 3: Manually normalizing audio with ffmpeg

Find out the gain to apply

Apply the volume filter:

Install ffmpeg on devuan

References

Weblinks

Option 3: Manually normalizing audio with `ffmpeg`