No. All processing is done on your device. No image, video, audio or chapters-file is ever uploaded.
All downloads, including zipfiles containing images, are generated on your device.
No, at the moment chapter generation only works with the cli app. Although ffmpeg.js exists, I doubt that it'd be worth the effort to implement it in the browser.
No, not at the moment but there are plans.
Yes, run npx --yes chaptertool@latest serve.
Yes, chaptertool supports PWA functionality. If you use a chromium-based browser (chrome, brave, edge, vivaldi) the option to install the app locally should appear on the right in the address bar
When analytics is enabled, only superficial usage is tracked. No filenames, chapter titles or any other actual user input is ever transmitted to google.
Generally use half of the percentage of the video canvas that actually changes:
For regular full motion videos, you'll have good results with a value of 0.5. A smaller value will create
more stills, a higher value less.
For slideshow videos that only use a portion of the screen for the actual content, the value must be smaller than fraction of the image that is used.
For example, this ... uses only ~20-27% of the video canvas for the actual slideshow, so any value scene-value above 0.27 will produce no snapshots at all because 73% of any two frames is identical. A scene value of 0.2 will produce some output while 0.1 produces a good amount. 0.1 is approximately half of the size of the smallest images used inside the video.
Likewise, for talks and lectures that show slides you should figure out how much of the screen a slide shown in the video takes up and use half of that percentage.
After the ffmpeg processing is done, the result is parsed and entries that are shorter that the given min chapter length are removed. Afterwards the images are deleted and renamed.
npx is a good alternative to npm i X -g. Using npx chaptertool@latest makes sure that you always use
the latest version without to remember to update your global npm install.
If you want to make sure that your process always uses the latest version without hanging on the install prompt, use
npx --yes chaptertool@latest
ffmpeg -i INPUT.mp4 -i FFMETADATAFILE.txt -map_metadata 1 -codec copy OUTPUT.mp4mkvmerge --chapters chapters.txt -o output.mkv input-file.mkv
# xml and simple formats work the same wayI honestly have no idea, just use the mkvmerge chapters or let me know if you know.
Paste the output in your video's description. Youtube will link the timestamps automatically. The first timestamp must be 00:00.
mkvmergesimple apparently pads the chapter index to 2 digits, while vorbis comment uses 3 digits.
I honestly don't know as the only reference I could find is NVEnc's documentation
I used the EDL format that is yielded by shutter encoder. I hope the format is compatible with other software, but I haven't tested it.
The podlove simple chapters are used in the podlove web player. The format is described in the podlove documentation.
I have no idea. Also the format is only partially supported as it has more features than chaptersjson (which is the internal format used by chaptertool) can express.
No clue, hopefully you can figure it out.
When enabled, clicking anywhere on the timeline will select the chapter below. When disabled you can click on the timeline without the chapter below being selected and scrolled into view.