Adding media to GNU Audio-Video

From LibrePlanet
Revision as of 08:44, 13 July 2017 by Johns (talk | contribs) (Step 3: Add description file (For RMS speeches only): *all* uploads need license files, not just RMS speeches)
Jump to: navigation, search

This page is for people interested in volunteering to help maintain the GNU Audio-Video public archive.

When RMS gives talks, people make recordings and then submit tickets to the GNU Audio-Video ticket tracker with the recordings linked or attached. Your job is to monitor the ticket tracker and put up new recordings as they come in.

The system is delicate, without version control to easily revert a mistake, so it's important to be careful. If part of the instructions isn't clear, or you have any doubt about a step, please email audio-video@gnu.org with "Help Needed" or similar in the subject line.

First time set up

Adding new files

Ok, you're ready to start putting up files. The first step is to explore the currently open tickets in the ticketing system and pick one. I usually recommend you pick the oldest open ticket first. Once you've picked a ticket, here are the next steps. I just want to stress, also, that the system is quite fragile, so make sure to create a backup where indicated and, if you have any concerns or questions, let me know and I'll help.

Step 1: Upload file

Upload them to dl.sv.gnu.org:/srv/audio-video, using the subdirectory audio-video/video or audio-video/audio as appropriate. There are multiple tools you can use for the upload (if you have access to the FSF's fencepost server, you can do it through there for faster results, but it's not necessary).

Do your best to continue all the name conventions you used on the audio-video server, including in the file name. Make sure that your username is your savannah.gnu.org username.

Here are some tools you can use for the upload:

rsync

Example command for uploading a video:

rsync -avz --progress FILE USERNAME@dl.sv.gnu.org:/srv/audio-video/video/

sftp

Note: this is currently not functional, but rsync still works.

Example command for establishing a shell connection to the server, from which you can use sftp commands to transfer and modify files.

sftp USERNAME@dl.sv.gnu.org:/srv/audio-video/video/

Step 2: Edit HTML page

Also edit the appropriate HTML page -- either /srv/audio-video/audio/index.html or /srv/audio-video/video/index.html -- to include links to the new video or audio recording. You can do this the same way you uploaded the initial media file, in this case overwriting the HTML file with a modified version with the exact same name. Note that there is no version control on the server for this, so *always* make a local backup copy of any HTML file you edit before uploading your modified file to the server.

You should see your change immediately reflected on <https://audio-video.gnu.org/audio/index.html> or <https://audio-video.gnu.org/video/index.html> .

Step 3: Add description file

Upload a small text file next to the video that includes the speech details, like time, date, location, title, and license. These details can be found in the fsf.org/events section. If it's an RMS speech, the license unless otherwise stated should be the CC BY-ND 4.0 license. Give the text file the same name as the audio or video file, but with a .txt extension appended after the original extension.

The purpose of the txt file is to keep the metadata permanently safe. As people reorganize the HTML files, some bits might get lost, so we want to have the important details in a .txt file next to the video file.

Step 4: Email Webmasters

Finally, it's generally appropriate to mail <webmasters@gnu.org> and let them know that the video or audio is there and should be linked to from the gnu.org nonphilosophy section. We'll consolidate eventually so that this cross-linking is not necessary, but for now it is.

Licensing

If a ticket in the tracker links to an original copy of the audio or video file that is not licensed properly for us to reuse, we have to contact the copyright holder and ask them to either:

  • Change the public license to CC BY ND. If necessary, a volunteer can do this.
  • Give us specific permission to redistribute it under one of these licenses. This should be done by a staff member.

Usually the latter will be easier. Here is example text for a message to copyright holders:

Hello,

I'm writing on behalf of Richard Stallman, who was featured in an audio program by RTS [1] in February. Richard asked me to put the recording up on https://audio-video.gnu.org/audio/. May we have permission to download it and post it there under a CC BY ND license [2]?

Thank you

1: http://www.rts.ch/info/sciences-tech/7463886--l-iphone-d-apple-est-un-ordinateur-prison-selon-le-pere-du-logiciel-libre.html 2: https://creativecommons.org/licenses/by-nd/4.0/

Video editing

This is usually not necessary, but we've included it here in case you need it. Loïc Cerf contributed to this section.

Here is how to call avconv version 9.20-6 ("libav-tools" in Trisquel 7's repository) to downmix the stereo audio of a video to mono, and add a title:

$ avconv -i stereo.webm -ac 1 -c:v copy -metadata title='INSERT-YOUR-TITLE' mono.webm

And here is how to turn two videos (with the original English audio and with its translation) into one video with two audio tracks in the same order, with a +2s88-offset to start the second video (as an example):

$ avconv -i en.webm -itsoffset 00:00:02.88 -i pt.webm -c copy -map 0:v -map 0:a -map 1:a -metadata:s:a:0 language=English -metadata:s:a:1 language='INSERT-YOUR-LANGUAGE' -metadata title='INSERT-YOUR-TITLE' combined.webm

However if published on the Web, it is better to publish the two videos, with one audio-track each, because Web browsers do not allow to switch between audio-tracks.

Deleting files

If you need to delete a file from the server, please email the FSF campaigns team at <campaigns@fsf.org> and they will remove it for you.