Creative Freedom Summit Video Edit Project

Closed Epic created 2 years ago by Máirín Duffy

Synopsis

We just wrapped up an awesome event! We need to clean up the videos though in a few ways:

Edit out "technical difficulties" 😅
Add title cards and nice intro music
Add outro cards and nice outro music
Split video files that have multiple talks into them into one file per talk
Reunite talks that got split across two video files
Add subtitles (using open source speech-to-text tools)
(If anybody's feeling fancy) Add overlays with the speaker's name etc as they start on camera

How You Can Help!

Read the instructions that apply to you:

I can edit videos, I'd like to help!

We recommend using kdenlive. If you're up to help, grab the assets below and the "Video Editing and Handoff" workflow below under "Logistics/Workflow."

Look for a ticket in the list that is tagged with the red "Needs video editing" label.

I'm up for installing kdenlive and running speech-to-text to generate captions

Refer to "Video Captioning and Handoff" under "Logistics/Workflow" below.

Look for a ticket in the list that is tagged with the orange "Needs subtitle generation"

I can help write short text snippets or edit subtitle files

Take a look at the "Video Metadata" section under "Logistics/Workflow."

Look for tickets in the list that are tagged with yellow "Needs subtitle editing" or green "Needs description written".

Assets

Here are some assets I put together for this:

Intro Cards

[The intro cards have the speaker's name and talk title - @jessicachitas created these. I will upload the one that goes to each talk to its individual ticket.)

Intro Music

CreativeFreedomSummit_IntroMusic

Outro Cards

Source SVG (Requires Montserrat) (Source SVG:

Outro Music

CreativeFreedomSummit_OutroMusic

Logistics / Workflow

Video Editing and Handoff

Grab an open ticket that's part of this epic (see below, under "Child issues and epics" - look for green tickets that are already open.) Leave a comment to let us know what you'll be doing - editing, or editing + subtitles - and we'll assign you to the ticket. Grab the raw video files (there will be links in each ticket.) Grab the outro card, the intro/outro music from here, and grab the intro card from the video ticket. Edit it down in kdenlive. Once you've got something, export to webm format.

Here's a quick table for timings that might help (noting the cropping, fade ins / outs on the audio are already handled in the files provided here):

Item	Duration	TL Start time	Fade in start / length	Fade out start / length	TL End time
Intro card	00:00:07,12	00:00:00	00:00:00 / 00:00,28	00:00:06,14 / 00:00,16	00:00:07,12
Intro music	00:00:08,09	00:00:00	N/A	00:00:04,25 / 00:03,14	00:00:08,09
Outro card	00:00:14,25	N/A	N/A / 00:00:03,26	N/A / 00:00:03,27	N/A
Outro music	00:00:16,17	N/A	N/A / 00:00:10,09	N/A / 00:00:03,14	N/A

Private message Máirín Duffy (@duffy:fedora.im) to obtain a device ID for syncthing to be able to upload your edited video.

Once you have the video edited, including intro/outro cards and music, you can upload it via SyncThing. Visit their website to get a copy (I use the Syncthing GTK flatpak).and go to "add new device" and scan or paste the device ID Máirín Duffy (@duffy:fedora.im) sent you.

Don't close the ticket. We'll keep it open until the video is finalized and posted to PeerTube.

Video Captioning and Handoff

If you have kdenlive, follow the instructions here to set up speech-to-text: https://docs.kdenlive.org/en/effects_and_compositions/speech_to_text.html

Kyle recorded a tutorial video on how to generate subtitles: https://tube.spdns.org/w/nCmzpWGs7gGzdbEpDcHYiA (PeerTube) https://youtu.be/Hqp28uPxBFY (YouTube)

Here are his instructions:

VM kept crashing, but I've taken screenshots of the whole process and need to arrange them. Here's a text version of some of what I figured out (Note: This is not a complete step by step guide at this point, but some helpful trial and error that accompanies the official instructions here):

Directly downloading via browser the 2.3G vosk-model-en-us-0.42-gigaspeech - @mairin notes this produces better output - and drag & dropping into the Settings -> Configure KdenLive -> Speech To Text window resulted in success on multiple machines for me - whereas the + button just spun forever.
The Edit Subtitle Tool icon can be hidden behind an arrow icon if your resolution is low enough (e.g. my laptop at 200%).
On some machines the Speech Editor tab (that should show up between Project Monitor and Project Notes didn't appear for me) but in any case it seems an unnecessary processing step to do that and the magic wand (as just the magic wand worked directly without the preceding Speech Editor step for me.
Once the Edit Subtitle Tool is clicked and shows up in the timeline, clicking the Magic Wand produces some confusing options. In my experience (limited, but running 5 times on 3 machines) the most straightforward success is:

Select the clip (click on it) in the timeline
Then Choose Selected clip (the 3rd option) in the Magic Wand Dialogue.
Note: When this completes you need to click Close as the window itself otherwise prompts you to click Process (which you already did and don't need to do again).

Using the other options in the dialogue requires you to adjust the thin blue bar just below the timeline to extend to the length of the whole track. Thankfully, Kdenlive has great keyboard shortcuts (I looked them up) and you can press the following sequence once you've clicked in the timeline to set the blue bar for the whole video:

HOME then i then END then o

These correspond to the below commands and should allow you to use the other options (in No. 5 above) if you're not getting any textual output in the 1st method.

If any of these run extremely fast - which is unlikely for a 30+ minute video it probably didn't work and you should try another method.

Once you have an srt file for the captions and/or a captioned version of the video, upload these new assets to the ticket via comment attachment. Let us know whether or not they are edited for accuracy or are raw machine-generated. Keep this ticket open until the video is posted to PeerTube.

Video Metadata

We'll need volunteers to help us put together easy copy/paste-able text blocks we can use for the video descriptions. It might require a little bit of research.

If you'd like to help with this, here's an example of how to do it that would work for my talk:

Come up with the talk title following this format:

Automation in Inkscape - Máirín Duffy - Creative Freedom Summit 2023

Come up with the talk description:
- Talk description from the schedule on creativefreedomsummit.com and/or the Fedora wiki page
- Talk bio from the schedule on creativefreedomsummit.com and/or the Fedora wiki page
- Any assets related to the talk (search the chat backlog for a link to the speaker's slides, if the speaker didn't provide it, reach out and try to get from them or ask one of the organizers for help reaching the speaker. If the speaker mentioned important resources / assets / links etc during their talk, it'd be helpful to gather those too.)
- Check the question block from the talk to see if there's any resources to add from that: https://hackmd.io/N_j36bIpSLCPXpYyXEb8MQ
- Combine this all in one text block, like so:

You've got to generate 100 graphics following the same template with different information and you're dreading it. Not to fear: you can automate it using Inkscape! Learn how in this session.

Resources for this talk:

Link to slides: https://example.com/slides

Link to the NextGenerator plugin: https://gitlab.com/Moini/nextgenerator

Link to templates provided in talk: https://gitlab.com/fedora/design/team/tutorials/inkscape-automation

Link to speaker's blog: https://blog.linuxgrrl.com

Speaker's Patreon: [I don't have one but many do]

About the speaker:

Máirín is a senior principal interaction designer at Red Hat. A recipient of the O’Reilly Open Source Award, Máirín has over 15 years of expertise in user experience and design working in upstream Free/Libre and Open Source Software (FLOSS) communities. Her portfolio includes a wide range of designs for FLOSS, from systems management tools; to desktop infrastructure and applications; to the cloud-based medical image processing project, ChRIS. Máirín has a masters' degree in Human-Computer Interaction and a B.S. in Computer Science from Rensselaer Polytechnic Institute.

Once you have this text drafted, post it as a comment to the ticket for the video it's meant for.

Video Uploading

A member of the Creative Freedom Summit team can upload videos to the channel. Here are those who currently have access:

Edited 2 years ago by Máirín Duffy

Linked items 0

Link items together to show that they're related or that one is blocking others.