+1 323 268 0395 kmw@caltek.net

SoCaLinux mailing list archives

Postprocessing workflow for S [...]

CalTek.Net, Kenneth Wyrick
- 03/14/2020 09:56:37

Postprocessing workflow for SCALE A/V
Application toolbar
Compose Reply Reply all
Reply-all options
Forwarding options
Delete Move Print Mark More
Message navigation
Message 2 of 173590 Previous Next
Subject: Postprocessing workflow for SCALE A/V
From Lan Dang <l.dang@ymail.com>
To Ilan Rabinovitch <ilan@socallinuxexpo.org>
, Kmw <kmw@caltek.net>, Michael Proctor-Smith <mproctor13@gmail.com>, bogovincent@gmail.com <bogovincent@gmail.com>, gilbert@socallinuxexpo.org <gilbert@socallinuxexpo.org>
Cc Jess Bermudes <jbermudes@gmail.com>
, capsulecorplab@gmail.com <capsulecorplab@gmail.com>
Date Today 02:45
Contact photo
Message Body
Hi all,

Post-processing is not well-documented as it is mainly exchanges between me and various people on the team.  So some
of this is for Jess and Sean's benefit.

The plan for post-processing was to complete the SCALE 17x editing/reviewing.  
Once all the streams are revieweed, I'm going to do a few JSON dumps and then
ask Vincent to load in SCALE 18x and announce it to the scale-av mailing list so
they can work on editing.

The actual cutting and upload process is decoupled and handled via a couple of scripts; we would prioritize SCALE 18x.

We are very close to finishing with the editing.  If you would now like to handle reviewing, you
can use the password "test_reviewer". 


The repo for the scale-av cutter app is here:

The scale-av cutter app has three roles:
  * Editor  - anyone.  Can only specify start/end time and mark a talk as Done or Unusable)
  * Reviewer - veteran A/V volunteer to quality control the edits.  Can override editor selections.
  * Admin - Find the appropriate YouTube day-long streams and add it to the app.

The post-processing workflow is as follows:

1) Use signs XML feed to populate the scale-av cutter application. 
2) Admin find the YouTube  ID for each day-long room stream; video transitions to (NEED CUT)
3) Editors click on (NEED CUT) rooms to evaluate the video for each room to see if it is worth cutting or Unusable and mark accordingly.
    If it is usable, they should specify a start/end time (clicking on the the dropper icon will pick up the current time from the video as it's playing)
    and mark the talk as Done.  When all talks are evaluated, video transitions to (NEED REVIEW)
4) Reviewers click on (NEED REVIEW) and evaluate the Editor's selection.  It can now be APPROVED or REJECTED.  We only want to mark talks as APPROVED if we are okay with having them published.
5) Once a day is done or all is done, we can make a JSON dump using filters.
We can form a URL to get a JSON dump based on some filters.  The http://scale-av-cutter.herokuapp.com/json allows accessing the raw data. it can be filtered with approved=1 and e.g. day=Friday, so only the cuts that are Done and Approved will be returned.  

6) The JSON output will be fed into the scripts that will download and split the videos.  It will also be used by the script that will upload the video.  The latter would be done by people who have access to the YouTube account.  

I haven't tested either of these steps yet, though Vincent has.  It's a TBD until we get through the SCALE 17x cutting and reviewing.