User talk:Doc James/Videos/RfC

Comments

I think there is a catch-22. If you make the videos highly editable, you're forced to use computer voice reading which is inherently disengaging (at present), and every edit will throw the timings off making the videos nearly useless. If you make them custom, the work required to create them is so high that people will either avoid editing them after creation, or not make them at all. One middle-ground might be to put a time limit on any one video like 45sec, and force a custom script. That way (a) it's a time limit that people will watch (b) if it needs editing, it's not a lot of work. For instance, a 10min video on Justin Beiber with a 30sec edit at time 0:30 will cause all of the timings to not work. If any video is 45sec max, changing the timings is easy. One article might have 10 videos associated with it. I think the RfC should have some options added to it about time, live/computer voice, work-flow, etc... to generate more discussion. Ian Furst (talk) 18:11, 15 February 2019 (UTC)[reply]

@User:Ian Furst: I agree with you that we should have short summary video (for eg 2 min summary videos.) What about a hybrid system that takes advantages of both the systems? TTS to edit and update the VideoWiki article but when you want to export the video to Wikipedia/Commons, you can add custom human voices. However, this hybrid system will best suit for 2 minute videos as you mentioned - the effort/output ratio has an optimum balance. Pratik.pks (talk) 08:09, 28 February 2019 (UTC)[reply]

User:Ian Furst these videos on youtube simple read Wikipedia with a machine voice and some have received more than 200K views.[1]

They generally keep them to under 10 minutes which IMO we should generally do aswell.

If you look at the script [HERE] you see a block of text associated with each image / short video. One can have the video / image present for variable periods of time. The video generator should keep everything timed fairly well when a new version is auto-generated. Doc James (talk · contribs · email) 06:28, 27 February 2019 (UTC)[reply]

Dishonest

This is a dishonest RFC and one that aims to get the community voting prior to real discussion that should occur. The technology is far far from being ready for use.

The RFC shows an example using professionally-created video merged with robotically generated audio from editable text. The videos are cut into slices per segment but they remain uneditable. James absolutely must offer this RFC with 100% user-created content.
Users wanting "rich content" means they would like to see videos of events and of people and of things, rather than just text and photos. This project does not solve the fundamental problems that we can't get hold of enough freely licensed videos. A robotic-narration series of power point slides with distracting transition effects is not what people had in mind. Wikipedia can already incorporate video clips.
The Youtube channel linked as having 20k+ hits is also misleading info. Look at Pedagogy. Well for a start, it is illegal as it takes Wikipedia content and does not give attribution. The owner claims Youtube will no longer let folk monetize text-to-speech videos. I suspect this is because they are crap and rip-off other people's work, adding little or no value. He admits it is "shitty". The first review comment is "Much faster and easier to read the article in Wikipedia". Another says "belch ...not 30 seconds into the video before I realize this is a waste of time". So basically, these "hits" are folks coming across the video when they search for something, and think they are going to get a great original video on the topic. The watch/listen for a few seconds before realising they've been duped. The owner says he's probably going to delete them. Doc James has no stats on videos like this that people want to watch from start to finish.
The RFC claims the concerns about previous videos have been addressed by Video Wiki. They really haven't. You have half-solved one problem: collaboratively editing audio, but introduced another: viewers have to listen to 3 minutes of robotic monotone.
It is simply dishonest to say "The video (the text and media) can be editable/updatable wiki style". Although the audio may be edited, you need to edit the video to keep it in sync, and we still have no way to create engaging videos and collaboratively edit them.
Commons is not a collaborative editing project and will no welcome your collaborative editing issues spilling over there.

There are many many more issues, noted at WT:MED and WP:NOTYOUTUBE. I sincerely hope this RFC is not going live any time soon, or we will see a car crash of negativity. -- Colin°^Talk 17:26, 1 March 2019 (UTC)[reply]