Video Player issues
dfourney at ryerson.ca
Thu Apr 25 14:48:19 EDT 2013
Please see my comments inline below.
David Fourney, Ph.D. Candidate
Department of Mechanical & Industrial Engineering
----- Original Message -----
From: Michelle D'Souza <michelled33 at gmail.com>
Date: Wednesday, April 24, 2013 5:51 pm
Subject: Re: Video Player issues
To: David Fourney <dfourney at ryerson.ca>, Anastasia Cheetham <acheetham at ocadu.ca>
Cc: fluid-work at fluidproject.org
> > 1. The video player breaks in Chrome in fullscreen. I downloaded an
> earlier version from January which works fine but the current one does
> Good catch - it looks like half the video is off screen. I've opened a
> JIRA ticket for this issue:
> Is that the same thing you saw, or did you see something different?
This is exactly what I see.
> > 3. If there is only one language available for the transcript, the
> grey box at the top still shows a dropdown menu. If there is only one
> language to choose from why have a menu? It just adds clutter. It is
> not obvious to me how to get rid of this dropdown menu and treat the
> language available as the "default".
> > Another issue with this grey box is that, in full screen mode, it
> scales up to become a much larger than needed box. This has a visual
> aesthetic impact.
> Good point. Any chance you'd be willing to open a JIRA ticket for this
> and put up a screen shot of what you are seeing?
> > 4. Is there a way to hardcode the UI Options? The options panel is
> very useful, but if you only want to show the page in a specific way
> to test specific views, is there an easy way to hardcode the options
> (e.g., turn on transcripts by default, turn on captions by default,
> turn on both by default, turn on a specific language option, etc.)?
> You should be able to easily over-ride the default settings for UI
> Options. Anastasia, I wasn't able to find some documentation on how to
> do this - do we have some?
Documentation with examples would be helpful.
> > 5. Does the transcript view support WebVTT? The demo videos all use
> JSON files for transcripts and WebVTT for captions. It would seem to
> me that if you have one or more WebVTT files, you would not want to
> recode them all into JSON, just reuse them in the transcript mode.
> While I like that the player supports Amara, WebVTT is the current W3C
> standard and should be supported for transcript views.
> Not yet, although we do plan to implement this at some point. http://issues.fluidproject.org/browse/VP-56
I noticed that this JIRA item was created last September. This matter may need to be pushed up the timetable because several browsers (Chrome, Opera, and Safari) now support WebVTT at various levels. As an immediate workaround, you may want to create a script that converts WebVTT into the expected JSON format.
> > 6. Is there a way to resize or reposition the transcript box? For
> some users, I can see the default width of the box being too small.
> They may want it to start a little bit more to the left. I can see
> other users wanting the transcript to appear on the left side of the
> video rather than the right (or even below or above the video).
> You can just override the default CSS by setting a different width
> using the 'fl-videoPlayer-transcriptArea' class.
This too would be useful to have documented along with the UI options.
> > 7. The transcript appears to get collapsed into one big runon
> paragraph creating some readability issues. How can this be broken up
> so that new lines appear at logical points?
> This is a great idea. Do you want to write a bit more about what you'd
> like to see?
At a basic level, the rendering system should expect that the presence of a new speaker (i.e., using WebVTT's voice labelling classes) would indicate the need for a new paragraph. For example, if Tom (labelled in the WebVTT file as <v Tom>) stopped speaking and was followed by Harry (labelled in the WebVTT file as <v Harry>), as per the following WebVTT code:
00:00:20.060 --> 00:00:23.060 position:15%
<v Tom>Hi. How are you?
00:00:23.080 --> 00:00:27.090 align:end
<v Harry>Doing great. Thanks for asking!
The current implementation would display this as one long runon paragrah:
Hi. How are you? Doing great. Thanks for asking!
But if we took into account the change in voice as meaningful, then the system could display this dialogue in the transcript view as two distinct lines:
Hi. How are you?
Doing great. Thanks for asking!
If the same speaker is going on for some number of sentences (e.g., like the video with Jutta), then the system should automatically introduce a break somewhere. What would be really nice is if the content creator could add flags to the WebVTT file indicating a logical break.
For example, WebVTT allows for positioning codes (position, align, line) to follow the timestamp, like so:
00:00:20.060 --> 00:00:23.060 position:15%
Now imagine if we could add a new flag (I'll use "break") a the end of all of the codes. Browsers would ignore it because they don't know it. The video player would hopefully not ignore it. So a timecode written like so:
00:00:20.060 --> 00:00:23.060 position:15% line:5 align:end break:here
Would then cause a newline to be inserted into the transcript view.
That, in a nutshell, is how I would envision this working.
Right now, I need to do some more testing with your existing system to see how well it behaves if I add whitespace to the JSON file. Otherwise, I will really need someone to fix the system because the ability for the system to obey newline requests will be absolutely needed for me to use the video player in my upcoming work.
More information about the fluid-work