[Re-adding the subtitles team with the right address, the original message can be seen at https://lists.debian.org/debconf-video/2019/06/msg00007.html] Hi Wouter and all, > I started working on a client-side dynamic HTML video player: > > https://debconf-video-team.pages.debian.net/videoplayer/ This is great! \o/ > > It is very much a proof-of-concept right now, not yet remotely feature > complete, but it already allows you to select a video by conference, and > to play it. Once I've got the bugs that currently exist ironed out, I > plan to also add support for subtitles in that player. This is even greater! \o/\o/ > > However, in order to be able to do that easily, it would be useful if > the SRT video subtitles that exist now would be converted to the WebVTT > format, which is a W3C standard originally based on SRT, and meant for > use with HTML video. There exist some SRT-to-WebVTT convertors out > there, so this can be automated; however, the WebVTT format has a few > extra features that you might want to look into, such as tagging a > particular text with a particular voice, which can then be styled > differently in the video playback (e.g., different colors), using CSS. > We might want to give that some thought (but I'll have to figure out if > it's easily possible to switch such a thing on and/or off without a page > refresh; otherwise people might get annoyed if they can't switch off the > subtitle angry fruit salad). For the existing subtitles, we'll probably stick to just converting them to WebVTT and look at those new features for the next subtitles. However, I think I'll tinker with WebVTT on the last talk I subtitled at MiniDebConf Marseille (we were three speakers, so the color thing could be useful). > > Also, I believe that all video players which support the SRT subtitle > format will also support the WebVTT one. We should double-check that; if > that is indeed the case, then it is probably more efficient to drop all > SRT subtitles and stick to WebVTT only. Agreed. > > While at it, it occurs to me that the current workflow of publishing > subtitles is somewhat suboptimal: > > - Someone needs to complete a subtitle file, which I gather is done on a > service called "amara". Amara is the "online" option, but people can also use offline tools such as gnome-subtitles or aegisub. Besides this detail, your summary of the publishing workflow is correct. > - Once the subtitle is complete, it needs to be downloaded from amara > and added to the salsa debconfsubs project (a manual action) > - Once that is done, someone needs to notify a member of the DebConf > video team that a new subtitle is ready so that it can be published, > which is then done by the team member (a manual action) > - Once the subtitle is published, it also needs to be added to the > archive-meta project by a member of the team (another manual action) > > That seems like a lot of manual work for just publishing a single file, > and I think we should automate as much as possible of this; that is, > marking that a subtitle file is complete (however this is done, be it by > marking them as such in "amara", or by adding them to the debconfsubs > project repository) should trigger all actions that are necessary to > publish them on the meetings-archie and get them listed in the > archive-meta project. > > Thoughts? Being the one that pings the DebConf video team each time a subtitle is ready, I would really love not being a SPOF anymore. Making the subs team a part of the video team (on Salsa) has already been discussed last year but we never actually did it. If that could help with making the process easier, then let's do it. :) Subtitles team, what do you think about it? Thomas
Attachment:
signature.asc
Description: OpenPGP digital signature