Wikimedia video plugin 
Please discuss at Wikimedia media plugin.
A more radical alternative for video 
I think the technical goals stated here are well motivated, but are missing a vital spark which I may be able to help provide.
There are so many issues I will need a discussion to reach a full solution, but I can start off with a few key points:
- Computers are still advancing rapidly, both in storage and internet bandwidth
- The typical internet device (by numbers) is now a mobile phone, not a PC
- Installed plug-ins are, by their very nature, slow to update and quick to become obsolete
- Cheap video serving is possible by making use of standard web caching from normal web servers
- There is a cross platform web-based solution which allows:
- Upload of video from cameras or phones
- Editing from a web browser on Windows, Mac or Linux PC
- Publishing for web streaming (and mobile phone viewing) via the web interface
- Access to anyone with a web browser to re-edit and publish the content
- I run the company which makes this solution
As we have created our own technology, IPR issues can be fixed. It may be appropriate for this solution to become preferred video solution for Wikipedia. I'd like to find out what the real issues are from a range of interested parties. Stephen B Streater 22:24, 8 May 2006 (UTC)
Presentation Practicalities 
For starters, if this solution were to become a vehicle for video in Wikipedia, some thought would need to be given to how an article incorporates / presents the video(s).
(1) New markup construct to accomodate a video applet
The drawback being the automatic downloaded of an applet when the article is opened in a browser, even if only one video were embedded into an article this may compromise browsing performance for modem users - let alone if multiple videos were present
(2) New markup construct bearing a static image that (when clicked) opens a new window in which the video applet loads and plays
I.e. no applet is loaded by opening an article up, only a static placeholder image per video.
(3) Plain old URL
Least work but so boring.
Option two surely? 18.104.22.168 09:43, 16 May 2006 (UTC)
More options 
As you can imagine, we've tried a few ideas out over the years.
Case (0) The video plays automatically when the article is opened. The can be quite annoying, particularly if there are multiple videos on the page, or people are at work and don't want sound blaring out suddenly.
Case (½) The best performance (for people who want to watch the video) is for the applet to load with the page, and to start downloading the video, but not start playing it until the play button is pushed. This means that the video plays back at higher frame rate / better quality than would otherwise be the case. Modem users get better video performance as it makes use of all the available bandwidth. This has the disadvantage that WP would be serving video which people aren't watching - which costs - and I think this alone rules it out.
Case (1) Video applet allowed in page - most simply by using the iframe tag - see this example. This is quite efficient as most browsers cache Java applets, and so any return to a page with a video in it (which is likely in my experience) would not be a big hit for either WP or the viewer. With broadband, the small applet downloads almost instantly, so when the player is downloaded doesn't matter very much. Forbidden has a very small applet to display a static image as quickly as possible and then chain the full player. This static image applet (which can also be cached) is minimal to allow playback to start quickly. In this scenario, video download does not start until the play button is pushed, and the player could be downloaded after a short delay to allow the rest of the page to be downloaded. This also leaves open the possibility of downloading the first couple of seconds of video ready to play as bandwidth costs come down.
Case (1½) Download only a small applet to display the image. This would give quick rendering of the image. Download of the video player would take place when the image was clicked on.
Case (2) An image with a clickthrough to an external link. Some people have pop-ups disabled, so this may have to open a new page, losing the article - not so good.
Case (3) This is what I do now in WP. This is not as good as an embeded image in the article, where it can be seen in context.
My preference is for option (1): cached applet downloads for people not playing the video, without having the video bandwidth hit, and reasonably responsive for those watching the video. If we start an applet up, we have a lot of choice about what it does, and can improve this easily later in the light of experience and/or improvements in internet bandwidth. Stephen B Streater 08:34, 20 May 2006 (UTC)
- "The typical internet device (by numbers) is now a mobile phone, not a PC" - that's just the same old marketing stuff. Most people still access the web from their PCs, virtually no one I know from his cell phone - because it's to slow, to cumbersome and to expensive. Wikimedia needs to improve its video capabilities, but this should start with providing adequate free windows software to convert .mov etc. to Ogg Theora, then we can integrate Bittorrent links next to the files in Wikimedia Commons to save bandwidth, then the file size limit on Commons has to be increased, and then things are changing. All this streaming stuff might be nice in the future, but are not of immediate concern. Longbow4u 21:41, 22 May 2006 (UTC)
- The mobile may be nearer than you think. Most consumer video is now shot on mobile phones. And WP will probably get its content from video shot by editors. Stephen B Streater 08:27, 23 May 2006 (UTC)
- Some problems with Ogg that are solved with Forbidden's solution are:
- Ogg won't play back for most users of WP in practice because it requires a plug-in which they don't have
- There is no simple cross-platform web-based tool to edit Ogg video collaboratively - which is what video on WP could be all about
- There is no simple way to upload mobile shot video over-the-air straight into a Ogg video editor
- Stephen B Streater 08:27, 23 May 2006 (UTC)
- Longbow4u, unless there is a vision for popularising video by making it more like viewing text / static pictures it will remain a minority sport. Fewer people viewing videos equates to the lack of incentive to publish them. What would be the state of affairs if, in getting Wikipedia off the ground, the same approach that is currently being taken to build a video capability was also taken to procure a computer / network hardware / software infrastructure capability? The answer is we would still be waiting - as we are still waiting for a capability to popularise video. Mark Kilby 12:24, 26 May 2006 (UTC)
These opinions may seem POV today, but many videos on WP are already shot on mobile phones, uploaded over-the-air to FORscene, edited on the web, and published in Forbidden's format. This is despite Forbidden not being endorsed by WP. FORscene is powerful and easy to use. Wouldn't it be better for WP to embrace Forbidden's Technology under some sort of Wiki-licence to allow free use within WP, to give better access to all WP editors and browsers? Stephen B Streater 08:27, 23 May 2006 (UTC)
Speaking as a Clesh/FORscene user I have lots of questions as to how all this could work. IMO it is an understatement to say Wikipedia has a lot to gain by adopting Java to extend its video capability, in particular the system under discussion given the proposal above.
Speaking as a visitor to Wikipedia, for all practical purposes Wikipedia is virtually a video free zone at a time when video is starting to blossom (explode?) elsewhere. This is only likely to change once it becomes easy and convenient to upload and administer video content.
Contrast the tools available on Wikipedia for collaboratively compiling text and static pictures with those for video - they are light years appart on Wikipedia at the moment. Video really is the poor relation.
IMO the average person does not find your average video content compelling enough to feel it's worth the effort (and potential security risk) of downloading specific video player software to watch it. Where is the incentive for authors to contribute video content that few people are ever likely to watch (because most people don't have the player installed).
Here's my first question - how would Clesh be best positioned?
(1) As a channel for serving ogg files in java format? I.e. less of an editing system and more as a video player in which anything uploaded in the ogg format automatically gets a Java stream.
(2) Option 1 plus as an editor for ogg files? I.e. all edited content is written back to ogg format as well as a Java stream.
(3) As a complete replacement for ogg with streaming and editing.
My view is Clesh is best positioned as a complement to ogg or similar.
Option one has to be the simplest and quickest, and so long as markup language allows embedding into articles it would massively increase the potential audience for video. Therefore increasing the incentive for authors to contribute content (could FORupoad/FORmobile echo content to ogg format?).
Option three implies unhinging ogg from Wikipedia which may not only be a lot of work but may also upset the people that use it.
Options two / three trigger too many questions in my mind for me to consider at this time of night...
As an aside, could the success of video be a curse for Wikipedia? I.e. would it present major issues in terms of cost of hosting footage especially? I don't know how efficient ogg format is. Would video need to be capped in terms of quality to preserve disk space?
Java for video on Wikipedia is a no-brainer in my mind. On top of this, a cutting edge solution provider is offering a technology on what appears to be very frindly / favourable terms. This proposal might not satisfy the die hard freedom purists but it sounds like a practical way forward - in the absence of anything remotely comparable.
Mark Kilby 23:49, 23 May 2006 (UTC)
We have been adding a number of output formats eg video iPod, MPEG 2. We already have Web Java player, Symbian mobile, Java mobile, EDL, XML and ALE. Ogg is more a container than a codec, so it seems that it is an alternative to a Java player. For example, some people are considering adding Dirac to the Ogg format. See this edit. This would suggest that Ogg will never have the advantages of an install-free Java player, and so has (technologically speaking) rather limited appeal. This probably rules out (1) above.
Option (2) is perfectly feasible, though why bother to write the Ogg files when almost everyone with Ogg also has Java, but not vice versa?
Option (3), although the most ambitious, is already technically solved. This is what FORscene and Clesh do already. We have a complete solution today.
As a matter of interest, does anyone knoow how many Ogg videos there are on Wikipedia, and how many people watch them?
Stephen B Streater 06:12, 25 May 2006 (UTC)
- Number of videos: The page here has a link to a video listings page here. This video directory has 183 listed. There are three other videos here Commons:Wikimania_Media_Competition_2005/Video making a total of 186, not including links to externally hosted video. To put this in context YouTube has in excess of 40m video clips (see recent BBC Article).
- Most videos on WP were uploaded by the team/person (Raul654) that built the current WP video capability, according to Raul654 "...I created the audio and video templates, and I've located and/or uploaded most of the full length songs on wikipedia and virtually all of the videos)..."). Raul654 went on to say "...We use ogg because we consider patented formats go against our core principles of making wikipedia free (as in speech)...".
- On another note. I don't see why format patents go against any core principles. The WP site can be divided into two streams; (1) content and (2) vehicles for content delivery. The former having the condition that contributions are free in the sense of "free speech" and "no charge". The latter being procured on other criteria such as least cost, which may or may not involve using technology that is patented such as computer hardware (hence the need for WP to ask for donations to pay for this). Furthermore the latter should have no issue at all concering whether a technology is patented or not (providing its use can be funded, or permission obtained to use it at "no cost"). If a technology to improve content delivery is available to WP, it should be used. I.e. it should not be rejected simply on the grounds a technology is patented. Mark Kilby 01:42, 28 May 2006 (UTC)
- I don't see why opinions cannot soften when a "Wiki-licence to allow free use within WP" is on the table. Exceptions to the 'rules' do exist in WP - for example granting permission for images (see "Images and uploaded files" here Wikipedia:Database_download) so why not for video formats too?
- Numbers of people watching videos: There are heaps of statistics (e.g. here) but no published statistics for video file downloads or indeed any media file (that I can find). The nearest I got was here and here. There are just over 600,000 media files. If there are only under 200 video files then, putting it mildly, video is under represented on WP. I can state for certain the Java videos I personally added in early 2006 have generated over 1000 viewings in just this short time. There are some data 'dumps' which are beyond my ability to decode, perhaps the information is buried somewhere within. Mark Kilby 01:24, 28 May 2006 (UTC)
- FORscene/Clesh would provide reporting of number of people watching videos too. It could also collect feedback on it to give other people an idea of whether it was worth watching. And of course, it would allow editing of it. Stephen B Streater 18:55, 29 May 2006 (UTC)
I really wish this discussion had be held in someplace public rather than in a forgotten corner of meta. There is so much factually incorrect and wrongheaded about the above that I don't even know where to begin. Patented formats are utterly unacceptable and are, fortunately, unneeded. We will have inline Theora playback soon, hopefully with wikibased basic editing (there is a question about how quickly that part of the software can be debugged). This will, hopefully, kill all arguments for using Stephen's proprietary software even without repeating the argument against proprietary viewer lockin and patented formats. --Gmaxwell 20:43, 20 August 2006 (UTC)
- I am in absolute 100% concurrence with Greg on this. All media on Wikimedia must, to the maximum extent possible, use formats that are unencumbered by restrictive patents. That means if there's a patent on the codec, it had damn well better be a defensive-only patent. Both encoding and decoding must be without payment of fee for any user and any purpose whatsoever. To do otherwise would spit in the face of our commitment to free, freely reusable and freely redistributed content. Kelly Martin 20:49, 20 August 2006 (UTC)
- Absolutely. There is no room for discussion about proprietary software here. This is completely and utterly irrelevant to our mission. Mr. Streater's proposal is unacceptable by WMF standards. Danny 01:24, 21 August 2006 (UTC)
- This is rather an old discussion which has now been superseded. I was planning to release sufficient software to enable this to work within the WMF system, but didn't manage to engage the right person on this subject - a potentially lost opportunity. I wasn't aware this was a forgotten backwater. This was not made clear by the person who told me it was the best place to discuss these issues. As it happens, FORscene can now output Ogg format videos, and a Java player is on its way (as stated above), so this meets most of my goals for video on Wikipedia. The only thing not available yet is wiki-style editing of video which currently exists, but not here. I'll work towards next. Stephen B Streater 19:59, 21 August 2006 (UTC)
That may not be enough. The software should be free in all systems, not just WMF. Are you committed to that? Is that the current situation of the software? Danny 09:34, 24 August 2006 (UTC)
- The current state is not what this discussion is about. The idea of the discussion here is to find out what to do to get people in Wikipedia actually using video. There are so many options, that it would be inefficient to do them all. For example:
- We could release existing technology under a free licence to get things started
- We could create new technology and release it under a free licence (it will probably turn out that the needs here are different from those of our customers)
- We could develop the existing free software WMF uses, as we do with Linux.
- Once people start using video, interest will increase, and the free development work will take off. The problem at the moment is that support is so poor and the technical work so hard that things are a bit stuck. I'm just breaking the log jam here. Stephen B Streater 13:07, 26 August 2006 (UTC)
- Then get back to us when the software is non-proprietary. If you want more information, I invite you to call the Wikimedia Foundation office. Danny 19:06, 27 August 2006 (UTC)
- I might do that, as it is not clear which of the above alternatives is best, or whether another solution would work better. There's not much point in coming back with the wrong free solution. I'd like to get a consensus involving as many people as possible first. Stephen B Streater 22:16, 27 August 2006 (UTC)
Public domain 
I saw this line "already available in the public domain (e.g. a lot of material at www.archive.org)" Is it OK to upload videos from Archive.org? I'm a regular contributor to Wikipedia (same handle) and wanted to know about uploading videos obtained from archive.org.
also is it ok to upload videos from youtube? I'm referring to the legalties etc. Thanks in advance. Idleguy 14:07, 11 November 2006 (UTC)
- Hi! I don't know about archive.org, but generally YouTube videos do not have a free licence, and cannot be legally uploaded to Wikipedia. Furthermore, YouTube uses a proprietory format, so any videos which were uploaded would need to be converted to a free format such as Ogg. It may be possible to add YouTube videos which you have made yourself and submitted to YouTube, as long as you own all the rights to them. Stephen B Streater 09:18, 26 January 2007 (UTC)
- Is a simple external link to a video on You Tube acceptable? There is a specific one there that I want to link to, and the preson who made it says it's ok for me to do this (and that the video will be on You Tube for some time to come). Is it ok by Wikipedia? I want to abide by the rules/conventions. Frankieparley 12:15, 19 November 2007 (UTC)
- Linking to YouTube pages is allowed. Jecowa 08:26, 1 December 2007 (UTC)
ffmpeg2theora audio 
I'm having trouble converting video files to Ogg Theora. The ffmpeg2theora program is not converting the audio channels of files it converts. Are there only certain audio codecs it can convert from? It freezes up every time I give it a video file with an audio channel. Jecowa 08:38, 1 December 2007 (UTC)
Preservation of original data 
Suppose my camera produces video that is patent-encumbered. I convert that to vorbis and upload it to wikipedia.
Next, my original video becomes unavailable. Perhaps I simply stop being a wikipedia contributer. Perhaps I drop dead. Perhaps my hard drive fails. Whatever the case, my original video is now unavailable forever.
A year later, somebody notices that I made some poor choices configuring the video conversion software. Oh well...
Another year passes, and a major bug is fixed in the video conversion software. Oh well...
Another year passes, and a better unencumbered format is invented. The new format doesn't have compression artifacts as much as theora does. Of course, the video already has theora compression artifacts burned into it forever. Oh well...
A couple more years go by, and the patents expire. We could enjoy the pristine original data, except we don't have it. Oh well...
22.214.171.124 07:20, 16 February 2009 (UTC)
unintended consequence 
I'm using purely free (Open Source, etc.) software on a Linux platform. I can't play the vast majority of videos associated with wikipedia articles.
An unintended consequence of the current policy is that wikipedia contributers upload video to Youtube instead of Commons. Reality crushes many dreams. :-/
Youtube requires a working flash player. Flash is quite proprietary. The free clones are slow, tend to crash my browser, and generally don't even work.
So I'm stuck now, unable to view nearly all of the video in wikipedia articles. The really sad thing is that I could actually deal with most of the proprietary video formats! Free software developers have done a pretty decent job on proprietary video, at least compared to flash.
Must we all install Adobe's proprietary flash player to fully enjoy wikipedia? Help! Our freedom is being oppressed!
126.96.36.199 07:56, 16 February 2009 (UTC)
Embedded Videos from Other Services and/or Kaltura 
What is the status/thinking on embedding videos from commercial services such as YouTube, Veoh, Vimeo, etc. Is there a fundamental prohibition on incorporating for-profit content or anything in the mission statement or other funding contingencies?
The downside is obviously that they may contain advertising. The upsides include: there is no storage issue (although a bandwidth issue), and they--not the Wikimedia Foundation--will have to deal with the copyright issues. Also, to the extent that users are not uploading content, the issue of using a proprietary format seems irrelevant (and a bit Stallman-esque). Users want to use video. Let's make it available. And when the bandwidth is cheap enough, the open source tools and formats are cheap enough, etc. Wikimedia can then offer users the upload option. Videos originally seen in embedded YouTube videos with ads, can be supplanted with open-source, GPLed, ad-free content. But waiting for a perfect solution until then seems silly, IMHO.
There are a number of extensions that have been written for MediaWiki that allow this type of embedding, e.g., Embed Video Extension.
Finally, what is the status of the partnership with Kaltura? Has it stalled -- or it plowing forward and making this discussion moot?
- If you read all the extensive discussion of using FORscene (or its consumer version Clesh) to edit and host the video, you will see people got banning threats for linking to video in that format. The gist of some of the arguments were that (1) the video was not IN Wikimedia, just accessed from it and (2) the accessibility of good video would discourage people from making the same video again and free. The idea was that it doesn't matter how long this takes. My view is that a lot of video is time dependent - it has to be shot at the time it happens - and content has been and will be lost forever if WP insists that it only allows free content in free formats. Also, some people are prepared to make video available but not freely available. In the long term, copyright and patents will expire, so current content lost for good would have been free in the long term. If they are thinking truly long term, they should allow all content to be uploaded to the archive, and make it accessible as rights expire. Stephen B Streater 15:59, 16 August 2009 (UTC)
why do we have to use a different format from industry norm? 
Why can't we just comply with Frunhaufer or whatever patents and use a normal format? I mean we have computers that have patents for parts of them. Why not just pick the best format for th reader? TCO 06:07, 28 January 2011 (UTC)