By Claude Almansi
Editor, Accessibility Issues
Non scientists should refrain from using scientific concepts as metaphors. I am fully aware of this, and actually, when a sociologist or other humanistic scholar thus hijacks terms or phrases like “black hole,” “big bang,” “DNA,” etc., I skip his/her text if possible.
Nevertheless, what little I understand of how the cellulase enzyme works for ruminants has been very instrumental in my first perception of how captioning videos helps all users digest their content, and underlies what I have written here so far about captioning. Hence the decision to come out explicitly with this subjective and uninformed perception of it.
Cows can digest and assimilate the grass cellulose because they ruminate it, but not only: humans could chew and re-chew grass for hours and hours, yet they would still excrete its cellulose whole without assimilating any because we lack something cows have: the cellulase enzyme that chops up the molecules of cellulose into sugar types so that they can be assimilated .
Interestingly, cows don’t produce cellulase themselves – bacteria in their stomachs do. While these bacteria only live in the stomachs of ruminants, cellulase and other enzymes are apparently sometimes added to high-fiber foodstuffs for non ruminants1 and used in the production of fruit juices2.
So maybe we do ingest some cellulase, at least indirectly, but probably not enough to implement the politically incorrect dream of the Venetian folksong:
My lover, she’s old
I keep her as a spare
when the grass starts growing
I send her grazing3
In Live Lecture Better Than Video Lecture?, Jim Shimabukuro reviewed a study conducted by David N. Figlio, Mark Rush, and Lu Yin – Is It Live or Is It Internet? Experimental Estimates of the Effects of Online Instruction on Student Learning – (National Bureau of Economic Research, June 2010). As he rightly pointed out- and as the authors themselves acknowledge – this study has limits, and it would be rash to give up the use of online resources just on its basis.
Nevertheless, David N. Figlio, Mark Rush, and Lu Yin venture an interesting hypothesis as to why, in their experiment, the students who could only use video versions of the lectures performed less well at exams than those who could only attend the lectures live, i.e., the former were tempted to postpone viewing the videos until it was too late.
Video consumption is bound to its medium. In a video, you can’t skim or go forward and backward to a given point as easily as you can with text. You could, but that would require inserting timed text markings on the video. The viewer could then navigate among the markings.
The paradox is that, as Roberto Ellero points out, “Ogni video parte sempre da un testo e al testo ritorna”, i.e., “Every video always starts from text and goes back to the text”4.
The starting text Ellero is speaking about here is the script for videos planned as such, e.g., tutorials. But his affirmation is also true for mere recordings of lectures: the words spoken by lecturers are text, the slides used – if any – are text, or are first conceived in words that are text. This also obtains for possible other “non verbal” contents, e.g., manipulations in an experiment done to illustrate a lecture.
Ideally – or rather, if accessibility regulations like Section 508 of the Americans with Disabilities Act (ADA) or the Web Content Accessibility Guidelines (WCAG) were actually implemented – all videos offered for learning would also provide the return to these texts via captioning and verbal audio descriptions of the non verbal contents, in a manner that allows users to switch on what they need or want, i.e., see DVDs where they can choose to have them or not.
Trying to digest long videos without text support is as difficult as trying to feed on grass without cellulase.
The problem with fully captioning and scripting an audio description, as advocated by Roberto Ellero, is that it takes time, and time is money. True, the YouTube automatic captioning via speech recognition software holds great promises, but it only works with English videos and is not quite satisfactory yet. There has been no noticeable improvement so far in the issues I described under “Short words are the rub” in Accessibility and Literacy: Two Sides of the Same Coin 11 months ago.
However, even if educational institutions cannot afford full captions and audio descriptions of the lecture videos they offer as learning materials, we live in a read-write culture based on collaborative online tools. So students could get together to caption video course materials themselves, sharing the work load. And if that’s still too much, they could use online captioning tools to at least take synchronized notes on the videos, and then share their notes, just as they could (or should) with the notes they take during a live lecture.
Since I wrote Three Video Captioning Tools, I have tried other tools. For example, Universal Subtitles, adopted by Creative Commons International for their videos, is particularly interesting for several motives. For subtitlers:
- Clear text instructions + video tutorial for each stage of the subtitling process.
- Timing accuracy via the second, synchronizing stage.
- Easy keyboard shortcuts (Tab for play or pause, Shift+Tab for skip back, Enter for new caption, Down arrow for next caption when fine-tuning caption timing), which means it can be used with only one hand and, possibly, with a screen reader as used by the blind and by people with print disabilities).
For users and consumers:
- Immediately viewable interactive transcript for existing subtitles in all languages that allows you to skip to a given point of the video. Unfortunately, this video navigation via transcript does not work with a screen reader yet, but at least the transcribed text does – whereas you cannot even view a YouTube or DotSUB transcript with a screen reader because you need to use a mouse to activate it.
- Possibility to download existing subtitles in all languages without logging in, in several formats, including a .txt file
without the time codes if you just want a transcript5.
This improved accessibility makes Universal Subtitles a very useful tool for studying video recordings, whether by collaboratively subtitling them or by hijacking them for notes.
Problem: copyright. While captions in the same language do not fulfill the creativity condition for a derivative work, as described by most national copyright laws, and could be considered as merely a technical process for viewing, copyright law also gives rights holders full control over the form in which their work is published. As captioned videos are not the same as uncaptioned ones, permission must be requested and obtained for captioning other people’s videos. And the terms of service of all video captioning platforms insist on that.
I thought the same would obtain for ICE’s YouTube video summary of its press conference about Operation in Our Sites 2, which I used in ICE’s Seizures of Domain Names Concern Us All, as US government communications are not protected by copyright6. Or so I naively believed. And while ICE had captioned the video, it had not enabled its interactive transcript, which I needed to quote from. So I re-captioned the video to get it.
However, Michelle Thorne, of Creative Commons International, pointed out in an e-mail to me that:
The only tricky thing is that US government works areindeed in the public domain – but only in the US. See the Wikipedia article https://secure.wikimedia.org/wikipedia/en/wiki/Copyright_status_of_work_by_the_U.S._government.
So, this might be a gray zone, depending on where one is arguably using the material (possibly the US, which is where Universal Subtitles is “based”, as much as internet-y things can be based anywhere).
So far, ICE does not seem to mind this Universal Subtitles transcript.
However, this is one more instance of the paralyzing paradoxes of the present state of copyright that Lessig exposed in his WIPO address (4 Nov. 2010):
[Copyright] has failed to assure the adequate incentives in the professional culture, and it has failed to protect the necessary freedoms in the amateur and critical or scientific culture.
It has failed at both of its objectives and its failure is not an accident. Its failure is an implication of the architecture of copyright as we inherited it. This architecture makes no sense in the context of the digital environment. The architecture, which triggers the application of copyright law upon the production of a copy, in a digital environment makes no sense. It regulates too much, and it regulates too poorly.
So think of a simple example of a book in physical space. If these are all the uses of a book in physical space, an important set of these uses are just technically unregulated by the law of copyright in physical in physical space.
So to read a book is not a fair use of the book, it’s a free use of the book, because to read a book is not to produce a copy. To give someone a book is not a fair use of the book, it’s a free use of the book, because to give someone a book is not to produce a copy. To sell a book is explicitly exempted from the reach of copyright law in many jurisdictions, including the United States, because to sell a book is not to produce a copy. In no jurisdiction in the world is sleeping on a book a regulated act because to sleep on a book is not to produce a copy.
These unregulated acts are then balanced by a set of necessary regulated acts, necessary to create the proper incentives to produce great new works. And then in the American tradition, there is a thin sliver of exceptions, acts that otherwise would have been regulated by the law but which the law says are to remain free so that culture can build upon those creative works in a way unhampered by the law.
Enter the internet, where because a digital platform, every single use produces a copy. And we go from this balance of unregulated and regulated and fair uses, to a presumptively regulated use for every single use, merely because the platform through which we get access to our culture has changed. This is the consquence of an architecture, an architecture of copyright law, an architecture of digital technologies.
It is that architecture that produced what Jessica [Litman] spoke of when she said, “a world where we can’t even go for an hour without colliding with copyright law”, and the collision is a problem not with some generation that can’t learn to respect the rules, it is a problem in the design of this system of regulation.
Lessig therefore asked WIPO to lead, as a long term goal, a fundamental revision of copyright law that would fit the conditions of our era of digital publishing: in fact, the 1996 WIPO treaties concerning digital works took years, in some cases decades, to be embodied in national copyright laws. As a short term remedy, he requested WIPO’s support for licenses by which right owners can confer users some rights while withholding others, as in Creative Commons and other copyleft licenses.
Whether WIPO grants Lessig’s plea remains to be seen. Even if it does, ICE’s Operation in Our Sites has demonstrated that some states are more apt to heed content industry requests than their own laws anyway.
And then, there are other obstacles concerning captioning. Some people who give unrestricted access to videos of their lectures object to having them “modified” by captions or subtitles. This is due to their lack of understanding of how closed captioning works.
True, in some cases, burning captions into a video may deface it. But closed captioning leaves the video as is and simply allows the user or consumer to synchronize a text file with it, on demand, to produce captions or subtitles in other languages.
Regardless of whether the objection to subtitles and captions is motivated by copyright or aesthetic reasons, it not only violates accessibility norms by excluding deaf people but also hinders serious and efficient study of video materials. Giving students unrestricted access to the videos of their lectures, but preventing them from captioning or annotating these videos, is like sending cellulase-less beings to graze grass: very little will get assimilated, as David N. Figlio, Mark Rush, and Lu Yin warned in Is It Live or Is It Internet? Experimental Estimates of the Effects of Online Instruction on Student Learning.
Unfortunately, getting people to appreciate this problem is not easy. Even educators may not take the time to fully understand this issue because they are otherwise engaged by their duties and inclinations.
So maybe the best strategy to remove barriers to subtitling and captioning by users would be the search engine facilitation argument. When a video is captioned or subtitled, search engines can index its verbal content, thus giving it more visibility.
1 See 2. Dietary additives for enhancing nutritional value of feeds in T.S.Johri’s Poultry Nutrition Research in India and its Perspective (FAO, 2004).
3 “Polenta e baccalÃ “. YouTube recording with Italian lyrics in the description: Orietta Berti – Polenta e baccalÃ¡, starting at 0:35.
4 Accessibilità e qualità dei contenuti audiovisivi – a tutorial for Accessibile, the observatory for the accessibility of the services of the Italian public administration, 2009-12-22 â transcript of the embedded video, with Italian captions and English subtitles, from 2:03.
5 See – and play with – the English and Italian subtitles and transcripts for Lawrence Lessig’s keynote address at the Facilitating Access to Culture in the Digital Age – WIPO Global Meeting on Emerging Copyright Licensing Modalities, Nov. 4 2010, done with Universal Subtitles and with DotSUB.