YouTube not for Everyone

I am starting my HCI course at Penn and coincidentally some of my last lectures in my Stevens courses deal with HCI issues.  One of my students in my Stevens Software Engineering course, Todd Bernstein, sent me this perceptive email that I thought I would share with you concerning YouTube and other on line video sites.

Does the 1998 amendment to Rehabilitation act extend to online videos? I am surprised that captions are not available on the many videos posted on the web. At the very least, I think the big networks sites should have them. I know many elders would benefit from this as well as the hearing impaired population. With the explosion in video on the web, I think it would make sense to have a inclusive reach to the audience and include language translations as well but that may be out of the scope of the Rehabilitation act.

Since I am in the midst of tests and papers I have not had a chance to investigate it but my bet is it does not since the Law pre-dates the site.  As Todd points out, this limits the population that can view the site and given the role these on line video sites played in the election, it makes accessibility difficult to a significant part of the electorate.  Of course this would be an expensive process, but it could also be a differentiator and a subtle way to advertise (close captioning provided by ...).  I would like to collect opinions on this.

By the way, if you would like to follow my lectures at Penn, you can download them at my home page here.

Hopefully since I am correcting a ton of logbooks and have some free time between semesters, the posts should increase again.   Later!

Comments

[Carl Frederick II]

It's likely that the sites like YouTube could provide technologies that convert the video words and sounds to text and description automatically.

Once these conversions are done, they can be shared with the person who submitted them, for any minor corrections, plus attributions.

These minor corrections could also serve to train the software in recognizing diverse types of voices better, including those with accents.

It's possible that people who constantly upload videos, will have algorithms that are optimized for their particular accents speach.

These will mean that features like "transcripts" and "reviewed transcripts" can be provided, even for hearing audiences.

Adding the closed-caption will also make the content more suitable for noisy environments, like at a restuarant / pub, perhaps increasing the prevalanece of venues where the content is displayed.

These transcripts will also make it possible to do other types of searching for video contents online. So, like find the lecture where the Professor talked about the origin of the word "Spam".

The automatic translation of the language to text will also enable the search results to identify the segment of the lecture where the content was found.

Applying linguistic analysis will also enable a recommended "starting point", perhaps at the beginning of a sentence, or a scene, based upon visuals and/or text and topic cues.

[Tochi Iheagwara]

This is something I never considered before but seems like there could be an easy and workable solution by integrating some speech to text technologies or applications by services provided by Google (e.g. Google Voice voicemail) and the like. Google Voice voicemail however is still premature its translation. You may see that your text translations are not always successfully interpreted. One would need to take into account linquistics (language, accents and dialiects). Accessibility to the web really expand the scope of the challenges to the viral technology applications like You Tube.