In November 2009, Australia’s Department of Broadband, Communications and the Digital Economy announced a discussion report on media access, viz captioning and audio description of TV, movies, DVDs, and online video. This document is my response.
This submission is permanently located at the address:
I shouldn’t have to be the first to inform the Department that RTF is not an “accessible format.” Neither is MS Word. I know you’re all wedded to your Windows computers and you spend all day every day using nothing but MS Office and Acrobat, and I know blind people in Australia are also in thrall to their Windows machines and thousand-dollar add-on screen readers, but the only genuinely accessible format for wide use in a consultation like this one is valid HTML. Mine will be the only submission in that format, hence the only actually accessible one.
The discussion paper is, at times, so badly written it actually makes mistakes. Authors of the paper are clearly unfamiliar with the topics at hand, by which I mean they lack lived experience.
Captioning is not “the presentation of the audio component of audio-visual content as text on screen.” Captioning is a transcription of the dialogue of a program along with notation of important sound effects.
Captions are only “coloured” in PAL countries like Australia; captions in North America have always had the colour option but barely ever use it.
The statement “Where speaker identification is not relevant, captions are usually in the form of white writing on a black backdrop” doesn’t just border on meaninglessness, it’s incorrect. Foreground and background colour are selectable in teletext, with limitations. If you’re trying to say that U.S.-made captions on U.S. shows arrive in the country white on black, that’s a different topic.
Closed captions are merely optional captions of any kind (a fact the discussion paper later admits in the context of Rear Window captioning). Closed captioning on analogue TV is indeed sent as teletext, but the transmission format is different for digital television systems, DVD, Blu-ray, and online video. (With a few overlaps here and there, every video format has its own caption format. They’re all closed captions because you have to turn them on.)
Open captions are merely captions everyone who can see the screen in the first place sees. You can’t turn them off.
The definition of audio description – “the presentation of the visual component of audio-visual content as additional verbal commentary that complements the underlying soundtrack” – borders on incomprehensible. Audio description is added narration that talks you through a program, describing in words what you can’t understand just from the main soundtrack.
“The audio description” – “the” audio description? – “is generally transmitted to the consumer via headphones” only in movie theatres and in some U.K. home configurations. Audio description on TV is just a separate audio track even in that U.K. example. (If the original statement were true, which it isn’t, how exactly is “the audio description” actually “transmitted”? Do I need a certain brand of headphones? Can my dog hear it?)
Captions do not “assist in the pronunciation of words” for ESL learners. Captions may provide a correspondence between the written and spoken form, but it’s really a stretch to claim that any form of written English assists in the pronunciation of spoken English. (If you and I read this paragraph out loud, the results wouldn’t sound remotely the same. And that says nothing about the unpredictable spelling of the English language.)
Captioning is useful “in a noisy environment” only when you can’t hear the TV audio. In “clubs and gyms,” TVs are usually on mute.
Colourblindness has no bearing whatsoever on reading print, save for the case of illegible foreground/background combinations (pale green on pale orange, dark red on black), which hardly ever come up. Nor does audio description have anything to do with colour deficiency. Nor does audio description help anyone who has “physical difficulties interacting with print and visual media.” These statements are flatly wrong in several ways and seems to be confusing audio description with alternate formats for print-disabled people, like talking books for dyslexics.
Time for an international comparison – because while no man is an island, no country is, either, not even Australia.
The Canadian Human Rights Act, at §5, states:
It is a discriminatory practice in the provision of goods, services, facilities or accommodation customarily available to the general public
- to deny, or to deny access to, any such good, service, facility or accommodation to any individual, or
- to differentiate adversely in relation to any individual, on a prohibited ground of discrimination
which includes disability. Undue hardship is a defence, but must be assessed in each case and is not solely about money.
The Canadian Human Rights Tribunal has assessed the application of the Canadian Human Rights Act to television captioning. In Vlug v. CBC, the Tribunal found that an absence of captioning constituted discrimination on the basis of disability. No undue hardship was found.
The two respondent networks, CBC Television and Newsworld (now CBC News Network), were ordered to caption 100% of their programming – of all kinds and at all times, including actual TV shows, promos, and unscheduled news flashes. Outside commercials were included in the order, a fact CBC has ignored while otherwise more or less complying with the order.
The Disability Discrimination Act (Australia), at §5, states that
a person... discriminates against another person... on the ground of a disability of the aggrieved person if, because of the disability, the discriminator treats, or proposes to treat, the aggrieved person less favourably than the discriminator would treat a person without the disability in circumstances that are not materially different, [additionally if] the discriminator does not make, or proposes not to make, reasonable adjustments for the person
An unjustifiable hardship, as separately defined in the Act and assessed in a case, is not a reasonable adjustment.
There is no material difference between the Canadian and Australian statutes, despite being worded differently. You can’t discriminate against people with disabilities, though some remedies might constitute undue hardship. It is not a great leap of the imagination to assume that if one national human-rights body can conclude an absence of 100% captioning constitutes discrimination or unequal treatment, a similar absence in a country with functionally equal laws would also constitute discrimination or unequal treatment. The only difference isn’t really a difference, since it applies in both countries: The undue- or unjustifiable-hardship test must be applied when it is advanced as a defence.
So I’m going to go out on a limb here and state that it is a settled matter of international human-rights law that anything less than 100% captioning constitutes discrimination. This will come as a big shock to everyone in Australia, except of course deaf people.
Since captioning involves technology and human decisions, errors will happen. As such, 100% captioning really means 99.999% or five-nines captioning, the standard used by HBO.The FCC engaged in a consultation concerning increased captioning requirements in the U.S. in 2005. In its filing, HBO stated (emphasis added):
HBO’s quality-control program monitors 16 categories of potential technical issues associated with video, audio and closed captioning, the occurrence of any one of which is considered a disruption to the service. HBO’s goal is to have each of the 30 linear programming feeds it originates experience no more than 5.5 minutes of programming disruption per year – a reliability factor of more than 99.999%. Over the past two and a half years, more than 25 of HBO’s programming feeds have met this reliability goal consistently. Those linear feeds that fell short of the goal missed it by an insignificant amount on an annual basis....
HBO has found that closed-captioning errors on its feeds account for less than 10% of all disruption events (i.e., less than 30 seconds per year). In fact, in HBO’s experience, the errors in closed captioning are fewer than the miniscule amount of audio discrepancies.
The actually viable tolerance for 100% captioning is 0.001% error. That figure is not a minimum allowance of uncaptioned programming; it is the maximum headroom necessary to account for the fact that computers sometimes crash and staff are sometimes asleep at the switch.
99.999% captioning means 0.001% of programming can air without captions. Out of 8,760 hours in a 365-day year, 0.0876 hours may air without captioning. That’s five minutes. The Vlug case allowed for such variation, admitting that “glitches” might happen. But “these should be the exceptions. The rule should be full captioning.”
Indeed it should. And, in Australia, it isn’t.
Australia is not actually interested in providing full accessibility of film, television, and video to deaf or blind people. Its recent history constitutes a Culture of Exceptions, in which human-rights bodies and government departments carve out policy that precisely matches industry’s policy of refusing to provide fully accessible programming. Under no circumstances do TV broadcasters, cinema owners, DVD distributors, or anyone else intend to caption and describe everything that needs to be captioned and described. And the Australian government backs them up. In fact, departments and agencies are first in line to sell out Australians with hearing or vision impairments by vitiating their legal rights.
Let’s list just those exceptions documented in the discussion paper:
Your system makes paternalistic assumptions about what programs deaf people should be able to watch and understand.
The Australian human-rights system, and the government generally, eagerly works in concert with private industry to cook up one excuse after another for the infringement of the rights of Australian citizens with disabilities.
As a further example of the Australian government’s eagerness to sell out the legal rights of its citizens, the discussion paper has the temerity to suggest that the only reason to make DVDs accessible is by creation of a “business-case analysis for including captions and audio description” on them. The relevant standard is accommodation short of undue hardship, not accommodation if it’s profitable.
Time and again, the discussion paper reiterates the excuses of industry as if they held water.
Foreign TV shows are hard to caption because they arrive almost immediately after airtime in their home countries (usually the U.S.), it is claimed. This is nonsensical at best. Since nearly the entire captioning industry, including Australian captioners, uses the Microsoft of captioning software, Swift by Softel, it is simply a matter of sending a single E-mail with attachment to distribute an original caption file. The Australian captioner doesn’t have to work with a tape; it can easily handle an FTPed digital file. All the Australian house has to do is adjust timings for some captions, typically those appearing immediately after commercial or other breaks. Software can handle retiming subsequent captions. Even if it can’t, it isn’t exactly difficult to retime captions on one episode of one show. This isn’t “recaptioning”; it’s reformatting.
Once that’s done, there is no need anymore to encode a submaster tape for broadcast, which would eat up even more time. A related Softel product can simply play out a selected caption file at airtime.
Got a really short deadline? Assign more than one captioner. That may strike you as unthinkable. After all, it costs more than nothing.
Now, all the foregoing assumes reliance on a monopoly software supplier, but that’s the present reality and it works. Of course you end up with captions of detectably foreign origin, but that’s a problem international standardization will solve. And none of the above has to happen for U.K., New Zealand, or other PAL-system countries; if an uncaptioned program shows up at your door, all you need is the original caption file, whose timings work without modification, assuming identical commercial breaks.
If Australian authorities had more technical expertise in captioning, they could see right through these excuses.
The discussion paper falls prey to broadcasters’ serial deceptions concerning real-time captioning. Here’s what’s really happening.
There is no such thing as “voice recognition” for captioning. The discussion paper does not make any explicit claim that there is, but that’s the implication. The only thing we have is speaker-dependent voice recognition for transcription of dialogue. Get your mental image straight, please: HAL 9000 is not sitting there listening to your Australian TV show and typing out what everybody says.
It would help if, in this discussion, the sole term used were the correct one, speaker-dependent voice recognition. Such systems are used only to save money, not to improve quality; a good stenocaptioner is always and without fail a better captioner than a good respeaker. Nor is speaker-dependent voice recognition used to alleviate the problem of too few court reporters. Every qualified court reporter gets a job, but since qualifications are high, so are salaries. You get what you pay for, and increasingly people feel that computers should be able to do this complex task and why are we paying this much money every month?
Broadcasting executives, who love their computers even while not understanding them in the slightest, want a set-it-and-forget-it captioning model, where some guy in the office can bring in an old 486 he’s had laying around in the garage for a couple years and set that up to do captioning 24/7. “Voice recognition” will handle the problem. It all gets done on computers now. Bang: Problem solved.
Meanwhile, broadcasters use every excuse in the book, most of which are outright lies, to justify the use of real-time or just scrollup captioning on fictional narrative programming. It is impossible to understand fictional programming when captioned by scrollup. It’s even worse with real-time captioning. In case you don’t believe me, I’ve got a few tapes I could send you.
Broadcasters would prefer to waste everyone’s time, and misuse a limited resource, by assigning real-time captioners (using any technology) to caption a prerecorded show. The usual excuses are “The show came in uncaptioned” (then get it captioned) and its subset, “The show came in too late for captioning” (then get more than one person to work on it). The real reason is it’s cheaper to hire a real-time captioner. (This stands rather at odds with the common complaint that real-time captioning is expensive. Actually, it’s underpriced.)
In any rationally managed system (Australia’s isn’t), the use of scrollup or real-time captioning for prerecorded programming would be banned save for emergencies, and even then limited to the very first airing of a program. Additionally, you wouldn’t be able to count a scrollup- or real-time-captioned prerecorded program as having been captioned; it wouldn’t go in the credit column of the captioning ledger.
Of course Australia will not suddenly begin to rationally manage its system; it’s just easier, and less embarrassing, to accept broadcasters’ excuses and lies at face value.
The discussion paper’s quoted claim that speaker-dependent voice recognition could let “[s]tenocaptioners... speciali[z]e in captioning more complex programs” is laughable. The entire purpose of speaker-dependent voice recognition is to save money. Its secondary purpose is to eradicate the high-cost profession of stenocaptioning from the face of the earth. Also, I have no reason whatsoever to believe that any manger at any Australian broadcaster (actually, anywhere outside NBC and CBS) sits there and makes informed decisions about which shows are “more complex” and deserve better, i.e. more expensive, captioning. It’s a complete joke.
One additional point about real-time captioning. The discussion paper provides the following howler:
Live stenographers are only able to accurately provide continual captioning for very limited periods of time without a break in proceedings. Sporting events, such as the long-form Bathurst motor races, may go for ten hours or more.
So? You really think exactly one person is going to sit there for the full ten hours captioning away?
How exactly does anyone in this process think 24-hour news channels get captioned? Let me tell you how: In shifts, by multiple people. Your ten-hour motor race can and will be captioned by more than one writer. Seriously, it’s a scandal that total nonsense like this even sees the light of day, let alone is quoted without rebuttal in a government discussion paper.
If the broadcast day is defined as beginning at 0600 and ending at 0000, then the highest percentage of captioned programming available is actually 75%, since 25% of the 24-hour day isn’t even regulated.
The definition of broadcast day is indirectly discriminatory in that it permits unfettered broadcast of programming inaccessible to people with disabilities. Even if such a definition were retained for other regulatory uses, captioning and description targets have to apply to the full measure of a broadcaster’s output. If a station runs n hours a day, then it has to caption and describe all n hours.
As some Australian readers will be aware, I have been beavering away for a number of years trying to put a pittance of money together to research, write, test, and publish a set of international standards for captioning and audio description (among other fields). The Open & Closed Project would additionally develop a training process and certify practitioners. Once all that was done, captioning and description could be carried out by certified practitioners according to a strict, evidence-based international standard. This beats the hell out of, say, Red Bee’s cornering the market or mom-and-pop shops’ opening up with the pledge to do better captioning at half the price. (The paper’s warnings that forcing broadcasters to caption more shows might “prioriti[z]e quantity over quality” are a bit of a laugh, as that describes the entire history of captioning and its present day.)
The discussion paper leads me to believe Australia will trod a well-worn path toward failure. The only thing you’re thinking of doing is convening industry meetings of some kind every now and again to jot down what the industry is already doing. This will then be presented as a voluntary standard; all that means is there will be two ways it won’t be worth the paper it’s printed on. And for audio description? An industry lobby group is working on a “scoping study.”
If anybody gets to work on captioning and description standards for your nation, Media Access Australia are the people who will want to do it. Like a daughter (or, I suppose, a very special son) dreaming of becoming a ballerina someday, unrealistic fantasies should be discouraged. Single-nation standards aren’t going to solve the problem because Australia is not engaged in single-nation broadcasting. The discussion paper spends a lot of time complaining about how hard it is to caption (“recaption”) foreign programming. Well, that’s going to stay hard to do if Australia captions one way and every other country captions 20 different ways each.
An Australian-made standard is not a stepping-stone to an international standard. It is the antithesis of an international standard. It defies international standards. What the Department, Media Access Australia, and the industry in general need to do is support our project. Funding would be nice, but I’d be happy to start with getting you all to give up the idea of a national standard in the first place. Your captioning and audio description is not strictly “national.”
If the reasons listed above aren’t enough for you, try this one: Your voluntary national standard will just get ignored anyway, so why waste everybody’s time?
In another recapitulation of established failures, the discussion paper discusses how to file a complaint about captioning. I’m sure other respondents will remark on the fact it can take 90 days to get a response, such response of course being in most cases “There was never a problem,” “Your equipment must have malfunctioned,” or the blandishment “It’ll never happen again.” When – inevitably – it does happen again, a new 90-day clock starts.
Complaints are pointless if it takes 90 days to get a response to a complaint about a half-hour show that has long since come and gone. I assume the complaints process is deliberately set up to frustrate and vitiate actual complaints. That is its result, in any event.
Meanwhile, it borders on impossible to actually file a complaint about an inaccessible program. You’re lying on the couch at 10:00 at night (or you’re busy in the kitchen right after school when the kids get home). What are you supposed to do – take detailed notes and call the TTY line for half an hour from work the next day? Captioning complaints are only barely possible when registered in person with tapes rolling to demonstrate the problem, let alone through any other method.
If you’re blind, by what actual method are you expected to file a complaint? (Can you use your Perkins Brailler to write it?)
Is it the responsibility of individual viewers to police the actions of multi-million-dollar broadcasting conglomerates, or is it the active duty of broadcasters not to screw up in the first place?
It is time to distance ourselves from a reactive and adversarial complaint process. And the only way to do that is to establish independent, researched, and tested standards to improve quality up front.
Broadcasters have proffered their usual lies about what kinds of shows are difficult or impossible to caption. Sticking with the English language for a moment:
Of course most digital platforms make closed audio description impossible. But you can produce the same effect on some digital distribution networks, i.e. digital cable or satellite providers, by creating a virtual duplicate channel for a network. On that virtual channel, you run main video plus captions and swap in the described audio track for the undescribed one. All you need is software and a feed of the description track, which is not difficult to source for the small number of broadcasters involved.
Far-fetched? Hypothetical? Impossible? None of the above: Bell TV (né Bell ExpressVu) in Canada has done it for years and does it today. It may or may not be technically possible on existing Australian services. If it isn’t, then the system needs to be upgraded to make it possible. Remember: Only undue hardship is a defence in this context, and system upgrades happen all the time anyway. (“We never thought of that and we don’t feel like trying to make it work” isn’t a reason not to do it.)
Australians already think captioning sucks. The discussion paper plainly states that, in surveys, “there was a general dissatisfaction with the current level of captioning among individuals and representatives of people with a hearing impairment.” Australians also know that audio description is nonexistent: “Representatives of people with a vision impairment note that there is no audio description on television.” Taken together, this means you’re screwing it up. Now, what is the government planning to do about it? According to the paper’s own evidence, nothing.
Here are all the “approaches” being “considered” that would result in precisely zero action or no outcome at all. (Emphasis added.)
No one in Australia is in a position to complain about low levels of captioning in first-run cinemas. It’s your own fault: You hitched your wagons to open captioning, a known failure everywhere it’s been tried.
First, let’s fix the discussion paper’s technical errors.
I’ve seen open-captioned films in Sydney (the old Insight Cinema burned-in open captions) and London (DTS projected open captions). I’ve also seen 68 closed-captioned films in Canada using the Rear Window system. Not only am I informed about the topic, I’m experienced. Barely anyone resident in Australia has seen a movie captioned with Rear Window.
In your country’s rush to get any kind of captioning on movie screens, you signed up for open captioning. People who hate captioning, or just don’t want to do it, or are doing it only because they’re forced to, will be the first to tell you that hearing people hate open captioning. This claim has never been tested anywhere, but people act like it’s true. They schedule only a few captioned screenings – by the discussion paper’s admission, “two to three” (2.7? 2.9999?) a week at certain independent cinemas.
Meanwhile, hearing people can and do go to any showing. Not only is this unfair, it’s blatant discrimination. But it’s your own fault: You chose to install open-captioning systems which always result in unequal, i.e. illegal, access to movies for deaf people. There is nothing technical that stops a cinema from running its open captions all the time. It’s just that they always have refused and always will refuse to do so.
You’re just like the British: You will go to any lengths to avoid admitting you made a fatal mistake. You’ll do this even while admitting that the closed-audio-description system that comes as a package deal with DTS open captioning makes every screening accessible to the blind. Audio description of cinema came to Australia later than captioning did, but it means blind people have greater access to cinema than deaf people where there’s a system installed in the first place. That’s because deaf organizations, cinema operators, and the Australian human-rights apparatus set it up that way.
For the better part of a decade you’ve had a viable closed-captioning system available to you. It’s so viable it uses the same driver hardware you’re already using. In fact, it’s so viable you’ve already used it twice, for those “demonstration” screenings that captioning opponents are always happy to set up. (Unlike the British, you weren’t bought off by DTS, which plastered the United Kingdom with “demonstration” installations that, surely by accident, resulted in a lucrative exclusive contract funded partially by lottery proceeds.)
Let’s talk about cost. Rear Window hardware doesn’t cost a lot. It’s impossible to take cost complaints seriously when cinema operators across your country are busily spending money on voluntary upgrades for 3D cinema (including 3D Imax cinema). The industry’s 3D upgrades have permanently invalidated any complaint about cost, which isn’t the barrier here, if there even is one. The problem is policy, and its attendant embarrassment. Australia decided by policy that it would use the cinema captioning method known to be a failure at providing equal access. And like any large institution, the last thing you want to do is admit you were wrong.
Rear Window captioning has drawbacks. It isn’t true you have to sit directly in front of the display (I know some people won don’t mind off-angle viewing), but it helps. You can’t sit really far away from the display. Cinemas never take good enough care of reflector panels, which end up scratched and dirty. You stick out like a sore thumb, at least while the lights are up. And, importantly, some people with visual impairments, including middle-aged people with presbyopia who don’t consider themselves visually impaired, really can’t use the system. (They cannot easily switch focus from nearby captions to faraway screen. For them, you still need additional open captioning.)
Also, nobody wants to install Rear Window systems in every auditorium, which itself guarantees unequal access when you consider all the auditoria in a specific movie house. But that problem is surmountable.
The fact remains that, for a lot of people, a well-configured Rear Window installation works great. Once you’re in your seat and fully dialled in, you can really enjoy a movie.
So: How do you fix the problem? By starting a massive installation drive of Rear Window systems across the country. This will cost money, but you’ve got money to spend. You’ll also need to license certain parts of the technology from WGBH so you can caption Australian and other movies using the system. (Don’t worry: Your existing Swift software works fine to create the captions.)
If you’re serious about making movies accessible to deaf people in Australia, which I don’t think you are, you’ll accept reality, admit you were wrong, and start the upgrade process.
Of course you won’t, though.
The discussion paper redundantly mentions “font type” as an issue in caption “quality.” You don’t have a choice of font in analogue TV captioning. Only the stupidest and most expedient choices for DVD bitmap captioning and digital bitmap captioning are ever made, viz the nonfont Arial and the typeface backed up by reams of junk science, Tiresias.
Australia can be relied upon to make all these same mistakes over again. Everybody else has, and everybody in the related industries uses Windows and doesn’t have any taste in or understanding of typography. The foregoing remains true despite the existence of an entire Web site devoted solely to typography of captions and subtitles.
– 30 –
Posted: 2010.01.28 12:41