I have here a booklet entitled HDTV/Digital Television and Captioning by the Canadian Association of the Deaf (James D. Roots, Kim R. Nichols, Henry Vlug). There’s no ISBN, but there is French on the flip side (TVHD/télévision numérique et le sous-titrage). The English runs 44 pages of overlarge laser-printed type (more on that shortly), the French 48 pages.
And if you’re a Canadian resident, you paid for it: “Financial support from Industry Canada to conduct the research on which this report is based is gratefully acknowledged.” More accurately, you paid $48,820 for the research; you have to mail $15 plus shipping to the CAD to get the booklet. But you needn’t bother. Even though I am a source whose writing was cited in the report, I can’t recommend it to anyone, let alone anyone who wants to learn more about the technical or operational details of captioning for digital TV in Canada. The report is almost relentlessly inaccurate, is thoroughly padded with dropped-in documents from other sources, and uses government funds to advocate an activist agenda unrelated to the stated topic of “HDTV/digital television and captioning.” To top it all off, the report ends with a recommendation that CAD be given a kind of veto power over all captioning in Canada and multi-year guaranteed funding.
Author James Roots stated in an E-mail that CAD received funding to publish 100 copies of the report “for distribution to parties specified in the application.” He saw no choice but to sell the report at retail in order to improve the cashflow of the Canadian Association of the Deaf. He explicitly rejected my suggestion that the report, already largely paid for by the government, could be a simple free download at zero marginal cost to the CAD.
Having bought the report myself, I see an even greater reason to put the full text online: Twenty of its pages are nothing but pasted-in reports, orders, and press releases from the CRTC and the FCC. You might think that all such information could be handled by simple hyperlinks on nice accessible Web pages, and you’d be right. (In that case, you’d have to use something other than the CAD’s noncompliant site.)
Moreover, there are only a couple of pages that do not contain a typo or spelling error (including the copyright page), and the whole thing is dreadfully “typeset” in the 12-point Times New Roman that is the dead giveaway of amateur use of an amateur tool, Microsoft Word for Windows. I expect that the perfect binding and coated cover stock (which quickly curls) are used to justify the price.
At times, the report is an activist screed. I welcome such things, since I have written quite a few of them myself, except I’ve never written an activist screed masquerading as a government-subsidized research report on a technical topic.
The report spends too much time for a government-subsidized research report on a technical topic treading over well-worn ground about deaf people’s right to watch captions. I’m not sure there’s anybody left in Canada who thinks deaf people do not have that right; everybody’s just arguing over details.
Roots confidently recounts the history of a few showings of open-captioned films in Canada in the 1970s as “establish[ing] the CAD as the premier expert on the captioning audience,” if he does say so himself. Given the range of the captioning audiences (plural) that Roots later mentions, I doubt that, for example, English as a second language learners or “senior citizens” are anxious to join a club of which Roots would have them be a member.
“Captioning consumers are no longer content to be mere bystanders,” Roots fairly shrieks, “as their right to access information and entertainment is toyed with by industry players.” Captioning “consumers” are clearly not “mere bystanders” when large numbers of television, home-video, and movie producers and distributors cater to them by providing captioning (in many cases voluntarily). Nor are those consumers mere bystanders if, by the report’s own admission, they are often successful at filing human-rights complaints that ultimately require captioning.
“It is not widely known outside the Deaf community that deaf people” – lower-case d in original – “are quite possibly the most disempowered minority in North America.... Virtually all of these well-meaning but disempowering people are not deaf themselves.... Captioning is one more way in which our daily lives are at the mercy of non-deaf persons.” Roots does not explain how captioning could even take place without hearing people, or what would happen to the shows he watches with captions if hearing people, who make up nearly all the staff and on-air talent on each show, were eliminated.
Given the ubiquity of captioning, it might be more useful to agitate for captioning on what isn’t captioned yet and better captioning everywhere, if “second-rate access” is truly the blight the CAD seeks to remedy: “Why must we grovel,” Roots cries, “to attain even second-rate access to the same programming and products that non-deaf people [sic] enjoy?” I’d say you can’t be doing a whole lot of groveling if the government of Canada pays you to write a technical report that you ultimately lace with activist complaints.
The CAD uses government funds to write a report that nakedly advocates that the CAD be given a kind of veto power: “Intervention... by providing resources through which a consumer organization such as the Canadian Association of the Deaf can participate as an equal partner in the development of technology and content are short-term measures that will pay off in the long term,” we are told. It would certainly pay off quite handsomely for the CAD (or “a consumer organization such as” the CAD).
If the CAD is serious about this, would it support some other consumer organization’s receiving these “resources”? What organization would that be? Given that the report itemizes a long list of groups in the captioning audience, who decides who the “consumers” are? Must they always and only be Deaf people, complete with capital D?
The CAD proposes that, if it’s brought to the table in this way, complete with “resources,” doing so will “remove the need to impose any intervention or regulation on the producers and distributors of DVDs and videos.” Essentially, if you give CAD greater access to the halls of power and write them a cheque, you won’t be regulated by government. The Association seems to have so little power it begs for a seat at the table, yet so much power it can promise to forestall legislation by Parliament.
Industry and government might be wise to consider their options carefully before handing over “resources” to the CAD for this kind of role. We’ll read more about these demands in the report’s recommendations.
I find the report’s examples shoddy and ill-researched.
Roots attempts to elucidate the difference between subtitles and captions through the geek-favourite example of Monty Python and the Holy Grail. (I love it too.) “The joke is that the knight” – it’s King Arthur – “can’t afford a horse, so he is keeping up appearances by having his lackey” – it’s Patsy – “bang together coconut halves to simulate the sound of horse’s hooves.” Such a sound effect is missing from subtitles but present in captions, which “explained ‘(banging coconuts like horse’s hooves).’ ” He adds “I am quoting from memory [of the VHS tape] and may not be 100% accurate”: Well, why not just check the DVD?
![Scene shows King Arthur and Patsy on a misty moor with the caption [ Coconuts Clacking Together Imitating Sound of Hoofbeats ]](HolyGrailClacking.jpg)
It certainly grew tedious to read (twice) the tale of the release of the first season of SCTV on DVD, which appeared without captions. After a letter-writing campaign, the second season was captioned, though “even these efforts could not persuade the producers to add captions to the second pressing of the first series.” It isn’t a question of “adding” captions, like some kind of decal on the package. The entire set of video assets must be remultiplexed using a Line 21 caption reference file in a specific format. The packaging must also be changed, though we’ve all had the experience of mislabeled DVD packaging.
I was frankly disappointed by the chapter entitled “Transition from Analogue to Digital TV Broadcasting: What Will Happen to Closed Captions?” by Henry Vlug, whom I quite like and respect and whose house I’ve actually visited. (We correspond by E-mail from time to time.) Dated October 2004, the chapter is a simple copy-and-paste job, with hundreds of words dropped in wholesale from government and regulator Web pages. (In fact, the URLs are unhelpfully and inelegantly plopped right down there on the page, too, complete with extremely swanky underlining. When was the last time you transcribed a two-line URL, complete with ampersands and equal signs, from the printed page to your browser?)
To save you the trouble, here are the links to the items Vlug recapitulates:
The final chapter, by (Mr.) Kim R. Nichols, carries the title “The Technological Research Report” and is replete with mistakes.
I still don’t know what the report means when it variously claims (as it does again in this section) that “DTV technology has the potential to place control of captioning into the hands of the consumers within a few years.” Exactly how much control do you need? A pressing issue to any captioning viewer is caption quality, which is not addressed at all by HDTV, except to complicate matters tremendously. What is the “control” that DTV technology will place into consumers’ hands, apart from selectable fonts, sizes, and colours? (That seems to be the case, given that Nichols copies and pastes a list of DTV features from Gary Robson’s Captioning FAQ, credited simply as “According to Robson of robson.org.”)
Nichols also seems to be talking about voice recognition, which we are (again) told will provide captioning that is “routine even on computers, personal device assistants (PDAs) [sic], and cell phones.” Nichols continues, bafflingly, “The possibilities are unlimited... yet today’s captioning progress lags behind the accelerated developments in HDTV.” Captioning is part of HDTV and, if captioning truly “lag[ged] behind,” we wouldn’t have a standard (CEA-708) documenting the captioning facility of HDTV.
Nichols confuses the concepts of “digital television” and DVD. “Sales of HDTV[s] have increased 50%, which has had the result of increasing the demand for DVD players, HD decoders and home-entertainment units,” Nichols writes.
Nichols confuses a satellite-TV service (DirecTV, as they write it) and a personal video recorder (TiVo), neither of which is obtainable in Canada. (The Supreme Court definitively ruled grey-market satellite dishes illegal in 2002. TiVo just isn’t sold in Canada, though a few Canadians have hacked together the program guides necessary to make TiVo work here.)
He refers to “satellite manufacturers (DirectTV, TiVo, ExpressVu),” none of which actually manufactures satellites. He explicitly writes that digital video recorders are “[o]therwise known as Direct to Home” and “are the set-top services such as DirectTV, TiVo, and ExpressVu.” Sadly wrong yet again, but in new and amusing ways.
Is “the speed of this data packet” in analogue 608 captioning really “960 bauds per second,” given that “baud” is a really old word that already means “bits per second”? (It isn’t a unit of measurement like centimetres; it’s more like hertz.) Hence does digital 708 captioning really run at “9600 bauds per second”? In fact, pretty much the whole discussion of transmission speed is a jumble.
HD does mean high definition, but SD in this context does not mean “serial definition”; it means standard definition, that is, analogue. The TVs you grew up watching are SD; a new TV you buy in the future may be HD.
Nichols has an incomprehensible paragraph about MPEG, and seriously fails to describe PSIP adequately. (Your program might be captioned, but if the Program and System Information Protocol doesn’t have exactly the right information, the home decoder will pretend the show isn’t captioned and will not display any captions. Like digital TV itself, it’s an all-or-nothing proposition. See “Implementing Closed Captioning for DTV” [PDF].)
Nichols does a good job of explaining the inconsistencies in FCC requirements, some of which force broadcasters to transmit analogue captions that digital decoders don’t have to accept.
The author takes literally the marketing bumpf of entrepreneur Mark Cuban, and Nichols copies and pastes Cuban’s description of his HDNet channel as “a DTV ‘conglomerate.’ ”
The Association’s report reaches a new low when it permits Nichols, in a chapter ostensibly on HDTV technology, to gripe endlessly about how cheap captioning is compared to Hollywood production budgets. I’ve been reading this argument for 20 years. Someday captioning activists will figure out it isn’t a question of cost but of willingness.
Using figures from NCI (which is suspect right there, as a competitor in the U.S. captioning market is not an unbiased source), Nichols copies and pastes a series of claims about five U.S. TV shows and their acquisition costs for syndication (e.g., “Turner paid $1.2 million per episode for E.R.”).
“What is the actual cost of closed captioning within the context of these total production costs?” Nichols asks, going on to paste in a table of typical production budgets alongside typical captioning budgets. But let’s answer the question as posed: Zero! Every episode of every series he lists (Chicago Hope, E.R., Ellen, The X-Files, and Walker, Texas Ranger) was captioned for first run. The syndicated broadcaster doesn’t have to pay anything to caption the shows; they’re already captioned.
If the syndicator wants to trim the episodes or insert different commercial breaks or time-compress the episodes, then the tapes may have to be sent out for syndication reformatting at a price no higher than 20% of the original captioning cost.
Nichols isn’t even asking the right question. If he’s worried about the cost of captioning vs. the cost of original production, he needs to worry about the cost of HDTV captioning, not analogue captioning. As of the report’s date (October 2004 for Vlug’s portion, February 2005 on the copyright page), there were almost no captioning systems that could natively produce 708 digital captions, and even those were rarely found in captioners’ offices. Very few hours of original programming have been originally captioned in original 708 formats. (I don’t know of any such programming at all. I cannot name a single show.) This cost needs to be addressed, and, given the Industry Canada funding for “research,” I would have expected to see an accurate and well-sourced discussion of this issue.
Instead, we are treated to a warmed-over diatribe about rich Hollywood producers being too cheap to caption. In truth, it’s very difficult indeed to find a Hollywood production that isn’t captioned, sometimes several times (for first-run movies, for DVD and VHS, and for TV, all in several formats each). If your concern is mere presence of captioning rather than quality, Hollywood productions have got the problem mostly solved.
The example of SCTV episodes released uncaptioned on DVD shows that it is B-list or edge cases that require higher scrutiny now. (And even the SCTV case was resolved!)
After all that time spent explaining that captioning really is quite inexpensive, we are informed that the magic bullet of voice recognition will make “captioning... seamlessly integrated into the production process and... almost cost-free. Certainly, 708 technology would render the financing of captioning as no obstacle whatsoever.” I thought financing already wasn’t an obstacle because it’s so cheap.
If you toyed with the idea of buying the booklet to receive accurate information on an obscure topic, let me disabuse you of that notion right now.
We are told that “[o]ver 3 million Deaf and hard of hearing Canadians [sic] depend upon captioning.” In fact, Statistics Canada estimates the total Canadian population with a hearing impairment at 1,038,140 for adults and 20,590 for children aged 5–14 (figures from 2001). That’s 1,058,730, not 3,000,000, and not all of them need captioning. CAD’s numbers are high by roughly two-thirds.
Further, the report lumps together into the captioning audience ESL and FSL learners, low-literacy persons, “senior citizens,” people with other disabilities like autism, and “people who are not normally captioning consumers but need captioning in order to ‘hear’ the TV over the noise of a bar or a party.” All very true. But there is no source for the report’s next claim, that “[t]he conservative estimate of the total captioning audience in Canada... is ten million people.” Do one-third of your friends watch captioning?
Neither is there a source for a later claim of an “audience of about 43 million people in North America alone who are deaf [lower-case d in original], deafened, or hard of hearing.”
Roots tells us that “captions can now theoretically be produced in different fonts, colours, positions, and even languages,” overlooking the fact that analogue Line 21 captioning can do all of that except switch fonts.
It gets better: “No longer are we stuck with white sans-serif fonts scrolling up in front of a rigid black background that blocks the lower third of the TV screen.” We’ve never been stuck with any of that.
Kim R. Nichols muffs it severely when he writes that “captioning in lower case includes serif fonts, which make it difficult for viewers to distinguish between ‘g’ and ‘q.’ ”
He continues: “On the other hand, ‘all-caps’ captioning is equally taxing to read and often rules out valuable information such as an emphasis on a specific word of dialogue.” He’s correct that reading extended capitals is slower, though research shows the speed deficit can be overcome with practice. The part about emphasis is flatly wrong; maybe on a typewriter (or in teletext captioning) we have to resort to something as primitive as CAPITALS for emphasis, but Line 21 captions have had italics and underlining since the beginning. Italics have always been in wide use.
The Association’s report is a pæan to the future of voice recognition. Unluckily for us, we don’t live in the future.
The report tells us that caption “consumers” assume “that, just because a movie or program is produced in digital format, it is automatically captioned.... [W]e are not yet at the stage where voice-recognition technology creates captions at the request of the consumer.” Yet the next page declares that “new technology promises to some day deliver a chip or a box that the consumer attaches to his/her set or monitor in order to transcribe audio programming into visual format.” (What kind of “visual format”? Sign language, captioning, or something else?)
Voice recognition has been a claimed magic bullet for captioning since the 1980s, and a succession of charlatans and dreamers, one of them Canadian, has been peddling this vapourware for that entire time. If speaker-independent voice recognition that’s at least as accurate as inept real-time court reporters ever appears in wide use in my lifetime, I’ll be surprised – and so will the CAD, if they’re honest with themselves. In fact, the report provides the blandishment that “a voice-recognition/translation approach to captioning... seems inevitable.” Maybe it is, but how many decades from now?
Having already dipped its toes into speculative fiction, the report now dives in headfirst, stating that “one avenue that should be explored is linking voice-recognition-based captioning with audio-description services. Not only could the two signals coexist in the SAP, they could also ‘feed back’ into each other to produce spoken and captioned texts in multiple languages as well.”
The authors fail to note that automatic voice recognition of television programs will essentially force all captioning to be real-time captions with a lag between speech and caption. That would not be very different at all from real-time captioning as we know it today on shows that really require it. However, it would be a huge drop in quality for every program that is not real-time captioned today, because that’s exactly how they’d end up being captioned in the future. Do you want all captioning everywhere to act like live captions?
Smart money bets on Nichols’s dreamed-of voice-recognition captions turning out as follows: Scrollup captions in capital letters. In other words, all HDTV captioning on programs that aren’t actually suited to live transcription would use it anyway. Those programs would standardize on the worst captioning we’re putting up with today. Frankly, I’d prefer to retain the human element if this is the alternative.
My own experience casts doubt on the report’s claim that “it is still possible today to find analogue TV sets with... no internal decoder chip” in Canada, and that hotel and motel TV sets are even more likely not to contain caption decoders. I’ve seen one TV set in the last 13 years that would be covered by the U.S. legislation yet still lacked caption decoding, and I go to some lengths to look for them. Canadian hotels in which I’ve stayed have modern TVs with captioning. The statement needs much more evidence to support it, evidence I think does not exist.
Kim Nichols writes an entire section entitled “OCR (optical-character recognition) translation via 708 technology.” In it, he notes that 708 can carry multiple streams of captions and that there is separate technology in development to automatically translate text in one language to another. How this involves OCR is beyond me, since the translation machine will receive tidy encoded text, not printed words. Nichols views this as a solution for French captioning, where he reiterates the oft-heard untruth of “the difficulty of captioning the French language.” French is captioned every single day. In any event, machine translation of captioning is already happening.
Further on that topic, Nichols provides the blandishment “With voice-translation or OCR translation, all languages would be on an equal footing”: Does that include sign languages? And if so, what device would output the sign language?
The report openly calls for the imposition of government requirements to caption “all television and movie programming sold in retail markets in DVD and/or VHS format.” Let’s overlook what the demand itself overlooks – digital delivery of programming. (If you really want universal accessibility, don’t hitch your cart to physical media.)
Let’s look instead at the report’s legislative justification – that “access to the mass media... is a human right that has been recognized and upheld by powerful institutions such as the Canadian Human Rights Tribunal, and implicitly... by the CRTC.”
The Tribunal has jurisdiction over federally-regulated matters, which home video isn’t, and the CRTC has narrow jurisdiction over broadcasting and telecommunications. To put into effect what Roots calls for would require expanding that jurisdiction. The report seems happy to have this happen, but the fact of expanding jurisdiction is never discussed or addressed, and the separate comparisons to “regulations governing the cutting and selling of meat” are rather irrelevant given that nobody could get sick and die from an uncaptioned movie.
It would have been easy for the report to argue that human-rights legislation and the Charter require companies to provide accessible products and services, but the report gladhands over this important issue. I doubt anyone working in the “DVD and/or VHS format” will be losing much sleep if this is the best argument the CAD can come up with.
The report’s recommendations tend to waver between self-evident motherhood statements and downright scary and baldfaced appeals for money – and for the CAD to hold some kind of veto power over a range of services.
The authors (really, this section reads like classic Jim Roots) advocate adoption of the 708 standard. No problem there. However, if “[t]he CRTC should clarify rules for counting captioning hours during the transition period from analogue to DTV technology,” perhaps the report should have taken the time to explain why that’s actually necessary. (Where’s the explanation of the problem? Is there a problem?)
CAD should work... to establish an industry-recognized body with representatives from various digital-industry segments in addition to consumers. The obligations of this body would include establishing quantitative and qualitative evaluation of all Canadian-based television captioning.
In short, the CAD wants to set up a panopticon that will watch and rate all the captioning in Canada – and, on the qualitative scale, rate it against criteria that don’t exist. (Actually, they say “qualitative measurement would be dependent on the evaluation of each TV stations’ quality control protocol” [sic], whatever that means.)
More? Yes, there’s more.
“The federal government, in collaboration with the CAD, should reorganize caption training schools”: I was not aware that there even were “caption training schools” plural in Canada, and that they were so disorganized they needed to be “reorganized.” If we need the feds to do it, who says the feds need any help? Why should the CAD be the sole collaborator? This is a call for federal regulation in another guise. If we need the mighty federal government, then we don’t need the puny Canadian Association of the Deaf.
Of course, when “voice translation software achieves 100% accuracy,” we can shut down the French-language schools, the report suggests. (We were earlier told that 99% accuracy would be good enough “as an intermediary for closed captioning.” French needs an extra 1%? Or rather, English doesn’t need 100%?)
“Industry Canada should provide the CAD with funding and resources” – finally we see the real goal – “to pursue accelerated research in voice-recognition/translation... using SAP mode for descriptive video with HDTV and ensuring translation into various languages.”
This recommendation seems to be a naked grasp of power, a thrust to become the biggest fish in several small ponds at once. They pretty much listed whatever they could think of, whether relevant to HDTV captioning or to a research report on that topic – or not.
In cased you missed the part where the CAD was being honest, note well that “[t]he funding needs to be multi-year and guaranteed.”
HDTV/Digital Television and Captioning was a waste of time. In fact, it’s the worst report on captioning I’ve ever read. It’s laced with appeals to the self-interest of the authors’ organization, the Canadian Association of the Deaf. Government funding was provided for “research,” yet the report is strewn with factual errors and off-topic musings. The report was heavily padded with dropped-in copy from other sources. The report is interested less in HDTV captioning than in the promotion of an unproven technology, voice recognition. The report is shoddily prepared and a poor value for money at every level.
All of the foregoing tend to undermine the Canadian Association of the Deaf’s claims of being “the premier expert on the captioning audience.” In fact, they’ve proved the exact opposite.
I asked Industry Canada for certain facts about the CAD report. My questions and their answers follow (with minor copy edits, including removing an off-topic paragraph with instructions on how to get a hardcopy of a certain document).
The Canadian Association of the Deaf (CAD) was awarded a Research Project Contribution of $48,820 under the Contributions Program for Non-Profit Consumer and Voluntary Organization for the fiscal year 2004–05 for the project “HDTV/Digital Television and Captioning.”
The Applicant’s Guide is available on our Web site, Consumer Connection.
The 2004–05 Awarded Amounts are available.
The Canadian Association of the Deaf (CAD) submitted its proposal under the Contributions Program for Non-Profit Consumer and Voluntary Organizations administered by the Office of Consumer Affairs, Industry Canada.
The Program’s Applicant’s Guide is sent to interested organizations in the Fall of each year. The CAD was on our mailing list and received a copy. The Guide was mailed to over 100 organizations. The submission deadline was December 1, 2003.
Although a suggested priority list for consumer issues to address under the Program is included in the Applicant’s Guide, organizations can select issues of their own, as did the CAD.
Organizations compete for the funding and their proposals are judged on the basis of merit. Research project proposals key evaluation criteria include: timeliness and policy relevance; methodology, capability and cost; intended use; and communications. CAD’s proposal was assessed by Industry Canada’s Office of Consumer Affairs and Assistive Devices Industry Office; the Broadcasting Policy and Innovation Branch of Heritage Canada; and the Knowledge Economy Section of Finance Canada. In total, CAD’s proposal received a high score and was among the highest-scoring research proposals and merited funding.
The terms and conditions of an award under the Program are set out in a Contribution Agreement. The Contribution Agreement is of standard format for all contributions. The only differences are the amount awarded, the project summary, and project budget.
We can provide a copy of the standard Contribution Agreement, CAD’s project summary, and amount awarded, which are in the public domain. A copy of the standard Contribution Agreement [was] attached [with their response]. CAD’s project summary submitted to Industry Canada’s Office of Consumer Affairs is available on our Web site. The amount awarded is in the answer to question one. [...]
The Office of Consumer Affairs reviews reports submitted by the recipient organizations to ensure compliance with Contribution agreements and that the work was completed for the project as required.
At the end of the project, and before making results public, organizations receiving Research Project Contributions under the Program are required to submit an independent review of project methodology, conclusions, and recommendations. A CAD project reviewer opinion is on file.
The reviewer hired by CAD to review the report was a former tenured assistant professor and program director, Teacher Preparation Program in the Education of Deaf and Hard-of-Hearing Students, Faculty of Education, York University. He is now retired.
He reviewed both the Vlug and Nichols papers contained in CAD report. In summary, he felt that Vlug’s paper adequately represents the then-available body of CRTC and FCC literature, and concluded that Nichols’s paper summarizes the reality of the changing video technology that we are facing, and translates into a challenge for us to ensure that the captioning technology changes with the change.
No other review was made of CAD’s “HDTV/Digital Television and Captioning” report.
From this response we can conclude:
You are here: joeclark.org → Media access → Captioning
Updated 2005.07.31