Ill Communication

So Malcolm Turnbull has just announced that captioning requirements will fall within the remit of the Government’s next adorably labelled “Red Tape Repeal Day ™” on 29 October. This is extremely troubling, as captioning requirements in Australia represent carefully designed, evidence-based policy which in their current form balance the needs of stakeholders, who include regular viewers, occasional viewers, content creators, broadcasters, captioners, advertisers and any data analysts who may use captioning output as metadata. Also, as some broadcasters hew tenaciously to the bare minimum of their legal captioning requirements, any loosening of the latter runs a very tangible risk of making your life worse if you use our services.

Not all red tape is bad you guys.

I’ll state at the outset that because these changes to captioning requirements form a small part of a much larger and fairly intricate repeal of broadcasting legislation, I’m relying on relevant sections of the explanatory memoranda authorised by Malcolm Turnbull and the initial report by Media Access Australia (who will no doubt have more to say in future). If I’ve misunderstood anything by not digging through the raw text of The Broadcasting and Other Legislation Amendment (Deregulation) Bill 2014 then please let me know.

The headline change is that free-to-air broadcasters will no longer have to report annually on their compliance or otherwise with captioning requirements. Instead, per Turnbull, “we are moving from annual reporting to a complaints-based approach”. There are a number of problems here. Broadcasters (and in some cases the third-party captioning companies they contract) are the ones best placed to collect data on their own captioning quality. Every time we have a loss, the supervisor at the front desk has to take action. To wake up the captioner, to find another captioner, to check the gateway, to scramble tech support, to contact the network – to do something. Once the crisis is over, they’re therefore perfectly situated to jot down a quick report. Time of day, amount of loss, reason, solution. Or when our accuracy slips below an acceptable threshold, we can refer back to our comprehensive text logs, as well as auditing them randomly to make sure all captioners are maintaining 97.5 or 98 percent. Networks will now only have to keep AV logs for 30 days. At our end we don’t keep comprehensive AV logs, but I assume the networks do anyway, for many reasons.

Why wouldn't they?

So the existing statutory annual report does not involve providers collecting any information we don’t already collect (a phrase Turnbull and Brandis seem to like). Annual reporting is just compiling the info we already have, and I would hope that my employers will continue to collect it anyway, so they can at least satisfy themselves as to their quality – never mind the regulations. As I’ve noted before, we already have a “complaints-based approach” as well. ACMA can hear and investigate complaints, and have done so a few times this year. So don’t be under the impression we’re adding any new types of scrutiny. This is purely about taking away accountability and eliminating records of shitty performance.

And there are some problems with a solely complaints-based model, which is a bit like dissolving the Tax Office and letting people just dob in their tax-dodging neighbours. Viewers don’t necessarily know their rights with respect to captioning. If a show isn’t captioned, they can’t necessarily be sure that it’s required to be. They might complain only to find it exempt from the requirements, or they might not complain because they don’t know they can, or don’t know who to contact, or what kinds of accuracy to expect. If like so many, for instance, you were under the impression we type fast on a normal keyboard, you might be so dazzled by our speed as to be unduly tolerant of bad performance, or bewildered by all the homophones and uncertain whether to complain. Programs with a civically-engaged viewership, like Mediawatch or The Project, might get dozens of complaints in the event of a sloppily captioned episode, but a dodgy Monster Trucks Almanac might get far fewer. We mustn’t be in the business of discrimination.

But instead the business of awesome.

The next part includes a kernel of sensible policy, but still causes some consternation. ACMA will be required to take into account whether a program is pre-recorded or live, or (most intriguingly) a combination of both, when determining whether captions are up to scratch. In itself this may provide a more nuanced analytical instrument. Our standards do and should vary between pre-recorded and live content – our workflow certainly does. The danger of introducing this in the context of red-tape-obliteration is that live (or “late” – where the media gets to us beforehand but too late for the more involved offline captioning process) could become a get-out-of-jail-free, where the networks justify substandard content by calling it live. There is also a proposal to make “engineering or technical failures” a cause of exemption. While I hate having my accuracy stats compromised by technical failures, the simple problem here is that the engineering, too, forms part of the networks’ responsibilities. Now they will have less incentive to build and maintain backup and redundancy plans, and to exceed their captioning requirements in case something goes awry. Caption an extra hour every day and you can afford an hour-long catastrophe, any day of the year. They may have less incentive to give us early access to files, or to make scripts and other resources available which improve live captioning. I just fear it will manifest as an overall loosening of the standard.

A similar concern accompanies the planned extension of the deadline to apply for exemptions to captioning standards. I’m not bureaucrat or wonk enough to know the implications of this – I just make captions. But prima facie it looks like a less stringent standard. Like more things may slide.

Things are definitely gonna slide, in all directions but up, as a result of the plan to let associated groups of sports channels “average” across their networks to meet their targets. As it stands now, each individual channel must meet its targets. In some circumstances, such as during the Olympics or World Cup, which may show for example on two of their five channels, they will voluntarily exceed their requirements on those channels. This change, though, would let them use that voluntary and commercially viable exceeding of the rules as a licence to neglect their other targets. Sports networks usually have a predictable weekly schedule and caption accordingly, so it will also likely increase viewer uncertainty about where to expect captions. A sport which required captions last week to meet the quota might not this week, solely because the Commonwealth Games is on the other network. And this too decreases incentives on networks to exceed requirements, in case of technical faults or other problems.

The next requirement is that new subscription TV channels will be granted a 12-month exemption from captioning requirements. This kind of just makes me sad. A start-up TV channel obviously has its work cut out for it and I understand the impulse to make it easier on them. Anything for a bit of media diversity. But a brand new channel is also such an opportunity. If they had even a modest captioning requirement from the start, they would be more likely to build-in efficient systems and procedures, which could then be scaled up. The worst captioning clients to work with are those for whom we are an afterthought (or non-thought), the best are those for whom we are part of the production process.

The last proposal is that programs previously captioned on a different subscription channel will no longer have to be captioned. Firstly, I don’t like it because it discourages networks from including caption files with programs they sell or repeat, which is totally a thing that can be done. Secondly, it sounds a bit like a butchered response to Media Access Australia’s suggestion that, since quotas are confusing, and tracking what is a repeat is difficult, simple 24-hour quotas should be substituted. The profit motive should ensure that the low-hanging fruit of repeat captions get picked anyway. Of course, there doesn’t seem to be anything here about strengthening quotas, so I guess the point is diminished responsibility.

“You wouldn’t like me when I’m disappointed.”

The net effect is an overall decline in standards, stringency and enforcement. I want to finish by illustrating, though, how this will not just potentially, but certainly, damage the quality of captions. As I’ve mentioned, this 98% accuracy standard I maintain for my respeaking means one word in 50 is wrong. At the start of this sentence, this paragraph was at 51 words, so imagine an error in every block of text that size. Now, I maintain it by usually having about 30 minutes of prep or standby time rostered for every 30-minute live session, or before each block of three or four “15-on, 15-off”cycles. Live is live – if I’m unprepared I can still do my thing. But for every minute less prep time, I have to cut corners. Since it’s NFL season, I’ll use the example of an American college football game. Let’s say there’s a game on between the Massachusetts Coffee-Preferers and the New Mexico Coverups. The captioner has had a crash, and their co-pilot isn’t available. A supervisor puts me on the air, and I immediately start talking. Now while on the air, I’ll engage my NFL house-style. I’ll google the teams and put their quarterbacks in my temporary macro slots, then do the same with the coaches. My Dragon won’t know the player names, and I can’t program them in while live (American football teams are huge). But with the playmakers locked in, at least some of what the commentators say can be transmitted without saying “this fellow threw the ball to that fellow, and this other fellow blew the whistle as the first fellow was tackled by yet a third fellow…”. All this googling also means I’m not watching my output, and so any Dragon errors will definitely go uncorrected. It’s good enough for emergencies, but hardly optimal.

Let’s imagine the same situation but with five minutes of prep. Now I can google teams, organise player names into easy-to-read lists and copy them in as required once I’m on the air. I’ll still start off with the quarterbacks and coaches ready to hand, and now I can watch my captions go out while I’m live. A quick sequence of passes will still be beyond me, as I can only juggle a few names at a time this way, but it’s an improvement.

10 minutes of prep, and I’m doing the above but also training the names (both surnames and full names) into my Dragon. So instead of saying “macro one” I’ll actually be saying “Jafarius Jones III”. Now we’re getting somewhere, though once on air I find that half the time it’s coming out as “Jeff ARIAs Joan’s a turd”.

20 minutes of prep, and I start getting time to practice each name off-air, maybe using in a sentence. I see what works, what needs re-training, what needs a house-style to correct an obvious homophone. Now the player names will be a pretty well-oiled machine (I still keep the list open in case something goes pear-shaped). But then they start talking about past player Steeve Stevesen, whose record is about to be broken that day, and back-room talks about who has been approached to replace flagging coach Nigel Blowhard, and I’m a bit lost.

So with the full 30 minutes, I’ve also skimmed a few sports pages, harvesting the names of other coaching candidates Felix Moneyball and John Lapsalot, and finding out the details of Stevesen’s record. I’ve also got the name of the stadium (the Starbucks-Sons of Liberty Arena) and found its nickname (Whipless-Mochachino-Freedom Park). That and only that is how to exceed 98%.

The captioning industry is competitive, and we already don't have enough prep time. Whoever can satisfy the requirements for the lowest price will get the contract, and captioning companies undercut each other on price. There is no possible world where reduced requirements in volume and accuracy won’t be passed on to captioning companies in the form of pricing pressure. Which must then be passed on to captioners in the form of decreased preparation time. Which must then be passed on to viewers in the form of unreadable nonsense. The only winner is the networks’ profit margins, and you, dear reader – I’m going to have so many mishaps to report in the future.

---

Disclaimer.