I installed and subscribed to #ParamountPlus, learned that #audioDescription isn't available on many pieces of content in the UK, and discovered that a US-based VPN doesn't work if you subscribed to the UK offering. I have now unsubscribed from and uninstalled Paramount Plus. #accessibility #exclusion
@jscholes Wait, so the exact same content has audio description in the US and not the UK? Unbelievable.
@simon Apparently so. This is unfortunately, dishearteningly common, and why I wish there was an equivalent of the Audio Description Project for non-US markets so that the disparity could be laid bare.
@jscholes I wonder if there are actual copyright issues surrounding audio description across borders. $US-Company does audio descriptions, licenses them for use in the US. $Streaming-Service makes the content available globally but can't make the description available. The 0.001% of people who care can't make enough noise to make a difference. One more thing that would be solved with in-house descriptions.
@simon That's one of the primary issues. Even when there are no copyright problems preventing AD from being shared, the practicalities of different distributors by country and a lack of accessibility prioritisation also contribute. Another issue is the differing frame rate between broadcast standards used throughout the world, making synchronisation between AD tracks across markets difficult or impossible. That is particularly the case when content is primarily made for TV, and happens to be streamed online too.
@jscholes @simon Another problem is that AD is often mixed with the audio track directly, and companies might not have access to the raw AD track. As far as I know, many companies specifically request mixed in AD from their contractors. This means that integrating that AD into content available in other countries is difficult if there are any changes to that content.
@objectinspace @jscholes @simon What I’d really like to see is platforms adopting the British solution and distributing the AD track separately from the main audio itself, letting the end user’s device mix the two according to the user’s preferences. This way, you could do things like redirecting AD to headphones while letting the main audio play on speakers (when multiple people are watching together and you’re the only one needing AD), or even play English AD on top of a dubbed track, if foreign language AD isn’t available. Audio description is often much easier to understand than dialog, with characters speaking quickly and with thick accents, so people who speak a foreign language at a fairly basic level would benefit from something like this.
@miki This was only ever the case on Freeview (DVB-T and DVB-T2), and never on satellite or cable where it is pre-mixed at the transmission side. Even on DVB-T and DVB-T2, most receivers implement the mixing algorithm, but provide no end-user control over the levels. Other parts of the spec, like panning, have never been known to be used at all. So while I agree with you about its potential, it is essentially a failed experiment in the UK too. @objectinspace @simon
@jscholes @objectinspace @simon This wouldn’t necessarily be the case if you were implementing and end-to-end solution with your own apps, though.
@miki No, but it would help for their to be a system to point to that had made a success of this, and highlight all of the positive user impact. Implementing that level of on-device mixing is nowhere near as straight forward as not doing it, and as you point out, strong reasoning is often needed to convince a company that the additional effort is worth it. @objectinspace @simon
@jscholes @objectinspace @simon True enough.
@miki Separately, AD companies should be doing a much better job of quality control, and accommodating non-basic mixing setups. Everything else aside, it is unacceptable to be paid tens of thousands of dollars, on a repeated basis, and still be sending back files that are accidentally in mono. @objectinspace @simon
@jscholes @objectinspace @simon We’ll probably see this massively improve in the next couple of years. I know people who know people in the Polish AD scene, and I’ve seen what Eleven Labs can do to audio description, and I genuinely wouldn’t have known it was Eleven if I wasn’t told before. Amazon already has an in-house tool to make AD that is up to scratch, requires nothing but a laptop and an intern in Mumbai to make, and takes hours instead of weeks to go from nothing to being ready for publishing. They do it with crappy Amazon Polly, but there are much better solutions available now, particularly if you spend a few grand to train a model. The cost savings are going to be so huge that it's unimaginable that other companies won't follow suit, either with their own tools or with a SaaS.
@miki @jscholes @simon Yeah I know that this is gonna be contravercial because it will unemploy some of the people who were previously doing it... I really don't care to be honest. If it will enable mass AD adoption at scale it's worth it. I suspect, though, that the tradeoff is actually different, so probaly they will just use ELeven Labs but that won't actually translate to more AD being used. But we'll see!