Global Video Accessibility: Multilingual Subtitling, Dubbing, and Audio Description [TRANSCRIPT]
NOAH PEARSON: Thank you all for joining today’s session, “Global Video Accessibility– Multilingual Subtitling, Dubbing, and Audio Description.” My name is Noah Pearson. I use he/him pronouns, and I am on the marketing team here at 3Play Media. And I will be moderating today’s webinar. And with all of that taken care of, I’d like to welcome today’s speakers, Lily Bond and Erik Ducker. And we’ll turn it over to you guys.
LILY BOND: Thanks so much, Noah. Really excited to have everyone here. Erik and I are going to be presenting on global video accessibility– multilingual subtitling, dubbing, and audio description. To introduce ourselves quickly, I’m Lily Bond. I use she/her pronouns. I’m the Chief Growth Officer at 3Play Media. I’ve been here for just over 11 years, and I’m very passionate about what we do. So looking forward to the conversation. And Erik, I’ll let you introduce yourself.
ERIK DUCKER: Hi, everyone. I’m Erik Ducker. I use he/him pronouns. I have not been at 3Play Media for nearly 11 years. But I have been in the video streaming space for probably nearly 11 years. I’ve been with 3Play for going on about 3 and 1/2 years, and excited to share some of the product side of this webinar with Lily. So I will let Lily take it from here– or sorry, I will not let Lily take it from here. I apologize. I’m doing the agenda.
LILY BOND: I was ready to go, Erik. I was ready to go.
ERIK DUCKER: So we talk every day, and we did not find time to rehearse that one hand-off. So the agenda today. Our goal is for you to leave with knowledge on what’s required to publish video content across primarily Europe and North America. I know we mentioned it’s global. We are constantly building our own knowledge base of all global accessibility regulations. But today, because of the mass amount in Europe that’s just released this summer, we’re going to be really focused on Europe today.
So Lily’s going to go at length and talk about why offer localized and accessible content, but also preparing your content for different countries. And then I’ll come back and wrap up with how 3Play Media supports you on this journey. So Lily, now I’ll let you take it away.
LILY BOND: Perfect. Thanks, Erik. Thanks, Erik. I’m going to start by setting some groundwork around why global video accessibility matters before we dig into some regulations. Obviously, there are legal requirements for these solutions. But there are a lot of other reasons why you should be considering them and how they can help expand your business and growth globally.
So just to level-set, what do we mean when we’re talking about accessible and localized video? When we’re talking about accessibility, we’re typically talking about closed captions and audio description. So closed captions assume the viewer can’t hear the video. And audio description assumes the viewer can’t see the video.
Particularly in Europe, you might also see closed captions referred to as subtitles for the Deaf and hard of hearing. The key is that closed captions or subtitles for the Deaf and hard of hearing include text for all spoken audio, as well as key auditory information, like speaker identification, sound effects, critical sounds that you might hear in the audio track. And audio description, for those of you who aren’t familiar, narrates critical visual information in a video and is a secondary audio track that describes those visual elements auditorily in between the spoken dialogue of the video.
And when we’re talking about localized video, we’re talking about subtitles and dubbing, which assume that the viewer can’t understand the dialogue’s language, [INAUDIBLE] that dialogue into additional languages. Subtitles are text versions, and dubbing is audio versions.
Beyond regulations, as I said, there are many great value-adds for these solutions. So the first one is that they can really help expand your audience. When we’re talking about captioning, we’re looking at the Deaf and hard of hearing population, where there are about 1.5 billion people globally who experience hearing loss. There are 2.2 billion people globally who are blind or low-vision and require audio description.
And then 15% to 20% of the global population is neurodiverse. And this population really benefits from additional options for consuming assuming content– so captions as a text version of the audio, audio description as an– of– like an auditory version of the visuals. Sorry, I struggled with that one. And then 80% of the global population is non-English-speaking. So when we’re thinking about subtitling and dubbing, this is a huge part of the population that you require subtitles or dubs to tap into.
These solutions also help improve viewer experience. There’s a lot of studies around this, particularly in the captioning and subtitling space. But there is a 12% lift in brand linkage or brand affinity when you have captions on your video. Facebook saw a 13% increase in view time. The University of South Florida, St. Petersburg, saw a 28% increase in transferability of knowledge and retention with captions and transcripts, which is really critical, particularly if you’re thinking about training content or educational content. You want people to retain the information in those videos and be able to apply it to their job or apply it to being able to use your software effectively.
And then Netflix has also seen a 120% increase in consumption of dubbed content annually. So we are seeing across the board, viewers want these solutions. And businesses are seeing a huge lift in engagement from using them.
So that’s all of the reasons why you should be thinking about these, anyway. And now that you might be ready to publish, let’s talk about some of the regulatory requirements. We’re going to focus in North America and Europe, as Erik said. These are areas where there has been a lot of legal action, and there are very clear accessibility laws.
So in Canada, we’re talking about the Accessible Canada Act. There are also provincial laws. The Accessibility for Ontarians with Disabilities Act is the landmark one, but many of the provinces in Canada have their own laws. In the US, we’re talking about the Americans with Disabilities Act, Section 504 and 508 of the Rehabilitation Act, and the 21st Century Communications and Video Accessibility Act, as well as FCC regulations for broadcast. And then in Europe, we’re talking about the European Accessibility Act, or the EAA, as I just referenced, as well as the Audiovisual Media Services Directive, which covers more advanced requirements for media content specifically.
So to break this down a little bit, in the US, most [AUDIO OUT], the only outstanding requirement is an update to Title II of the Americans with Disability Act, which affects public entities. This update phases in in April of 2026. So many public entities and universities are starting to think about compliance for this rule now.
Across all of the laws in the US, we’re talking about compliance requirements for public entities and places of public accommodation, federal and federally funded programming, and all broadcast and streaming content. And all of these rules require slightly different things. But generally, they’re all referencing the Web Content Accessibility Guidelines, level AA, and FCC caption quality standards on the FCC and CVAA side.
In Canada, most of the requirements are active. The core requirements of the Accessible Canada Act are already active, but it does have additional requirements phasing in through 2040. And provincially, all of the regulations cover public and private organizations. The ACA covers all federal and federally regulated programs, and then there are some broadcast requirements as well. And most of these rules also reference WCAG AA requirements.
In the EU, most requirements are active as of a month ago, in June of 2025. The only outstanding requirement under the EAA is that any backlog content is exempt from the initial requirement, but does have to be compliant by 2030. So when we’re talking about the EAA, we’re talking about anything created after July 2025 has to be accessible moving forward. And then anything created before it has to be compliant by 2030.
This is affecting certainly all broadcast and streaming content, and then any businesses selling products or services to the EU, regardless of headquarters. So you can be a US-based company, but if you serve customers in the EU, you do have to comply with the EAA.
And the EAA is a directive. So it basically requires each member nation to transpose the directive into national law. All 27 member nations have done so. And we have breakdowns of each nation’s requirement that we can share with specific questions that you might have. Similarly, we’re looking at Web Content Accessibility Guidelines, level AA, and the European Harmonized Standards for Accessibility.
And what does that mean? That means closed captioning and audio description on all content. And then there might be some language requirements as well. So in the US, closed captions and audio description are required under all laws. And then you may have business requirements for subtitles and dubbing. For example, we have a large Spanish population in the US, Spanish-speaking population in the US. Adding Spanish subtitles or dubs may be a business requirement for you, or if you are looking to distribute your content in other regions around the world, you would want to subtitle and dub your content to reach those global audiences.
In Canada and the EU, again, closed captions and audio descriptions are required. But you have to intersect the accessibility laws and the national language requirements. So in Canada, there are parts of Canada that require French Canadian as a required language. So you have to provide closed captions, and then you have a national language requirement for English and French Canadian. You have to provide those captions in both English and French Canadian. Similarly, in the EU, if you are delivering content in English, Spanish, German, you need to make sure that the closed captions and audio descriptions are available in English, Spanish, and German.
And in terms of enforcement, in the US, it’s really lawsuit and litigation-based. You can have independent litigation or a lawsuit directly from the Department of Justice. In Canada, there’s a Chief Accessibility Officer who is responsible for investigations, financial penalties, and compliance orders. Financial penalties can be up to a million. And then in the EU, again, these are country-level regulations and enforcement policies that vary. And we can share any specific country-level requirements you’re concerned about. But there are a combination of lawsuits, financial obligations and penalties, and some kind of criminal punishment.
So that is it for the legal requirements. And now I will hand it over to Erik to talk through how you can actually implement some of these solutions and requirements.
ERIK DUCKER: Great. And Lily, you can breathe now.
[LAUGHING]
I hope everyone burns that last slide into your memory. I think it’s a really, really simple way to remember what you need based on the country that you’re working in. As Lily alluded to, something that we’ve been doing for the last 15 years as a company is making sure that we’re publishing educational content around these rules and regulations. So we will have country-level guides on, how do you publish video in Spain? How do you publish video in France? And how do you get ready for that?
So we’re going to have that information for you. And if there’s a country that’s not being talked about today, raise your hand. And we won’t be able to probably address it today. But it’s something that we want to make sure that, wherever you’re publishing video, whatever country that you’re really trying to get your content into, you’re prepared. And so that’s really what our team is here from a marketing perspective– to be a partner with you in terms of ensuring that your strategy is compliant and you can execute it seamlessly with your stakeholders.
So you need to understand what support you need. So it starts with your content. And really, the core of everything that we do at 3Play is making sure that we adhere to the content needs that you have. So the first question that I always like to ask is, does your video have dialogue? Of course, right? Not all video has dialogue. Some video is just there to portray visual information, in a training sense or in an entertainment sense.
So if you don’t have dialogue in your video, some of the things you have to pay attention to is if there is no critical visual information, you likely don’t need to do anything. You likely may want to consider if it needs to be published. But if you do have critical visual information in your non-dialogue video, now you need to question, are you just purposely distributing this content to multiple countries?
So if you are just trying to reach that English-speaking audience in the United States and someone accidentally finds your video in Spain, that’s not that important. What’s important is if you have an audience and you’re trying to do business with people in the US and your video has critically important information, you need to make sure that you have audio description for America.
If you’re sending that same video also to France, you can’t assume that the blind and low-vision population knows how to speak English. And so you’re also going to need to provide a French audio description track as well. So wherever that video is being distributed for business purposes, whether it’s a training material or entertainment, that the audio description is available so that the individual who is consuming that video can actually process and understand the material.
So that’s a unique, new thing. And we’re encountering this at 3Play, where we’re actively working with customers who have non-dialogue video, especially kids content, where there’s a lot of visual information, but not a lot of dialogue. And we’re actually using audio description in multiple languages to make sure that it’s published and compliant for EAA regulation.
So going back up to the top, we have dialogue in our video. So now the question is, once again, does it have critical visual information? If the answer is no, you likely just need to provide closed captions in each source language that you provide audio.
And critical visual information could be anything that drives the meaning and the comprehension of that piece of content. So if you have just a single speaker and it’s a talking head for five minutes, likely do not need to do audio description. But if you have a single speaker with lots of slide content and lots of information on those slides, you likely will have critical visual information.
So you need to jump over to the other side here, which is if you are providing audio, for each language, you’re going to need to provide a matching audio description track for each audio language that you provide. So if you only provide English, you just need to provide audio description in English. But if you provide English and a dubbed in French and a dubbed in German and Spanish– I’m teasing something for later– you’re going to need to provide audio description in each one of those languages.
So this is a great decision tree to have in your mind around, how does this compliance potentially relate directly to the individual files that I am publishing to the populace? Lily talked about the high-level requirements. But the actual file-level requirements, this is really important to understand as well.
So Lily, how do we do this? That’s the next question. So if we move forward, the 3Play platform is really an all-in-one global marketplace of– sorry, a global marketplace for you to order and receive services in all of subtitling, dubbing, audio description, and closed captions. So we make it super easy, if you’ve used us before or have not used this before. We work in a cloud-based system where you self-serve an order. And you get all the files back that you need. And you don’t have to lift a finger after that.
So in the 3Play platform, we’re going to just walk through, at a high level, how a source video, and we’re going to use English as a source video, can move through our system to apply all of the different service types on top of that. If you want a more in-depth demo of some of the tools that I’m going to show screenshots of, we do have a webinar that we did back in the spring between me and our co-founder and CEO, Chris Antunes, where we shared the vision of this global marketplace. This is a little bit of a, hey, here’s what we’re doing now with real needs in the marketplace.
So we start with our English file. And we can produce closed captions. So we can produce closed captions. So if we move forward–
[LAUGHING]
Once again, we rehearsed this one really, really thoroughly.
So we produce the closed caption track. This is that near 100%, 100% accurate closed caption track. And what we’re doing here is we’re getting the speaker labels so we know exactly who’s speaking. We’re getting all of the microsecond timing, so we know exactly when someone starts and stops speaking, so we know when there’s gaps in the dialogue, and so forth. And we’re getting all of the source words accurately transcribed. This is what 3Play has done for 15 years, bread and butter service of 3Play.
Once you have that closed caption track and all the data associated with it, now you can start making derivatives of that content, whether it’s localized or audio-described. So for this example, you could use subtitles. We’re not going to actually produce subtitles this time. We’re seeing a lot more demand in terms of, I want to produce a dub and then use the output of the dub to create a closed caption track for each individual target language. Because from a user experience, if you have subtitles, which are translated in a different optimization than dubs, you’re going to have mismatch of text and audio if someone were to click those into the player. So we’re seeing a choice between subs or dubs. We’re going to go forward with dubs so that we can show some of the audio description stuff.
So our dub process is straightforward. It could be, aptly, an AI dubbing service. But I want to walk through the three primary steps that are involved in our dubbing process. So in step 1, we’re really focused on script preparation for each target language that we’re dubbing in. So we just already produced 1 through 3 here– sorry, the first three bullets in script preparation. We have our closed caption file. And now we’re going to use machine translation infused with your glossary of terms to output a baseline file that can be used for editing and providing the dub mix back.
In step 2, we’re going to do a bunch of audio processing. We’re going to do background separation automatically. We’re going to analyze the voices. And we’re going to do a voice match to identify– to match voices. This is where you could, in theory, have voice selection. So if we wanted– if there were concerns around voice-matching technology, we can also assign specific native voices that we choose together for each character.
And then once we’ve compiled the full timing and mix of the dub, we actually do another AI-driven review using LLMs, which is going to look for a few things. One, it’s going to modify translations to make sure that it’s grammatically correct. So we’re going to focus on gendered language, making sure that things are appropriate for the target language. But also, we’re going to look at timing. And we’re going to suggest edits and revisions to make sure that the segment in which you have available to translate matches the original time that you have.
So translations, for example, can be really long in German relative to English. And so we need to make sure that we modify German outputs to fit within the confines of time in that dub. So that all happens automatically. And we do steps 1 through 2, except for the caption part, with no humans involved.
Then the question– the dial comes in. So you’re seeing this little purple dial go left and right, visually stark, very weird. But it’s there for a reason– one, to remind me. But two, this is really where 3Play Media gets creative. We’re able to really identify and work with you to identify how much human review does your content really need and make sure that we fit within the budget that you have in front of you.
So our standard process is going to provide translation review, make sure that we’re adhering to your glossary. And we’re going to do some audio QA on your dubbing. That means that we’re going to make sure that pronunciation is right, and so on and so forth. And then we’re going to automatically audio mix those files back together.
But maybe you don’t have a ton of needs. Maybe you don’t need the translation be super, super specific to your brand. It’s very generic language that you can use. You can turn down the dial. You can make sure that you’re only spending what you really need to to get the output that’s going to provide ROI for this content. So that’s a huge part of our system. This exists across the board for 3Play, but definitely most evident in our dubbing process, given all the different bells and whistles in terms of human-level review that we can provide.
And to just illustrate what our dubbing editor actually looks like to our marketplace of contractors, we’re constantly fine-tuning how we make sure the quality of the service maintains from file to file, no matter what your requirements are. So you’ll see, for example, we color-code every single segment by speaker. So we know that yellow always means one speaker. These are just small visual cues that we’re really paying attention to to make sure that we adhere to the consistency from file to file, no matter who that linguist is in the background doing the work.
So once we have our dub, now we need to provide audio description. And audio description has historically been this workflow of, well, I’ll give you the dub, I’ll give you this output, and you’re going to go find an audio describer who specializes– a Spanish speaker who specializes in audio description, or a German speaker who specializes in audio description. As you can imagine, especially right now, given English has been the predominant consumer of audio description, there are not a lot of resources who are specialized in audio description in all of these other languages.
So we’re solving two problems in one here. One, we want to make sure that there’s consistency from each audio description file that we produce, no matter what language you’re in. And then two, we want to reduce the costs dramatically to make sure that you can reach all of your viewers with the right audio description.
So we’re doing this right now with a major publisher in the space who is using our audio description translation service that effectively takes an English template that we’ve created. And we’re translating that into multiple languages, whether it’s German, French, Russian, Turkish, you name it. And we’re doing it at a significantly lower cost than trying to find individual resources for each one of those audio description languages’ requirements.
So to give you just a sneak peek on what that editor looks like, you’ll see that it looks very similar to our dubbing editor. But it’s similar because it needs to be simpler. We’re just focused on translating the description track. And one of the things that’s really neat about how we visually present this is that waveform file.
Audio description cannot overlap with the dialogue of the content. So we’re providing frame-level accuracy to ensure that the audio description does not overlap with existing dialogue, no matter what language we’re in. So we’re using linguists to modify descriptions in the target language to fit within the actual segmentation available for audio description. And we’re doing this at awesome scale already. And I think this is going to be really, really powerful for folks in adhering to EAA regulations across the marketplace.
So with that, we have now produced, effectively, the dub, the audio description, the closed captions for a variety of languages. And all you’ve done is provided us a single MP4 file. So that’s really the magic at the end of the day with 3Play– is, yes, you’re going to maybe have some complicated workflows. This simplifies things a little bit. But ultimately, you’re handing us a file. And we’re producing all of the assets that make it possible for you to be ready to publish, wherever you need to be across the globe.
So are you ready to jump in? So we have a fun offer. And this has been going on for a month or two now. And we’re going to continue to extend this for the next couple of months. But really, we want you to have a chance to jump-start compliance and localization for yourselves.
This means it’s a limited offer to help meet your compliance goals now– and really working with you to craft the right workflow in the short term, but also for the long term so that as you continue to increase your content flow, we can increase our scale, and you don’t even have to think about it.
The program, ultimately, is lower prices. Yay! So available for all non-English language needs– we still have our base English services and still standard pricing and bulk discounting available. But this is really special pricing for the non-English language needs that you might have– and really making sure that we work together on the right level of turnaround requirements so that you can get started now. But then as we ramp up together, we can increase those turnaround requirements as necessary.
So with that, I want to open it up to Q&A. As noted in the chat, the jump-start program is available to sign up. And you’d probably get in contact with me, most likely, by default. But otherwise, you can get in touch directly with Lily or I. Our emails are Lily– L-I-L-Y– @3PlayMedia.com and Erik– E-R-I-K– @3PlayMedia.com.
NOAH PEARSON: I’ve got some good attendee questions for you two. First question here– how accurate do the captions need to be in order to be compliant?
ERIK DUCKER: I like having Lily answer this one. I’m good at it, but Lily’s better.
[LAUGHING]
LILY BOND: That’s a great question. So there are unfortunately no specific regulations around, but there is a very consistent industry standard of 99% or higher. And in the US, at least, there has been multiple lawsuits that reference caption quality, and specifically automatic captions, or ASR, as not being sufficient for compliance and directing in the consent decrees that captions should be 99% or higher.
So it’s pretty well acknowledged in the industry that that’s the standard people should be shooting for. And then there are other guidelines, like the FCC caption quality standards have very strict quality rules. And in the EU, individual countries are starting to explore quality requirements for subtitles and subtitles for the Deaf and hard of hearing. Spain is actively pursuing updates to quality requirements right now. So this is definitely a topic of interest, particularly with AI. And I think the TL;DR is that you should shoot for 99% accuracy. But you’re not going to find it directly in the law.
NOAH PEARSON: Amazing. All right. Next question here. One example of a video we’ve done is a slideshow where the audio pretty much describes what is happening. We only provide closed captions. Is that OK?
LILY BOND: Yeah. In that case, if the speaker is describing all of the critical visuals, then there’s no need for a secondary audio description track. And you can coach speakers to this, for sure, if you’re trying to proactively avoid audio description. But there are many scenarios where that’s not possible– for example, like motion picture or a streaming television show. You can’t have the actors describing the visuals. So it really depends on the scenario you’re in. But if the speaker has described everything visually, then you’re safe.
NOAH PEARSON: All right. Great to know. Next question here. If you’re a US-based company, does the backlog prior to July 2025 apply for the EU directives to be delivered by 2030?
LILY BOND: Yes. So the requirements under the EAA apply to anyone that is delivering content or products or services to the EU. So if you’re US-based and you do not have a consumer base in the EU, you don’t have to worry about it. If you’re a US-based and you do have a consumer base in the EU, you have to have the backlog compliant by 2030.
NOAH PEARSON: OK. Awesome. A similar question here. Do these compliance guidelines apply to Australia and New Zealand?
LILY BOND: If you are, similarly, delivering content, products, or services to consumers in the EU, yes. Australia and New Zealand have their own national regulations that we can definitely follow up on. We use the US as an example with the EAA, but it applies to any region around the world that is delivering content to the EU, must be compliant.
NOAH PEARSON: All right. Awesome. This attendee is US-based. Do we need audio description if there is no room in the video to put audio description without overlapping with dialogue? I do a lot of documentary-style videos with tightly packed interviews that tell the story with B-roll over the top that supports the dialogue.
LILY BOND: That’s a great question. And I’ll say that the answer depends. So when we’re talking about Web Content Accessibility Guidelines, this is what is referenced in most laws and litigation, what we would call extended audio description, where you extend the space available for description in the video source to allow space for description. That is a level AAA requirement, which is typically not written into the law, but may be required by request, for example, or if you are in a publishing scenario where you are able to publish extended audio description, you would want to provide a version that supports that.
There are publishing experiences where it’s very easy to publish extended audio description, and then publishing scenarios where you’re restricted. So if you’re delivering to broadcast, for example, you cannot provide a secondary audio track for extended audio description that would extend the length of that file, whereas if you’re publishing to the web on your website, having that video go longer than the original source wouldn’t be a publishing problem.
So it’s a nice-to-have. And it’s certainly an accommodation-based requirement for those who require it. But depending on your publishing scenario, there’s a little bit of leeway there.
NOAH PEARSON: Sure. All right. Next question here. Are the audio description tracks using AI-generated voice talent, or do you rely on human voice talent for these?
ERIK DUCKER: I can jump in here. So I also read the second question from the panel– sorry, from the participant. So I’m going to combine them into two, which is, does our AD meet regulatory standards? Great. But how is the reception from the low-vision/blind community?
3Play Media has been using synthetic or AI voices for probably almost 10 years now, specifically for audio description service. We’ve been deployed against numerous education and university use cases and corporate use cases using synthetic voices. In the media and entertainment industry, we’re starting to see the shift to be open to synthetic voices.
So the example of the global publisher who’s translating audio description into the multiple languages– are using AI-generated voices. The quality has gotten so much better than 10 years ago. And that has been well-received from a customer perspective. It’s a little too early to tell if the audience has any issues. But we have been doing– we do our own testing with some segment of the low-vision/blind community. And we’ve really not heard any concerns around the quality of the newer, latest release of voices that we’re seeing.
NOAH PEARSON: Awesome. All right. Next question here. Are there options for selecting voices that match the tone or energy of the original speaker?
ERIK DUCKER: I’m going to assume that this is for dubbing. So for dubbing, there’s really two competing styles of voice selection. There is the, let’s match closely the original voice. And then that’s going to pick up on some of the– sorry, on some of the nuances of their intonation and their personality. But it’s not necessarily going to be a clone, that is, it’s a unique model. So from a legal perspective, to be clear, it’s not a clone. So there’s that matching capability.
And then you can also select native speakers. And obviously that mimics more of the traditional legacy dubbing workflows of the voiceover does not match Tom Cruise, but it is Tom– it represents Tom Cruise in certain countries. So that is also an option. And we have pretty big flexibility to support that, whether it’s unique voices that we work to curate together or standalone, off-the-shelf voices that exist today with the variety of voice vendors that we work with.
NOAH PEARSON: All right. Amazing. Well, it looks like we’re about running out of time here. So I will wrap it up. If you do have any more questions for Erik and Lily, their emails are on the screen there for you to reach out to them– but big thanks to Lily and Erik for presenting today and sharing some great information with us, and big thanks to our audience here for joining and asking some great questions. And with that, I hope everyone has a great day and an even better weekend. Thank you so much for joining.
LILY BOND: Thank you, everyone.