« Return to video

The 3Play Way: Unlock Global Reach and Revenue with Subtitles [TRANSCRIPT]

JACLYN LAZZARI: So thank you all for joining us for today’s session, entitled “Unlocking Global Reach and Revenue with Subtitles.” My name is Jaclyn, and I’ll be moderating today’s session. I use she/her pronouns, and I’m on the marketing team here at 3Play. Very happy to be here today. OK, so with all that being said, I’m happy to welcome today’s speaker, Jesse Arris. Thank you for being here today, Jesse. I’ll pass it off to you for what I’m sure will be a wonderful presentation.

JESSE ARISS: Amazing. Thank you so much. It’s great to have everyone here. My name is Jesse. I’m really excited to talk about localization. I know that might seem crazy, but I’ve been working in online video for most of my career, and I love sharing some of the things that I’ve learned.

We’re going to cover why you should be adding subtitles to your video content, some of the things you should be looking out for. And I’m going to move through the content quickly, because I want to make sure that at the end of this 30 minutes– or before the end of this 30 minutes that we have some time to answer questions. So buckle up. We’re going to get right into it, crash course on subtitles.

So I’d like to level set when I talk about subtitles with some definitions. Subtitles are, quite simply, translations of what is being said on screen displayed as text. Now, even for me, this is where it starts to get a little bit tricky. Captions and subtitles are not the same thing, yet they’re infamous for being confused with each other. A great way to remember it, at its simplest, captions are transcriptions. Subtitles are translations.

But to make things a little more complicated, there’s even multiple types of subtitles, most notably would be non-SDH and SDH, which would mean subtitles for deaf and hard of hearing. SDH subtitles make an assumption that the user who’s watching the video cannot hear the dialogue, and so those subtitles include important, non-dialogue information such as sound effects that are happening or speaker identification.

But for the purpose of what we’re talking about today, we’re really going to be talking about non-SDH, which is basically something that you would see if you’re watching Squid Games in Korean on Netflix, for example, but using the English subtitles. If you have any questions, write them down or send them in the chat, and we’ll make sure to make this interactive. But at its core, that’s what subtitles are.

Now, subtitles are incredibly important. And even though it may seem like just a small change, it can make a huge impact. On the left here, we’ve got Maria. She’s a marketing executive, and she prefers to watch webinars on the go. I certainly do. I will often be on the subway, or I’ll be on the bus, and I’ll be fumbling for my AirPods, or I just want to look at it and see what’s happening.

Subtitles are great for her and for me, in that case, to stay engaged and get those key insights without sound. And because the videos are accessible in her language with subtitling– her native language– she watches them even longer, and that allows her to take in crucial information.

And we’ve got Liam. Liam depends on Facebook for driving engagement for his brand. And he knows, just like Maria and myself, people are not watching Facebook videos with sound. In fact, nearly 90% of videos that are being watched on Facebook, and even on Instagram, are watched without sound. You’ve probably seen on Instagram a lot of videos have words bouncing all over the place, and that’s because those creators understand that you are probably not listening with sound so it allows your viewers to get more engaged, draw in more engagement and more view time, which ultimately boosts SEO.

And apart from that, adding subtitles is the right thing to do. On the left is Danielle. She’s a freelance developer in Germany. She works with an American company, but because they’re providing subtitles available in German for their internal content, she can better understand the nuances of what’s being said. That’s allowed her to expand her career, and it also allows that American company that she works for to get more productivity out of her.

And then, of course, compliance, right? This is super important. Compliance for Liam here on the right, he’s a compliance officer, and subtitles are more than just a convenience. They’re actually a necessity. Many of the workers in the factory that he oversees speak Spanish and Vietnamese as their primary language, and subtitle-supported training in those languages helps ensure safety compliance, and bridge that gap, and help reduce accidents and legal consequences.

But it’s not all just corporate. It’s also very exciting, right? Some of you may know Mr. Beast. Mr. Beast is, obviously, a YouTube phenomenon, and a large part of his success is due to his commitment to localization. So a couple of years ago in the midst of the pandemic, Mr. Beast expanded his reach by investing in localizing his video content into multiple languages, Spanish, Portuguese, and even Japanese. Then he established separate YouTube channels for each language where those videos are dubbed by actors, they also have subtitles, and the goal is to cater to non-English-speaking audiences.

And this exploded his success. His localized channels started to get millions and millions of subscribers in a short period, unlocking access to these markets that he was never able to get into before. It broadened his audience. He was able to do more merchandise in those markets, more brand partnerships in those markets. So by investing in that high-quality localization, he transformed his North Carolina, United States, brand into a global entertainment enterprise.

But why did he decide to go in on localization? Well, we understand that, but why did he invest in localization, dedicate an entire team to it? And the answer, quite simply, is localization is hard. I’ll give you a few examples. What you’re looking at on the screen right here is from the TV show The Office.

What’s happening here is there is a typo because Ryan, the actor, the character in the office is answering the phone and saying, “Dunder Mifflin.” Well, if you’re a computer that recognizes speech– and this is how a lot of these captionings are done, through what’s called ASR or automatic speech recognition– you may not know what the heck a Dunder Mifflin is. That’s maybe something you’ve never heard before.

But bless that ASR’s little heart. It’s tried its hardest. But you can see how it all starts with a strong transcription, because if you were to translate “D mlin” into French, well, that’s very confusing, and that just wouldn’t work, And the things would really, really start to fall apart if your transcriptions are not high quality or go unchecked.

Now, this isn’t always the case. ASR, automatic speech recognition, generated captions, might work for you. But I would say if you’re creating content that’s vital and customer-facing where your message, your reputation of your brand, and your bottom line are all at stake, you should really ask yourself, does it need to be perfect or is that 90% accuracy from those robots good enough?

And just for fun, I’ve put on the screen a few examples of that 10% error that it can sometimes get wrong. Now, if you’re a brand, if you’re an educator, even if you’re just the type of person who doesn’t like when your name gets spelled wrong, you can understand why you’d likely need a solution that goes beyond ASR.

It’s not just phonetics either. There are massive cultural implications you should be considering when you look at localizing your content. This is a great one. We saw a video where the speaker said, “Wow, he really hit it out of the park this time.” It wasn’t a baseball video. It was just an expression that we’re all familiar with here in North America.

Now, if that phrase was translated directly to French, it’d be no problem. We can do that. Even machines can do that. But baseball isn’t really that well known in France. In fact, if– you saw it wasn’t at the Olympics, but it’s probably about as popular there as cricket is here, right? So in that case, the expression that’s talking about a home run that we get, it just doesn’t make sense, even though it was actually translated correctly.

So one thing to keep in mind is the folks who are helping you with your localization efforts should have a strong understanding of local cultures and customs. In this case, the phrase was not just translated but also adjusted to resonate with the local audience, so “Wow, he scored an exceptional goal.” That gets the message across– it actually gets the exact same message across in a way that resonates with the audience. So this is really, really key and something to keep in mind as you look towards localizing your videos.

Now, as you start to explore localization, which you’re doing because you’re here, you’re going to– and by the way, thank you for joining. You’re going to be talking to some folks. Maybe it’s a team internally that you have. Maybe it’s an external vendor that you’re working with. And some things to keep in mind to not just qualify them as the right partner for yourself but to also help you understand your own localization goals are listed on this page.

So as someone who’s now been working in this space for a while, if I was looking for a vendor for translation localization, one of the first questions I would ask are, are humans involved in this or is it all AI? AI is great. I love AI. Don’t get me wrong. I got ChatGPT up on the other window right now. But it’s not going to replace the need for human quality checks. However, we still think it should be used in this process, because it will still speed up the entire process. I just do believe that there is a spot for human-in-the-loop, and I’m going to talk about that concept of human-in-the-loop in just a couple minutes.

Another thing I’d ask is, what kind of outputs am I going to get? Is it going to be SDH if I wanted it? The subtitles for deaf and hard of hearing that we talked about earlier. Am I just going to get the subtitle file back as a text file or am I going to get the whole video? Do I have to upload it? Does it work with YouTube? All those questions about the process, those are going to be really important.

And then again, where are the translations or the translators coming from? Do the translators that you’re working with really understand those cultural nuances that we talked about, or are they just really glorified spell checkers for Google Translate? And then finally, of course, have they done this before? What is their reputation? Have they done this in mission-critical environments where a single error could spell disaster?

I want to share just quickly, before we get into Q&A, just a few more things about how we at 3Play address some of those challenges that I’ve talked about and how some of the technology that we’ve been working on and continuously driving has improved over the nearly 20 years that we’ve been helping folks in the transcription, translation, and subtitle space.

It starts with that human-in-the-loop that I talked about. Human-in-the-loop, or– I want to say HITL. Good luck to the ASL interpreter on that. I think it’ll HITL is going to become a thing. Let’s make HITL happen. HITL is human-in-the-loop, where we combine the speed and scalability of AI with that critical role of human error correction to get as perfect as we possibly can.

I’ll give you a real-world example that I absolutely love. And I wanted to lead with this, but I think it’s really, really fun. So we had seen a video online about Lionel Messi, and it was obviously run through an AI speech-to-text generator, because the output was “Lion all messy,” which, obviously, is hilarious, but it’s also incorrect. And you can see how those mistakes can very, very easily happen without the observance by a real human person.

With the human-in-the-loop process that we use here at 3Play, the output is reviewed to ensure that the incorrect text is caught and fixed before that video is published to make sure that you get the right output the first time every time.

We’re also big fans of workflows that work with your environment. At 3Play, we have lots of corporate customers, small businesses, mom-and-pop shops, all over the place, and they all use different types of systems. And you don’t really need another system to log in, to upload. Do I have to save that to my desktop? No, wait, did it download to the Downloads folder? Why does it do that? We work how you work and where you are. So if you use a video platform like Brightcove, well, with just the click of a button, you can get subtitles, captions, whatever you need added to your video. No downloading, no uploading.

Maybe you’re a little bit more old school and just want to email it to us. No problem. Just send us an email. FTP more your speed? Sure. Have something even more complex than that, like some custom API requests? We’ll work with you. We want to focus on taking the frustration of managing your subtitles off your plate so that you can focus on your actual job.

So the big takeaways today that I think you’ll benefit from as you get into your localization journey are, number one, start to think about globalization effort as markets, not just languages. It’s one thing to translate to French, but it’s another thing to translate for French people. And so that means understanding the nuances of a culture, and that can only be powered by translators that are speaking and living that language every day.

I’d also suggest that you ask your vendor some of these questions around content confidentiality, things around trust. Ask if they use your videos to train their AI models. Look closely at the terms or, again, just straight up ask. You’d be surprised who’s maybe trying to make a buck off your hard work. And also make sure that your vendor that you’re working with knows their way around legislation and can show proof of examples of that.

With the rise of AI, unfortunately, you’re going to need to do a little bit more diligence than ever before to ensure that you’re working with a professional and not just someone with a .com domain and a ChatGPT pro license, right? Now, there’s a wide range of folks you can work with all over the place from there to a full, hands-on solution like 3Play Media. But my suggestion for you is to do your diligence and figure out exactly who you need to work with.

So that’s it for me. Real quick, I just want to recap. I wanted to share with you some of the learnings I’ve found. I hope those have been helpful for you. I hope you consider those as you go on your localization journey, and I’m just thrilled to have had the opportunity to speak with you for the last 20 minutes or so, but we do have a few minutes left, and I’ve promised some Q&A, so I’ll pass it back over to Jaclyn to help me out with that. Thank you.

JACLYN LAZZARI: Thanks, Jesse. That was a great presentation. Yeah, we do have some questions and we have some time, so let’s dive in. First question is, are there any laws or compliance standards– in the US at least– that kind of pertain to subtitling, and can you speak to that?

JESSE ARISS: Well, yes, there are a lot. And can I speak to them? I can speak to them at a high level, how about that? Because I am certainly not a lawyer by any means, but I can take you through some key legislation where actually us at 3Play are experts, right? I’ll give you an example. The ADA mandates that all public-facing content, including videos, must be accessible to people with disabilities.

One of the simplest ways to show your commitment to accessibility is to make sure that your subtitles are available and to make audio content available to deaf and hard of hearing individuals. There’s also a section, section 508– I hope that’s right– of the Rehabilitation Act. That’s just for federal agencies and contractors, but that states that they also must make their elected information and technology available to people with disabilities. You’ve got the W-C-A-G, WCAG, which is for– it’s sort of an international web standards that you’ll want to be aware of if you’re in Europe. That’s a whole ton of accessibility laws that I couldn’t even get into on this.

And if you’re in Ontario where I am, Ontario, Canada, then you’re also going to want to be aware of things like the French Language Services Act, again, just for public services, but also the AODA, the Accessibility for Ontarians with Disability acts, which requires businesses to make their content accessible. So yes, there’s a ton of stuff. I can just hit it at a surface level, but we have a lot of content available on our website if you’d like to learn more.

JACLYN LAZZARI: Awesome. Thank you, Jesse. And now, you kind of talked about some of the different scenarios in which subtitling might be helpful for the different people like Maria or Liam. Can you kind of talk about what kind of content do subtitles usually work best for? If you have a certain kind of content where you might say, wow, subtitles would really take it to the next level?

JESSE ARISS: Yeah, I find that, obviously, subtitles are going to be beneficial for all types of content. And as I mentioned, we’re probably already consuming subtitles today without even realizing it as we watch foreign language films. So that’s a great use case, but I just think it goes so far beyond that. Imagine you are responsible for internal learning at your organization and you want to get out a course about some of the features of your new product.

Well, if your product name, product information– we’re using Canva for this presentation. Canva, if they were educating their internal employees on different things, they would want to make sure that terms like Canva and other internal terms are actually done correctly in the new language that it’s being put in, so just keep that in mind as you create your subtitles.

But in any space– I would say education is great. Folks can absorb the content so much easier when they can read it as well as even listen to it. You could combine subtitles with dubbing to actually create an even more engaging experience, and that’s something we’re also able to do with our AI Dubbing Solution. But regardless of the content, I don’t think there’s a limitation in terms of what could be subtitled effectively.

JACLYN LAZZARI: Great. Thank you. And then just because you mentioned it, you mentioned dubbing. So can you maybe quickly give the difference between subtitling and dubbing?

JESSE ARISS: Yeah, thank you. Sometimes I get ahead of myself because there’s so much– there’s so many terms in this space, but subtitles, as we mentioned, is the translation of what’s being said into text on the bottom of the screen. Dubbing is actually an audio track that replaces the audio that’s being spoken in the original language with audio in your native language.

So when you watch– again, I’ll use Squid Games as the example, because who doesn’t love Squid Game? Season two coming out soon, by the way. You have the choice of listening to it dubbed, where it’s spoken in English, or subtitled, where you read the English, because, again, the source language for that TV series is Korean, which I certainly don’t understand.

JACLYN LAZZARI: Me either. [LAUGHS] Well, thank you, Jesse. That’s all the time we have for today. Thank you for the presentation, and thank you to our audience for joining us today. And thank you, again, to everyone. We appreciate your time and hope you have a great rest of the day.

JESSE ARISS: Thank you so much.