PARIS—Ateme, a provider of video compression, delivery, and streaming solutions, today announced it is leveraging Google Cloud’s generative AI capabilities to automate large-scale, multilingual subtitle generation, a move Ateme says could drastically reduce costs and production times while enhancing content accessibility and value for audiences worldwide.

Ateme’s file transcoding application, TITAN File, is containerized as microservices and deployed through Google Kubernetes Engine. Media files stored in Google Cloud Storage are processed by Vertex AI, Google Cloud’s unified AI development platform. This distributed, flexible, and scalable infrastructure has allowed for a smooth transition and seamless integration into client workflows, according to Ateme.

“This innovation addresses a key challenge for content creators and distributors who have often been limited to translating content into a single language, thereby restricting international distribution and accessibility for non-English speaking or hearing-impaired audiences,” said Yohann Guilloux, VP Global Partnerships at Ateme.

“In the face of exploding content volumes, it’s no longer enough to have good technology: you have to be able to innovate quickly, at scale, and with the right partners,” said Jules Mermet-Husson, Cloud Solutions Architect at Ateme. “Google Cloud’s generative AI capabilities enable us to automate what was once a challenging process. Together, we are building an agile, open, and scalable platform that transforms every technical constraint into an opportunity for value creation.“

Over the past several years, there has been a surge in interest and demand for non-English video content. Viewership is booming, especially among younger audiences in primarily English-speaking markets like the U.S., U.K., Canada, and Australia.

In a recent study from Ampere Analysis, 54% of internet users in those markets reported watching international (i.e., non-English) TV and movies “very often” or “sometimes” — up from 43% five years ago. Likewise, 66% of 18-to-34-year-olds in those markets reported watching international content regularly. And, according to the report, the most popular way to consume that content was through subtitles.

These findings underscore just how important subtitles and captions have become. Delivering properly localized subtitles is now essential for making content accessible to a broader audience. Without a solid subtitling and captioning solution, broadcasters and video service providers risk being left behind in the race to secure viewers and keep them happy.

The Complexity of Implementing Proper Captions and Subtitles
International content presents a huge monetization opportunity for video service providers, but many say that the complexity of localized captioning makes it hard to capitalize on that opportunity.

Ensuring synchronization between audio and text is critical, as captions must match dialogue precisely to maintain clarity and viewer engagement. Synchronization is especially tough for multilingual or fast-paced content, where timing discrepancies can happen easily. Manual captioning is labor-intensive and error-prone, requiring multiple reviews to meet guidelines for accuracy, completeness, and display metrics.

Additionally, captions must be carefully placed to avoid obscuring important on-screen visuals, often necessitating frame-by-frame inspection or advanced object detection. Translating humor, idioms, and cultural references and nuances further complicates the process. Literal translations can fail to capture the intended meaning, requiring nuanced adaptation so that audiences can relate.

Finally, complying with varying regional accessibility and legal standards adds another layer of complexity. It can take up to 10 hours to caption one hour of video content manually, and even then, there’s no guarantee the captions will be compliant.

Automated Captioning for the Win
Automated, software-based captioning tools streamline the process by converting spoken language into text (speech to text), segmenting it appropriately, and ensuring captions are accurately timed and placed within the video frame. Modern systems are designed to maintain high standards of accuracy and synchronization, and they often include features for multilanguage support, error correction, and compliance checks. Such characteristics make these systems essential for today’s global and fast-paced video streaming environment.

These tools also perform comprehensive quality control checks, such as verifying reading speed, display duration, and caption segmentation, and can automatically adjust placement to avoid blocking important elements like scoreboards or faces. This level of precision ensures captions meet legal requirements, such as those set by the FCC in the U.S. or Ofcom in the U.K., and adhere to best practices for accessibility.

By automating much of the workflow, automated captioning systems significantly reduce captioning time and effort vs. manual processes. That 10-hour captioning task can now be completed in real time or near-real time, cutting production time by 40% to 50% and labor costs by 30% to 40%.

Automated captioning systems also consolidate tasks such as transcription, review, and compliance checks within a single interface, minimizing errors and streamlining collaboration between automated processes and human reviewers. This efficiency is crucial for delivering high-quality, accessible content quickly across multiple languages and formats.

How AI Improves the Process and Elevates the Viewer Experience
The latest captioning tools use AI-driven technologies such as automatic speech recognition (ASR) and natural language processing (NLP) to deliver superior results. ASR rapidly transcribes spoken audio into text, even in challenging conditions with background noise or diverse accents.

NLP further processes this text, interpreting context, idioms, slang, and tone to produce captions that are not only accurate but also natural and easy to read. These technologies also enable features like automatic speaker identification, punctuation correction, and the grouping of words into meaningful segments for improved readability.

AI and machine learning advancements have transformed captioning by enabling adaptive learning and multimodal analysis. For example, computer vision can detect faces, on-screen text, or important visual cues, allowing the system to reposition captions dynamically to avoid covering critical information.

Machine translation models, now enhanced with large language models, provide more context-aware and culturally appropriate translations, ensuring captions resonate with local audiences. Multimodal AI can also identify nonverbal cues, such as laughter or gestures, and incorporate them into captions, resulting in a richer and more immersive viewing experience.

As these systems process more data, their accuracy and adaptability continue to improve, reducing the need for manual intervention and ensuring consistent quality across diverse content.

Cultural Sensitivity: It’s About More Than Translation
When it comes to captions and subtitles, cultural resonance and compliance with local standards can determine a title’s success. Today's mainstream translation engines excel at converting languages. As sophisticated large language models become fully integrated into readily available translation engines and captioning tools, they will be able to natively incorporate deep cultural sensitivity too.

That means AI-driven captioning solutions are poised to go far beyond simple translation. These future systems are expected to use contextual understanding to capture the subtleties of language — such as humor, sayings, and regional expressions — that are essential for truly authentic localization.

The aim is for these advanced solutions to one day adapt captions to align with local customs, regulatory requirements, and viewer expectations, ensuring content is not only accessible but also culturally relevant and accurate. This includes adjusting translations to preserve the original intent and tone, moving beyond literal word-for-word rendering.

As these more nuanced AI capabilities mature and become integrated into common tools, human expertise will remain indispensable. Human reviewers will play a vital role in guiding, validating, and finessing AI-generated captions, ensuring they accurately reflect the necessary subtle nuances and meet cultural appropriateness standards.

By combining the emerging power of AI for cultural adaptation with diligent human oversight, the next generation of captioning tools will be better equipped to help global audiences experience content as intended, fostering inclusion and engagement across linguistic and cultural boundaries.

Conclusion
Advanced captioning tools enable providers to meet diverse audience needs, ensure accessibility, and deliver culturally relevant content efficiently and accurately for every viewer everywhere. They expand the reach of regional content and boost international connectivity through video, making them essential for success in the global streaming market.

LOS GATOS, Calif.—After greatly expanding its global production and offering programming from around the world, Netflix is also making the viewing experience more multilingual with a new “Language Picker” feature that allows viewers to select from more than 30 languages for subtitles and dubs when watching programs on TV.

In launching the feature, the streamer explained that nearly a third of all viewing on Netflix is for non-English stories and that its catalog includes a wide variety of TV shows and movies in over 30 languages, from critically acclaimed films to binge-worthy series.

In addition, some of the most popular TV and film titles available on the streaming platform include Squid Game (Korea), Berlin (Spain), Lupin (France), Who Killed Sara? (Mexico), Troll (Norway), and Oscar winner All Quiet on the Western Front (Germany).

In response to viewer demand to make its programming available in more language, Netflix has launched a new features that allows viewers to pick from the full list of available languages for any title when watching Netflix on a TV. Netflix noted that this much-anticipated feature carries over the experience members already enjoy on mobile devices and web browsers.

Netflix also explained that many members use Netflix to learn a foreign language thanks to the many languages available on the service and it is now offering additional features to support language learning, including the ability to customize subtitles, and the Browse by Language feature on PC computers. These helpful features enhance the accessibility of shows and movies, while supporting those set on being multilingual.

More details are available in this YouTube video

Viewers in English speaking countries—including the U.K., US, Australia and Canada—are viewing more non-English language film and TV shows in recent years, according to a new study from Ampere Analysis.

Regular viewing of international (non-English language) TV shows and movies has increased by 24% since the first quarter of 2020, among 18 to 64-year-olds in those countries. More than half (54%) of Internet users in these markets claimed to watch non-English language content “very often” or “sometimes”—up from 43% since the start of 2020. This is despite titles from primarily Anglophone markets like the U.S. typically making up the bulk of global streaming libraries, Ampere said.

Korean TV shows and movies have seen a 35% rise in frequent viewing in English-speaking markets in the last four years. Frequent viewing is up from 16% of 18 to 64-year-olds to 22% over the last four years, with titles including Squid Game and Parasite driving awareness of the high-quality shows and movies produced in South Korea. Netflix in particular has increased spend on TV shows and movies produced in South Korea.

Foreign language content is particularly popular among 18 to 34-year-olds in English-speaking markets, with 66% of this age group regularly watching. However, growth is strongest among older age groups, with frequent viewing increasing by more than one-third among 45 to 64-year-olds in the last four years (up from 30% to 41%). The move towards foreign language content consumption by older consumers reflects growth in the adoption of streaming services, which offer greater volumes of international content than their broadcast counterparts.

When it comes to how viewers watch foreign language content, subtitling is most popular. More than one-quarter (28%) of people in English-speaking markets enjoy watching this way. In comparison, just under one-fifth (19%) enjoy dubbing.

The situation in non-anglophone markets is different, however, according to the researcher. While growth in interest in foreign language content is minimal, there is already a high baseline willingness to watch international content—with 88% of consumers regularly engaging with foreign-language TV shows or films.

The way foreign language content is watched also differs in some markets. In France, Germany, Italy, and Spain, there is a strong preference for dubbing, which is reflected in the availability of dubbed content in these languages on video services. Yet while certain markets show preferences for dubbed content, others—like the Nordic territories and the Netherlands—have a strong preference for subtitling. Ampere says this reflects a mix of the lower historical availability of TV shows and movies dubbed into local languages, but also—in many cases—strong English-language skills.

“The increased viewing of international programming in English-speaking markets shows that as content producers diversify production regions, viewers are ready and willing to transform their viewing habits,” said Annabel Yeomans, Research Manager at Ampere Analysis. “This offers multiple advantages for streamers facing economic pressures. They can investigate markets with lower production costs and focus on productions in newer streaming markets to grow subscriptions while catering to their existing subscriber base. Developments in AI technology for subtitling and dubbing make it easier than ever for platforms to offer TV shows and movies on a global scale.”

NEW YORK—A new study from YPulse suggests that subtitles are being widely adopted by younger viewers as they watch TV, with 59% of Gen Z viewers and 52% of Millennials reporting that they prefer to watch TV with subtitles. 

The results highlight the importance of accurate subtitling for both general audiences and viewers with special needs and hearing issues. 

In its latest TV and Entertainment report, YPulse asked 13-39-year-olds about the TV and movie content they watch and their viewing experiences.

The finding that a majority of Gen Z and Millennials prefer to use subtitles when watching TV indicate that the use of subtitles has gone beyond the traditional use by people with special needs to a larger audiences, the researchers noted. 

“[M]ore than half of young people tell YPulse they prefer to use subtitles, and it’s not just because they need them; the gen makes use of reading text while watching movies/TV to keep up with murmuring dialogue, to distinguish thick accents (looking at you Peaky Blinders), and some say just to get a head start on a scene and go back to looking at their phone,” the researchers noted in a blog post. “Subtitle use is especially popular with Gen Z males: 61% say they prefer to use them.”

The post also stressed that “this popularity is beneficial to those who do need them because media brands like Netflix are putting more effort into making subtitles as immersive an experience as the audio itself….Netflix is the number one source of entertainment Gen Z and Millennials are using, and the one they feel offers the most inclusive content, which extends to hearing accessibility.”  

YPulse also cited children as a factor in the popularity of subtitles among Millennials. “According to a Kids Industries survey of parents with kids 5-15-years-old (a.k.a. Millennial parents) in the U.S. and U.K., the majority of children in the U.S. watch content with subtitles or closed captions all or most of the time, and half of kids in both countries need subtitles to enjoy the content they are watching,” the blog post noted. 

“Watching content with closed captions can reportedly improve literacy, vocabulary, and the speed of reading,” the researchers added. “For kids’ media brands, the widespread use of closed captions should be a sign to improve accuracy and make sure subtitles are available for all programs.”

Interest in foreign language shows is another factor, with the success of “Squid Game” on Netflix being only one example of a foreign language film attracting large U.S. audiences. 

The YPulse’s Local/Global Citizenship report data found that 76% of Gen Z and Millennials agree with the statement: “It doesn’t matter what country a TV show / movie is from, if it’s good I’ll watch it.” 

Another key factor is the popularity of subtitles on Gen Z’s favorite social media platform TikTok, YPulse reported. 

More information can be found here.  

SAN JOSE, Calif.—XL8 Inc., a Silicon Valley tech company that provides AI-powered Machine Translation technology, has announced new sets of Context Awareness (CA) language pairs that are designed to increase the translation and subtitling accuracy for localization service providers (LSP) using their technologies. 

XL8’s language pairs have been found to deliver new workflow efficiencies for LSPs in terms of less post-editing work and a 32% reduction in project delivery times, the company said. 

These newest models bring the company’s total number of context-aware language pairs to 40. By the end of 2022, all of XL8’s 73 engines will be context-aware, XL8 reported. 

XL8’s application of Context Awareness in its Machine Translation technology is designed to enhance the immersive experiences for audiences watching live or pre-recorded streaming or broadcast content, the company explained. 

XL8’s Context Awareness engine does what was previously the sole domain of humans, accurately considering the context of a conversation and evaluating the subtle differences of gender, slang, formalities, multiple word meanings, and other language intricacies, the company said. 

XL8’s Context Awareness technology focuses on “source language to source language” to create different and specific language pairs and, as a result, achieve higher levels of accuracy.

XL8 said that a third-party committee of content localization service providers tested the new models for translating from English to Latin Spanish using several categories of programming (e.g., sci-fi, comedy, food, travel, drama). The tests were conducted with and without the XL8 Context Awareness model applied.  While both sets performed well, the accuracy of XL8's CA model averaged 95.5% and the normal model average was 91.2% (a percentage change of +4.3%).

Overall, the Context Awareness model was more accurate regarding gender and formality consistency among multiple subtitles. While both performed well at providing coherent sentences, even when faced with misspelled words or odd phrasings, the Context Awareness model was more accurate with certain categories like food, where dishes were described in extreme detail with long lists of ingredients.

Linguists involved with the testing have regularly noted the improvements in their performance and accuracy by using the XL8 technology.

“XL8’s new Context Awareness engine has turned out to be a great ally for me,” said Gisela C., a linguist with 10 years of localization experience. “I was surprised to see how terms, genre and treatment are consistent throughout the files. The Context Awareness engines definitely had a positive impact on my performance, I can do more in less time! I certainly recommend it.”

The XL8 models also significantly improve project delivery times, especially important when LSPs need to deliver content in multiple formats and languages to an increasing number of platforms and viewers worldwide in less time and with fewer resources.

"The XL8 Context Awareness engines help my work as an editor, especially when it comes to quick translation and consistency throughout the file,” said Rido M., a linguist with seven years of localization experience specializing in Bahasa Indonesian and English. “It also helps with my time-efficiency since many short and long sentences need no more editing. We now have a reliable tool that's very helpful for today's fast-paced environment. I definitely recommend it.”

BOSTON—3Play Media, a provider of video accessibility services, including captioning, subtitling and audio description, announced today that as part of its acquisition of Captionmax, a provider of live and recorded captioning, localization, and audio description, it has been fully integrated into the 3Play Media brand. National Captioning Canada (NCC), the largest live captioning provider in Canada and a subsidiary of Captionmax, was also acquired and has been rebranded as 3Play Media Canada, the company said.

“We are thrilled to bring Captionmax and NCC into the 3Play Media family. From the very beginning, it was clear that all 3 of our businesses were aligned around a few critical values: delivering peace of mind, meeting our customers where they are, and being trusted advisors in the space of media accessibility. This shared DNA enables us to deliver the most robust media accessibility platform on the market today and has exciting ramifications for our customers, particularly those in the media and enterprise sectors, including Canada, where we are able to deepen our expertise and relationships,” Josh Miller, co-founder and co-CEO, 3Play Media said. “With all our collective capabilities under the 3Play Media umbrella, no other company in the market can deliver the scale and reliability of the 3Play offering, cementing our position as the leader in quality media accessibility services in North America.”

With the integration, 3Play Media says it now offers “the widest breadth of offerings for media accessibility services in North America,” providing closed captioning, live captioning, transcription, broadcast scripting, audio description, subtitling & translation services, and an interactive media player. The brand integration’s first phase includes the launch of a partner program and the addition of post-production tools like voice artist audio description, expert caption styling, and broadcast scripting to the 3Play platform.

“By incorporating Captionmax into the 3Play Media brand, we are now able to deliver a high-end, highly customizable and scalable solution for post-production and broadcast live services to our media clients,” Chris Antunes, co-founder and co-CEO, 3Play Media, said. “For the first time, all the tools broadcasters and post-production studios need will be available in a centralized, platform-based solution, allowing 3Play Media to provide the most innovative solutions to support the media and entertainment industry.”

3Play Media Canada will remain focused exclusively on the Canadian market, providing the region with its own business development and account teams that understand the inherent needs of Canadian broadcasters, corporate entities, government agencies, and educational institutions.

“3Play Media Canada will continue to follow the path set by NCC for more than 30 years when it started delivering accessibility solutions to Canadian broadcasters,” said Melina Nathanail, President of 3Play Media Canada. “Having run NCC as a family business for that period of time, I couldn’t be more excited to adopt the 3Play brand, joining a team that feels like a family and has the same passion for accessibility that has driven us all these years to deliver the best to the Canadian market.”

“The synergies between 3Play Media and Captionmax - and how both of our customer bases benefit from this relationship - are obvious, making the rebrand a natural transition,” Truck Morrison, SVP of Corporate Development, 3Play Media, said. “3Play’s deep expertise in technology-driven accessibility at scale combined with our own history with live broadcast creates a compelling, complete media solution that cements 3Play as the most robust media accessibility provider in North America.”

To recognize and celebrate the integration of both brands, 3Play Media will host a virtual public event open to anyone on September 29, 2022. For more information, please visit: https://www.3playmedia.com/rebrand-celebration/.

FREMONT, Calif.—Digital Nirvana has unveiled an upgrade for its Trance self-service SaaS application, a tool that supports transcription, closed captioning and subtitle generation and automatic translations to localize content within media companies’ workflows.

Enhancements offered in Trance 4.0 include elaborate account management, modular rights for users and advance pro-captioning window for enterprises and individual users, Digital Nirvana said.

With its new modular application approach, Trance 4.0 can be used as a combined or individual tool for transcription, captioning, text localization or conformance of existing captions, it said. 

The latest version’s new and improved transcription window can be used as a standalone app rather than being part of a three-step transcription, captioning and subtitling workflow that only outputs after each step is completed.

The company has made major changes to the pro-captioning window to make it more user-friendly and efficient. Machine-generated timecodes can be adjusted using spectrogram and manual inputs and the window now enables user to upload media, generate automatic captions and display using desired presets defined by the user or organization, it said.

An extension of the pro-captioning window, text localization can now be used as a Trance module to focus on localizing captions. Users can now import existing captions and quickly use the machine translation feature to generate highly accurate localized text specifically designed for subtitling purposes, the company said.

The caption-conformance app allows users to import text with new presets to conform with various publishing platforms. It automatically reviews text and highlights any nonconformities with selected guidelines, it said.

Digital Nirvana also has upgraded Trance’s management functions. The dashboard now allows better real-time visibility into which users are online and the progress of jobs.

The new version offers enhanced user control over processing jobs with the addition of the ability to choose from available jobs, import a transcript and create captions with existing or new presets, it said.

Trance 4.0 now allows enterprises to assign work items to internal staff and to share work items with Digital Nirvana’s service team to receive fully qualified, human-curated output using the same application, it said.

More information is available on the company’s website.

During Netflix’s Q4 2021 earnings call, COO and Chief Product Officer Greg Peters revealed the company “subtitled 7 million run-time minutes in '21 and dubbed 5 million run-time minutes” of content to reach their 222 million subscribers worldwide. That’s 116,666 hours of subtitles and 83,333 hours of dubs they produced before releasing titles anywhere across their 192-territory footprint. That’s a massive undertaking.

Those familiar with the process know “subs and dubs” are done by separate teams of people simultaneously. These teams include translators, editors, proofreaders, managers, producers, quality control, and others responsible for ensuring stories are translated accurately into various languages and are appropriate in as many diverse cultures worldwide as possible.

Kudos to Netflix for investing in the content that requires localization and taking on the challenge of preparing it for global markets. Despite the popularity of “Squid Game” last year, inaccurate translations of stories can adversely impact the viewer’s experience. With that said, linguistic skills are only part of the equation. Familiarity with local culture is crucial when conveying scene or dialogue nuance in the final product. Still, not everyone has that knowledge, and issues are not always caught before a title is released.

Netflix’s remake of “Perfect Strangers” is an excellent example of how being aware of cultural or contextual issues can affect audience acceptance of a title. Released in the Middle East on January 20, the film immediately drew fire from many in the region. Because it included a gay character, some claimed critics felt the film promoted homosexuality and “moral degradation.” One Egyptian critic on Twitter said Netflix “has an agenda to spread ‘deviant ideas’ to ‘normalize false ideas.’” 

The conflict ultimately led lawyers in Egypt to call for an outright ban of the film or face a lawsuit against the country’s Culture Ministry. In its story on the controversy, the Hollywood Reporter said, “few would have anticipated the immediate wave of controversy it would provoke.”

Given the history of cultural sensitivities and censorship in the Middle East to LGBTQIA+ and promiscuous sexual references, this controversy could have been anticipated. Awareness of these concerns in this and other markets is not as widespread as it needs to be despite dozens of examples. Improvements are necessary to incorporate what we call “culturalization” into the localization process.

Netflix now has to deal with the negative press and reputational hit as a company “that spreads poisons.” What does this mean for future titles Netflix wants to release in the region, and how willing will local regulators be to allow them to exhibit? From both a legal and marketing context, it’s better to know when problems may arise before they happen.

Unfamiliarity with local culture and customs may have been what Netflix CFO Spencer Neumann alluded to during their earnings call, “Entertainment is still fundamentally pretty local around the world. So, it's global and local, and we need to figure that out.” The good news is that tools and techniques are available to make localization more accurate, culturally respectful, and appropriate. It is easier to avoid being hit by the proverbial bus when you know the route.

TEL AVIV, Israel—OOONA, a global provider of professional management and production tools for the media localization and subtitling, has announced that it has reached an agreement with memoQ Translation Technologies to offer the memoQ toolset inside OOONA’s subtitle editing platform. 

memoQ specializes in the development of translation management systems designed to increase productivity as well as enhance consistency and quality in translation workflows, the companies said. The collaboration will build in an extra layer of automation to the process.

OOONA co-founder and CEO Wayne Garb noted that working with memoQ “will allow us to bring the best translation management into subtitling. Our goal is to equip our users with more powerful tools to address the market’s stringent requirements for faster workflows and larger volumes of content localization.”

“memoQ is in business to help translation professionals thrive,” added memoQ Co-CEO, Peter Reynolds. “One way we can do this is by cooperating with great partners such as OOONA. We are very excited at the work they are doing to integrate our best of breed translation technology within OOONA’s best of breed high-end subtitling platform.”

TEL AVIV, Israel—OOONA, a global provider of professional management and production tools for the media localization industry, has announced that it has merged with Spain-based subtitling training provider GOSUB. 

The combined company, operating as GOSUB OOONA, will be managed from OOONA’s existing headquarters in Tel Aviv.

“We are very pleased to have GOSUB on board,” says Wayne Garb, OOONA co-founder and CEO. “Kelly is a talented trainer with years of experience advising translators who wish to specialize in subtitling and the use of localization software and resources. Kelly will continue to provide the same training support and courses as before, complementing our role as a developer of localization products tailored for the media vertical.”

“I am looking forward to working with our OOONA colleagues, offering practical and thorough online training for each element of their industry-leading localization platform,” Kelly O’Donovan adds. “Demand for skilled subtitlers and captioners conversant in leading subtitling software is growing exponentially even in countries that do not have a tradition in subtitling. The goal of our training program is to prepare students fully to work directly with language service providers.”

More information is available at the rebranded website.   

Today, content creators are increasingly looking to expand across geographies, but one of the biggest challenges they face is overcoming language barriers. Delivering content to non-native speakers is possible through captions and subtitles, or audio dubbing. Yet, quality control (QC) checks are imperative to ensure that everything flows smoothly. This article will examine why audio dubbing is better than providing subtitles, the key capabilities to look for in a QC system and how machine learning (ML) and artificial intelligence (AI) technologies are simplifying QC workflows for audio dubbing.

WHY AUDIO DUBBING MAKES SENSE VS. SUBTITLES

Subtitles are an easy and affordable way to deliver content to non-native speaking viewers, but they have limitations. For example, using subtitles, it can be difficult to deliver long dialogue scenes within limited screen times. Furthermore, subtitles can be distracting to viewers, since it makes them focus on the text vs. important details in the scene. Another reason why subtitles aren’t ideal is because they don’t express the emotions being delivered in the dialogue.

Audio dubbing to regional languages is an alternative option that translates a foreign language program into the audience's native language. Dubbed tracks are created by adding language-specific content to the original audio and have become a cost-effective solution for content creators to reach audiences in different areas of the world. When creating dubbed audio tracks, content creators must ensure high quality and synchronization with the video or original audio track.

HOW TO SIMPLIFY AUDIO DUBBING QC WORKFLOWS

Recent advancements in ML and AI technologies have simplified the way that content creators perform quality control on audio dubbed tracks, automating the process of verifying languages and assuring synchronization.

There are several key capabilities content creators should look for in an ML/AI-powered QC solution for dubbing workflows in order to ensure the utmost efficiency, accuracy and quality of experience for viewers.

One important feature is the ability to verify dubbing packages with complex structure. The QC solution should be able to verify complex dubbing packages. Oftentimes, a dubbing package consists of multiple MXF and .wav files as opposed to a single file. The MXF file contains video tracks along with original audio tracks and .wav files corresponding to multiple dubbed audio tracks. Sometimes .wav files represent individual channels of a 5.1 audio track, or there are multiple audio tracks or channels encapsulated in container formats instead of .wav files. The QC solution should be able to handle package variations and verify the multiple dubbed audio tracks properly.

For audio dubbing, metadata verification is a must-have capability in a media QC solution. The system needs to verify the number of audio tracks or the channel configuration of multiple dubbed tracks, along with the duration of the original audio track compared with the dubbed audio tracks.

Language identification is an integral component of a QC system to ensure that each track has the correct intended language. Over the years, machines have become more intelligent and computing power more affordable, making accurate, automated language detection a reality. If the content creator has access to a few hours of audio content in the target languages, they can be used to train the ML models for language prediction purposes. Content creators can then verify that the language is correct by checking it against metadata.

During the audio dubbing process, content creators must make sure synchronization between video and dubbed tracks exists. This can be challenging, as there is no way to map the correct audio sequence for any video screen with dialogue. However, the majority of video content contains black frames and color bars that are designed to help in meeting requirements like synchronization. The corresponding audio sequence for black frames and color bars is silence and test tone, respectively. Choosing a QC solution that can verify the presence of black frames in video tracks as well as silence while color bars appear with test tone in dubbed audio tracks is critical.

Synchronization must also be present between original and dubbed tracks. This is no easy feat since audio data in master and dubbed tracks are completely different. It’s likely that common background music or effects were not used for the master track and dubbed track. These can be separated from the audio track using mechanisms like band pass filter.

Checking for loss of synchronization between background beds of dubbed audio tracks and the original audio track can easily be performed by a QC system with ML and AI technology. The challenge is ensuring a proper separation of background bed from audio tracks. One way this can be achieved is by comparing loudness curves. The technology will compare loudness curves, checking for mismatch between loudness values of original and dubbed tracks.

CONCLUSION

Through audio dubbing, content creators can reach audiences all around the world, providing content in their local language without the limitations of subtitles. Automating QC dubbing workflows speeds up this process immensely, while also introducing increased accuracy. When ML and AI technologies are added to the workflow, dubbed packages can be created even more rapidly, with reduced manual intervention. It’s a win-win situation for content creators and viewers. Content creators can expand their brand into new regions of the world, and viewers are assured a high-quality television experience in their native language.

Niraj Sinha is the principal engineer at Interra Systems.

SEATTLE—Amazon Web Services (AWS) has released a Live Streaming with Automated Multi-Language Subtitling solution for automatically generating multi-language subtitles for live streaming in real time.

AWS Live Streaming encodes and packages content for adaptive bitrates across multiple screens, then converts the audio to text using AWS Lambda, Amazon Transcribe and Amazon Translate.

The solution can be used out-of-the-box, customized for specific use cases or paired with AWS Partner Network (APN) to implement an end-to-end subtitling workflow.

Subtitle generation starts when MediaLive output is sent to the solution’s Amazon S3 bucket. The CaptionCreation Lambda function takes the manifest files from the bucket, extracts unsigned pulse-code module (PCM) audio from the TS video segments and saves the PCM audio to Amazon S3.

The TranscribeStreaming function then converts the audio stream to text in real time, before sending the transcript back to CaptionCreation. If multiple languages are required, the CaptionCreation function calls Amazon Translate to translate the transcript.

MediaPackage ingests the files and packages them into formats that are delivered to four MediaPackage custom endpoints.

PALM SPRINGS, CALIF.—This year’s HPA Tech Retreat will once again see a gathering of industry-leading thinkers who will discuss the latest trends in the creation, management and dissemination of video content. The weeklong Palm Springs event, which is capped at just 700 attendees and begins Monday, Feb. 11, will feature a main conference program, the “TR-X Tech Retreat Xtra” focused on eSports, as well as breakfast roundtables and a curated innovation zone, along with other opportunities for industry leaders to meet and exchange ideas.

The event officially gets underway with the Tuesday Supersession, which will highlight Next-Gen Workflows and Infrastructure, featuring insight from Tom Graham of Dolby on Dolby Vision from set to consumer; Tina Eckman of Dolby on Dolby Atmos from set to consumer; and Scott Kramer of Netflix on immersive sound from capture through post production.

HDR AND THE MOVE TO 8K

Numerous sessions on Wednesday including an HDR “Mini” Session, will address the importance of HDR in video production. There will be an introduction to HDR, how it is being used in anime and how it can push the limits of motion in video production. These sessions should highlight how 4K and UHD are not interchangeable terms, and that resolution is just part of the picture with HDR playing an important role in UHD, according to HPA president Seth Hallen, who will host the “Intro to HDR” session.

“UHD implies not only resolution, but other things such as HDR and its wider color gamut,” said Hallen, who is also head of global strategy and business development at Pixelogic. “Studies have shown that the human brain can detect a greater range of light to dark than it can picture resolution. HDR thus is a massive improvement for overall picture quality. But it can also ensure that the content is delivered as closely as possible to the filmmaker’s intent. Hopefully the consumer will appreciate that.”

None of that is to say that resolution isn’t important, and even as 4K content is only now starting to pick up steam the industry is clearly looking ahead to what the next generation—8K—can offer. Pete Putman of ROAM Consulting will host two sessions, as well as take part in a breakfast roundtable—firstly offering his thoughts on HDMI 2.1 and Display-Port before providing an overview of products in the world of consumer electronics that were introduced at the recent 2019 CES in Las Vegas. Putman will then dive into what we can expect in 8K in 2019 and beyond in the session, “8K: Whoa! How’d We Get There So Quickly?”

“The 8K talk will be an overview of everything from cameras to displays, based on current information,” said Putman, who will discuss various aspects of the technology including appropriate codecs for compression and data transfer rates. “It should bring people up to speed on just how fast 8K is coming to market.”

TAKING IT ON THE ROAD AND BRINGING IT HOME

Mark Chiolis, director of business development for the Mobile TV Group will host Thursday’s “Remote and Mobile Production Panel,” which will discuss the changing world of remote production that now includes not only live sports and concerts for broadcasters, but addresses how this is fitting into the world of streaming services and the growing eGames/eSports market.

“With so many live events you start out with a blank canvas and the set up for a production crew is very tight,” said Chiolis. “We’ll discuss the whole gamut of what is involved in preparing for a live production and what we’ve seen come out in 2018 and what we expect to help with productions in the next couple of years. Part of the discussion will include the role of REMI applications [REMote Integration], IP workflows, virtual productions and of course, eSports. It should be an exciting time for live production.”

Thanks to larger TV screens with greater resolution joined by surround sound systems, seeing movies at home has gotten closer to what filmmakers have intended, and this aspect of what the viewer actually sees will be discussed in “Downstream Image Presentation Management for Consumer Displays.” Moderated by Michael Chambliss of International Cinematographers Guild, the session will feature Steven Poster of the American Society of Cinematographers as a panelist.

“We, as the ASC, started participating at the HPA Tech Retreat seven years ago, and at the time we were asked why we were here,” said Poster. “Now they have asked us to give our presentation, which we see as important as it discusses the importance of the creative intent of the filmmaker.”

That intent of what the director and cinematographer has improved a lot from the days when movies were in essence “butchered” via pan and scan, but today even with HDR and greater picture resolution, there still needs to be more progress. Poster, who first formed a technology committee when he was ASC president 20 years ago, had a vision whereby the TV set could ensure that the picture is up to the standards of the filmmaker. However, he admitted that the technology wasn’t quite there at the time.

Now two and a half decades later that is finally becoming a reality.

“Our TV set technology has become quite good,” said Chambliss. “So why can’t we have a little bit of metadata that is used in the production making it way to the sets. So far Netflix and Sony have implemented what is called calibrated modes so that the set closely mirrors the monitors we used in the production. Instead of leaving this in a 1950s ‘it is up to the viewer to adjust the set,’ the smart TVs could optimize the picture. The sets are capable of doing this, so let’s automate the process.”

CAPTIONING AND SUBTITLES

Subtitling and captioning will get its moment in the spotlight at the Retreat with “IMSC 1.1: A Single Subtitle and Caption Format for the Entertainment Chain,” hosted by Pierre-Anthony Lemieux of Sandflow Consulting.

Lemieux will highlight IMSC, the W3C standard for worldwide subtitles/captions, which came about as the result of an international collaboration—the initial version of which was published in 2016.

“Technology has made it possible, along with flexible delivery for subtitles and captions to be provided in so much more content,” said Lemieux. “It has been in the broadcast world where the United States was the exception for providing subtitles for the hearing impaired, but worldwide there have been little mandates for accessibility. But that is changing.”

In addition, thanks to streaming services—notably Netflix, which provides content from international markets—as well as changing viewing habits, subtitles and captions are becoming much more important.

“Even people that aren’t hard of hearing are enjoying content in bed late at night or on buses and airplanes and having the sound on isn’t an option, so captions are necessary for the message to be conveyed,” added Lemieux. “It is easy to forget that just 10 years ago subtitles were literally burned on a film with a laser in France, and that was an expensive and difficult process. Today technology—notably IMSC—has made it easy for subtitles to be added to the content. Subtitles and captions were an afterthought, but now in the world of audio/video production these will be first class citizens.”

Predicting everything that attendees will see would be akin to looking into the future, said Mark Schubin, program chair at the HPA. “There’s a wild array of possibilities on both the acquisition and presentation ends. How do we deal with the new world?”

For more information, visithpaonline.

STOCKHOLM—Edgeware today said it is buying Swedish subtitling expert Cavena Image Products, acquiring 100 percent of the company’s shares for 8 million Swedish Krona (or about $900,000 U.S.) on a cash and debt-free basis.

The move broadens Edgeware’s origin offering to broadcasters by enabling the company to integrate Cavena’s subtitling solutions.

Adding subtitling technology will allow broadcasters to build a higher-value origin system that can enable personalized advertising, support any type of device and automate the subtitling process, the company said.

“Our growing numbers of broadcaster customers are demanding the highest quality from their OTT services,” said Edgeware CEO Karl Thedéen. “We already work alongside Cavena’s systems in some of our largest deployments, such as myTV SUPER in Hong Kong.”

Integrating Cavena technology to serve the needs of broadcasters “makes sense,” he said, adding that the acquisition also offers “the opportunity to introduce our respective technologies to each other’s customers.”

Cavena, which has sold its subtitling products for more than 10 years to broadcasters, counts among its customers Sky, Canal+ and HBO Asia. Its technology delivers high-quality subtitles across any screen format to any client in any character set.

The company’s technology uses Optical Character Recognition to convert subtitles from image to text formats, which provides a higher-quality experience for OTT viewers. The system can also transcode different subtitle formats within milliseconds, scaling to support hundreds of channels.

Cavena-generated text is then passed on to Edgeware’s TV Repackager system, to create OTT content. The combined solution supports any language, including more complex Asian character sets.

According to Edgeware, the acquisition will have only a minor impact on its 2019 forecasted revenues and profitability. The deal is expected to close during Q1 2019. More information is expected to be available Feb. 1 with the presentation of Edgeware’s 2018 year-end report.

More information is available on the Edgeware website.

GENEVA—At its most recent meeting, the DVB Steering Board approved the new specification for DVB TTML subtitling systems. The specification enables the delivery of next generation subtitles in transport streams for broadcast television services, based on W3C Timed Text Markup Language. The new specification is designed to complement the bitmap based subtitle solutions.

With the new TTML delivery specification, service providers will be able to transition over time to a common TTML subtitle format for both broadcast and internet delivered services.

“The approval of the transport stream carriage of the TTML specification adds a very important component to the DVB specifications, most notably providing support for UHD subtitles,” said Kevin Murray, DVB technical module chair. “This also provides a common subtitling mechanism between transport streams and DASH delivered content and, being a textual based format, opens up a range of possibilities for enhancing accessibility of services.”

The specification has been published as BlueBook A174 and has been forwarded to ETSI for standardization as a European Norm.