LMT Chair: Language Metadata Table Helps Everybody Manage Their Media Assets

The Language Metadata Table (LMT) was created to provide a unified source of reference of language codes for use throughout the media and entertainment industries, and it continues to win support and make important strides, according to Yonah Levenson, manager of metadata strategy and terminology governance at WarnerMedia/HBO and chair of the Language Metadata Table (LMT).

“The Language Metadata Table is going to help everybody manage their media assets,” she said July 2 during the session “How the LMT Helps the World Manage Our Media Assets” at the Global Media & Entertainment Day event presented live, virtually, from London.

In addition to WarnerMedia, LMT Committee members and contributors include Amazon, Discovery, Disney, Fox, Lionsgate, NBCUniversal, Paramount, Showtime, Sony and Turner, she pointed out.

“We have more and more organizations that are getting on the LMT bandwagon,” Levenson said.

That signals the clear need there is for LMT in the industry. “What do we know? We know that there are lots of languages everywhere,” and, if you speak more than one language, it is fun to be able to find content in languages you are looking for, Levenson noted.

“But one of the problems is that there really hasn’t been a single unified standard of language codes, and that’s where the LMT comes in,” she told viewers, adding it is also important because language metadata is used by every aspect of media and entertainment, publishing and other content-centric organizations.

She created LMT at HBO in 2017 “to make sure we had a unified source of reference for language codes, and what has happened is that it’s expanded and now it’s being used for the media and entertainment industries — and it also is applicable really for anybody that needs a language code,” she said.

LMT started with 128 languages and the Media & Entertainment Services Alliance (MESA) serves as its sponsor, she noted, adding it was presented publicly in 2018 at a MESA event, a working group was formed that year, and LMT version 1.0 was published that year also.

LMT version 3.0 was published early this year and the number of languages has grown to “200-plus,” she said.

There is, meanwhile, a “vetting process [for LMT] to make sure that the codes we’re using are going to meet everybody’s needs, so this way it’s truly a consensus and truly a global taxonomy,” she told viewers.

“We really only focus on languages” with LMT, but “we include notation for script or text,” as well as endonyms (the name of the language in a particular country’s language) to make sure it is accessible to everybody,” she said. Also included are exonyms (the language name as spoken in other countries) because “we’re trying to take away any ambiguity and make sure that the language that is written down in the LMT is the language that people who are implementing can understand,” she explained.

Levenson recalled she was on a flight to London about 1-2 years ago and was browsing through the movie offerings. The audio and subtitle language abbreviation “PO” was an option and she was not sure what language that was. She guessed Polish and was wrong. It was actually Portuguese even though PO is not a valid standards abbreviation for either Polish or Portuguese, she noted. It validates the reason she created LMT.

The LMT codes can be applied to audio and timed text for content, visual or written languages, accessibility for the visually and hearing impaired, rights and licensing localization and distribution territories, she said, making note of the distinct differences between Spanish used in Latin America versus other regions.

“There are language codes that are needed for each one of these areas,” Levenson said.

By having a working group that vets the terminology across the industry, she said, “we make sure that we can identify and have a language code for the distinctions between the spoken and written languages, as well as the localization.”

Also, “by having consistent codes between the service providers, the clients and the content owners, it really reduces the Tower of Babel and makes it easier for everybody to understand,” she noted, adding: “It also reduces the amount of development work that needs to be done because everybody’s working off the same table. You don’t have to maintain mapping tables and have different payloads for different distributors.”

While discussing LMT use cases and examples, she said: “It facilitates the licensing of international content”; it is for the distribution of non-English content”; LMT addresses accessibility requirements; and is for end-user localization preferences.

And “this is really the heart of it,” she told viewers: “The codes can be applied… for audio, as well as the closed captions, burned in or forced narratives [and] the accessibility.” LMT includes codes for sign language, she noted. LMT is also applicable for those involved with acquisition rights and electronic sell-through partners.

Another key point: “We are not inventing the wheel” with LMT, she pointed out, noting it adheres to Internet Engineering Tax Force (IETF) BCP 47, as well as other industry standards.

While MESA continues to sponsor and promote LMT, “one of the things we know is that, as LMT continues to grow… we really need a technical home,” she went on to say.

And the Society of Motion Picture and Television Engineers (SMPTE) has “stepped up to the plate and they’re going to provide the tools and infrastructure for the LMT,” she said. LMT is also partnering with other industry standards, including the InterSociety Digital Cinema Forum and MovieLabs, she said.

“We’re working on creating a standardized register of the LMT codes” that will “adhere to” the SMPTE Public Committee Draft (PCD) mechanism, she added. Later this month, LMT will start more detailed work on the new Register process and the PCD is expected a couple of months later, she said.

Prior to standardized Register approval, the Register will be available to test for humans and machines at the SMPTE-RA website, she noted. The whole PCD process is expected to take anywhere from six months to three years, she estimated.

However, she explained: “Obviously we’re hoping for a much shorter time period and that’s what we’re aiming for since we’ve already been working on it for a while. But the reality is that standards can take time. You want to make sure you really kick the tires and that the process is robust so nothing breaks.”

The fourth annual M&E Day event, presented by MESA, featured mainstage panels and more than 15 breakout sessions, covering the latest it data, cloud, IT and security across the media and entertainment technology ecosystem.

The event was presented by Caringo, with sponsorship by Convergent Risks, Cyberhaven, Richey May Technology Solutions, RSG Media, Signiant, Whip Media Group, Zendesk, Tape Ark, Sony New Media Solutions, 5th Kind, ATMECS, Eluvio, Tamr, the Audio Business Continuity Alliance (ABCA), the Entertainment Identifier Registry (EIDR) and The Trusted Partner Network (TPN).