Connections

Cloud-based artificial intelligence (AI) and machine learning (ML) solutions are continuing to transform and accelerate virtually every aspect of content creation and distribution, according to Digital Nirvana.

The technologies are bringing new levels of accuracy, efficiency, compliance and cost savings to broadcast operations.

During the Licensing & Distribution breakout session “Revolutionising Media Creation and Distribution Using the Science of AI & Machine Learning” at the Feb. 25 Content Workflow Management Forum, Digital Nirvana highlighted real-world examples of media enterprises that are transforming every aspect of their workflows, from subtitles creation to the distribution of content, by implementing the use of AI and ML technologies.

In the area of content acquisition, content creation and production, it first explored a speech-to-text and captioning workflow that Digital Nirvana co-developed with a “well-known entertainment, news and information outlet,” according to Ed Hauber, director of business development at Digital Nirvana.

For that client, the “most pressing requirement was the ability to turn around and deliver accurately captioned content within a very short time frame,” he explained.

“The single biggest driver was time or the extreme lack of time,” he noted, adding: “Like many operating in the news arena, deadlines are tight and time is in short supply. Our client had a two-hour window in which to ingest, edit, caption and deliver its finished product to” an over-the-top (OTT) provider in the expected format.”

Adding to that challenge was a “large quantity of daily, raw contribution content from the field, all of which had no metadata or context of any kind,” he pointed out.

Digital Nirvana used automatic speech recognition (ASR) technology to help solve that challenge, according to Russell Vijayan, head of AI products and services at Digital Nirvana.

Digital Nirvana provided the client with the ability to go to specific places in a piece of video and, by clicking on search terms, find all the clips they needed much faster, allowing it to produce content within the timeline that they were given, explained Vijayan.

The secondary challenge for that client was the need to deliver accurately captioned content within the time frame allotted to it, Hauber said. Adding to the challenge was the need to deliver content in full compliance with the OTT service provider’s “stringent style guidelines for captioning and subtitles,” he noted.

The client did not have time to use a third-party resource to send out the content, have it captioned and turned around, Hauber said, adding the client wanted an in-house workflow where they could automatically use speech-to-text for captioning and assess quality, and also publish this to the streaming platform the content was being distributed on.

Digital Nirvana used a combination of technologies to overcome those challenges quickly, Vijayan explained. Digital Nirvana created transcripts that could be converted into subtitles or captions.

It then applied natural language processing (NLP) that would help split the captions so that the user didn’t need to spend much time making any changes, he said. The captions created using this system have 98-99% accuracy, he added.

The third challenge was something we see sort of across the board today: Everybody needs to localise their content – specifically the captions – into other languages, in this case, English and Latin American Spanish, Hauber said.

In this case, Digital Nirvana needed to create Spanish captions after the English captions were created, Vijayan said, adding it created a set of algorithms that would dictate how much content needed to be translated to get a much more accurate result. The second use case was to import existing English captions and quickly translate them, he added.

Moving on to distribution and video metadata, Hauber briefly discussed applications of AI in the content delivery space, “primarily the mining of insights from within content, which allow our clients to analyse and report on production product placements, perform ad classification and uncover potential compliance issues such as identification of explicit content.”

Hauber pointed to use cases with two multichannel video programming distributors. In both cases, they had original content but most of the distribution was other peoples’ content — 300-400 channels worth of it, he said, noting both companies used compliance logging and monitoring technology to record distributed content that had been aired. The goal was extracting data from the distributed content, he noted.

Digital Nirvana solutions provide: Customised video-based metadata for better searchability; the ability to accurately identify spots based on content; identify and replace brands within content based on consumer data; and a reduction of effort in identifying explicit content, according to the company.

As an example, Vijayan noted there may be a Microsoft Surface tablet on the screen and a company might want to replace it with an Apple ad.

Click here to access video of the presentation. Click here to download the presentation slide deck.

The Content Workflow Management Forum was produced by MESA, the Audio Business Continuity Alliance, Content Localisation Council, Smart Content Council, and the Hollywood IT Society, with sponsorship by Iyuno Media Group, Richey May Technology Solutions, Whip Media Group, Deluxe, Digital Nirvana, Meta, Vubiquity, EIDR, Keywords Studios, Los Angeles Duplication & Broadcasting, Nexus TV, OOONA, Signiant and Titles-On.